Integrating Health Care Systems to Maintain Quality Care and to Manage Cost.

Chapter 1: Introduction to the Study

Introduction and Background

The Department of Defense (DoD) is responsible for overseeing the Military

Health System (MHS), the means for providing care to military members, retirees, and

eligible family members. The MHS is the primary source of rendering health care in

military treatment facilities (direct care system), which is augmented by care provided

under the Civilian Health and Medical Program of the Uniformed Service

(CHAMPUS)/TRICARE in the private sector (purchased care) in the United States and

overseas. Outside the United States and its territories, except for the District of Columbia,

CHAMPUS/TRICARE is referred to the TRICARE Overseas Program (TOP), which is

the same CHAMPUS/TRICARE program administered in the United States, with the

overseas cultural differences and nuisances for every country being taken into

consideration.

The DoD is concerned with escalating health care costs throughout the MHS. In

general, the rising costs of health care can be attributed to factors including advanced

technology, people living longer, fraud and abuse, inflation, changes in physicians’

philosophy in treating their patients, and malpractice suits. This research focused on the

exorbitant rise in health care costs in the Philippines. According to Scott (2006), the

excessive increase in health care costs in the Philippines caught the attention of the

DoD’s Office of the Inspector General (DoDIG), prompting an investigation into

improper payments to providers and third-party billing agencies and waiving of

beneficiaries’ deductibles and cost shares. The DoDIG found that the Defense Health

Agency (DHA), formerly TRICARE Management Activity (TMA), paid $2.87 million in

fiscal year (FY) 1988 and $64.19 million in FY 2003 for health care in the Philippines

(Scott, 2006). These amounts reflected a 2,135% rise in health care costs over a 6-year

period to a population that had remained relatively stable throughout 5 plus years (Scott,

2006). The DoDIG recommended that DHA implement provisions of accountability for

claims submitted by providers and third-party billing agencies that submit claims on

behalf of the providers and initiate steps toward ensuring that providers do not waive the

cost shares and deductibles of beneficiaries (Scott, 2006). The DoDIG conducted an audit

to examine inaccurate reimbursements of TRICARE claims and the necessity for the

establishment of negotiated rates for health care services rendered overseas (Prinzbach,

2008).

The DoDIG also conducted an audit from July 2009 through June 2011 to assess

the certification of Philippine providers who rendered health care to military retirees and

their families and the process for adjudicating the claims associated with health care

(Carey, 2011). The audit found that the TOP contractor who was responsible for provider

certification failed to “provide adequate documentation to support 28 of 63 certifications

of medical providers,” and the overseas claims processor failed to verify the addresses of

TRICARE beneficiaries before mailing reimbursement checks (Carey, 2011). The

DoDIG reported that beneficiaries were in jeopardy of receiving care from providers who

were not certified and licensed and that beneficiaries did not receive reimbursement from

TRICARE because their addresses were never confirmed (Carey, 2011).

The DoDIG audits recommended several areas in which the DoD could reduce the

cost of health care in the Philippines. As a result, the DHA implemented several

administrative controls in the Philippines that lowered the cost of health care somewhat,

but other cost containment measures are still necessary. Therefore, the DHA devised a

plan to implement the DoD TRICARE Philippine Demonstration Project (Philippine

Demonstration) as a cost containment measure and fraud deterrent.

The following sections include the problem statement, research questions and

associated hypotheses, the theoretical basis for the study, the nature and scope of the

study, and its assumptions, delimitations, and limitations. The chapter concludes with a

section on the professional and social implications of the study that addresses how the

study contributes to the delivery of an alternative health care model and social change.

Problem Statement

TMA (2008a) identified a major concern in the Philippines with rising health care

costs. The Philippines has a reputation for ubiquitous health care fraud within TRICARE,

so the DHA of the DoD implemented administrative controls to deter fraud in the

Philippines (TMA, 2008a). In 2013, the DHA implemented the DoD TRICARE

Philippine Demonstration because the administrative controls were not enough to deter

fraud and decrease health care costs.

The purpose of this quantitative cross-sectional research study was to determine

the efficacy and acceptability of a different method to deliver health care to increase

access to health care and decrease out-of-pocket costs while maintaining quality of care

for TOP Standard beneficiaries who receive health care under the Philippine

Demonstration.

Gaps in the Literature

Scholars have addressed country-specific health systems, rising medical costs,

cost containment strategies, reimbursement methods, and fraudulent claims submissions

in extensive literature writings. I am not aware of any literature with controlled studies

that integrates the MHS with foreign health care systems testing a different method of

providing health care. The Philippine Demonstration is the first time the DoD has tested

an alternative method of providing care in a foreign country. The DoD conducted a

demonstration testing an alternative method of delivering health care to beneficiaries and

active service members in the 1990s in 11 states under the CHAMPUS Reform Initiative

during the managed care era (Anderson, Hosek, & Bloomfield, 1994). This

demonstration resulted in the addition of a health maintenance organization option and a

preferred provider organization (PPO) option to CHAMPUS.

Purpose of the Study

The cost of health care continued to rise in the Philippines, while the number of

TRICARE beneficiaries has remained the same. According to TMA (2008a), the

Philippines is known for its widespread health care fraud against the TRICARE program.

The DHA implemented administrative controls such as proof of payment, prepayment

review, boots-on-the-ground validation and certification, a government-directed foreign-

fee schedule, and attestation, but these controls alone have not been sufficient to address

the problem (TMA, 2008a). As an alternative method of health care delivery, the DoD

TRICARE Philippine Demonstration was conceptualized as a resolution to provide

beneficiaries using the TRICARE Standard option access to quality care as well as to

control health care costs without jeopardizing access to quality care.

Therefore, the purpose of this quantitative cross-sectional research study was to

determine the accessibility and acceptability of health care received under the Philippine

Demonstration for TOP Standard by beneficiaries, retirees, family members, and

TRICARE for Life beneficiaries who reside in the Philippines.

Nature of the Study

This study was of a quantitative nature. I used the quantitative research design to

address questions regarding access to care, cost, and quality of care for the TRICARE

Standard population who reside in the Philippines and receive care in designated

Demonstration areas (Creswell, 2009). Essentially, I adopted a postpositivist view (see

Creswell, 2009) using Andersen’s (1995) behavioral model of health services use and

used Donabedian’s (1990) quality model to assess patient experience and satisfaction.

Research Question and Hypotheses

I derived the following research questions (RQs) from the DoD TRICARE

Philippine Demonstration Beneficiary Satisfaction Survey:

RQ1: How has the Demonstration affected the beneficiaries’ financial burdens

and confidence that their health care needs will be met?

H01a: There is no statistically significant relationship between financial

burdens and satisfaction with the Demonstration overall.

Ha1a: There is a statistically significant relationship between financial burdens

and satisfaction with the Demonstration overall.

H01b: There is no statistically significant relationship between financial

burdens and confidence that health care needs will be met.

Ha1b: There is a statistically significant relationship between financial burdens

and confidence that health care needs will be met.

RQ2: What is the relationship between gender, age, and health care finder

functionality for accessing quality health care under the Demonstration in Metro

Manila, Cavite City, Cavite, and Iloilo City, Iloilo?

H02a: There is no statistically significant relationship between gender, age,

and health care finder functionality access to quality health care in Metro

Manila, Cavite City, Cavite, and Iloilo City, Iloilo.

Ha2a: There is a statistically significant relationship between gender, age, and

health care finder functionality access to health care in Metro Manila, Cavite

City, Cavite, and Iloilo City, Iloilo.

RQ3: What is the relationship between a closed network, actual beneficiary out-

of-pocket cost, and quality health care in Demonstration areas?

H03a: There is no statistically significant relationship between actual

beneficiary cost and a closed provider network.

Ho3a: There is a statistically significant relationship between actual

beneficiary out-of-pocket cost and a closed provider network.

H03b: There is no statistically significant relationship between quality health

care and a closed provider network.

Ha3b: There is a statistically significant relationship between quality health

care and a closed provider network.

Beneficiaries’ Actual Out-of-Pocket Costs

In order to determine if beneficiaries out of pocket costs decreased, it is

imperative to conduct an analysis of their out of pocket costs for calendar years 2011,

2012, 2013, 2014, and 2015 in Metro Manila, Cavite City, Cavite, and Iloilo City, Iloilo

decrease or increase. Prior to the Philippine Demonstration, beneficiaries paid a

deductible and cost share plus balancing billing. After the implementation of the

Philippine Demonstration, beneficiaries paid a deductible and cost share. It is my

assumption that beneficiaries’ overall out-of-pocket costs would increase initially

because greater than half the TRICARE beneficiaries were not using the TRICARE

benefit. As more beneficiaries use the TRICARE benefit and receive care from approved

demonstration providers, the cost should eventually decrease.

Theoretical Base

As a theoretical framework, I used Andersen’s behavioral model of health

services use and Donabedian’s quality model are used to explain the study results. The

goal of Andersen’s (1995) model was “to provide measures for access to medical care”

(p. 4) to include potential, realized, equitable, and inequitable access. Andersen enhanced

the model to include measures important for health policy and reform. Andersen’s model

in the third phase measures for effective access as shown by improved health status

through use or beneficiary satisfaction. Efficient access is based on health status and

increased satisfaction in relation to the consumption of health care (Andersen, 1995).

Donabedian’s (1980) quality assessment of health care was based on the experience of

the patient. Donabedian’s (1980) quality assessment model looks at how the patient

defines quality based on his or her values and expectations.

The theoretical foundation for this study, presented in depth in Chapter 2, was the

Andersen’s behavioral model access utilization and Donabedian’s quality assessment. For

this study, access was defined as the entry point when a patient enters the health system.

Operational definitions for other concepts used in this study are listed in the following

section.

I present a more detailed description of the theoretical framework as well as the

conceptual model that grounds this study in Chapter 2.

Definition of Terms

Definitions of terms used in this study were as follows:

Access: The point of entry where a patient enters the health care system.

Approved Demonstration provider: A provider in the Philippines who has agreed

to participate in the DoD TRICARE Philippine Demonstration. This provider has agreed

to accept the TRICARE-allowed amount that includes the beneficiary cost share and

deductible and submit claims for the TRICARE beneficiaries for medical services

provided. The provider has also agreed to meet the terms of on-site verification,

certification, and credentialing according to TRICARE regulations and policies. An

approved Demonstration provider may opt to require TRICARE beneficiaries to pay their

deductible and cost shares up front for services provided.

Certified Philippines provider: A provider who has met the on-site verification,

certification, and credentialing requirements. This provider has not agreed to accept the

TRICARE allowed amount that includes the beneficiary cost share and deductible or to

submit claims for TRICARE beneficiaries for medical services provided. This provider

can require TRICARE beneficiaries to pay the entire bill in full up front for services

provided.

Civilian Health and Medical Program of the Uniformed Service (CHAMPUS):

Medical program established to provide care to family members and retired military

members in the private sector.

Health care fraud: Providers intentionally bill services at a higher level

suitable for the services rendered to the patient.

Power purchasing parity: Calculating the exchange rate of currency for different

countries in order to purchase the same goods in one country at the same rate in another

country.

Quality of health care: Consists of structure, process, and outcome as defined by

Donabedian (1966).

TRICARE beneficiary: A military service member, retired service member, or

eligible family members of the military or retired service members. For the purposes of

the DoD TRICARE Philippine Demonstration, TRICARE beneficiaries were retired

service members and eligible family members of the military and retired service members

living in the Philippines.

TRICARE program: The DoD managed care program that allows for competitive

selection of contractors who are willing to take financial risks for health care delivery

under CHAMPUS (TMA, 2009). The implementation of TRICARE was an enhancement

to CHAMPUS introducing TRICARE Prime (DoD health maintenance option),

TRICARE Extra (preferred provider option), and TRICARE Standard (basic CHAMPUS

fee-for-service).

TOP: The TRICARE health care program administered overseas and in the U.S.

territories, except for the District of Columbia. There is one option available overseas for

nonactive service beneficiaries, the TOP Standard option.

Assumptions

TRICARE Standard beneficiaries residing in the Philippines compose a small

segment of the Filipino population accessing health care. The DoD TRICARE Philippine

Demonstration hinges on providers in the civilian sector joining a closed network of

providers to provide care to TRICARE Standard beneficiaries and accepting what

TRICARE allows as payment in full. I assumed that some TRICARE beneficiaries

residing in the Philippines would be resistant to receiving care under the Demonstration,

while other beneficiaries would welcome the alternative method of delivering health care

for a decrease in out-of-pocket costs and an increase in access to quality health care. I

expected that age and gender would have an impact on the health care functionality.

Limitations

The accuracy of the CHAMPUS/TRICARE beneficiary contact information was a

major limitation for the Beneficiary Satisfaction Survey. For example, a beneficiary may

not have updated their address or telephone number in the Defense Enrollment Eligibility

Reporting System (DEERS), the official system of record for CHAMPUS/TRICARE

beneficiaries. These beneficiaries input their personal contact information, such as

mailing address, telephone number, and e-mail address (this is not a required field) in

DEERS. Like any database, the information is only as good as the data that goes into the

database. If beneficiaries did not update their information, their chances for participating

in the survey were decreased. Outreach to the beneficiaries in the Philippines informing

them to update their personal information via the DHA’s Web site or the TOP

contractor’s Web site should have minimized this limitation.

The country code listed in DEERS for Philippines addresses is “PHL”; however,

the country code listed for an Army Post Office or Fleet Post Office (APO/FPO) address

in the Philippines is “US.” The “US” country code automatically excluded a legitimate

APO/FPO address in the Philippines, thus eliminating potential respondents for the

survey. The inclusion of the zip code for the APO/FPO address as a data element should

have resolved this limitation. Although Tagalog is the native language in the Philippines,

English is a familiar language in the Philippines. Some beneficiaries who do not speak

English can opt to have someone in the household translate. If beneficiaries required a

translator, the interviewer should have required the beneficiary to authorize the

interviewer to speak with the individual in the household as a translator. The interviewer

should have documented the authorization to show there was no breach of the Health

Insurance Portability and Accountability Act of 1996 (HIPAA). There is a possibility in

which beneficiaries with one e-mail address per household may not have received a

survey to avoid violating HIPAA. The only way the beneficiary could receive a survey

via e-mail was to notify the DHA official appointed to oversee this survey and provide

authorization to send a survey via e-mail if the beneficiary met the criteria to participate

in this survey.

Telecommunications barriers exist in the Philippines because some beneficiaries

may only have a mobile telephone. In cases such as this, the beneficiary may not have

answered the telephone or had the telephone set up not to accept incoming calls because

the call would use the minutes on the telephone plan. If the beneficiary had a usable e-

mail address, the survey could have been sent via e-mail. There were other barriers, such

as an incorrect telephone number in DEERS that could have been be remedied by the

beneficiary updating their contact information in DEERS.

Delimitations

This study was delimited to TRICARE Standard beneficiaries who lived in the

Philippines and received health care from approved Demonstration providers in

designated Demonstration areas from January 1, 2013, through December 31, 2015. The

purpose of this study was to evaluate the patient’s experience to determine the efficacy

and acceptability of the Philippine Demonstration, focusing on outcome measures for

access to quality of health care and beneficiaries’ financial burdens. I used secondary data

from a Beneficiary Satisfaction Survey directed to respondents who received health care

during all phases of the Demonstration. This research confirmed if the DoD met its

objective, decreasing out-of-pocket costs for beneficiaries while ensuring that the

beneficiaries maintained access to quality health care in the Philippines.

Significance of the Study

According to Schoen et al. (2010), the health insurance design affected access,

cost, and experience of care in various countries. The Philippine Demonstration will

allow the DoD to determine whether it is possible to control costs, reduce aberrant claims

activity, and eliminate balance billing issues while providing high-quality, safe, and

effective health care to TRICARE Standard Overseas beneficiaries who reside in the

Philippines and receive care in designated Demonstration areas (TMA, 2008). The

alternative method will change the reimbursement methodology, reducing out-of-pocket

costs and eliminating payment in full of services up front, thus alleviating financial

burdens for beneficiaries.

The Philippine Demonstration would have a major impact on health care delivery

for TRICARE Standard beneficiaries residing in the Philippines and providers rendering

services to these beneficiaries. The engagement of cultural and economic differences in

two very distinct health care systems, the MHS and the Philippines health care system,

would merge, creating social change.

Professional Application

The DoD will assess the efficiency and acceptability of the Demonstration during

and at the end of the 3-year Demonstration. If it is successful, there is a possibility that a

closed network of preferred providers may be implemented throughout the Philippines. A

closed network of preferred providers could become the premier health care delivery

model for TRICARE Standard beneficiaries who receive health care overseas.

Access to providers in the Philippines was never a concern because TRICARE

Standard beneficiaries had the option to see any TRICARE certified provider. The

implementation of a closed network of preferred providers guarantees access to quality

providers who are committed to providing safe, quality health care to TRICARE

beneficiaries at a lower cost. Health care costs could eventually decrease due to the

acceptability of the fee schedule by providers versus the DHA paying billed charges in

various countries. Furthermore, decreasing the number of providers should result in better

oversight of health care that could eventually lead to a decline in fraud.

Philippine Demonstration Blueprint

The DHA established specific criteria to use a phased approach for launching the

Philippine Demonstration in designated areas. The TOP contractor who administers the

TRICARE program overseas was responsible for developing a closed network of

physicians, hospitals, and other practitioners that could provide quality care to TRICARE

beneficiaries (TMA, 2008a). The TOP contractor was responsible for determining the

ratio of medical providers and specialties based on the population of TRICARE Standard

beneficiaries (TMA, 2008a). The TOP contractor selected a limited number of providers

who met the credentialing and licensing criteria imposed by Philippine law and U.S. law.

The selection criteria excluded providers who were on prepayment review (TMA,

2008a).

The TRICARE medical benefit under the Philippine Demonstration did not

change from the original TRICARE medical benefit (TMA, 2008a). The TOP contractor

is responsible for utilization management, case management, and quality management for

beneficiaries receiving care under the Philippine Demonstration (TMA, 2008a). The TOP

contractor is also responsible for functioning as health care finder to assist beneficiaries

in locating approved Demonstration providers (TMA, 2008a). The TRICARE Standard

beneficiary is still responsible for obtaining preauthorizations for specific care as outlined

in the TRICARE Policy Manual (TMA, 2008a). Although access to care standards is not

part of the TRICARE Standard option, the approved Demonstration providers are

required to meet TRICARE access standards for appointments. Beneficiaries seeking

appointments for urgent care should be seen the same day, for routine care within 7 days,

and for specialty care within 30 days (TMA, 2008a).

The reimbursement method changed for providers under the Philippine

Demonstration. The change in reimbursement required providers to accept the

government-directed foreign-fee schedule as payment in full to eliminate balance billing

for TRICARE beneficiaries (TMA, 2008a). The approved Demonstration provider can

request the TRICARE Standard beneficiary to pay their cost share and annual deductible

at the time the medical services are provided. Beneficiaries who receive care in

designated Demonstration areas must receive care from approved Demonstration

providers or obtain a waiver to see a certified provider to avoid paying 100% of the cost

for covered services (TMA, 2008a).

Implications for Positive Social Change

The Philippine Demonstration would have a major influence on health care

delivery for TRICARE Standard beneficiaries residing in the Philippines and providers

rendering services to those beneficiaries. The blending of the MHS and the Philippines

health care system, two very distinct health care systems, would create positive social

change.

The use of the foreign fee schedule reimbursement methodology using the

purchasing power parity conversion factor allows the DoD to purchase the same amount

and type of health care services in the Philippines of equal value. Before the Philippine

Demonstration, the DoD reimbursed medical care based on the foreign-fee schedule, but

the providers were not required to accept the foreign-fee schedule as payment in full. To

be an approved provider under the Philippine Demonstration, the provider must agree to

accept the TRICARE maximum allowable charge based on the foreign fee schedule as

payment in full (TRICARE Operations Manual, 2008a). The financial burden would be

decreased for the TRICARE Standard beneficiary because they would be liable only for

the annual deductible, individual or family, and cost shares and balance billing would be

eliminated under the Philippine Demonstration. In some instances, the beneficiary would

not be burdened with having to make upfront payments before hospital admissions or

high-cost outpatient services.

There should be a significant reduction in unusual claims activity. A provider who

desires to become an approved Demonstration provider must not be on prepayment

review, agree to accept the foreign-fee schedule, agree to file the claim on behalf of the

beneficiary, and agree to list their name in the approved Demonstration provider database

(TRICARE Operations Manual, 2008a). Implementing a PPO model with approved

demonstration providers limits the number of providers so that an appropriate patient to

provider ratio can be obtained using a beneficiary-sizing model. When the pool of

providers is small, the TOP contractor can monitor the billing activities of the providers

more efficiently, decreasing or eliminating fraudulent activities.

If the government determines that it met the expected objectives, there is a

possibility that this alternative method of delivering and reimbursing health care may

become the health care model in other countries for all TRICARE Standard beneficiaries

where accessing safe, quality care is problematic, and the government reimburses billed

charges to health care providers.

Summary and Transition

This chapter presented the major constructs of the study, the RQs, and gaps in the

current literature. Chapter 2 poses major themes from the relevant literature, the

theoretical base of the research, and the methodology for statistical analysis. Chapter 3

provides information regarding the sample, data, independent and dependent variables,

and methodology selected to explore relationships between and within variables. Chapter

4 presents the statistical analysis of data and key findings. Chapter 5 concludes with a

discussion of the implications of the findings for access to high-quality care in the

Philippines while containing health care costs.

Chapter 2: Literature Review

Introduction

The purpose of this literature review is to assess articles on access, quality care,

and cost containment mechanisms incorporating the health system for one country

(United States) with another country’s health system (Philippines) to determine if an

alternative method for providing medical care would be efficient and acceptable.

Search Strategy

A comprehensive search strategy was used to locate relevant articles related to

access to health care, health care financing, and quality of care based on the experience of

patients. Multiple databases were used to locate published studies that were conducted in

the literature on health services, health care fraud, quality, political science, business, and

economics for the years 1966–2015. Databases included Academic Premier, CINAHL

Plus, Emerald, OVID, ProQuest, PubMed, SAGE, ScienceDirect, Thoreau, and Google

Scholar. Keywords used included access to care, quality of care, patient experience,

health care fraud, managed care, CHAMPUS, TRICARE, health care reimbursement,

Philippine health care, and patient satisfaction. Any recent research that used patient

experience, quality of care, and access to care theories to explain patient satisfaction were

included. Only articles that were peer reviewed and appeared in scholarly journals were

included.

Theoretical Frameworks: Access to Health Services and Quality Health Care

Andersen’s behavioral model for health care laid the foundation for assessing

health care access and utilization. The initial framework, behavior families’ use health

services, has undergone five modifications, wherein scholars and Andersen have made

modifications to fit the assessment of health care. During this period, cost containment

was not an issue. Rather, the model focused on situations that enabled or obstructed the

utilization of health care (Andersen, 1995). The purpose of this theoretical framework is

to evaluate the effectiveness of the Philippine Demonstration’s impact on health care

delivery access using a PPO.

Andersen’s Behavior Families’ Use of Health Services

The behavior families’ use health services model was developed in the 1960s to

(a) determine the reason a family unit would access medical care, (b) construct a

definition for equitable access as well as determine the variables to measure equitable

access, and (c) incorporate the policy aspect for equitable access to care (Andersen, 1974,

1995, 2008). Andersen (1974, 1995, 2008) posited that a family’s demographic and

socioeconomic factors and health policy could elucidate the use of medical care.

Furthermore, Andersen determined that the family unit’s head of household gender, race,

ethnicity, education, and beliefs about health illuminate the use of seeking care (Aday &

Awe, 1997; Andersen, 1974, 1995, 2008). Health care policy, insurance, income, and

availability of facilities either enable or obstruct the use of care by the family (Andersen,

1974, 2008). The family unit’s environment dictates the need for medical care, for

example, the location where the family lives and their health status (Aday & Awe, 1997;

Andersen, 1974, 1995, 2008). Since the conception of Andersen’s original model, several

iterations have taken place to build upon the behavioral model improving its usefulness.

Phase 2. Andersen and Aday incorporated a health care system because of the

significance of national health policy and resources and organizational structure for

accessing care, and consumer satisfaction as an outcome measure for using the health

system (Aday & Andersen, 1974; Aday & Awe, 1997; Andersen, 1995, 2008). Aday and

Andersen (1974) acknowledged that the use of medical care falls within the scope of

politics for making changes to the health care system, and according to Aday and Awe

(1997), Andersen’s model was the first national survey assessing access to care in the

United States.

Phase 3. Andersen and Newman (1973/2005) integrated societal determinants,

health systems services, individual determinants, and health services utilization in

Andersen’s behavioral model (see also Aday & Awe, 1997). Andersen and Newman

(1973/2005) posited that individual determinants are affected by societal determinants

(technology and norms) and the health services system (resources and organization). The

societal determinants of this model posit that advancement in technology as well as

societal norms and values have affected the use of health services (Andersen & Newman,

1973/2005). For example, the shift in surgical technology has evolved in such a way that

it changed the length of hospital stays patients usually experience for complex surgeries.

Advances in therapeutic medication and the location where patients receive treatment, as

shown in the reduction of mortalities because of infectious disease, and the reduction of

conditions requiring inpatient treatment treated on an outpatient basis are additional

examples of the effects of advancements in technology. The societal norms dictate the

financing of health services, and societal values play a role in financing some health

services (Andersen & Newman, 1973/2005).

Andersen and Newman’s (1973/2005) health services system’s components

included resource and organization. The resource element included the dedicated finances

for providing health care and some resources to provide the appropriate provider to

patient ratio. The organization element examined the requirements for individuals to

access medical care and the identification and resolution of barriers interfering with

individuals accessing medical care. The cost of care in the form of out-of-pocket

expenses by individuals, wait time for some specialty treatments, and the different

qualifying conditions (benefits) can delay patients accessing needed medical care.

Andersen and Newman believed that access to health care would increase if patient out-

of-pocket expenses decreased because of public and private insurance, decreased wait

times for medical care, and an increase in the number of conditions (benefits). The

structure element under the organization component is the most complicated element

(Andersen & Newman, 1973/2005). The structure element links to other elements for

accessing care. The process for treating patients after entering the health system includes

evaluating the type of provider first administering care, required ancillary services,

referrals for specialty care, and necessary hospital admissions.

The individual predisposing factor includes sociodemographic factors, whereas

structure (age, gender, past illnesses), social structure (education, cultural factors,

religion), and beliefs (values regarding health, attitudes toward obtaining medical care,

and knowledge of different diseases) dictate the propensity for accessing care (Andersen

& Newman, 1973/2005). The individual enabling factor includes family (income, health

insurance coverage, accessibility of care) and community (provider [individual and

facility] to patient ratio, the cost of care, and place of care) components that might affect

utilization of medical care (Andersen & Newman, 1973/2005). The illness level

perceived (disability days away from work or school, symptoms of the individual, health

status of fair, good, or excellent reported by the individual) and evaluated (diagnoses of

illness caused by symptoms) determined the need for medical care (Andersen &

Newman, 1973/2005). According to Andersen and Newman, individual determinants,

predisposing, enabling, and illness level, are important in determining utilization patterns.

The illness level component (perceived and evaluated) under the individual determinant

ranked the highest for the relative importance of use (Andersen & Newman, 1973/2005).

Andersen and Newman (1973/2005) emphasized the health services use

component. The significant factors under the health services use component are type,

purpose, and unit of analysis. The type of health services use is affected by societal

determinants for various providers (individual, facility, and dental) of care. The purpose

of care is broken out into categories of primary care (prevention of illness), secondary

care (treatment of illness), tertiary care (stabilization of chronic conditions), and custodial

care (provides personal needs only). The unit of analysis accounts for contact (initial

contact or number of services during a period of time), volume (number of time patient

accesses medical care and type of provider), and episodic care (specific diagnoses) for the

use of health services (Andersen & Newman, 1973/2005).

Andersen (1995) included consumer satisfaction with health status based on the

perspectives of the population and providers as a measurement of health outcomes.

According to Andersen, effective access had been achieved when the population’s health

status or satisfaction with the health care provided improved. Additionally, efficient

access is based on health status and increased satisfaction about the consumption of

health care measures relevant to health policy and reform (Andersen, 1995).

Phase 4. The fourth iteration of the behavioral model of health services use was

modified to show several factors for use of health services and health status (Andersen,

1995). This model indicates that future predisposing, enabling, and need characteristics

have the propensity to be affected by outcomes (Andersen, 1995).

Phase 5. Modernization of Andersen’s current model, the behavioral model of

health services use, incorporated the reasons for the population using medical care

through contextual characteristics and individual characteristics. The components of this

model’s framework consist of contextual characteristics, individual characteristics, health

behaviors, and outcomes (Andersen, Davidson, & Baumeister, 2010). This model

examined contextual characteristics to determine (a) the illnesses that caused the

community to access care, (b) the factors that enabled or obstructed the community’s

access to care, and (c) the circumstances that were noticeable by the patient or provider

that medical care was needed.

Contextual characteristics. The contextual predisposing characteristics

identified the conditions under which care was accessed as demographics, social status,

and beliefs (Andersen et al., 2010). The contextual enabling characteristics are health

policies, financial resources, and organization structure in the community. Health policies

at the federal, state, and local levels can affect access to health services. The impact of

federal health policies and international health policies affected access to health services

and the quality of health care for TRICARE Standard beneficiaries in the Philippines.

Financial resources available for purchasing health care, cost of care, method of

reimbursing providers, and health insurance coverage were measured. The structure of

the organization consisted of the number of facilities and providers, a mix of providers to

include the types of services in the community, use management, quality assessment, and

mechanisms for educating purchasers and providers of health care (Andersen et al.,

2010). The contextual need characteristics included measurements of the population

health indicators that are linked to the community but not necessarily linked to the

environment (Andersen et al., 2010). The rate of mortality, morbidity, and disability

conditions are measurements of population health.

Individual characteristics. The individual predisposing characteristics are

demographic, social status, and health beliefs (Andersen et al., 2010). The individual’s

age and gender play a part in the propensity for accessing health services. The social

status of the individual, such as level of education, cultural background, and type of

work, can enable or obstruct individuals’ access to health services. The values, beliefs,

and experiences of individuals with health services can affect their future health services

use.

The mutability of enabling determinants for the family and community factors is

high, whereas the mutability of predisposing determinants for sociodemographics is low,

and beliefs are medium (Andersen & Newman, 1973/2005). The distribution of health

services is affected by specific factors that can undergo change. Decision makers, health

care administrators, and political leaders can use health care policy to improve access,

use, and quality.

Andersen’s (1995) behavioral model of health services is a prominent model used

by researchers to assess access to care (Ricketts & Goldsmith, 2005). Andersen’s

behavioral model, initially conceptualized in 1960, focused on policy development for

access to care, reasons associated with families’ consumption of medical care, and on

clarifying and evaluating access to care. Throughout the years, Andersen’s behavioral

model underwent several iterations. Other scholars perfected the second phase of

Andersen’s model in the 1970s, introducing the health care system and national health

policy and consumer satisfaction and use as important indicators for measuring access to

care. Andersen perfected the third phase of the behavioral model, focusing on

individuals, measures for effective access as shown by improved health status and

improved beneficiary satisfaction, and efficient access based on health status and

increased satisfaction in relation to the consumption of health care measures, which are

important for health policy and reform. Karikari-Martin (2011) asserted there is no

consistency between the different theoretical frameworks for measuring access. Karikari-

Martin examined Penchansky’s model, which measured availability, accessibility,

accommodation, affordability, and acceptability, whereas the Institute of Medicine model

measured access impediments, use of services, and use of mediators, and the behavioral

model of health measured the influence of community-level and individual-level

characteristics on access. Andersen’s behavioral model of health services was best suited

for this research.

Donabedian’s Quality of Health Care

Donabedian (1980, 2003) pointed out that before assessing the quality of care,

there must be a shared understanding of what constitutes quality. Donabedian found

defining quality in health care to be challenging but necessary. Donabedian divided

quality in health care into two distinct elements: (a) technical management, which is

science and technology, and (b) interpersonal management, which is the application of

science and technology. Donabedian (1980) added amenities as the third element of

quality in health care, which could be an aspect connected to interpersonal management.

The innovation of science and technology coupled with health professionals using the

knowledge derived from science and technology to take care of patients produces

attributes of quality in health care.

The attributes of quality in health care are efficacy, effectiveness, efficiency,

optimality, acceptability, legitimacy, and equity (Donabedian 1980, 2003). Quality in

health care can be defined “as the degree to which health services for individuals and

populations increase the likelihood of desired health outcomes and are consistent with

current professional knowledge” (Institute of Medicine, 1990, p. 44).

Structure, process, and outcome. Donabedian explained health care quality

through the concept of structure, process, and outcome. Donabedian’s (1980, 2003)

quality theory uses structure, process, and outcome to assess the quality of care.

Donabedian (2003) made it clear that “structure, process, and outcome are not attributes

of quality” but are three approaches used to presume “whether quality is good or not” (p.

47). According to Donabedian (1980, 1988, 2003), a functional relationship must exist

between structure, process, and outcome to make any presumptions regarding quality.

Structure depicted a place of health care delivery, human resources,

reimbursement methodology, and instrumentality to perform health services. Process

depicted the actions of patients to access health care and the actions of health

professionals to provide health care. Outcome depicted “the effects of care on the health

status of patients and populations” (Donabedian, 1988, p. 1745). Donabedian’s (1980,

1988, 2003) quality model shows the importance of the existence of a functional

relationship between structure and process as well as process and outcome to evaluate the

quality of health care.

Health Care in the United States

The U.S. health care system is a complex, mixed system with public- and private-

sector health schemes. The public sector scheme is composed of health systems that are

managed by the government. The Bismarck model, a social insurance system, created in

1883, and the Beveridge model, a universal system, formed in 1946, are the basis for

Medicare, a federal program established in 1965 (Kovner & Knickman, 2011).

Employees and employers contribute to Medicare. Medicare was created to provide

medical care to individuals who turned 65 and individuals with disabilities regardless of

age. The Balanced Budget Act, amended in 1967, created Medicaid and the Children’s

Health Insurance Program, extending access to care for children whose parents did not

have the means to buy care in the private sector (Longest, 2006). The federal and state

governments fund Medicaid. The U.S. Veterans Health Administration and DoD health

systems are based on the Beveridge model and are funded by taxes, copayments, and cost

shares from eligible beneficiaries.

Health expenditure 2011. The United States is classified as a high-income

country with a population of 316,094,000 (World Population Statistics, 2013).

Expenditures for medical care in the United States are constantly rising in the public and

private sectors. In 2010, the United States paid $2.7 trillion for health care, and the total

national health expenditure capped at 17.9% of the gross domestic product from 2009

through 2011, with a spending growth rate of 3.9% (Centers for Medicare and Medicaid

Services, 2013). The cost of health care will outgrow the gross domestic product, and if

not contained, health care costs will use up a significant portion of the gross domestic

product; therefore, it is prudent to take measures to contain health care costs. The

contributors to financing health care for 2011 were households at 28%, the federal

government at 28%, private business at 21%, state and local government at 17%, and

other private revenue at 7% (Centers for Medicare and Medicaid Services, 2013).

Regulating health care. Regulatory practices in health care are geared toward

access to care, quality care, and cost containment (Field, 2007). The Hill–Burton Act of

1946 allowed the construction of hospitals and expansion of hospitals to reduce the

geographical impediments that created barriers to accessing care (Field, 2007). In 1956,

CHAMPUS was established under the Dependent Medical Care Act to provide care to

service members’ families (Longest, 2006). The Comprehensive Health Planning and

Services Act of 1966 was amended in 1974, establishing national Health Systems

Agencies to regulate building or expansion of medical facilities at the state level. This act

mandated states to start a certificate-of-need program in which hospitals identified

medical and technological requirements in their health plans (Field, 2007). The Balanced

Budget Act of 1997 established many requirements under Medicare, such as revising the

reimbursement system, reducing payments for inpatient and outpatient services, as well

as expanding care to children for families who could not afford to purchase medical care

(Field, 2007; Longest, 2006).

The Health Maintenance Organization Act of 1973 mandated that health

maintenance organization options be included in the selection of health insurance

provided by employers (Field, 2007). Established in 1951, the Joint Commission on

Accreditation of Healthcare Organization, accredits hospitals, ensuring hospitals comply

with established quality stands (Field, 2007). In 1990, the National Committee for

Quality Assurance was created to provide oversight of quality for managed care

organizations in the private sector.

The Federal Anti-Kickback laws were implemented in 1972 to deter individuals

from receiving financial inducements for referrals or purchasing of goods and services

reimbursed under a federal health care program (Matyas & Valiant, 2006). Title 32 Code

of Federal Regulations Section 199 regulates CHAMPUS/TRICARE programs,

providing guidance on benefits, claims processing, accreditation, and fraud, among other

aspects of purchasing direct care (Matyas & Valiant, 2006; TMA, 2008a). The Federal

Appropriations Act of 1993 stipulated that CHAMPUS beneficiaries cannot be billed

115% over the CHAMPUS allowed charge (TMA, 2008c). The Balanced Budget Act of

1997 afforded the same balance billing protection for Medicare beneficiaries (Field,

2007). The balance billing protection law only applies in the United States.

Health Care in the Philippines

The Republic of the Philippines is a low middle-income country with a population

of 98,007,000 as of August 2013 (Population Reference Bureau, 2014). The Philippines

health system is a decentralized National Health Plan model in which the Philippine

Health Insurance Corporation (PhilHealth) administers the public sector health benefit

and collects the payments, whereas the Department of Health, Philippines oversees health

care policies (Boslaugh, 2013). Many Philippine citizens do not understand the benefits

administered by PhilHealth (Boslaugh, 2013).

Health expenditure 2011. In 2011, the Philippines’ total health expenditure was

PhP431 billion (National Statistical Coordination Board, 2013). Funded by various

sources, the amount of funding from each contributor were PhP116,433 billion from the

government, PhP39,126 billion from social insurance, PhP272,009 billion from private

sources, and PhP3,478 billion from grants (National Statistical Coordination Board,

2013). Therefore, private sources spent 63.1% on health care; the government spent 27%

on health care, social insurance contributed 9.1% to the cost of health care, and grants

contributed 0.8% to the cost of health care (National Statistical Coordination Board,

2013).

Selected Republic Acts. The Philippines health system underwent major changes

to adopt regulations to provide access to quality medical care by implementing various

Republic Acts (Republic of the Philippines Department of Health [PDH], n.d.). The

Hospital Licensure Act of 1965 was implemented to ensure that every hospital has a

license to operate (PDH, n.d.). The Medicare Act of 1969 was superseded by the National

Health Insurance Act of 1995, expanding access to care to citizens of the Philippines,

specifically, the impoverished (PDH, n.d.). The Pharmacy Act of 1987 governs the sale

of medicines in the Philippines (PDH, n.d.). The purpose of the 1999 Health Sector

Reform Agenda was to improve the facets of the Philippine health system at different

levels (PDH, n.d.). The Universally Accessible Cheaper and Quality Medicines Act,

implemented in 2008, keeps quality medicines at an affordable price for everyone (PDH,

n.d.).

Cost Containment

Several mechanisms have been implemented to reduce health care expenditures.

These mechanisms to contain rising costs in health care differ between the public sector

and private sector. In the public sector, under the Medicare program, the prospective

payment system (PPS) and the resource-based relative value system (RBRVS) are used to

control costs. The PPS controls rising costs for hospitals by establishing reimbursement

rates based on the patient’s diagnosis, whereas the RBRVS controls physicians’

reimbursement rates based on the level of expertise needed to treat a condition. Under the

Medicaid program, states have opted to determine the benefits provided to beneficiaries

who are eligible for the program as well as regulate reimbursement rates to providers. As

a cost containment measure, managed care was implemented to contain rising medical

costs (Mohaghegh, 2007; Rodwin, 2010; Weiner, Famadas, Waters, & Gikic, 2008;

Zwanziger, Hart, Kravitz, & Sloss, 2001).

Managed care organizations. Cost containment, access to care, and quality care

are concerns in the health industry (Mohaghegh, 2007). The development of managed

care was originally intended to control the rising cost of health care in the United States

(Mohaghegh, 2007; Rodwin, 2010; Weiner et al., 2008; Zwanziger et al., 2001).

However, Koop (1996) clarified that access to care was the catalyst for managed care, not

cost containment. Koop believed that physicians wanted the latitude to provide care to

their patients without concern if the patients could afford to pay for the medical services.

After the implementation of managed care, the focus was placed on medical practices and

disease management that produced better outcomes and utilization management that

eliminated medical procedures the patients did not need (Koop, 1996). In essence, cost

containment was an outcome of managed care.

The rising cost of health care, access to health services, and quality care were not

only concerns in the United States but also concerns in other countries (Weiner et al.,

2008). According to Schoen et al. (2010), the design of the insurance plan influences

access to medical care and costs in the United States and internationally. Vargas,

Vázquez, Mogollón-Pérez, and Unger (2010) also found that insurance design impacts

access to medical care. Patients were required to pay user fees, creating barriers to care,

affecting access to much-needed medical care. Other countries were evaluating ways to

contain their health care costs and provide oversight of medical services, so the concept

of managed care in the United States was exported to other countries (Rodwin, 2010;

Weiner et al., 2008). Some countries had difficulty with managed care, for example, the

first time the Philippines tried to implement managed care, managed care techniques

were not successful until later.

Schoen et al. (2010) conducted a computer-assisted telephone interview survey

March through June 2010 by the Commonwealth Fund to explore the differences in

healthcare access, cost, and care experiences based on income with different cost sharing

provisions in 11 countries including the United States. Schoen et al. found that providers

were reimbursed the same regardless of the beneficiary’s income in all countries except

the United States. In fact, Schoen et al. found that insured adults in the United States in

comparison to the other countries were more than likely to forgo medical care because of

cost, financial difficulty paying for medical care, problems with their insurance company,

and although insured, still have medical expenses.

Rodwin (2010) noted physicians provided a high amount of medical services

under the fee-for-service (FFS) payment system; however, capitation was implemented to

pay for each patient, which ended entrepreneurship and fee-for-payment conflicts.

Scholars believed that health care reimbursement differed based on the country; whereas

some countries use a FFS system, capitation, and a PPS (Rodwin, 2010; Weiner et al.,

2008), Kwon (2011) found that some Asian countries with universal health care use an

FFS reimbursement system effectively for reimbursing providers, and other Asian

countries that use FFS reimbursement systems should use a different reimbursement

system for effectiveness.

The government may finance the health care system as seen in New Zealand, but

in Singapore, the bulk of health care expenditure is privatized (Weiner et al., 2008).

Wagner et al. (2011) recommended passing public policy in which individuals could have

access to medical care without jeopardizing their quality of life. In various countries,

individuals have been forced to choose between receiving medical care and paying for

household goods, such as rent or food (Wagner et al., 2011).

Understanding the difference in managed care organizations (MCO) is important,

especially when medical care is delivered through an MCO instead of the traditional

insurance plan (Wagner & Kongstvedt, 2007). Each type of MCO has specific rules

beneficiaries must follow.

Health maintenance organizations. Beneficiaries must enroll in HMOs,

whereby they will be assigned a primary care manager who will coordinate their medical

care. HMOs could be reimbursed using a sliding scale, fixed fee schedule, or diagnosis-

related groups (Wagner & Kongstvedt, 2007). HMOs must ensure beneficiaries have

access to high-quality care. Cost containment mechanisms, such as utilization

management, case management, and disease management, are part of the services

provided in an HMO plan. There are staff model HMOs, group model HMOs, and

independent practice model HMOs with varying differences.

Preferred provider organizations. PPOs contract with providers to provide

medical services to beneficiaries. PPOs have a negotiated rate that includes a fixed fee

schedule, all-inclusive per diem rates, or diagnosis-related groups (Wagner &

Kongstvedt, 2007). The size of the PPO may be limited. Beneficiaries covered under

PPOs can use non-PPO providers, but coinsurance rates, such as copayments, will be

higher, as will deductibles (Wagner & Kongstvedt, 2007). As a cost containment

mechanism, utilization management programs are in place to contain cost.

Point of service. Wagner and Kongstvedt (2007) thought of the POS model as a

hybrid of the HMO and PPO models. Under the PPO model, the provider would be

reimbursed using a capitation method or performance-based method. The HMO model

would allow the beneficiary to seek care outside the HMO network, but the beneficiary

would be responsible for a higher deductible and coinsurance.

MCOs utilized cost constraint mechanisms, such as disease management, case

management, utilization management, and financial incentives, to contain costs (Granger,

Boyer, Weiss, Linton, & Williams, 2010; Mohaghegh, 2007).

Heath Care Fraud

Fraud contributes to the escalation of health costs. Sutherland (1940) described

fraud as “white-collar criminality” that is present in various businesses and different

professions. Health care is a lucrative industry making health systems around the world

appetizing to white-collar criminals. Savedoff and Hussmann (2005) estimated that $3

trillion or more is spent on health care globally, which makes the health care industry

more susceptible to fraud and abuse.

The Philippines is known for the widespread health care fraud and abuse activities

that contributed to the increased health care expenditures. In response to the fraudulent

and abusive activities, the DHA implemented fraud and cost control initiatives in the

hope of controlling health care costs and eliminating fraudulent billing practices in the

Philippines. These initiatives included the following:

1. Provider validation and certification: In the Philippines, the DHA identified a

significant number of claims submitted by providers who did not exist. The

DHA implemented a requirement that DHA must certify all Philippine

providers who serve TRICARE beneficiaries. The TOP contractor is

responsible for verifying that providers meet the Philippine government’s and

its licensing board’s established standards for credentialing and licensing. The

providers must also meet some of the U.S. standards. The certification process

requires on-site verification of the providers’ offices and inpatient facilities to

ensure they exist and are capable of providing the specified level of care

billed.

2. National drug coding (NDC): DHA required providers who exceeded the

$3,000 limit for pharmacy services in a year to submit pharmacy charges with

the NDC.

3. Prepayment review: As one of the most effective antifraud controls, the TOP

contractors reviewed claims before paying claims.

4. Government-directed foreign fee schedule: In November 2008, DHA

implemented the use of the purchasing power parity indexed fee schedule in

the Philippines. Using the country-specific index for the Philippines reflects

the actual medical costs for services rendered. The foreign-fee schedule

controls costs and reduces fraud and abuse.

5. Education: The identification of inappropriate behaviors for providers and

beneficiaries who submit claims results in educational letters sent to the

providers and beneficiaries explaining the inappropriate behaviors and the

consequences of the behavior.

6. Data mining tools: Data mining provides TRICARE with the technology to

transform complex and voluminous health care transactions into useful

information to detect fraud and abuse.

7. Payments and explanation of benefits mailed to providers versus billing

agents: In the Philippines, third-party agencies (TPA) submitted inflated

claims to TRICARE unbeknownst to the providers as the payments went

directly to the TPAs. As a result, TRICARE only allows for payments and

explanations of benefits mailed to the providers at the location identified on

the claims, not to the TPAs.

8. Supporting documentation: In November 2008, the DHA allowed providers

and beneficiaries to fax claims and supporting documentation to the TOP

claims processor.

9. TRICARE certified pharmacies: Providers filled prescription medications in

their offices and billed TRICARE for an office visit. DHA implemented

certification requirements for pharmacies in the Philippines, whereas

prescription medications from certified retail or hospital-based pharmacies

may be cost shared.

10. Proof of payment: In September 2012, DHA implemented a proof of payment

requirement for all health care received overseas and paid for by the

beneficiary (Carey, 2011; Scott, 2006; TMA, 2008a).

Department of Defense Health System

The DoD has the oldest, largest medical system, dating back to 1799, when

Congress enacted legislation to care for disabled seamen and, in 1884, extending the

health care entitlement to family members at no cost (Granger et al., 2010). Active-duty

family members could obtain care in the civilian sector after the Emergency Maternal and

Infant Care Program was passed in 1943 and the Dependents Medical Care Act was

passed in 1956, giving DoD the authority to enter into contracts with nonmilitary

providers (Barton, 2007; Coppola, Harrison, Kerr, & Erckenbrack, 2007; Granger et al.,

2010). Retirees and their family members could obtain medical care in the civilian sector

after the Military Medical Benefits Amendment was passed in 1966 (Granger et al.,

2010). As a result, in 1967, CHAMPUS was formed providing a mechanism for non-

active-duty beneficiaries to receive care in the civilian sector (Granger et al., 2010).

Approximately 500 military installations closed between 1988 and 2005 under the

Base Realignment and Closures Act, decreasing access to care in military treatment

facilities (MTF) for retirees and their family members as an effort to decrease rising

medical expenditures in the MTF (Coppola et al., 2007; Granger et al., 2010). The need

to purchase medical care in the civilian sector doubled health care expenditures, leaving

DoD to examine other ways to deliver medical care to this population. Medical care

expenditures were still out of control, so DoD conducted the CHAMPUS Reform

Initiative (CRI) from 1988 to 1993, to provide medical care using a managed care model

that mirrored managed care in the private sector. The CRI managed care model offered

three optional plans for beneficiaries: the CHAMPUS Prime option, which resembled a

health maintenance organization; CHAMPUS Standard, the original basic FFS option;

and CHAMPUS Extra, a hybrid of the private sector preferred provider option (Granger

et al., 2010; Zwanziger et al., 2001). This initiative was the beginning of the

transformation of CHAMPUS to TRICARE.

Military Health System

The MHS is responsible for providing health care to 9.7 million service members,

retirees, eligible family members, National Guard and Reserve members, and former

spouses all over the world (TMA, 2012). The MHS is a combination of direct care, care

received in MTFs, and purchased care, or care received in the civilian sector (Granger et

al., 2010; TMA, 2012).

The MHS mission is “to enhance the Department of Defense and our nation’s

security by providing health support for the full range of military operations and

sustaining the health of all those entrusted to our care” (TRICARE, 2012, para. 3). The

vision of the MHS is “a world-class health care system that supports the military mission

by fostering, protecting, sustaining, and restoring health” (para. 4).

TRICARE Health Plan Options

TRICARE Prime is DoD’s health maintenance organization–like option.

TRICARE beneficiaries must take action to utilize this option. Beneficiaries desiring to

use TRICARE Prime must enroll by completing an enrollment application and selecting a

primary care manager. Through enrollment in TRICARE Prime, beneficiaries will

receive medical care in MTFs or the private sector (TMA, 2009). Depending on the

beneficiary category, there will be no cost or reduced out-of-pocket cost (TMA, 2008c).

Active-duty family members enrolled in Prime are not subjected to an enrollment fee and

copayments, whereas retirees and their family members enrolled in TRICARE Prime will

be subjected to an enrollment fee and copayments (TMA, 2008c). Copayments are only

applicable when retirees and their family members receive care in the private sector

(TMA, 2008c). Beneficiaries enrolled in TRICARE Prime must follow the rules of

TRICARE Prime, for example, a beneficiary cannot self-refer to a specialist without

obtaining a referral from the assigned PCM and authorization provided by the contractor

for the region in which the beneficiary is enrolled. If beneficiaries receive care without

the proper referral and authorizations, they are encouraged to use their point of service

option. The TRICARE Prime option is not available worldwide (TMA, 2009).

TRICARE Standard is the basic FFS option in which CHAMPUS laid the

foundation for the benefits currently provided with enhancements (Granger et al., 2010;

TMA, 2009). The TRICARE Standard option allows the most freedom to select providers

but costs more money. The beneficiaries will be responsible for deductible and cost

shares based on beneficiary category (TMA, 2008c). There is no enrollment, and

generally, referrals and authorizations are not required with a few exceptions, for

example, inpatient treatment, skilled nursing care (TMA, 2008a, 2008b). Beneficiaries

must receive care from TRICARE-authorized providers for TRICARE to reimburse on

their claims. According to the balance billing law, providers cannot balance bill

TRICARE Standard beneficiaries more than 15% above the TRICARE maximum

allowable amount; the balance billing limitation is only applicable in the United States

(TMA, 2008c). TRICARE Standard is available all over the world.

TRICARE Extra is the preferred provider option in which TRICARE

beneficiaries use providers in the TRICARE network. Beneficiaries do not enroll in this

option; they are TRICARE Standard beneficiaries opting to see network providers on a

case-by-case basis in which they receive a 5 percent discount (TMA, 2008b, 2008c,

2009). Providers cannot balance bill the beneficiaries because they are network providers

who have signed contracts agreeing to accept what TRICARE allows in full (TMA,

2008b, 2008c). TRICARE Extra is only available in the United States (TMA, 2008b).

Evolution of the Military Health System

According to Granger et al. (2010), the DoD MHS is the oldest and largest health

care system, dating back to 1799, when Congress enacted legislation to care for disabled

seamen. In 1884, Congress extended the health care entitlement to family members at no

cost. The Emergency Maternal and Infant Care Program was passed in 1943, giving

active-duty family members the opportunity to obtain care in the civilian sector (Granger

et al., 2010). After the Dependents Medical Care Act was passed in 1956, DoD had the

authority to enter into contracts with nonmilitary providers (Granger et al., 2010). Active-

duty family members were given priority over retirees and their family members in

MTFs, resulting in a shortfall of appointments. As a result, the Military Medical Benefits

Amendment had passed in 1966, allowing retirees and their family members to obtain

medical care in the civilian sector (Granger et al., 2010).

Military Health Care Demonstration: CHAMPUS Reform Initiative

The health system for the military faced the same problems as other health care

systems throughout the United States and internationally. Thus the evolution for

CHAMPUS to incorporate TRICARE as an enhancement to the MHS started in 1988

with the CRI. The CRI was designed to decrease the costs of purchased care and direct

care, reduce the utilization of health care in the private sector, improve coordination of

care between the MTFs and civilian providers, and increase patient satisfaction

(Zwanziger et al., 2001). The DoD conducted the CRI 5-year Demonstration Project in

Hawaii and California (Zwanziger et al., 2001).

The sample frame consisted of randomly selected active-duty members and

retirees, stratified (Zwanziger et al., 2001). Eligible beneficiaries had the option to select

CHAMPUS Prime, the HMO-like option, the CHAMPUS option, the original FFS basic

program, or CHAMPUS Extra (Zwanziger et al., 2001). Secondary data, data from two

CHAMPUS beneficiary surveys and claims data for beneficiaries surveyed, were used to

evaluate cost, utilization, and patient satisfaction (Zwanziger et al., 2001). A multivariate

analysis was employed to evaluate overall cost, and multivariate regression models were

used to evaluate continuous (linear) and categorical (logistic) variables (Zwanziger et al.,

2001).

It was concluded that the CRI Demonstration was successful; although the cost of

health care increased, beneficiary satisfaction and access to care in the MTF also

increased (Zwanziger et al., 2001). The research could not determine if CRI impacted

quality.

Enhancements in TRICARE 1999 Through 2011

The DoD health care program provided services for approximately 8.3 million

beneficiaries in 1999 (TMA, 1999). In 1999, beneficiaries enrolled in TRICARE Prime

were automatically reenrolled; action was taken only if they did not want to reenroll

(TMA, 1999). Prime enrollees were no longer subjected to balance billing; thus, reducing

out-of-pocket expenses for enrollees. TRICARE beneficiaries no longer paid multiple

copayments for ancillary services. TRICARE aligned provider reimbursements with

Medicare reimbursement rates (TMA, 1999). During 1999, active-duty service members

geographically separated from units with MTFs were deemed TRICARE Prime Remote

so they could receive care in the private sector close to where they worked. In 1998, six

TRICARE Senior Prime Demonstration sites were launched to provide care for seniors

over 65 (TMA, 2000).

Enhancements in 2000

The DoD health care program provided services for approximately 8.2 million

beneficiaries in 2000 (TMA, 2000). Several demonstrations were established to provide

access to care for seniors who had served their country. The Federal Employee Health

Benefits Program, a 2-year Demonstration, January 2000 through December 2002,

TRICARE Senior Supplement, a 2-year Demonstration, April 2000 through December

2002, Pharmacy Redesign Pilot Program network retail and mail order pharmacy, but

beneficiary must have Medicare Part B (TMA, 2000). The Base Realignment and

Closures resulted in limited access to MTFs that affected retirees. Beneficiaries enrolled

in TRICARE Prime were assigned primary care managers by name. Case managers were

assigned to care for complex cases (TMA, 2000).

Enhancements in 2001

In 2001, the DoD health care program provided services for greater than 8 million

beneficiaries (TMA, 2001). TRICARE Prime Remote for Active Duty was extended to

the family members living with the service members stationed at geographically

separated units (TMA, 2001). The TRICARE Dental Program for family members and

the TRICARE Selected Reserve Dental Program provided enhanced benefits. TRICARE

for Life reinstated TRICARE benefits for retirees age 65 as long as they had purchased

Medicare Part B (TMA, 2001). As of 2001, active-duty family members enrolled in

Prime did not have to pay copayments (TMA, 2001). In select MTFs, active-duty service

members could receive chiropractic care. The catastrophic cap for retirees was reduced to

$3,000, the initiation of a nutritional program, Women, Infant, and Children Overseas,

was put in place, and the Pharmacy Data Transaction Service started to improve patient

safety (TMA, 2001).

Enhancements in 2007

Military medicine had to change from treating members during peacetime to

treating members during wartime. Military members are indoctrinated to fight wars, so

they must be indoctrinated back into the civilian sector once their tour of duty is over.

Military providers and civilian providers had to be equipped to treat posttraumatic stress

disorder (PTSD), traumatic brain injuries, and amputees capable of returning to work

(TMA, 2012).

Enhancements in 2011

Fast forwarding to 2011, the implementation of the TRICARE Young Adult

(TYA) benefit was in response to the Patient Protection and Affordable Care Act of 2010

that required private medical insurance companies to allow parents to provide medical

coverage for their children up to age 26 (TMA, 2012). TRICARE benefits typically

covered children up to age 21 or age 23 if in college full-time or older if disabled (TMA,

2009). The beneficiary has the option under TYA to enroll in TRICARE Prime or use

TRICARE Standard, but the beneficiary must pay a premium (TMA, 2012). The contract

for the TRICARE Dental Program was awarded to a different contractor resulting in

lower premiums for all beneficiaries and allowing those beneficiaries and survivors not

previously enrolled to enroll (TMA, 2012). Generic pharmacy drugs copayment reduced

to zero in October 2011 through the TRICARE Pharmacy Home Delivery. The

enhancements listed above are not all-inclusive.

Assessment of Patient Satisfaction

Evaluations must be conducted to determine if policy enhancements or

demonstrations achieved the expected outcome. One way to determine if the expected

outcome has been achieved is by conducting a satisfaction survey. Patient satisfaction is

an important dimension for measuring access, quality, and cost. Studies have been

conducted to evaluate patients’ satisfaction with the type of plan option selected, like the

survey assessing patient satisfaction with services provided by the primary care clinics

that participated in the primary community care network (PCCN) Demonstration Project

in Taiwan (Lin, Lin, & Lin, 2010).

This survey also evaluated patient satisfaction for patients receiving care from

non-PCCN clinics. Lin et al. (2010) found that there was no difference in patient

satisfaction among PCCN member patients and non-PCCN patients for the care quality

measures. The survey also indicated that PCCN member patients would recommend the

use of the PCCN clinics to others, unlike the nonmember patients, because of their

satisfaction with the quality of their relationship with the doctor (Lin et al., 2010). Lin et

al. also found that PCCN providers who had trusting relationships with their patients

could recruit patients to become members of the PCCN.

Lin et al. (2010) and Calnan and Rowe (2006) believed that trust relations in

health care between patients and providers were important. The survey depicted that the

responsibility for erosion of trust between patient and provider was the health care

provider. Calnan and Rowe conceded that the concept of trust correlates with patient

satisfaction.

Owusu-Frimpong et al. (2010), in their study, focused on public and private

health care users in the United Kingdom: The patients in the public health care sector

were not satisfied with the relationship or attention they received from their providers.

Friesnor, Neufelder, Raisor, and Bozman (2008) found that business processes modeled

after the concept of continuous process improvement could be implemented by health

care organizations to increase the satisfaction of patients who are already satisfied by

evaluating patient concerns other than medical, for example, assessing registration

processes and parking facilities. Owusu-Frimpong et al. (2010) used a mixed method,

qualitative (semistructured face-to-face interviews) and quantitative (cross-sectional

survey, 5-point Likert scale) methodology to evaluate patient satisfaction with access to

care. Owusu-Frimbong et al. employed an analysis of variance (ANOVA) to explore and

compare the respondents’ perceptions of access to care to their experience with access to

care. The Pearson chi-square statistics and associated p-values were used to establish the

extent of correlations between overall patient satisfaction and the patients’ demographics

(Owusu-Frimpong et al., 2010).

Price et al. (2014) conducted a review of the literature on “the associations

between patient experience measures and other indicators of health care quality” (p. 525)

to determine the logical connection between patient experiences and measures of

structure, process, and outcomes. In total, 422 articles were searched, but 34 articles,

from 1992 to 2013, were reviewed and used for this research. The articles reviewed

methodologies that evaluated the linkage between “patient-reported experiences and

processes and outcomes of care” as well as “articles reporting on CAHPS surveys” (Price

et al., 2014, p. 525).

The literature review found a link between positive patient care experiences and

patient adherence to treatment and influenced adherence to the treatment process for

clinical outcomes for inpatient settings. It was also determined that there is a correlation

between positive patient experiences and better hospital patient safety culture, and lower

unnecessary health care utilization (Price et al., 2014). Measuring the experience of

patients in patient-centeredness care environments is important as it increases

accountability and quality improvements based on reported patient experiences.

Russell, Johnson, and White (2015) conducted a case study from July 1, 2011,

through June 30, 2012, using patient satisfaction data from Portage Health in the western

Upper Peninsula of Michigan. The multifacility health care system Portage Health

comprises “a 36-bed acute care hospital, several family practices, numerous specialty

medical practices and multiple hospital-owned clinics; including a university center and

express care clinics” (Russell et al., 2015, p. 1162). Press Ganey administered a 35-

question survey using a 5-point Likert scale that evaluated five areas of quality:

1. access (a1–a7), including helpfulness, promptness, courtesy, convenience,

ease of reaching the clinic by phone, and ease of scheduling appointments

2. moving through the visit (v1–v7), including speed of registration; wait times

in waiting room, in exam room, and at clinic; information about delays; and

comfort and pleasantness of waiting room and exam room

3. nurse/assistant (n1–n2), including friendliness, courtesy, and the concern of

the nurse/assistant

4. care provider (cp1–cp10), including friendliness, courtesy, and concern of

care provider; explanations, information, instructions, and clear language;

time with patient; and patient confidence

5. personal issues (i1–i6), including cleanliness, safety, security, privacy,

sensitivity to patient needs, and pain control (Russell et al., 2014, p. 1162).

The survey was mailed to 6,824 patients after an encounter; 1, 514 patients

responded, in which 129 patients submitted surveys with 10 or more questions

unanswered, so 1,385 surveys were considered usable. Some of the remaining surveys

had missing data for demographics which were not corrected, and the other missing data

were corrected with average responses (Russell et al., 2014). Press Ganey used the

statistical package SAS/JMP to run the regression models for the overall measures of

patient satisfaction. “ANOVA F-tests and the coefficient of determination were used to

test the significance of the models” (Russell et al., 2014, p. 1166). The dependent and

independent variables, except the demographic variables, were measured using the 5-

point Likert scales. Four questions (independent variables) were included under the

Overall Assessment section of the survey: “overall cheerfulness of our practice (o1), how

well the staff worked together to care for you (o2), overall rating of care you received

during the visit (o3) and likelihood of you recommending our practice to others (o4)”

(Russell, Johnson, & White, 2015, p. 1164). The other overall assessments were

considered as dependent variables, except for “overall cheerfulness of our practice”

(Russell et al., 2015, p. 1164).

Russell et al. (2015) determined that variables for access, moving through the

visit, nurse/assistant, care provider, and personal issues did have an impact on overall

assessments of care quality. It was determined that gender and type of care provider did

not have an impact on overall patient satisfaction. It was noted that older patients, 55

years or older, responded to the mail surveys. The demographics of the nonrespondents

are unknown. Therefore, Russell et al. recommended using different methodologies for

surveys.

A study was conducted to assess access to care and medicines and household

economic burdens (Wagner et al., 2011). Wagner et al. stated, “Each year, an estimated

44 million households suffer severe financial hardship and 25 million are pushed into

poverty because they need to pay for health care” (p. 151). Individuals use their savings,

decrease the monies spent on food, and sell their assets to pay for health care, or they

may opt not to receive the needed health care (Wagner et al., 2011). This study was

conducted across 70 countries with varying degrees of income—22 low, 18 lower middle,

10 upper middle, and 20 high income—using the criteria established by the World Bank

in 2003 (Wagner et al., 2011). Better access to acute care and chronic care decreased the

possibility of catastrophic health care expenditures.

This particular study revealed that although most insurance companies covered

inpatient services and outpatient physician services, they did not cover the medications

prescribed during those episodes of care. The study also highlighted the correlation

between access and health care costs financial challenges. Assessment of patient

satisfaction will result in process improvement (Owusu-Frimpong et al., 2010), change in

public policy (Wagner et al., 2011), and other patient concerns among other

improvements (Friesner, 2009).

Alkhwaldeh et al. (2014) argued that in developing countries people were living

longer. Alkhwaldeh et al. conducted a cross-sectional study to evaluate the utilization

patterns for primary health care factors and predictors of utilization and nonutilization of

health care for older adults Jordan within the past 1,6, and 12 months. The sample frame

consisted of 190 participants age 50 and older, 79 participants in the South, 85

participants in the center, and 26 participants in the North, in the Irbid governorate of

Jordan (Alkhwaldeh et al., 2014).

Alkhwaldeh et al. (2014) used Andersen’s behavioral use model to assess the

predisposing, enabling, and need factors to determine the patterns associated with the

utilization of primary health care services. The dependent variables were primary health

care service utilization in the past 1, 6, and 12 months. The research question was “Did

you visit the primary health care center in your region during the past month? Past 6

months? Past 12 months?” (Alkhwaldeh et al., 2014, p. 2). The independent variables

were predisposing (age, gender, tobacco use, employment status, education level, and

marital status), enabling (monthly insurance and health insurance coverage), and need

(chronic illness self-reports; Alkhwaldeh et al., 2014). Alkhwaldeh et al. used the Elderly

Cognitive Assessment Questionnaire instrument to measure older adults’ cognitive

impairment.

A Likert scale 1 to 10, 1 represented the “worst I have ever felt,” and 10

represented the “best have ever felt,” was used to measure the perceived general health of

the participants (Alkhwaldeh et al., 2014, p. 2). The participants were asked, “What

number would best represent your general health today? 6months ago? 12months ago?”

(pp. 2–3).

Alkhwaldeh et al. (2014) also used the 12-item Short Form Health Survey version

2 to measure the perceived general health status for the past 1 month. The domains

identified were “physical function, role-physical, bodily pain, general health, vitality,

social functioning, role-emotional, and mental health” (p. 3) labeled as physical

component summary (PCS) and mental component summary (MCS).

The data were analyzed using descriptive statistics, chi-square test of association

for categorical variables, Pearson correlations to measure associations between

independent and dependent variables, and other correlation tests (Spearman’s rho, point

biserial 𝑟, phi coefficient (𝜙), and Cramer’s V methodology; Alkhwaldeh et al., 2014).

Alkhwaldeh et al. created binary logistic models to analyze the predictors of primary

health care service utilization in the past 1 month, past 6 months, and past 12 months.

The University of Jordan School of Nursing and the Jordan Ministry of Health

Ethical Committee granted Alkhwaldeh et al. (2014) permission to conduct this study.

Alkhwaldeh et al. informed the participants that their responses would be confidential

and their identity would remain anonymous. Alkhwaldeh et al. obtained informed verbal

consent from all those involved in the study.

The limitations disclosed were the self-reported use of health care services,

chronic illnesses, and participants’ health status, and generalization of the results for

using primary health care services in other governorates or Jordan could not be done

(Alkhwaldeh et al., 2014). Alkhwaldeh et al. had to take precaution to eliminate recall

bias.

Data were analyzed for 190 older adults in which the average age of the

participants was 64.6 (6.7 SD), and 57.4% were male and 42.6% female. Of the 190

participants, 88.4% were married, 36.8% were not formally educated, 42.1% had a

primary school education, and 21.1% had a secondary education or higher; 93.7% of

participants reported they had health insurance when 6.3% participants reported they did

not have health insurance (Alkhwaldeh et al., 2014). The income mean per month per

participant was 218.2 (88.7 SD) Jordanian dinar currency; 27.9% participants reported

they did not have a chronic illness, when 72.1% reported they did have a chronic illness,

and 96.8% participants reported they had no cognitive impairment, when 3.2%

participants reported they had cognitive impairment (Alkhwaldeh et al., 2014). The

perceived general health status in the past month PCS average was 41.28 (11.0 SD), and

the MCS average was 50.46 (7.3 SD; Alkhwaldeh et al., 2014). The perceived health

status in the past 6 months was 6.46 (1.3 SD) and in the previous 12 months was 6.64

(1.3 SD; Alkhwaldeh et al., 2014).

Alpha levels of 0.05* and 0.01** were used for this study. Older adults who used

significantly more primary health care services in the past 1 month (a) had no formal

education or had a primary school education, p = −0.220**, (b) were nonsmokers, p =

−0.162*, (c) had a chronic illness, p = .453**, (d) had perceptions of poorer health status

today, p = −0.272**, and (e) had symptoms or poor physical health, p = −0.377**

(Alkhwaldeh et al., 2014).

The factors significantly associated with increased primary health care services

utilization in the past 6 months were (a) increasing age, p = 0.229**, (b) being

unemployed or retired, p = 0.178*, (c) having no formal education or only a primary

school education, p = 0.200**, (d) being nonsmokers, p = −0.166**, (e) having a chronic

illness, p = .633**, (f) those having poor self-rated general health status today, p =

−0.355**, and (g) those having a poor self-rated general health status in the past 6

months, p = 0.409** (Alkhwaldeh et al., 2014).

The predictor variables associated with significantly greater primary health care

utilization in the past 12 months were (a) increasing age, p = 0.205**, (b) being

unemployed or retired, p = .218**, (c) having no formal education or only a primary

school education, p = −.240**, (d) being nonsmokers, p = −.145*, (e) having a chronic

illness, p = .650**, (f) those having poor self-rated general health status today, p =

−.373**, and (g) those having poor self-rated general health status in the past 12months,

p = −.306** (Alkhwaldeh et al., 2014).

Alkhwaldeh et al. (2014) concluded that although chronic illness was the most

significant predictor of primary health care services utilization, all the factors could be

linked to the use of primary health care services.

Van de Ven (2014) conducted a secondary analysis of longitudinal data from

patient experience surveys carried out in 1996 and 1997 by the Picker Institute in Boston,

Massachusetts, and data from the Healthcare Organization Survey (HOS). Van de Ven

wanted to determine the relevant attributes linking satisfaction to patient experience of

care as well as the causal connection between “dimensions of care to clinic size,

economic performance, and employee job satisfaction” (p. 131). Data collected from the

HOS in 1996 and 1997 was used to assess the attitudes of providers (Van de Ven, 2014).

The third data set collected for this study comprised data from records audited for

productivity and the second year profits for both clinics (Van de Ven, 2014). Van de Ven

examined the contextual factors (clinic size and clinic employee job satisfaction), process

factors (activities performed by healthcare providers and staff), and outcome factors

(overall patient satisfaction and indicators of clinic performance—physician productivity

and clinic profitability). The goals of this research were

1. to empirically identify and determine the relative importance of different

aspects of the patient care experience as they relate to overall satisfaction of

patients served by primary care clinics

2. to determine how the various dimensions of care experienced by patients

relate to patient characteristics, clinic size, and economic performance as well

as the attitudes of clinic physicians, nurses, and staff toward their work as

clinical care providers. (Van de Ven, 2014, p. 132).

The sampling frame for the patient experience survey consisted of 8,363 patients

in the first year and 7,733 patients in the second year. The patients participating in the

study received medical services from the primary care clinics that were part of a large

medical group in an integrated health system located in two states in the Midwest (Van

de Ven, 2014). The patients could receive services such as obstetrics and gynecology,

family practice, and internal medicine at the clinics (Van de Ven, 2014).

After removal of the patients’ personal identifying information, Van de Ven

(2014) evaluated 23 questions from the survey. The independent variables for the patient

survey were participative provider care, staff courtesy, health status, staff follow-up,

waiting, and medical explanations, and the dependent variable was overall satisfaction

(Van de Ven, 2014). Based on the responses of the participants, the study showed strong

correlations between participative patient care and staff courtesy (r = .41 for 1996 and

1997; Van de Ven, 2014, pp. 133, 136). The previous two variables were found to be

directly related to staff follow-up (r ≥ .27 in 1996 and ≥ .24 in 1997) and medical

explanations (r ≥ .29 in 1996 and ≥ .28 in 1997; Van de Ven, 2014, pp. 133, 136). The

results showed participative provider care, and staff courtesy had an indirect association

with waits and delays (r = −.33 in 1996 and −.30 to −.32 in 1997; Van de Ven, 2014, pp.

133, 136). However, correlations between patients’ self-reported health status and the

other elements were less than .08 in magnitude.

Van de Ven (2014) examined the correlation between specific patients’

demographics (gender, age, and education), patients’ “willingness to recommend the

clinic to family and friends, satisfaction with the purpose of the visit, and overall visit

rating” (pp. 133–134), and patient care experience. Van de Ven reported that patient

demographics and satisfaction measures had a direct and significant correlation with

participative provider care, staff courtesy, staff follow-up, and medical explanations. Van

de Ven reported an indirect correlation between participative provider care, staff

courtesy, staff follow-up, and medical explanations in regard to waiting time. However, it

was reported that there was not a correlation to the self-reported health status (r ≤ .08)

(Van de Ven, 2014). Van de Ven did report a finding of some exceptions to correlations

between patient age and health status (r = .24 in 1996 and .27 in 1997) and patient

education and health status (r = −.23 in 1996 and −.22 in 1997). Van de Van found that

gender and level of education did not play a role in “determining patients’ assessments of

their care experience” (p. 133); however, most of the correlations fell below .06 and were

statistically insignificant.

Van de Ven (2014) used factor analysis to determine the variance in patient

responses, alpha coefficients to determine internal consistency, and because of the

assumption of correlations between satisfaction elements, the principal component

analysis with oblique rotation was used. A multiple regression analysis was employed to

determine the importance of the six care experience elements based on the perspective of

the patient’s visit using a 5-point Likert scale with a range from poor to excellent (Van de

Ven, 2014).

The sampling frame for the HOS consisted of 69 managers, 153 physicians, 429

nurses and clinicians, and 318 support staff in the first year and 83 managers, 207

physicians, 524 nurses and clinicians, and 334 support staff in the second year (Van de

Ven, 2014). Van de Van employed correlations and regression analysis to evaluate the

relationship between the patient experience of care and job satisfaction (doctors, nurses,

and staff, clinic size, and performance). Van de Van averaged the patients’ responses for

clinics that provided the medical services. Van de Van also averaged responses to the

HOS of the employees who worked in those clinics. Van de Van merged the data for size,

profitability, and physician productivity for the clinics. After cleaning the data, the

analysis was performed for 42 clinics. Van de Van reported “patient satisfaction was not

highly correlated with employee job satisfaction (r = .04), patient satisfaction was

negatively correlated with clinic size (r = - .40), and patient satisfaction was not

correlated to clinic profitability (r = .04) or clinic productivity (r = .02)” (pp. 136–137).

Clinics’ increase in profits, employees’ job satisfaction, clinic size, and the productivity

of the providers at clinics did not influence patient satisfaction (Van de Ven, 2014).

Causes for Rising Health Care Expenditures

The United States finds rising health care expenditures challenging to control.

Bodenheimer (2005a) explained the opposing roles of each actor, which was described as

a “battleground” between the actors (purchasers and insurers) who expended dollars on

health care and the actors (providers and suppliers) who received dollars for providing

health care services. The providers and suppliers are against containing health care costs,

whereas purchasers and insurers are for containing health care costs (Bodenheimer,

2005a). Bodenheimer examined several perspectives for the reason of rising health care

costs and cost containment.

Bodenheimer (2005a) found that some scholars had a difference of opinion

regarding the emphasis placed on rising health care costs. Some scholars see rising health

care costs not as severe as some might think because of the increase in jobs for the

economy and increase in healthy outcomes (Bodenheimer, 2005a). Other scholars see

rising health care costs as a primary concern because health care becomes unaffordable

and impacts the purchasers of health care patients, local and state governments, and

employers (Bodenheimer, 2005a). Bodenheimer believed that forces outside the scope of

health care factored into the increasing health care costs. Bodenheimer argued that the

economy is a driving force for growing health care costs. When a country is wealthy, the

country pays more for health care per capita. Bodenheimer also looked at the

demographics of an aging population as a possible cause for rising health care costs.

Research has shown an increase in the elderly and the rate of growth for health care

spending for this population, but these increases are slow, so the impacts on rising health

care costs are minimum (Bodenheimer, 2005a; Meara, White, & Cutler, 2004).

Policy gurus thought that health care costs would continue to rise without market

competition and patients cost sharing on their medical care (Bodenheimer, 2005a). They

argued that the cost of health care would be set by supply and demand in a competitive

market versus an individual entity (Bodenheimer, 2005a). This concept assumed that

purchasers would be able to make astute decisions when purchasing health care. One

major contributor to rising health care expenditures is technologic advances

(Bodenheimer, 2005b). New technology used to treat medical conditions can improve

health outcomes where the old technology lagged. Cutler and McClellan (2001) evaluated

the benefits and costs of technologic advances. The scholars examined five conditions:

(a) breast cancer, (b) cataracts, (c) depression, (d) heart attacks, and (e) low birth weight.

The cost of the technology for treating cataracts, depression, heart attacks, and low birth

weight did not outweigh the cost of the benefits for treating these conditions (Cutler &

McClellan, 2001). However, the cost of the technology used to diagnose and treat breast

cancer and the health outcomes balanced each other.

Bodenheimer (2005b) found that administrative costs contributed to rising health

care expenditures. Public and private insurers burden providers with administrative

responsibilities such as billing for patient services. Administrative overhead is higher for

private insurers than public insurers (Bodenheimer, 2005b). Bodenheimer argued that

simplifying the insurance system would help contain cost. Scholars asserted that the

absence of strong cost control measures such as expenditure caps and global budgets

contribute to health care expenditures rising (Bodenheimer, 2005b). Expenditure caps and

global budgets have their drawbacks. When expenditure caps are implemented, providers

have a tendency to increase the quantity of services provided to patients. This is known as

“volume creep” (Bodenheimer, 2005b). When physicians in Canada increased the volume

of services rendered, there was a reduction in the fee for the services provided to avoid

exceeding the yearly expenditure cap (Bodenheimer, 2005b). In the United States,

Medicare implemented a sustainable growth rate mechanism that tracks providers’

expenditure targets (Bodenheimer, 2005b). With this mechanism in place, physicians

who exceed the expenditure target would have their fees for the following year reduced.

Bodenheimer argued that global budgets would be most effective under a single payer.

Administrative costs would be reduced, and some providers may be allowed to make

decisions on how the budget is spent (Bodenheimer, 2005b). A disadvantage of global

budgets is that the budget can be too small, impacting access to high-quality care by

limiting access to new technology (Bodenheimer, 2005b).

Scholars argued that when providers of care have market power, the cost of health

care increases rapidly, and when health care purchasers have market power, the cost of

health care increases slowly. Market power for the provider is when an insurer signs a

contract with the provider for the rate requested resulting in lucrative reimbursements

(Bodenheimer, 2005c). Market power for the insurer is when the insurer does not sign a

contract with the provider for the rate requested (Bodenheimer, 2005c). Bodenheimer

asserted that the market power for providers could be controlled by offsetting the power

of providers and insurers as well as implementing regulations.

Bodenheimer (2005c) examined the relationship between high health care

expenditures, provider fees, and increased utilization of services. Comparison of the cost

of health care in the United States to the cost of health care in other countries for similar

medical conditions and procedures reveals that they cost more in the United States than in

other countries (Bodenheimer, 2005c). Health care economists believe that the method in

which providers are reimbursed affects rising health care expenditures. Economists

argued that physicians who were reimbursed under the FFS reimbursement method

performed more services than providers who were reimbursed under the capitation

reimbursement method. Providers reimbursed under the capitation reimbursement

method share financial risk for services provided to the patients so that they would limit

the number of services (Bodenheimer, 2005b). Physicians in other countries were slow

using new technology unlike physicians in the United States (Bodenheimer, 2005b). For

example, physicians in the United States may increase the number of services such as

magnetic resonance imaging (MRI) scans; whereas, other countries would not increase

the use of MRI scans or have an MRI scanner.

Cost containment strategies can be implemented after identifying contributing

factors to increasing health care costs. As noted by Bodenheimer (2005a, 2005b, 2005c),

administrative costs, an aging population, lack of market competition, patients not cost

sharing on their medical expenses, technological advances, and the nation’s economy

contributed to growing health care costs. However, Bodenheimer (2005a) and other

scholars found that the elderly population contribution to increasing the cost of health

care was minute. Technologic advances were seen as a major contributor to rising health

care costs.

Critical Analysis

Critical Analysis of Theoretical Frameworks

Understanding the complexity of access to quality care is paramount but often

misunderstood by those trying to access a system that provides quality care. Andersen’s

behavioral model of health services use has undergone several modifications but is

frequently used as a framework to determine and assess the individual needs to accessing

and utilizing health care. The Andersen’s behavioral use model incorporated the

importance of national policy on access to care as well as provided various individual

determinants that may influence the access to care. Andersen’s model does not inform

decision makers if the individuals who need care received the care when care was needed

as well as where care was needed. However, Andersen’s behavioral use model is useful

for providing a point of entry to the health care system.

The point of intersection of the Andersen and Donabedian theoretical models

plays a crucial role in accessing health care and developing a health care system that

provides quality care. Donabedian argued that the availability of facilities not be an

assurance of access to care (Aday & Andersen, 1974; Donabedian, 1972, 1980). The

health care system with the appropriate number of facilities with the right mix of

providers to patient ratio is essential to “protecting and promoting the quality of care”

(Donabedian, 1980, p. 82). Therefore, as Donabedian pointed out having the facilities in

place does not necessarily mean that individuals are receiving quality care, they must

have access to enter into the system to receive care.

Using Andersen’s and Donabedian’s theoretical models together places emphasis

on the need for a change in policy for access to quality care and the impact of that policy

changes factor into patient experience.

Critical Analysis of Literature on Methodology

Based on the literature review, researchers have used descriptive statistics,

correlation, and regression analysis. Alkhwaldeh et al. (2014) used descriptive statistics,

chi-square test of association for categorical variables, Pearson correlations to measure

associations between independent and dependent variables, and other correlation tests

(e.g., Spearman’s rho, Point Biserial 𝑟, Phi coefficient (𝜙), and Cramer’s V methodology)

in a cross-sectional study to evaluate the utilization patterns for primary health care

factors and predictors of utilization and nonutilization of health care for older adults

Jordan. As previously mentioned, Alkhwaldeh et al. concluded that although chronic

illness was the most significant predictor of primary health care services utilization, all

the factors could be linked to the use of primary health care services.

Owusu-Frimpong et al. (2010) employed an ANOVA to explore and compare the

respondents’ perceptions of access to care to their experience with access to care. The

Pearson chi-square statistics and associated p-values were used to establish the extent of

correlations between overall patient satisfaction and the patients’ demographics (Owusu-

Frimpong et al., 2010). Owusu-Frimpong et al. found significant problems with access to

care.

Van de Ven (2014) analyzed the correlation between patient experience

(participative provider care, staff courtesy, health status, staff follow-up, waiting, and

medical explanations) and patient satisfaction in the first and second year. Correlations

between participative patient care, staff courtesy follow-up, and medical explanations

were strong and positively correlated to patient demographics and satisfaction elements

(Van de Ven, 2014, pp. 133, 136). Correlations between participative provider care and

staff courtesy were negatively correlated with waits and delays (Van de Ven, 2014).

Correlations for gender and education were not statistically significant (Van de Ven,

2014). Van de Ven employed a multiple regression analysis to determine the importance

of the six care experience elements based on the perspective of the patient’s visit using a

5 point Likert scale with a range from poor to excellent. The results indicated the most

powerful predictors of patient satisfaction were participative provider care and staff

courtesy (Van de Ven, 2014).

Zwanziger et al. (2001) conducted a study using secondary data; data from two

CHAMPUS beneficiary surveys and claims data for beneficiaries surveyed were used to

evaluate cost, utilization, and patient satisfaction for a 5-year CRI demonstration.

Zwanziger et al. used multivariate analysis to examine overall cost, and multivariate

regression models were used to assess continuous (linear) and categorical (logistic)

variables. Zwanziger et al. concluded that the CRI demonstration was successful:

Although the cost of health care increased, beneficiary satisfaction and access to care in

the MTF also increased.

The literature review found a link between positive patient care experiences and

patient adherence to treatment and influenced adherence to the treatment process for

clinical outcomes for inpatient settings. Price et al. (2014) determined that there was a

correlation between positive patient experiences, better hospital patient safety culture,

and a decrease in unnecessary health care utilization. As mentioned earlier, measuring the

experience of patients in patient-centeredness care environments is vital as it increases

accountability and quality improvements based on reported patient experiences.

The statistical tests discussed are relevant to this descriptive cross-sectional study.

The tests would provide a summary of the data, information whether to accept or reject

the hypothesis, the correlation between and within the variables, predictors for accessing

care, and relevant elements for patient experience that could increase patient satisfaction.

Summary and Transition

In summary, this chapter provided information on the theoretical frameworks,

Andersen’s behavioral use model, and Donabedian’s quality of care model, used for this

research, and explained the health care system in the United States and the Philippines.

This section also provided information regarding the MHS and the evolution of the

TRICARE benefit as well as major themes from the relevant literature regarding

managed care and the importance of measuring patient experience and the statistical

analysis for assessing quality outcomes. Integrating two health care systems from two

countries is complex and challenging. As shown, there are different laws that must be

followed when the U.S. health care system is integrated with a health care system in a

different country. To achieve buy-in from providers in a host nation country, the United

States must devise a strategy to incentivize host nation providers to accept its

reimbursement methodology. Implementing a closed provider network Chapter 3 will

provide information regarding the sample frame, an overview of the data plan, covariates,

independent and dependent variables, ethical considerations for human subjects, and the

statistically methodology selected to explore relationships between and within variables.

The assessment of the variables identified in Chapter 3 should help determine if the

expected outcomes were or were not met and the acceptability and efficiency of an

alternative method for delivering health care to TRICARE Standard beneficiaries who

live in the Philippines and receive care in the designated Demonstration areas.

Chapter 3: Research Method

Data Analysis Plan

This cross-sectional quantitative study evaluated patient experience for TRICARE

Standard beneficiaries who received medical care under the DoD TRICARE Philippine

Demonstration using secondary data provided by the DHA under the DoD. The

Philippine Demonstration tested the efficacy and acceptance of an alternative approach to

health care rendered by certified Philippines providers by establishing a closed network

of approved Demonstration providers in designated Demonstration areas selected by the

DoD. The goals of the Philippine Demonstration were to control rising health care costs,

reduce aberrant billing activity, and eliminate balance billing without jeopardizing access

to high-quality health care for beneficiaries. The TOP contractor under contract with the

DoD and their subcontractor conducted the 3-year Philippine Demonstration using a

phased approach. There are three distinct phases under the Philippine Demonstration:

1. Phase 1 of the Philippine Demonstration started on January 1, 2013, in Metro

Manila and its surrounding areas: Angeles City, Pampanga; Olongapo,

Zambales.

2. Phase 2 of the Philippine Demonstration started on December 1, 2013, in the

province of Cavite, comprising General Trias, Naic, Bacoor, Imus, and Cavite

City.

3. Phase 3 of the Philippine Demonstration began on July 1, 2014, in Iloilo City,

Iloilo.

This study concentrated its efforts on the following Demonstration locations:

1. Phase 1 of the Philippine Demonstration started on January 1, 2013, in Metro

Manila and its surrounding areas.

2. Phase 2 of the Philippine Demonstration started on December 1, 2013, in

Cavite City.

3. Phase 3 of the Philippine Demonstration began on July 1, 2014, in Iloilo City.

This chapter includes the data analysis plan, methodology, and ethical considerations for

the project, and a concluding summary.

Operational Definitions of Variables

Independent Variable

The respondents’ gender was identified as male or female. Respondents whose

age range was between 18 and older than 62 years were included in the survey. The

TRICARE beneficiary categories are TOP Standard, TRICARE for Life, TRICARE

Retired Reserve, TRICARE Reserve Select, and TYA. These beneficiaries could be a

military retiree, family member of a military retiree, family member of an active service

member, or reserve member. Satisfaction with health care and overall satisfaction with

the Demonstration helped to assess as well as inform which attributes are important

regarding patient experience.

The coefficient variable was the designated Demonstration area (the city) in

which the care was rendered:

1. Phase 1: Metro Manila and its surrounding areas

2. Phase 2: Cavite City, Cavite

3. Phase 3: Iloilo City, Iloilo

Dependent Variable

Financial burdens assess if the beneficiary’s out-of-pocket cost exceeds the annual

deductible and cost share per visit and confidence that the TRICARE program can meet

their health care needs. This variable could also be indicative that the approved

Demonstration provider is not balance billing the beneficiary. Locating an approved

Demonstration provider assesses the difficulty in finding a provider.

Population

Approximately 10,000 TRICARE Standard beneficiaries live in the Philippines;

however, based on claims data, only 4,909 beneficiaries living in the Philippines used the

TOP Standard option 2 years prior to Phase 1 of health care delivery on January 1, 2013.

The targeted population for the study consisted of TRICARE Standard beneficiaries who

lived in the Philippines and accessed health care in designated Demonstration areas. The

DoD’s selection of respondents for the study was nonbiased as the survey was a

consensus survey and all beneficiaries who met the inclusion criteria had the opportunity

to respond to the survey. The respondents did not receive compensation for their

participation. Because the Philippine Demonstration was voluntary and beneficiaries

living in other provinces in the Philippines could receive care in designated

Demonstration areas, it was difficult to estimate the number of TRICARE beneficiaries

who would make up the accessible population of the study.

Sampling Frame Inclusions and Exclusions

The sampling frame included TRICARE Standard beneficiaries comprising

military retirees, family members of military retirees, and family members of active-duty

service members who lived in the Philippines and accessed health care in designated

Demonstration areas: Metro Manila and its surrounding areas; Cavite City, Cavite; and

Iloilo City, Iloilo. TRICARE beneficiaries enrolled in TOP Prime and TOP Prime

Remote, TRICARE Standard beneficiaries under 18 years of age unless they were

considered a TRICARE beneficiary in their own right, and TRICARE Standard

beneficiaries whose home address was not in the Philippines were excluded from

participating in the survey. The home address was determined by the address listed on the

TRICARE DoD/CHAMPUS Claim Form–Patient’s Request for Medical Payment (DD

Form 2642). The web-based survey was administered to TRICARE beneficiaries using

the e-mail addresses in M2, the information system owned by the DoD.

Sample Size

The sample size requirement was calculated for logistic regression using a

Pearson correlation with a modified power using G*Power 3.1.7. This correlation

assumed a medium effect size and a modified power of (0.80/0.91) 0.88. If statistical

significance was assessed within a 95% confidence interval (α = .050), the sample size

required to assure empirical validity was 102 (Faul, Erdfelder, Buchner, & Lang, 2013).

Accessing Data

The DHA Data Sharing Agreement Application was submitted to the DHA

Privacy and Civil Liberties Office for approval to access the government data files from

the program office that owns the data. Following the approval process, the DHA Decision

Support Division provided deidentified data on an encrypted computer disk. Obtaining

deidentified data for this research presented a problem, as the original plan was to include

various predisposing and enabling factors that would increase the understanding drivers

for access to health care. The data were collapsed to remove all 18 Health Information

Portability Accountability Act identifiers. I had to work closely with the analysts

providing data to obtain usable data to complete this research.

Level of Measurement

The demographic variables of interest were gender, age, and Demonstration Phase

I, II, and III. The predictor variables were (a) locating an approved provider, (b)

satisfaction with requesting a waiver, (c) financial burden and confidence that TRICARE

meets health care needs, and (d) satisfaction with health care and overall satisfaction with

Demonstration.

Methodology

This cross-sectional quantitative descriptive study included some nonparametric

analysis of relationships between specific variables. I calculated frequencies and

percentages for any categorical variables of interest, such as gender. Means and standard

deviations were calculated for any continuous variables of interest, such as age (Howell,

2010).

Data were analyzed for statistical significance using analysis of variance

calculated in SPSS Version 22. Prior to analysis, data were screened for validity, missing

data, patterns of missing data, and outliers. I used ANOVA to assure that the data met the

assumptions of independence, normality, and homogeneity variances. The Wilks–Shapiro

statistic was used to assess the assumption of normality. The Levene’s homogeneity-of-

variance test was used to assess the assumption homogeneity variance. Multicollinearity

was assessed by testing tolerance and the variance inflator factor within SPSS. Due to the

nonparametric nature of the Spearman rho correlation, none of the restrictive assumptions

typically associated with a correlation test needed to be met (Morgan, Leech, Gloekner,

& Barrett, 2007). I created standardized values for each scale level research variable and

examined for values that fell above 3.29 and those that fell below −3.29, which identified

outliers (Tabachnick & Fidell, 2012). Cases with missing data were also examined for

nonrandom patterns. The DHA conducted a survey evaluating the Philippine

Demonstration from which secondary data were obtained to answer the RQs for this

study. Surveys for participants who did not provide the phase in which they received care

were excluded.

Research Question 1

RQ1: How has the Demonstration affected the beneficiaries’ financial burdens

and confidence that their health care needs will be met?

H01a: There is no statistically significant relationship between financial

burdens and satisfaction with the Demonstration overall.

Ha1a: There is a statistically significant relationship between financial burdens

and satisfaction with the Demonstration overall.

H01b: There is no statistically significant relationship between financial

burdens and confidence that health care needs will be met.

Ha1b: There is a statistically significant relationship between financial burdens

and confidence that health care needs will be met.

To assess RQ1, I conducted two Spearman rho correlations. The first Spearman

correlation assessed the relationship between the dependent variable of financial burden

(i.e., Item 11) and the independent variable of overall satisfaction (i.e., Item 13). Item 11

asked participants to compare their out-of-pocket expenses for health care before the

Demonstration with the same expenses after the Demonstration and ranged in responses

from (a) higher now to (c) lower now. Item 13 asked about their overall satisfaction with

the Demonstration and ranged from (a) very satisfied to (e) very dissatisfied. The second

Spearman correlation assessed the relationship between the dependent variable of

confidence that health care needs were met (i.e., Item 12) and the independent variable of

overall satisfaction (Item 13). Item 12 asked participants whether the Demonstration

increased their confidence that TRICARE was able to meet their health care needs and

ranged in responses from (a) strongly agree to (e) strongly disagree. Each survey item in

the set of analyses represented an ordinal response variable. The Spearman rho

correlation was the appropriate analysis when the goal of the research was to assess the

strength or existence of a statistically significant relationship between two ordinal

variables. I used Cohen’s d (Cohen, 1988) to evaluate the correlation coefficient to

ascertain the magnitude of the difference.

Research Question 2

RQ2: What is the relationship between gender, age, and health care finder

functionality for accessing quality health care under the Demonstration in Metro

Manila and its surrounding area, Cavite City, Cavite, and Iloilo City, Iloilo?

H02a: There is no statistically significant relationship between gender, age,

and health care finder functionality access to quality health care in Metro

Manila and its surrounding areas, Cavite City, Cavite, and Iloilo City, Iloilo.

Ha2a: There is a statistically significant relationship between gender, age, and

health care finder functionality access to health care in Metro Manila and its

surrounding areas, Cavite City, Cavite, and Iloilo City, Iloilo.

To examine RQ2, I conducted a series of regressions. Because the relationships

between the predictors, gender and age, and the dependent variable, health care finder

functionality, were in question, I conducted a regression analyses in two stages on the

overall data set to include Metro Manila and its surrounding area, Cavite City, Cavite,

and Iloilo City, Iloilo. In the first stage, I conducted the regression on the overall data set,

and this regression included data from Metro Manila and its surrounding area, Cavite

City, Cavite, and Iloilo City, Iloilo. In the second stage of analysis for RQ2, I conducted

regressions for Metro Manila and its surrounding area, Cavite City, Cavite, and Iloilo

City, Iloilo. Findings from both the descriptive statistics and from the normality testing

for each of these regressions indicated the possibility of acquiescence bias and resulted in

a series of secondary analyses through binary logistic regression to confirm the findings.

To accomplish this, the satisfaction variable was transformed into a dichotomous

outcome, where zero represented any category less than very satisfied, and one

represented a response of very satisfied.

To examine RQ3, I conducted descriptive statistics on responses to Items 3 and 4,

as applicable. Item 3 asked participants whether they had submitted a waiver to receive

care from a nonapproved Demonstration provider. If participants responded that they had

submitted a waiver, the participant was asked to respond to Item 4, which asked how

satisfied they were with the provider and had response categories ranging from (a) very

satisfied to (e) very dissatisfied. Thus, responses to Item 3 were presented to show the

number who did not submit a waiver, while Item 4 identified how satisfied the proportion

of participants who did submit a waiver were with their nonapproved provider. I

conducted means and standard deviations to detail the spread and central tendency of

responses, while frequencies and percentages provided the proportionality of response

rates.

Research Question 3

RQ3: What is the relationship between a closed network, actual beneficiary out-

of-pocket cost and quality health care in Demonstration areas?

H03a: There is no statistically significant relationship between actual

beneficiary cost and a closed provider network.

Ho3a: There is a statistically significant relationship between actual

beneficiary out-of-pocket cost and a closed provider network.

H03b: There is no statistically significant relationship between quality health

care and a closed provider network.

Ha3b: There is a statistically significant relationship between quality health

care and a closed provider network.

To examine RQ3, three questions from the survey were chosen as representations

of the beneficiary out-of-pocket costs and quality health care. The first question asked

participants to rate their out-of-pocket costs in relation to those same costs prior to the

Demonstration. To examine the quality of health care, participants were asked to rate

their satisfaction on a scale from 0 to 10. Finally, participants were asked to compare

their current experience with receiving health care to their experience before the

Demonstration. However, to test whether these responses were significantly higher or

lower than their mid-level response categories, a series of one-sample t-tests was

conducted.

Threats to Validity

Threats to validity and reliability should be considered when designing research.

Trustworthiness and credibility come to mind when gathering data for a study. All

researchers should want to ensure that their research is reliable and valid. Bowling (2010)

asserted that a research instrument must be assessed using test–retest, interrater

reliability, and internal consistency to determine reliability. The research instrument

should be adequately tested for validity numerous times using the same population it was

intended to test before being considered valid (Bowling, 2010). A research instrument

can be reliable without being valid, but a research instrument that is not reliable is not

valid. Researchers using the same research instrument should be able to replicate the

results of previous research using the same criteria and environment (Creswell, 2009;

Field, 2007). When the research instrument measures what it is designed to measure, the

research instrument is considered to be valid (Creswell, 2009; Field, 2007). The survey

instrument that was used for this descriptive cross-sectional quantitative study was

developed, and secondary data were analyzed; therefore the instrument must be assessed

for reliability and validity. Creswell (2009) recommended evaluating the instrument by

determining if there is a connection between the variables, RQs, and items on the survey

instrument. Cleaning the data increased reliability, help with consistency, and accuracy.

The data cleaning process should be performed with care when removing outliers and

duplicate data. Prior to cleaning the data, a copy of the original should be saved and

security measures should be taken to protect the data. The performance of a statistical test

could determine internal consistency: Perform a reliability analysis, and check the

statistical assumptions for violations.

Statistical Conclusion Validity

Violation of a statistical assumption poses a threat to statistical conclusion

validity (Creswell, 2009). Check statistical assumptions to avoid making the wrong

conclusions. The statistical assumptions are assumptions of independence, normality,

homogeneity of variances, interval, and multicollinearity (Field, 2007). According to

Miles and Shevlin (2001), a violation of the assumption of independence can be difficult

to detect. However, when autocorrelation occurs, the assumption of independence has

been violated, which can happen with a time series design or clustered sampling design

(Miles & Shevlin, 2001). The Wilks–Shapiro statistic can evaluate the assumption of

normality. The Levene’s homogeneity-of-variance test can assess the assumption

homogeneity variance. When the distance between intervals is not equal, there is a

violation of the assumption of interval. Multicollinearity can be assessed with the

tolerance statistic and the variance inflator factor. The appropriate actions should be

taken to mitigate the threat of conclusion validity when there is a violation of any

statistical assumption.

Ethical Consideration

The study was submitted to the Walden University Institutional Review Board

(IRB) for approval before requesting data files and data analysis. After receiving

conditional approval from Walden University IRB, a copy of Walden University’s IRB

conditional approval was submitted to the DHA Human Research Protection Board,

which determined that human subjects would not be used for this research. After

receiving the determination from the DHA Human Research Protection Board, a Data

Sharing Agreement Application was submitted to the DHA. Upon receipt of the approved

Data Sharing Agreement Application 17–1810, the final package was submitted to

Walden University IRB for final approval. After receiving Notification of Approval to

Conduct Research from Walden University IRB (approval no.01-31-18-0082516), the

request was submitted to the Decision Support Division for access to the TRICARE

Philippine Demonstration raw survey and claims data files. Although this study consisted

of a secondary analysis, the policies and regulations for Human Research for the

Protection of Human Subjects were followed. All data have been maintained on an

encrypted removable storage disc under lock and key accessible only by the primary

investigator. After analysis, the data were stored on an encrypted removable storage disc

under lock and key accessible only to the primary investigator and will be properly

destroyed after 5 years.

The principal investigator (PI) for the Philippine Demonstration Beneficiary

Satisfaction Questionnaire who conducted the original research used a combination of

Web and telephone surveys and provided documentation of approval to conduct the

research from Liberty’s IRB. The PI did not target any members from vulnerable

populations. The survey participants did not receive financial incentives or gifts for their

participation. The survey was conducted over the telephone, by which the interviewers

obtained informed consent from the participants. The interviewers informed the

participants that participation in the survey was voluntary. If the participant did not want

to participate in the survey, the interviewer informed the participant that his or her

TRICARE benefits would not be affected. Additionally, the participant had the option to

stop the survey if he or she became uncomfortable. The participants were given a

telephone number to contact the PI if they had any questions as well as a different

telephone number for technical difficulties for the Web survey. The investigator did not

have any conflicts of interest.

Summary

This chapter discussed the methodology, data analysis plan, and ethical

considerations used in this cross-sectional quantitative analysis for evaluating TRICARE

beneficiaries’ satisfaction under the Philippine Demonstration. Assessing statistical

significance within a 95% confidence interval (α = .050), the sample size for the study

with a medium effect size required to assure empirical validity was 102. The database

contained approximately 8,000 TRICARE Standard beneficiaries living in the

Philippines, of whom 4,909 beneficiaries had accessed their TRICARE benefits. Chapter

4 presents the results of the descriptive and inferential analysis for the RQs and

hypotheses presented.

Chapter 4: Results

Introduction

In response to TMA’s (2008a) identification of a major concern in the Philippines

with rising health care costs, as well as the Philippines’ reputation of ubiquitous health

care fraud with TRICARE, the DHA, DoD, implemented administrative controls to deter

fraud in the Philippines (TMA, 2008a). Although the administrative controls were

implemented, the cost of health care continued to rise. Therefore, the DHA implemented

the DoD TRICARE Philippine Demonstration in 2013 to determine the efficacy and

acceptability of using an alternative method to deliver health care, increase access to

health care, and decrease beneficiary out-of-pocket costs while maintaining quality health

care. The purpose of this quantitative cross-sectional study was to address questions

regarding TOP Standard beneficiaries’ ability to access quality health care and their out-

of-pocket costs for health care under the Philippine Demonstration.

The following chapter includes a description of the data collection procedures

pertinent to the sample used for analysis. In this section are a description of the sample

and a comparison to the population of interest, intended to clarify the generalizability of

the results. The results of the analyses for RQ1, RQ2, and RQ3 follow, with an

explanation of the assumptions required for each assessment, as well as the probability

levels of the findings and an explanation of what they signify. The chapter closes with a

summary of the findings and description of what will be included in the following

discussion chapter.

Data Cleaning

I removed the responses such as “suppressed,” “don’t know,” and “unsure” unless

ordered correctly. For example, an unsure between agree and disagree is applicable,

while a separate category outside of Likert ranking is not. I reversed items for better

interpretability (appended with “_Reverse”):

• Q1 was recoded for higher scores indicative of higher satisfaction.

• Q4 was recoded for higher scores indicative of higher satisfaction.

• Q10 was recoded for higher scores indicative of better experience.

• Q13 was recoded for higher scores indicative of higher confidence.

Data Collection

Results

The final sample for analysis consisted of responses from 180 individuals in the

Philippines who received care under the TRICARE Demonstration and completed a

survey on their experiences with the Demonstration and health care thereafter. Of these

respondents, a majority were male (53.3%), though an additional 11.1% did not provide

this information. Exactly half of the available sample was retired, though 11.1% of the

sample provided no information regarding their beneficiary group either. Finally, though

the same 11.1% did not indicate an age, the sample consisted of a slight majority of those

within the 60–79 years of age group. Many of the sample’s participants were also within

the 40–59-year age group (n = 30, 16.7%) or 80 years or older age group (n = 28, 15.6%).

With a minority of the sample below the age of 20 (6.1%), this sample was highly

representative of older individuals. Demographic information for each possible grouping

of these three categories of descriptive information is available in Table 1.

Table 1

Demographic Information for Final Sample

Variable

Gender

Female

35.6

Male

53.3

No response

11.1

Beneficiary group

Retired

50.0

Retired dependent

38.9

No response

11.1

Age (years)

0–19

6.1

20–39

0.0

40–59

16.7

60–79

50.6

80+

15.6

No response

11.1

Note. N = 180.

Research Question 1

RQ1: How has the Demonstration affected the beneficiaries’ financial burdens

and confidence that their health care needs will be met?

H01a: There is no statistically significant relationship between financial

burdens and satisfaction with the Demonstration overall.

Ha1a: There is a statistically significant relationship between financial burdens

and satisfaction with the Demonstration overall.

H01b: There is no statistically significant relationship between financial

burdens and confidence that health care needs will be met.

Ha1b: There is a statistically significant relationship between financial burdens

and confidence that health care needs will be met.

To assess RQ1, two correlations were necessary. These two correlations follow

the associated hypotheses, focusing on the relationship between financial burden and

satisfaction and financial burden and confidence. To test a Pearson correlation with

confidence in their statistical validity, the assumption of linearity requires assessment

(Pallant, 2007). However, due to the nonparametric nature of the Spearman correlation,

this is not a requirement. The Spearman correlation was chosen as appropriate due to the

ordinal response categories for both variables in either pair in this analysis. In addition,

the use of Spearman correlations would circumvent any validiy concerns based on the

distribution of the responses concerning satisfaction or confidence, which appeard to be

skewed toward a more positive response (Lehmann, 2006). This is based on the ranking

used in Spearman correlations, which takes place in the first stage of analysis and

removes any distributional properties of the data by replacing raw values with ranks, as

needed for use in this form of correlation. For this reason, there were no assumptions that

required assessment, and the analyses could be considered valid regardless of the

distribution of the variables used (Lehmann, 2006). The results of both correlations were

significant and positive (satisfaction, r = .41, p < .001; confidence, r = .44, p < .001),

which indicated that, among those with higher financial burdens, their satisfaction with

the Demonstration overall and confidence that their health care needs would be met were

both correspondingly higher. Both of these correlations were statistically significant,

indicating a great degree of confidence that these correlations were accurate depictions of

reality. These findings suggest that those with a higher level of financial burden are not

only happier with the Demonstration but left feeling more confident.

Research Question 2

RQ2: What is the relationship between gender, age, and health care finder

functionality for accessing quality health care under the Demonstration in Metro

Manila, Cavite City, Cavite, and Iloilo City, Iloilo?

H02a: There is no statistically significant relationship between gender, age,

and health care finder functionality access to quality health care in Metro

Manila, Cavite City, Cavite, and Iloilo City, Iloilo.

Ha2a: There is a statistically significant relationship between gender, age, and

health care finder functionality access to health care in Metro Manila, Cavite

City, Cavite, and Iloilo City, Iloilo.

In response to RQ2, I conducted a series of regressions. Because the relationships

between the predictors, gender and age, and the dependent variable, health care finder

functionality, were in question for each region of the Demonstration, I conducted

regression analyses in two stages. In the first stage, I conducted the regression on the

overall data set, and this regression included data from all three locations. By assessing

these relationships, I could determine whether there was evidence of a predictive effect of

gender or age on health care finder functionality regardless of location.

Prior to analysis, I considered the assumptions of the regression. The first

assumption was the absence of multicollinearity. Though not expected due to the nature

of the independent variables, I assessed the variance inflation factor between gender and

age for any interference in the interpretation of results. A variance inflation factor above

5 is cause for concern and might indicate that the independent variables are too highly

related (Tabachnick & Fidell, 2012). This could possibly be the case if too many of one

gender were clustered within a specific age group, but it was found not to be problematic,

as the variance inflation factor’s value was 1.04. The next assumption was normality,

which concerns the normality of the distribution of error in the regression. Though the

univariate distribution of satisfaction responses regarding healthcare finder functionality

was skewed toward a more positive perception, it was important to understand the

residual normality as well, as univariate normality is not a direct assumption of the

regression analysis (Tabachnick & Fidell). This assumption is typically assessed with a

normal P-P plot, which should ideally display data points close to a hypothetically

perfect diagonal line (Tabachnick & Fidell). As seen in Figure 1, the data points did not

follow the ideal line but were not so deviant as to cause concern. However, the

assumption test’s results indicated that this analysis should still be interpreted with some

caution, as the slight deviation may cause the model estimates to be less valid. This

deviation may also have been due to the skewed distribution of satisfaction with

healthcare finder functionality. The final assumption is homoscedasticity, which is

testable through a standardized residual scatterplot (Stevens, 2016). This plot should not

show any notable pattern and is cause for concern when there is apparent funneling. As

seen in Figure 2, none such patterning was identified, and the assumption of

homoscedasticity was met.

Figure 1. Normal P-P plot for overall regression of care finder functionality.

Figure 2. Standardized residual plot for overall regression of care finder functionality.

As the results of assumption testing were questionable, the regression could be

conducted with little cause for concern as long as the results could be replicated with an

adjusted dependent variable value. To this end, satisfaction with care finder functionality,

which was skewed toward a higher satisfaction, was transformed into a dichotomous

variable for use in secondary testing. However, the originally proposed regression

analysis was conducted first. Results of the originally proposed linear regression

indicated a nonsignificant model, F(2,154) = 2.76, p = .067, R2 = .04, and did not provide

evidence that gender or age was significantly predictive of a Demonstration recipient’s

perception of health care finder functionality. This was confirmed with the findings seen

in Table 2, which show that neither predictor was significant, though gender did approach

significance. However, as predictors should not be evaluated in a nonsignificant model,

these results are not meaningful.

Following the originally proposed model, based on the understanding that the

skew in care finder functionality could have been due to acquiescence bias, the

satisfaction scale for care finder functionality was transformed into a binary variable

where zero represented any response less than very satisfied, and one represented a

response of very satisfied. This method of analysis returned similar results, with no

evidence for significance between the predictors of gender and age and the dependent

variable of satisfaction with care funder functionality (χ2(2) = 5.12, p = .077). These

findings confirmed the results of the linear regression, and thus none of the predictor

variables required interpretation.

Table 2

Overall Regression of Care Finder Functionality on Gender and Age

Source

Gender (ref: female)

−0.42

0.22

−0.16

−1.97

0.051

Age

0.18

0.11

0.13

1.66

0.100

In the second stage of analysis for RQ2, I conducted regressions for each region

of the Demonstration individually. As a preliminary data assessment, the sample was

assessed for subsample sizes in each region. Table 3 shows these subsample sizes and

indicates that the Metro Manila subsample was much larger than both the Cavite City and

Iloilo City subsamples. Though these subsample sizes were not all-sufficient for the

regression analyses, they were conducted for completeness and based on the possibility

that effect sizes of predictive relationships within a subsample were so strong that they

lowered the required effect size to achieve significance.

Table 3

Phases of Philippine Demonstration Represented in the Data Set

Phase

Phase 1: Metro Manila

109

60.6

Phase 2: Cavite City, Cavite

21.7

Phase 3: Iloilo City, Iloilo

6.7

No data

11.1

As with the preliminary regression, assumptions were assessed for each individual

model for the three regions. Variance inflation factors for each regression were

acceptable, with a value of 1.03 for the regression conducted on Phase 1, 1.11 for Phase

2, and 1.15 for Phase 3; these values indicated that multicollinearity was not problematic

for any subsample. Normal P-P plots, seen in Figures 3, 4, and 5, were similar to the first

regression’s assumption plot within the Metro Manila regression but deviated more from

the ideal normal distribution within the Cavite City and Iloilo City subsamples. These are

likely artifacts of the small sample sizes (Cohen, 1988) and cannot be adequately

corrected with transformations or nonparametric analyses, which would require much

larger sample sizes to detect significance. In contrast to these findings, homoscedasticity

was not problematic, as the standardized residual scatterplots within Figures 3 and 4 were

satisfactory. However, homoscedasticity could not be assumed for the subsample of Iloilo

City, though this was again a likely result of the small sample size. Though these

limitations should be noted, the regression analyses were continued for completeness and

to explore any possible patterns within the data. However, in addition to these analyses, a

second series of binary logistic regressions were conducted in an effort to better

understand the results and determine whether they could be confirmed or denied through

subsequent testing. This testing took place similarly to the regression of satisfaction with

care finder functionality, with the use of a dependent variable transformed such that zero

represented any response less than very satisfied, and one represented a response of very

satisfied. The binary nature of this variable both addressed the possibility of acquiescence

bias and the non-normal distribution of residuals for each analysis.

Figure 3. Normality and standardized residual scatterplots for Metro Manila.

Figure 4. Normality and standardized residual scatterplots for Cavite City.

Figure 5. Normality and standardized residual scatterplots for Iloilo City.

Metro Manila. Results of the first regression were not significant but did

approach significance, F(2,104) = 2.93, p = .058, R2 = .05, suggesting that there is a

possibility for relationships between the independent variables of gender and age and the

dependent variable of health care finder functionality within the Metro Manila

subsample. Again, these findings were not significant, and should not be considered for

interpretation, but they were still examined in an exploratory manner. Assessment of the

individual predictors showed that the only potential influence resulted from

Demonstration recipients’ gender. Though these predictors should not be interpreted due

to the overall regression results, there is some evidence that male Demonstration

recipients were less satisfied with the health care finder functionality, p = .027, B = −.61,

and were .61 units less satisfied on average. No effects were evident from the age

variable. Results of a secondary analysis through binary logistic regression confirmed this

lack of significance (χ2(2) = 4.44, p = .109), and suggested that the nonsignificant

findings were not likely due to the influenced of nonnormality or acquiescence bias.

Cavite City. Results from the second regression were also nonsignificant, F(2,

35) = 0.19, p = .827, R2 = .01, suggesting that there was no evidence for a relationship

between the independent variables of gender and age and the dependent variable of care

finder functionality. Further assessment of the predictor variables supported this lack of

significance. Results of a secondary analysis through binary logistic regression confirmed

this lack of significance (χ2(2) = 0.68, p = .712), and suggested that the nonsignificant

findings were not likely due to the influenced of nonnormality or acquiescence bias.

However, it is possible that these findings were due to the smaller than ideal sample size.

Iloilo City. Results from the third regression were also nonsignificant, F(2,9) =

1.41, p = .294, R2 = .24, suggesting that there was no evidence for a relationship between

the independent variables of gender and age and the dependent variable of care finder

functionality. Further assessment of the predictor variables supported this lack of

significance. Results of a secondary analysis through binary logistic regression confirmed

this lack of significance (χ2(2) = 3.94, p = .140), and suggested that the nonsignificant

findings were not likely due to the influenced of nonnormality or acquiescence bias.

However, as with the regression of Cavite City, it is possible that these findings were due

to the smaller than ideal sample size, which was 12 for this regression. Results of these

three subsample regressions can be found in Table 4.

Table 4

Individual Regressions of Care Finder Functionality for each Phase of Demonstration

Source

Phase 1: Metro Manila

Gender (ref: female)

−0.61

0.27

−0.22

−2.24

0.027

Age

0.17

0.14

0.12

1.27

0.207

Phase 2: Cavite City, Cavite

Gender (ref: female)

0.20

0.39

0.09

0.52

0.605

Age

0.03

0.19

0.03

0.15

0.879

Phase 3: Iloilo City, Iloilo

Gender (ref: female)

−0.43

0.27

−0.49

−1.57

0.152

Age

0.00

0.49

0.00

0.999

As a follow-up to the regressions conducted to determine the combined predictive

effect of gender and age on health care finder functionality, a final regression was

conducted to examine the effect of these independent variables on satisfaction with

submission of a waiver to request to see a nonapproved provider in a designated

Demonstration location. Within the data set, a total of 33 Demonstration recipients had

requested a waiver and were the only participants to provide data regarding their

satisfaction with this process. Out of these 33, 29 also had data available regarding their

age and gender, which was necessary for use in the regression. These were the only

participants who provided input on the process of submitting such a waiver and were the

only participants used in the regression of waiver submission satisfaction.

Assumption testing again started with an assessment of multicollinearity for this

subsample of 29. With a variance inflation factor value of 1.19, the assumption was met.

As seen in Figure 6, normality slightly deviated from the perfect normal line, which could

be cause for concern when interpreting the results. The residual scatterplot (see Figure 7)

revealed two clusters of responses on the outcome variable, satisfaction with request for a

waiver, though the two clusters were similarly sized and did not indicate problems with

homoscedasticity. The results should nonetheless be considered with caution, as the

normality was questionable and the sample size was not ideal. For this reason, in addition

to these analyses, a binary logistic regressions was conducted in an effort to better

understand the results. This testing took place with the use of a dependent variable

transformed such that zero represented any response less than very satisfied, and one

represented a response of very satisfied. The binary nature of this variable both addressed

the possibility of acquiescence bias and the non-normal distribution of residuals.

Figure 6. Normal P-P plot for regression satisfaction with waiver.

Figure 7. Standardized residual plot for regression of satisfaction with waiver.

Results of this final regression were also nonsignificant, F(2,26) = 1.22, p = .313,

R2 = .09, and provided no evidence for a relationship between a Demonstration

recipient’s gender or age and his or her satisfaction with request for a waiver. Subsequent

testing through a binary logistic regression provided similar results (χ2(2) = 2.75, p

= .253), indicating no evidence for a relationship between the predictors of gender and

age and the outcome of satisfaction with the request for a waiver. Because these results

were not significant, the individual predictors should not be examined further; however, a

quick visual assessment of the two predictors confirms the overall findings. These results

can be found in Table 5.

Table 5

Regression of Satisfaction With Waiver on Gender and Age

Source

Gender (ref: female)

−0.99

0.67

−0.31

−1.50

0.147

Age

0.06

0.29

0.04

0.20

0.844

Research Question 3

RQ3: What is the relationship between a closed network, actual beneficiary out-

of-pocket cost, and quality health care in Demonstration areas?

H03a: There is no statistically significant relationship between actual

beneficiary cost and a closed provider network.

Ho3a: There is a statistically significant relationship between actual

beneficiary out-of-pocket cost and a closed provider network.

100

H03b: There is no statistically significant relationship between quality health

care and a closed provider network.

Ha3b: There is a statistically significant relationship between quality health

care and a closed provider network.

To examine RQ3, three questions from the survey were chosen as representations

of the beneficiary out-of-pocket costs and quality health care. The first question asked

participants to rate their out-of-pocket costs in relation to those same costs prior to the

Demonstration. As seen in Table 6, most felt that their costs were either higher now

(36.1%) or about the same (35.6%), with a minority reporting lower costs following the

Demonstration (10.0%). To examine the quality of health care, participants were asked to

rate their satisfaction on a scale from 0 to 10. As seen in Table 6, many responded with a

score of 5 or above. Finally, participants were asked to compare their current experience

with receiving health care to their experience before, and the two largest response groups

consisted of those who felt their experience was about the same (28.9%) or much better

(29.4%). However, to test whether these responses were significantly higher or lower

than their mid-level response categories, a series of one-sample t-tests was conducted.

101

Table 6

Responses to Questions Regarding Beneficiary Cost and Quality Health Care

Response

Out-of-pocket expenses following Demonstration

Higher now

36.1

About the same now

35.6

Lower now

10.0

No response

18.3

Rating of satisfaction with health care quality where 0 is worse and 10 is best possible

5.0

1.7

2.2

4.4

7.8

5.0

15.0

26.1

13.3

16.7

No response

0.6

Comparison of current experience receiving healthcare before Demonstration

(1) Much worse

11.7

(2) Somewhat worse

9.4

(3) About the same

28.9

(4) Somewhat better

10.6

(5) Much better

29.4

No response

10.0

102

Each of these three questions was first assessed to determine which response

category would take the role of comparison value. For out-of-pocket expenses, the

category of about the same now was chosen, and results would thus indicate whether

participants responded with a significantly higher or lower value on average. Similarly,

the rating of 5 was chosen for the rating between 1 and 10, as this indicated a middle

response, and testing would thus indicate whether participants responded significantly

higher or lower than the objective middle category. For the last question, which asked

participants to compare their experience receiving health care before and after the

Demonstration, the category of about the same was chosen, as results would thus indicate

whether participants tended to respond close to the much worse or much better anchor of

the scale. Examination of the distribution for responses to all three of these items

indicated no evidence of acquiescence bias or skewed distributions, suggesting that there

was no need for a binary transformation, and that t tests would be the best suited method

for statistical analysis.

Results for all three tests were significant, indicating that participant responses

tended to be significantly different from their middle response categories. The responses

to out-of-pocket expenses compared to a value of 2, or about the same were significant at

the p < .001 level, t(146) = −5.68. Examination of the mean difference showed that

participants tended to respond that the expenses were higher now with a frequency more

common than would be expected at random. However, the satisfaction value with

response categories ranging from 0 to 10 also consisted of responses significantly

different from the middle response of 5. This difference was also significant at the p <

103

.001 level, t(178) = 15.27. Results of this analysis showed that the Demonstration

recipients tended to respond an average of 3.04 units higher than the middle response of

5, suggesting that they responded significantly closer to the best possible anchor of the

satisfaction scale. Finally, the comparison of satisfaction from before to after the

Demonstration was also significant at the p < .001 level, t(162) = 3.77. Examination of

the mean difference between the sample’s responses and the response of about the same

showed that participants responded significantly higher, or closer to the much better

anchor of this scale. See Table 7.

Table 7

One-Sample t Tests for Expenses and Satisfaction Following Demonstration

Variable

t (df)

Mean difference

Out-of-pocket expenses

−5.68 (146)

<0.001

−0.32

Satisfaction (0–10)

15.27 (178)

<0.001

3.04

Satisfaction compared to predemonstration

3.77 (162)

<0.001

0.41

Beneficiary Actual Out-of-Pocket Costs

How much were the actual out-of-pocket costs for beneficiaries for calendar years

(CYs) 2011, 2012, 2013, 2014, and 2015 in Metro Manila, Cavite City, Cavite, and Iloilo,

Iloilo? Prior to the Philippine Demonstration, beneficiaries paid a deductible and cost

share plus balance billing. After the implementation of the Philippine Demonstration,

beneficiaries paid a deductible and cost share. There is an assumption that beneficiaries’

overall out-of-pocket costs would increase initially because greater than half the

TRICARE beneficiaries were not using the TRICARE benefit. As more beneficiaries use

104

the TRICARE benefit and receive care from approved Demonstration providers, their

costs should eventually show a decrease.

To evaluate beneficiary actual out-of-pocket costs, an analysis of claims data from

January 1, 2011, through December 31, 2015, was conducted to assess the impact of the

Philippine Demonstration on beneficiary financial burden comparing beneficiary out-of-

pocket cost prior to the Philippine Demonstration and during the Philippine

Demonstration. All numbers for the analysis were rounded, less than .50 were rounded

down, and .50 and higher were rounded up.

The analysis was based on CYs January 1, 2011, through December 31, 2015.

This analysis was based on Phase 1 Metro Manila and its surrounding areas, Phase 2

Cavite City, Cavite, and Phase 3, Iloilo City, Iloilo. An analysis of the claims data from

January 1, 2011, through December 31, 2012, used the same locations selected for the

Philippine Demonstration and analyzed using special processing codes: Special

Processing Code 1, Special Processing Code 2, Special Processing Code 3, and Special

Processing Code 4. To protect the identities of TRICARE beneficiaries, the 18 categories

of HIPAA identifiers were removed. Therefore all personally identifiable information

was removed from the claims data.

Inpatient Beneficiary Out-of-Pocket Costs

Table 8 and Figure 8 show the grand total for inpatient beneficiary out-of-pocket

costs using all special processing codes for CY 2011 as $147, 320, CY 2012 as $143,

682, CY 2013 as $140, 224, CY 2014 as $223, 820, and CY 2015 as $231,506. A closer

examination of the inpatient beneficiary out-of-pocket costs for Phase I for Metro Manila

105

shows beneficiaries out-of-pocket costs for CY 2011 as $133,510, CY 2012 as $133,276,

CY 2013 as $133,192, CY 2014 as $196,199, and CY 2015 as $195,521. Phase 2 for

Cavite City, Cavite, shows beneficiaries’ out-of-pocket costs for CY 2011 as $5,440, CY

2012 as $5,672, CY 2013 as $1,347, CY 2014 as $6,626, and CY 2015 as $8,386. Phase

3 for Iloilo City, Iloilo, shows beneficiaries’ out-of-pocket costs for CY 2011 as $8,371,

CY 2012 as $4,733, CY 2013 as $5,684, CY 2014 as $20,995, and CY 2015 as $27,598.

Table 8

Inpatient Beneficiary Out-of-Pocket Costs All Special Processing Code

Demo city/phase

CY2011

CY2012

CY2013

CY2014

CY2015

PH1

$133,510

$133,276

$133,192

$196,199

$195,521

PH2

$5,440

$5,672

$1,347

$6,626

$8,386

PH3

$8,371

$4,733

$5,684

$20,995

$27,598

Grand total

$147,320

$143,682

$140,224

$223,820

$231,506

Figure 8. Inpatient beneficiary out-of-pocket costs all special processing code.

$50,000

$100,000

$150,000

$200,000

$250,000

CY2011 CY2012 CY2013 CY2014 CY2015

PH1 PH2 PH3 Grand Total

106

Table 9 and Figure 9 show the grand total for inpatient beneficiary out-of-pocket

costs for using Special Processing Code PH for CY 2011 as $0, CY 2012 as $0, CY 2013

as $65,529, CY 2014 as $172,507, and CY 2015 as $184,158. For Table 9, the dollar

value for CY 2011 and CY 2012 will not be discussed because the Demonstration had not

commenced and claims were not processed using Special Processing Code PH. Phase I

for Metro Manila shows beneficiaries’ out-of-pocket costs for CY 2013 as $63,911, CY

2014 as $149,911, and CY 2015 as $149,014. Phase 2 for Cavite City, Cavite, shows

beneficiaries’ out-of-pocket costs for CY 2013 as $196, CY 2014 as $6,162, and CY

2015 as $8,103. Phase 3 for Iloilo, Iloilo, shows beneficiaries’ out-of-pocket costs for CY

2013 as $1,422, CY 2014 as $16,435, and CY 2015 as $27,041.

Table 9

Inpatient Beneficiary Out-of-Pocket Costs With Special Processing Code PH

Demo city/phase

CY2011

CY2012

CY2013

CY2014

CY2015

PH1

$63,911

$149,911

$149,014

PH2

$196

$6,162

$8,103

PH3

$1,422

$16,435

$27,041

Grand total

$65,529

$172,507

$184,158

107

Figure 9. Inpatient beneficiary out-of-pocket costs with Special Processing Code PH.

Table 10 and Figure 10 show the difference between Table 8 and Table 9 for

inpatient beneficiary out-of-pocket costs, keeping in mind there is no difference for CY

2011 and CY 2012 because the Philippine Demonstration had not been implemented.

Thus the grand total for inpatient beneficiary out-of-pocket costs difference for CY 2013

is $74,695, CY 2014 is $51,312, and CY 2015 is $47,347. Phase 1 for Metro Manila

shows a difference for inpatient beneficiary out-of-pocket costs for CY 2013 as $133,192,

CY 2014 as $196,199, and CY 2015 as $195,521. Phase 2 for Cavite City, Cavite,

difference in out-of-pocket costs for CY 2013 is $1,152, CY 2014 is $464, and CY 2015

is $283. Phase 3 for Iloilo City, Iloilo, difference in out-of-pocket costs for CY 2013 is

$4,262, CY 2014 is $4,560, and CY 2015 is $557.

$50,000

$100,000

$150,000

$200,000

CY2011 CY2012 CY2013 CY2014 CY2015

PH1 PH2 PH3 Grand Total

108

Table 10

Difference of Inpatient Beneficiary Out-of-Pocket Cost for Table 8 and Table 9

Demo city/phase

CY2011

CY2012

CY2013

CY2014

CY2015

PH1

$133,510

$133,276

$69,282

$46,288

$46,507

PH2

$5,440

$5,672

$1,152

$464

$283

PH3

$8,371

$4,733

$4,262

$4,560

$557

Grand total

$147,320

$143,682

$74,695

$51,312

$47,347

Figure 10. Difference of inpatient beneficiary out-of-pocket cost for Table 8 and Table 9.

Outpatient Beneficiary Out-of-Pocket Costs

Table 11 and Figure 11 show the grand total for outpatient beneficiary out-of-

pocket costs using all special processing codes for CY 2011 as $368,029, CY 2012 as

$232,041, CY 2013 as $187,933, CY 2014 as $197,438, and CY 2015 as $218,544. An

examination of the outpatient beneficiary out-of-pocket costs for Phase 1 for Metro

Manila shows beneficiaries’ out-of-pocket costs for CY 2011 as $211,087, CY 2012 as

$20,000

$40,000

$60,000

$80,000

$100,000

$120,000

$140,000

$160,000

CY2011 CY2012 CY2013 CY2014 CY2015

PH1 PH2 PH3 Grand Total

109

$169,253, CY 2013 as $130,576, CY 2014 as $154,243, and CY 2015 as $175,780. Phase

2 for Cavite City, Cavite, shows beneficiaries’ out-of-pocket costs for CY 2011 as

$144,960, CY 2012 as $45,562, CY 2013 as $37,870, CY 2014 as $15,286, and CY 2015

as $15,025. Phase 3 for Iloilo City, Iloilo, shows beneficiaries out-of-pocket costs for CY

2011 as $11,983, CY 22012 as $17,225, CY 2013 as $19,487, CY 2014 as $27,909, and

CY 2015 as $27,740.

Table 11

Outpatient Beneficiary Out-of-Pocket Costs Including Special Processing Code PH

Demo city/phase

CY2011

CY2012

CY2013

CY2014

CY2015

PH1

$211,087

$169,253

$130,576

$154,243

$175,780

PH2

$144,960

$45,562

$37,870

$15,286

$15,025

PH3

$11,983

$17,225

$19,487

$27,909

$27,740

Grand total

$368,029

$232,041

$187,933

$197,438

$218,544

Figure 11. Outpatient beneficiary out-of-pocket costs including special processing code.

$50,000

$100,000

$150,000

$200,000

$250,000

$300,000

$350,000

$400,000

CY2011 CY2012 CY2013 CY2014 CY2015

PH1 PH2 PH3 Grand Total

110

Table 12 and Figure 12 show the grand total for outpatient beneficiary out-of-

pocket costs for using Special Processing Code PH for CY 2011 as $0, CY 2012 as $0,

CY 2013 as $91,521, CY 2014 as $157,590, and CY 2015 as $192,797. For Table 12, the

dollar value for CY 2011 and CY 2012 will not be discussed because the Demonstration

had not commenced and claims were not processed using Special Processing Code PH.

Phase 1 for Metro Manila shows beneficiaries’ out-of-pocket costs for CY 2013 as

$89,100, CY 2014 as $127,467, and CY 2015 as $152,897. Phase 2 for Cavite City,

Cavite, shows beneficiaries’ out-of-pocket costs for CY 2013 as $992, CY 2014 as

$13,541, and CY 2015 as $14,173. Phase 3 for Iloilo City, Iloilo, shows beneficiaries’

out-of-pocket costs for CY 2013 as $1429, CY 2014 as $16,581, and CY 2015 as

$25,727.

Table 12

Outpatient Beneficiary Out-of-Pocket Costs With Special Processing Code PH

Demo city/phase

CY2011

CY2012

CY2013

CY2014

CY2015

PH1

$89,100

$127,467

$152,897

PH2

$992

$13,541

$14,173

PH3

$1,429

$16,581

$25,727

Grand total

$91,521

$157,590

$192,797

111

Figure 12. Outpatient beneficiary out-of-pocket costs with Special Processing Code PH.

Table 13 and Figure 13 show the difference between Table 11 and Table 12 for

outpatient beneficiary out-of-pocket costs, keeping in mind there is no difference for CY

2011 and CY 2012 because the Philippine Demonstration had not been implemented.

Thus the grand total for inpatient beneficiary out-of-pocket costs difference for CY 2013

is $96,412, CY 2014 is $39,848, and CY 2015 is $23,747. Phase 1 for Metro Manila

shows a difference for inpatient beneficiary out-of-pocket costs for CY 2013 as $41,476,

CY 2014 as $26,775, and CY 2015 as $22,882. Phase 2 for Cavite City, Cavite,

difference in out-of-pocket costs for CY 2013 is $36,879, CY 2014 is $41,745, and CY

2015 is $852. Phase 3 for Iloilo City, Iloilo, difference in out-of-pocket costs for CY

2013 is $18,058, CY 2014 is $11,328, and CY 2015 is $2,013.

$50,000

$100,000

$150,000

$200,000

$250,000

CY2011 CY2012 CY2013 CY2014 CY2015

PH1 PH2 PH3 Grand Total

112

Table 13

Difference of Outpatient Beneficiary Out-of-Pocket Cost

Demo city/phase

CY2011

CY2012

CY2013

CY2014

CY2015

PH1

$211,087

$169,253

$41,476

$26,775

$22,882

PH2

$144,960

$45,562

$36,879

$1,745

$852

PH3

$11,983

$17,225

$18,058

$11,328

$2,013

Grand total

$368,029

$232,041

$96,412

$39,848

$25,747

Figure 13. Difference of outpatient beneficiary out-of-pocket cost.

Summary

Chapter 4 contained the statistical findings resulting from an examination of the

survey responses from Metro Manila, Cavite, and Iloilo. The findings centered on three

RQs and an analysis of claims data organized as such. Results of the three RQs provided

evidence for a positive relationship between Demonstration recipients’ financial burdens

$50,000

$100,000

$150,000

$200,000

$250,000

$300,000

$350,000

$400,000

CY2011 CY2012 CY2013 CY2014 CY2015

PH1 PH2 PH3 Grand Total

113

and satisfaction with the Demonstration as well as their confidence that their health care

needs would be met. However, when assessing age and gender, there was little evidence

to suggest that Demonstration recipients had different opinions of the health care finder

functionality or satisfaction requesting a waiver based on their age or gender. Conversely,

this indicated that men and women of all ages were similarly satisfied with the health

care finder functionality and ability to receive a waiver to contact a nonapproved

Demonstration provider. The results also showed that participants were highly satisfied

with the Demonstration overall and were much more satisfied with their health care

following the Demonstration. However, these same participants also felt that their out-of-

pocket expenses were higher than before the Demonstration. An analysis of the claims

data for the beneficiary out-of-pocket costs for inpatient and outpatient care showed that

the beneficiaries’ out-of-pocket costs increased in demonstration areas and decreased in

non-demonstration areas; thus, indicating a preference for a closed network. In Chapter 5,

these findings will be explored in terms of what is already known, compared to my

expectations, and assessed for interpretation and suggestions for further research.

114

Chapter 5: Discussion, Conclusion, and Recommendations

Introduction

The aim of this research was to examine the efficacy and acceptability of

employing an alternative method for the delivery of health care by increasing access and

reducing beneficiary out-of-pocket costs while maintaining quality health care. The DHA

implemented a closed network under the DoD TRICARE Philippine Demonstration.

Eligible beneficiaries for the Demonstration—TOP Standard beneficiaries, consisting of

retirees and their family members, TRICARE for Life, family members of active service

members, and TYA, Reservist, and National Guard who resided in the Philippines—

received care in the designated demonstration areas and used approved Demonstration

providers. Through this research I evaluated the following Demonstration areas: Metro

Manila and its surrounding areas of Cavite City, Cavite, and Iloilo City, Iloilo.

Interpretation of the Findings

The results of the study show that that men and women of all ages were similarly

satisfied with the health care finder functionality and ability to receive a waiver to receive

health care from a nonapproved Demonstration provider. However, these same

participants also felt that their out-of-pocket expenses were higher than before the

Demonstration.

According to Andersen et al. (2010) and Andersen and Newman (1973/2005), an

individual’s age and gender could play a role in the propensity for accessing health

services. However, when assessing the correlation of age and gender for the

Demonstration in Metro Manila, Cavite City, Cavite, and Iloilo City, Iloilo, there was

115

little evidence to suggest that Demonstration recipients had different opinions of health

care finder functionality or satisfaction requesting a waiver based on their age or gender.

Conversely, this indicated that men and women of all ages were similarly satisfied with

the health care finder functionality and ability to obtain a waiver to receive health care

from a nonapproved Demonstration provider.

The results also showed that participants were highly satisfied with the

Demonstration overall and were much more satisfied with their health care following the

implementation of the Philippine Demonstration. However, these same participants also

felt that their out-of-pocket expenses were higher than before the Demonstration.

Interestingly, an analysis of beneficiaries’ out-of-pocket costs for inpatient and outpatient

health care services for CYs 2011 and 2012 showed that beneficiaries’ out-of-pocket

costs were higher in the years before the Philippine Demonstration. After the

implementation of the Philippine Demonstration, beneficiaries’ out-of-pocket costs for

inpatient and outpatient health care services decreased for CYs 2013, 2014, and 2015.

Beneficiaries’ actual out-of-pocket costs based on inpatient and outpatient claims

data January 1, 2011, through December 31, 2015, showed a decrease in the amount

beneficiaries paid out-of-pocket in Metro Manila and its surrounding areas, Cavite City,

Cavite, and Iloilo City, Iloilo to providers not accepting TRICARE. Before the Philippine

Demonstration, the providers in the Philippines did not accept what TRICARE

reimbursed, which subjected beneficiaries to balance billing. Beneficiaries’ out-of-pocket

costs collectively were $147,320 (CY 2011) and $143,682 (CY 2012) for inpatient

116

services, shown in Table 8, and $368,029 (CY 2011) and $232,041 (CY 2012) for

outpatient services, shown in Table 11.

Schoen et al. (2010) argued that the design of the insurance plan influenced access

to medical care and costs. Upon implementation of the Philippine Demonstration, the

approved Demonstration providers agreed to accept the TRICARE reimbursement as

payment in full, eliminating balance billing for beneficiaries. After implementing the

Philippine Demonstration, beneficiaries’ out-of-pocket costs collectively were $65,529

(CY 2013), $172,507 (CY 2014), and $184,158 (CY 2015) for inpatient costs, as seen in

Table 9, and $91,521 (CY 2013), $157,590 (CY 2014), and $192,797 (CY 2015) for

outpatient costs, as seen in Table 12. Beneficiaries who chose not to receive care from

approved Demonstration providers’ out-of-pocket costs collectively were $74,695 (CY

2013), $51,312 (CY 2014), and $47,347 (CY 2015) for inpatient services, as depicted in

Table 10, and $96,412 (CY 2013), $39,848 (CY 2014), and $25,747 (CY 2015) for

outpatient services, as depicted in Table 13. Table 12 and Table 13 are rather revealing in

a few ways. These results are indicative that the beneficiaries’ out-of-pocket costs

actually decreased. Andersen and Newman (1973/2005) predicted that access to health

care would increase if the patient out-of-pocket expenses decreased. As the Philippine

Demonstration matured, more beneficiaries started accessing care from approved

Demonstration providers, indicating the efficacy and acceptability of a closed network

changing the health care model for delivering care and for reimbursement methodology.

117

Limitations of the Study

In reviewing the literature, there was no research found that integrated the MHS

with a foreign health care system to create an alternative for providing health care. This

research was limited to the patient experience. As previously mentioned, based on claims

data, out of approximately 10,000 TRICARE Standard beneficiaries living in the

Philippines, only 4,909 beneficiaries used the TOP Standard option before implementing

the Philippine Demonstration. The predisposing factors used to evaluate the Philippine

Demonstration were limited to age, gender, and location. The beneficiary categories were

collapsed so the TOP Standard categories or options could not be determined, rendering it

hard to decide which beneficiary category was receiving the most care. Additionally, it

was impossible to determine the number of beneficiaries accessing their TRICARE

benefit under the Philippine Demonstration as well as the beneficiary category.

Recommendations

Further research should be done to determine the full effects of the Philippine

Demonstration, evaluating the impact of predisposing factors for TOP Standard

beneficiaries accessing health care and trending actual health care expenditures paid by

the DoD, beneficiaries, and other health insurance. It should promote wellness programs

for TOP Standard beneficiaries that would result in better health outcomes. An approach

to the study from the perspective of the providers could also prove beneficial.

Implications for Positive Social Change

The TRICARE beneficiaries in the Philippines experienced a positive social

change at the individual level after effectively blending the MHS and the Philippines

118

health care system, which created a closed network, resulting in decreased out-of-pocket

costs and increased access to quality health care. The contractor responsible for

administering the Philippine Demonstration recruited providers who agreed to accept

what TRICARE reimbursed, which included the beneficiary’s deductible and cost share

as payment in full, alleviating balance billing. Furthermore, the providers agreed to

undergo a stringent credentialing process that assured providers provided quality health

care. This research has promising implications for the DoD and beneficiaries who live

and receive health care overseas. There is a high probability that health care expenditures

could decrease for the DoD and beneficiaries.

Positive social change for increasing access to quality health care and decreased

out-of-pocket costs can be accomplished when the DHA implements a closed network of

providers throughout various countries outside the continental United States, whereby

beneficiaries are not subjected to balance billing and have access to providers who

provide quality health care and accept what TRICARE reimburses as full payment.

Regardless of the balance billing law, the standards for health care, and access to

numerous providers in the United States, military retirees and their family members and

family members of active service members are subjected to higher out-of-pocket costs

and decreased access to quality health care in countries outside the United States.

Therefore, the DHA’s role in social change is to develop a strategy and legislation to

decrease out-of-pocket costs and increase access to quality health care for integrating the

MHS with foreign health care systems. Socializing this thought process will take time to

effectuate social change worldwide.

119

Conclusion

In conclusion, the results of the three RQs provide evidence of a positive

relationship between Demonstration recipients’ financial burdens and satisfaction with

the Demonstration as well as their confidence that their health care needs would be met

under the Philippine Demonstration. Although some beneficiaries believed that their out-

of-pocket costs increased, an actual analysis of the claims data showed a decrease in

beneficiaries’ out-of-pocket costs. Therefore, a closed network of approved

Demonstration providers proved to be beneficial for increasing access to quality health

care while decreasing beneficiary out-of-pocket costs.

Integrating Health Care Systems to Maintain Quality Care and to Manage Cost.

Students also viewed