Assignement introduction Demographic Characteristics and Response Preparedness

Section 1: Foundation of the Study and Literature Review

Ebola virus disease (EVD) outbreaks have been a major public health concern for

decades because of the high morbidity and mortality associated with the disease (Centers

for Disease Control and Prevention [CDC], 2018). For example, during the 2014-2016

EVD epidemic in West Africa, 11,310 fatalities were confirmed by the CDC, and it is

possible there were many additional cases that were not confirmed (2017). Though new

research is emerging, currently there is no cure for the infection (CDC, 2018). It therefore

remains the responsibility of those in public health to attempt to prevent—or at the very

least limit—the spread of future EVD outbreaks.

Since there is no cure for EVD infection, only supportive therapy such as IV

fluids, antidiarrheal medication, and pain medication can be used (Chertow, Kleine,

Edwards, Scaini, Giuliani, & Sprecher, 2014). This therapy can be costly, especially if

the outbreak is located in a country that already has minimal healthcare resources.

Without sufficient esources, countries are forced to ask for assistance. During the 2014-

2015 outbreak in West Africa, outside aid workers helped control the spread of the

disease (Chertow et al., 2014). Though necessary, this added both additional cost and

additional risk in controlling the outbreak, since many of these workers became ill

themselves (Chertow et al., 2014). This cost, in addition to the loss of workers during the

outbreak, seriously damaged many countries’ economies during the epidemic (Gostin &

Friedman, 2015).

External aid workers and travelers also have the potential to carry the disease

back to their own countries (Rothstein, 2015). Many countries instated quarantine

measures during the 2014-2015 West African EVD epidemic, but some workers and

travelers who were exposed to the virus had already traveled back to their home country

by the time these measures were in place (Rothstein, 2015). Diseases can spread more

rapidly from one country to another as modern modes of transportation, such as flying

and trains, become more ubiquitous throughout the world (Rothstein, 2015).

In America specifically, eleven individuals who had traveled from West Africa

were treated with EVD, and of those individuals some had returned to the United States

before developing symptoms (CDC, 2017b). It is important to remember that it is

possible to spread EVD before symptoms begin or when only mild symptoms are

presenting (CDC, 2017b). Several Americans died as a result of the 2014-2016 outbreak,

including a doctor and other aid workers (CDC, 2017b).

While there is still a lot that remains unknown about EVD, it is well understood

that the virus is spread through body fluids (CDC, 2018). Individuals who protect

themselves from contaminated body fluids through proper hygiene are more likely to

avoid infection (CDC, 2018). Increasing proper hygiene campaigns in at risk countries

like Liberia may be an appropriate step to prevent another deadly epidemic. To do this

effectively, it is important to ascertain how much individuals already know about both the

virus and the measures that can be taken to prevent the spread of infection.

A survey conducted in Monrovia, Liberia, during the 2014-2016 outbreak asked

individuals what they knew about EVD and what they knew about hygiene measures that

could be taken to avoid infection (Humanitarian Data Exchange, 2017). Within, this

population, those who were employed will be looked at specifically in this study

(Humanitarian Data Exchange, 2017). Around 97% of the population in Liberia is

employed (CIA, 2018). Employed individuals are more likely to be exposed to the

disease because they are leaving their homes, and they are a specific population that can

be targeted through future educational campaigns (CDC, 2018).

Ultimately, if there is another EVD outbreak in Liberia, improving preventative

measures among employed Liberians may protect both individuals and the economy by

limiting the spread of disease and thus reducing morbidity and mortality associated with

it. Basic hygiene improvements may even reduce the morbidity and mortality associated

with other diseases (CDC, 2018). Finally, employers may realize from this work that they

have an opportunity to reduce the spread of disease through workplace education.

This chapter is organized into six sections. The first section provides the

background of this study. The second section provides the problem statement, and the

third section provides the purpose of the study. The fourth section includes both the

research questions and the hypotheses. The fifth section examines the theoretical

framework of the study. The sixth section covers the nature of the study. Final parts of

this section include the definition of terms, assumptions, scope and delimitation,

limitation, significance, and summary.

Background

Historical Perspectives on EVD

The first major outbreaks of EVD occurred during the 1970s in the Democratic

Republic of the Congo (DRC; Li & Chen, 2014). It was nearly 2 decades later before

another outbreak was reported (Li & Chen, 2014). Since the 1990s, outbreaks have

occurred in several countries, including the Democratic Republic of the Congo, Gabon,

and Uganda (Li & Chen, 2014). Most outbreaks have been isolated and have occurred in

countries near or around the equator in Africa (Li & Chen, 2014). There have been

isolated cases in other countries, including the U.S. (Li & Chen, 2014).

These cases are rare and have typically occurred when a sick animal has been

transported from Africa for research purposes (Li & Chen, 2014). A table showing the

location of previous EVD outbreaks up to the present epidemic, as well as case numbers

and fatality rates, can be seen below in Table 1.

Table 1

List of EVD Outbreaks

Country Cases Deaths Year

Democratic Republic

of the Congo (DRC)

Ongoing Ongoing 2018

DRC 8 4 2017

DRC 66 49 2016

Multiple 28652 11325 2014-2016

Uganda 6 3 2012

DRC 36 13 2012

Uganda 11 4 2012

Uganda 1 1 2011

DRC 32 15 2008

Uganda 149 37 2007

DRC 264 187 2007

South Sudan 17 7 2004

Republic of Congo 35 29 2003

Republic of Congo 143 128 2002

Republic of Congo 57 43 2001

Gabon 65 53 2001

Uganda 425 224 2000

South Africa 2 1 1996

Gabon 60 45 1996

Gabon 37 21 1996

DRC 315 250 1995

Ivory Coast 1 0 1994

Gabon 52 31 1994

South Sudan 34 22 1979

DRC 1 1 1977

South Sudan 284 151 1976

DRC 318 280 1976

Note. List of EVD Outbreaks. CDC. (2018b). EVD Distribution Map: Cases of EVD in

Africa 1976 to 2017. https://www.cdc.gov/vhf/EVD/history/distribution-map.html

Ecology and Transmission of EVD

The primary means of initial transmission of EVD during an epidemic has been

zoonotic (Li & Chen, 2014). Swine, monkeys, and humans are all known carriers and

fruit bats have been confirmed as reservoirs for the virus (Li & Chen, 2014). Often,

humans come into contact with these animals when seeking a food source. This is termed

a spill over event which means that the virus can now cross species into humans (CDC,

2016b). Monkeys in particular are a common food source in many African countries

(Rizkalla, Blanco-Silva, & Gruver, 2007). Also called “bush meat,” monkeys are found

in abundance in some areas and are relatively easy to hunt because of their large size

(Rizkalla, Blanco-Silva, & Gruver, 2007). In some instances, bats may also be consumed

as food (Rizkalla, Blanco-Silva, & Gruver, 2007).

Known Risk Factors for Ebola Virus Disease

Consumption or interaction with bush meat or wild animals in Africa is one

known ways that EVD may be introduced into a population (Rizkalla, Blanco-Silva, &

Gruver, 2007). Hunters who kill and consume or sell bush meat may not realize the

animal is sick, since some animals can be contagious without displaying symptoms.

However, some hunters may still sell or consume a sick animal out of desperation for

food or income. Individuals who buy these animals in a market may not see any visual

signs that the meat is diseased (Rizkalla, Blanco-Silva, & Gruver, 2007).

Caregivers for the sick are another at risk group. Specifically, women are more at

risk for contracting EVD (Suwantarat & Apisarnthanarak, 2015). This is because,

culturally, women are given the task of tending to the children and also often assume the

responsibility of caring for those who are unable to care for themselves within the

extended family (Suwantarat & Apisarnthanarak, 2015). Extended family often includes

primarily the older, but may also include younger siblings in some situations (Suwantarat

& Apisarnthanarak, 2015).

Proper precautions to prevent transmission may not be taken by caregivers for

several reasons. In many areas, access to healthcare is limited or costly (Boozary, Farmer,

& Jha, 2014). Second, because EVD mimics the symptoms of many other diseases, it is

possible that some caretakers assume their patient is sick with something other than EVD

(Beeching, Fenech, & Houlihan, 2014). Even if EVD is suspected, limited resources may

inhibit a caretaker’s ability to protect themselves (Gatherer, 2014).

Caregivers outside of the immediate family may also be more at risk for

contracting EVD. Healthcare workers, such as doctors and nurses, those working in

public health, and home healthcare providers such as midwives, are more at risk for

exposure. This population is more at risk compared to the general population simply

because of the quantity of ill people that they care for on a daily basis. This is to be

expected during outbreaks anywhere, but it is important to note that certain additional

challenges existed in many areas where the 2014-2016 outbreak occurred. It can be

challenging and time consuming to put protective measures in place in areas where the

healthcare infrastructure is not well developed. Bleach, for example, is one of the few

disinfectants that can effectively kill EVD and not all small clinics in West Africa had or

have enough bleach on hand to deal with an outbreak (Gee & Skovdal, 2017). PPE or

proper protective gear for doctors and nurses should ideally include a complete body

covering, including shoes, double gloves, and a face shield (Gee & Skovdal, 2017). These

are not items that are commonly available or routinely worn when available in most

clinics throughout Africa (Gee & Skovdal, 2017). Therefore, these steps for protection

may not be in place when the first patients arrive (Gee & Skovdal, 2017). Sometimes

individuals are misdiagnosed as well, thus leading to an even further delay in healthcare

providers taking correct protective measures and thus increasing the chance of

nosocomial infections (Gee & Skovdal, 2017).

Finally, some people are at risk who are exposed to more body fluids than others.

Children who breast feed are more at risk of contracting the disease from the mother

(Bausch, et al., 2007). Men who engage in intercourse after recovering from the disease

are more likely to spread the disease to their partners through their semen (Bausch, et al.,

2007). Healthcare workers who tend to the body fluids of patients can be more at risk,

such as nurses who change bedding or bed pans (Suwantarat & Apisarnthanarak, 2015).

Similarly, anyone who prepares a body for a funeral and comes into contact with body

fluids through that process may be more likely to contract the disease (Pandey, Atkins,

Medlock, Wenzel, Townsend, Childs, & Galvani, 2014).

Problem Statement

In 2014, an EVD outbreak erupted in Western Africa (CDC, 2017). Three

countries, Liberia, Sierra Leone, and Guinea, experienced the highest number of cases

from the epidemic with the greatest degree of morbidity and mortality coming from

Liberia (CDC, 2017). At one point during the epidemic in Liberia, over 800 individuals a

week were being diagnosed with EVD (CDC, 2017). Though EVD is difficult to contain,

the spread of the disease may be slowed down through public health education campaigns

(CDC, 2018).

One possible public health campaign could have targeted working adults in

Liberia. In Liberia, over 97% of the adult population report working in some capacity

(CIA, 2018). A large portion of the population could, therefore, be reached through

workplace EVD education initiatives. Currently, there is no research that shows how

much workers know about EVD and about preventative hygiene.

However, there is research showing how non-pharmaceutical public health

campaigns have reduced the spread of disease during other pandemics. In an article by

the CDC (2006), the effectiveness of public health campaigns in major disease outbreaks

were explored. The article looked only at pandemics in the last century and how public

health responded to them. A pandemic is a disease outbreak that effects a large

geographic area including more than one country (CDC, 2006). Specifically, the CDC

looked at the 1918 flu pandemic and SARS pandemic in Eastern Asia. The 1918 flu

pandemic was explored in the article because it was the worse pandemic since the Black

Plague. The 1918 flu public health campaign was also the first public health campaign in

America that specifically looked at methods of public health to slow a pandemic (CDC,

2006). The SARS pandemic was explored because it is one of the most recent influenza

pandemics, and because extensive public health measures were used to prevent it from

becoming worse (CDC, 2006).

What the article found is that in some past epidemics, such as during the 1918 flu

pandemic, quarantine measures were not enough to contain the disease and keep it from

spreading to other areas (CDC, 2006). The article suggests that more than just

quarantining has to be used to stop an outbreak of any kind (CDC, 2006). With SARS

and other recent flu outbreaks throughout the world, research has investigated exactly

what public health measures may work to reduce spread of disease beyond just

quarantining. One multivariate case control study in Hong Kong found that those who

washed their hands more than 10 times a day were less likely to contract respiratory

illnesses (CDC, 2006). EVD is not considered to be a respiratory illness, but like SARS

and other respiratory illnesses, the spread of EVD has been proven to be reduced through

hand washing (CDC, 2018). Hygiene education and public health education about

diseases and how they spread may therefore be an essential part of stopping

communicable disease outbreaks.

Another aspect of this problem is that such a high rate of illness not only resulted

in high mortality rates, it also had a damaging effect on the country’s economy (Adegun,

2014). Economies suffer during epidemics for many reasons. Businesses lose workers,

people are afraid to get out into the public, and often tourism and trade also decline

(Adegun, 2014). During the 2014-2016 outbreak, an estimated $2.2 billion was loss in

GDP in Liberia, Sierra Leone, and Guinea (CDC, 2016). Significant financial losses

occurred in the areas of private sector growth, agricultural production, and cross- border

trade (CDC, 2016). To ensure the stability of Liberia’s economy in case of a future

epidemic, it is imperative that measures be taken to protect the working class. These same

measures may have global implications, since no country is totally immune to the

possibility of an epidemic.

The endemic countries in Africa were not the only countries whose economies

were affected by the outbreak. The United States alone donated over $2 billion to efforts

in Western Africa to contain the disease (CDC, 2016). Countries like the United States,

Great Britain, and Germany paid out large amounts of money to ensure that the affected

countries had the resources they needed (CDC, 2016). Without these resources, the

disease could have spread from Africa to the donor countries, so it was seen as a

necessary expense (CDC, 2018).

Widespread food shortages were another problem that occurred during the EVD

outbreak of 2014-2016 (CDC, 2016). As mentioned, many individuals were unable to

work during the epidemic, leading to a loss in agricultural production (CDC, 2016). The

loss of agricultural production was severe enough that it led to food insecurity throughout

the affected regions (CDC, 2016). Food insecurity persisted as a serious issue long after

the epidemic ended (CDC, 2016). It has taken these countries a long time to recover in

this particular area (CDC, 2016).

It is also important to note the impact the EVD outbreak had on healthcare

systems throughout Western Africa. During the outbreak, 881 healthcare workers became

infected with the disease, of which 513 died (CDC, 2016). It is estimated that overall

healthcare services were reduced by up to 50% (CDC, 2016). This created multiple

problems. First, during the outbreak it was difficult to adequately treat the large number

of incoming EVD patients in healthcare facilities when the facility was short staffed (Gee

& Skovdal, 2017). Secondly, other diseases became less of a priority, leading to a surge

of deaths caused by HIV/AIDS, malaria, and tuberculosis. Some of these deaths may

have been because some patients with other diseases were afraid of contracting EVD if

they sought care at a facility that also saw EVD patients (Gee & Skovdal, 2017). For

patients with other diseases who did seek care, it is believed that these patients were no

longer getting adequate care because care facilities were short staffed. It has been

estimated that an additional 10,600 lives were lost to HIV/AIDS because of inadequate

care during this time (CDC, 2016).

Finally, this epidemic had a huge impact on children in Western Africa. Nearly

20% of all EVD cases were children (CDC, 2016). It is also estimated that after the

epidemic ended, over 17,000 orphans were left as a result of the disease (CDC, 2016).

During the outbreak, children suffered academically with some areas cancelling school

for close to 30 weeks (CDC, 2016). Children also missed out on routine health checks

and vaccinations during this time, making them more vulnerable to other infections

(CDC, 2016).

This research filled a gap in understanding by specifically looking at whether or

not those working in Liberia during the epidemic had adequate knowledge about EVD

and its prevention. Literature searches yielded no results when looking for data specific to

this population group in Liberia. However, researchers have cited the need for improved

education in Liberia as gaps in current research (Christie, Davies-Wayne, Cordier-

Lasalle, Blackley, Laney, Williams,... & Ladner, 2015). Specifically, Christie et al.

(2015) found that knowledge about EVD spreading through semen needed to be

improved in Liberia. In Nigeria, volunteer health advisors who were already working in a

healthcare field during the epidemic showed gaps in EVD knowledge (Patel, Pharr,

Ihesiaba, Oduenyi, Hunt, Patel, ... & Ezeanolue, 2016). It is possible that there is a similar

deficit of knowledge among workers in another West African nation, Liberia.

Literature searches also revealed that a similar approach to understanding EVD

knowledge in the United States yielded interesting results. First, those in the United

States with more knowledge about EVD were considered to have a lower risk of

contracting the disease if it came to the United States (Rolison, & Hanoch, 2015).

Second, research found that increasing knowledge about EVD led to individuals taking

the disease more seriously (Rolison, & Hanoch, 2015).

Social change may occur when businesses realize that they have an opportunity to

educate a large portion of the Liberian population. Secondarily, social change may occur

if individuals follow the guidance/training offered by their organizations, resulting in less

illness and fewer businesses losing workers in the future. When businesses are able to

function during an epidemic, the economy suffers less loss and thus many vulnerable

populations who rely on economic assistance remain unharmed.

Purpose of the Study

The aim of this study was to investigate the impact of employed adults’

socioeconomic and demographic factors on the use of EVD preventative measures. This

study could be used to inform future public health campaigns among working adults.

Workplaces offer a way to get information to large groups of individuals at a time (Gee &

Sokval, 2017). Workplaces can also serve as a great place to begin to enforce proper

hygiene practices that may then be done at home as well, once the individual knows what

to do (Gee & Sokval, 2017). Seeing others participate in these programs can also serve as

a form of positive peer pressure (Gee & Sokval, 2017). This study was conducted to

reveal whether or not working adults were more or less educated about EVD than those

who did not work and may therefore offer insight into whether or not workplace

education is being done.

Research Questions and Hypotheses

This study was guided by the following two research questions:

1. Is there an association between socioeconomic and demographic factors

(education, age, gender, and occupation) and frequency of employed adults

who report the use of preventative hygiene practices (use of hand sanitizer and

bleach) in Monrovia, Liberia?

a. H1

: There is no statistically significant association between

socioeconomic and demographic factors and frequency of employed adults

who report to use preventative hygiene practices in Monrovia, Liberia.

b. H1

: There is a statistically significant association between socioeconomic

and demographic factors and frequency of employed adults who report to

use preventative hygiene practices in Monrovia, Liberia.

2. Is there an association between socioeconomic and demographic factors

(education, age, gender, and occupation) and frequency of employed adults

who report understanding the basic nature of the EVD (common symptoms

and how it is spread) in Monrovia, Liberia?

a. H1

: There is no statistically significant association between

socioeconomic and demographic factors and frequency of employed adults

who report understanding the basic nature of the EVD in Monrovia,

Liberia.

b. H1

: There is a statistically significant association between socioeconomic

and demographic factors and frequency of employed adults who report

understanding the basic nature of the EVD in Monrovia, Liberia.

Nature of the Study

A cross sectional survey was used as the secondary data source for this research.

The survey was conducted by a Liberian NGO, called Parley, which is a partner with the

United Nations which, in turn, runs the Humanitarian Data Exchange (2017). According

to the Humanitarian Data Exchange (2017), the survey selected 77 communities from 15

wards in the capital of Monrovia (). The number of communities chosen from each ward

was proportional to that ward’s population size, so no ward was under- or

overrepresented in the survey (). Households surveyed from each community were

randomly chosen (). In the survey, individuals were asked about their awareness of EVD

as well as the steps they had taken to prevent EVD infection (). The study also recorded

the individuals’ occupation, gender, and education ().

The survey was conducted from December 2014 to January 2015 (Humanitarian

Data Exchange, 2017). At this point, EVD had been present in Monrovia for

approximately six months (CDC, 2018). The survey asked questions about hand hygiene

and the nature of the EVD virus, for example, how it is spread, how to know if you have

it, etc. (Humanitarian Data Exchange, 2017). These results were divided based on

education, age, occupation, and gender to see if there was a correlation between any of

these sociodemographic/socioeconomic factors listed and knowledge about EVD and the

prevention of EVD through improved hygiene.

Population

The population of the survey included representation from every community in

Monrovia (Humanitarian Data Exchange, 2017). Monrovia is the capital of Liberia and

has a population of over 1 million, nearly a fourth of the country’s total population (CIA,

2018). Those living in Monrovia have faced many challenges, including civil wars (CIA,

2018). The most recent war ended in 2003, but when the EVD epidemic struck Monrovia

in 2014, the city had not yet fully recovered (CIA, 2018). The war combined with other

health issues, like high fertility rates, likely contributed to the fact that a majority of the

population (almost 60%) was under the age of 25 (CIA, 2018).

Currently, Liberia is considered a low-income nation, though a majority of the

population does work (CIA, 2018). Major industries include mining and agriculture (CIA,

2018). A majority of the population also lives without access to improved sanitation,

which is a factor being explored (CIA, 2018). In urban areas like Monrovia, it is

estimated that 72% of the population does not have access to improved sanitation like

waste disposal and clean water (CIA, 2018).

Theoretical Foundation for the Study

The theory of reasoned action/planned behavior is applicable to EVD prevention

(Fishbein, 1979). This theory assumes that most health-related decisions are rational

decisions that lead to planned behavior (Bensley & Brookins-Fisher, 2003). Preventative

hygiene and understanding how disease is spread may lead to better informed citizens

and thus better planned actions/behaviors (Bensley & Brookins-Fisher, 2003). Therefore,

the first step is to know which demographic groups of workers need to be a focus of new

interventions so that beliefs about EVD can be changed. Second, it is important to know

which demographic groups of workers are at risk, so EVD infection may be prevented

through targeted educational campaigns designed to meet the specific issues (CDC,

2018).

Under this theory, specifically the situational factors, attitudes, intentions, and

subjective norms about hygiene practices among workers will be explored to determine

which demographic groups of workers need to be targeted (Bensley & Brookins-Fisher,

2003). First, certain occupations may have situational factors that cause those groups to

be more at risk for contracting the disease than others (CDC, 2018). Miners, for example,

are in close proximity to bats and may be more at risk of contracting the disease, and

funeral and healthcare workers may come into more contact with body fluids than other

occupations (CDC, 2018). This may mean that these individuals know more about EVD

and thus use more preventative hygiene techniques then other occupations which are less

likely to be exposed.

Attitudes and intentions may also vary across different demographic groups of

workers based on individual level of education. Some occupations require educational

experience while others do not. A study by Oladimeji, Gidado, Nguku, Nwangwu, Patil,

Oladosu,... & Musa, found that healthcare workers with more education, such as doctors,

were more likely to practice good hygiene (2015). Healthcare workers with less

education did not practice hygiene techniques adequately (Oladimeji et al., 2015).

Finally, subjective norms about hygiene may vary among different ages, genders,

and different occupations. Occupations that earn less money may not have as a high of a

standard for preventative hygiene, since it can be costly to purchase things like soap,

clean water, and hand sanitizer (Doocy & Burnham, 2006). Whether or not hygiene

practices are more of a norm among a certain gender or age was explored as well. In one

scenario, bathroom handwashing rates in a service station were recorded (Judah, Aunger,

Schmidt, Michie, Granger, & Curtis, 2009). It was discovered that men in particular were

less likely to wash their hands than women and thus needed to be targeted differently

with public health campaigns (Judah, Aunger, Schmidt, Michie, Granger, & Curtis,

2009). Men were more likely to wash their hands if a picture reminding them to wash

their hands contained disgusting facts about germs (Judah, Aunger, Schmidt, Michie,

Granger, & Curtis, 2009). Women did not need to be disgusted in order to wash their

hands more often? (Judah, Aunger, Schmidt, Michie, Granger, & Curtis, 2009). For

women, an advertisement reminding them to wash their hands was sufficient (Judah,

Aunger, Schmidt, Michie, Granger, & Curtis, 2009). Finally, different ages may be more

at risk for disease due to socioeconomic factors. According to Maharaj (2012), the older

may be more at risk for disease due to poverty and lack of access to healthcare.

Literature Search Strategy

For the literature review, two research strategies were used. First, several

databases (list them all here, database1, database2, database3, and Google Scholar) were

used to find peer-reviewed scholarly articles on the topic. The following keywords were

used: EVD in Liberia, the history of EVD, EVD, the history of Liberia, the Theory of

Reasoned Action/ Planned Behavior, and the culture of Liberia. All searches were limited

to the years 2008 to 2020, though some primary sources outside of this range were

included. Little research was found on how much employed adults knew about the EVD

virus in Liberia. Therefore, I explored how these factors were associated with EVD virus

fatality in previous outbreaks in other countries in Western Africa.

Secondly, the CDC, CIA World Factbook, and the World Health Organization

were accessed directly. The CDC and WHO both have pages dedicated to information on

EVD that can be searched by country. The CIA World Factbook has a page on Liberia,

from which some information was derived. All of these sites are updated regularly.

Information from these sites was not older than 2014.

Finally, census data were used from the 2008 household census in Liberia. These

were collected and published online through the Liberia Institute of Statistics and Geo-

Information Services (LIS-GIS). This census was government sponsored and therefore

had reliable oversight in its distribution and publication.

Introduction

EVD has been cited by the WHO (2018) as one of the ten most serious diseases in

the world today. This disease has baffled scientists for decades and there still remain

many unanswered questions about the disease (WHO, 2018). Fortunately, there are some

basic facts about the nature of the disease, how it is spread, and how it may be prevented

that offer some insight into possible prevention techniques (WHO, 2015). All of these

topics were explored in further detail throughout this section.

About EVD

EVD is a violent and often lethal infection caused by a virus from the family

Filoviradae (Beeching, Fenech, & Houlihan, 2014). There are two types of hemorrhagic

fevers included in this family: Marburg and EVD (Li & Chen, 2014). Though similar,

EVD has been the virus of most concern over recent years since the 2014 outbreak in

West Africa (Li & Chen 2014). There are five known species of EVD viruses: Zaire,

Sudan, Reston, Tai Forest (which was also known as Cote d’Ivoire EVD virus until

2010), and Bundibugyo (Li & Chen, 2014). The Sudan and Zaire strains are known to be

the predominant species associated with recent epidemics (Li & Chen, 2014). The fatality

rates for all five strains vary but may be as high as 90% in some instances (Li & Chen,

2014).

In order to contract the disease, one must come into contact with infected body

fluids from an EVD patient (Beeching, Fenech, & Houlihan, 2014). Once exposed, it

takes 1- 21 days to develop symptoms depending on age, health, and strain of the virus

(Beeching, Fenech, & Houlihan, 2014). At first, symptoms may include high fever,

fatigue, nausea, abdominal pain, and vomiting (Beeching, Fenech, & Houlihan, 2014).

This makes the virus difficult to distinguish from many other diseases including malaria

and cholera (Beeching, Fenech, & Houlihan, 2014). After these symptoms manifest, a

patient may then develop unexplained bleeding or bruising (Beeching, Fenech, &

Houlihan, 2014). Bleeding and bruising are indicators that the patient has a hemorrhagic

virus and further testing must be done to determine what hemorrhagic fever has presented

(Beeching, Fenech, & Houlihan, 2014).

A patient is considered contagious as soon as he or she has presented any of the

previously mentioned symptoms (Beeching, Fenech, & Houlian, 2014). Once symptoms

develop, the virus multiplies inside the body rapidly (Bausch, Towner, Dowell, Kaducu,

Lukwiya, Sanchez, & Rollin, 2007). The virus may be present once a patient is

symptomatic in the saliva, mucous, tears, breast milk, and semen (Bausch, et al., 2007).

The virus may survive in body fluids even after death and has been proven to survive in

semen for up to 40 days after symptoms have ended (Bausch et al., 2007).

The 2014- 2016 EVD Outbreak in West Africa

In December 2013, Guinea reported its first cases of EVD to the World Health

Organization, and in March 2014 the WHO released a statement about the outbreak that

would go on to kill thousands across primarily three countries (Gatherer, 2014). The

disease spread west from Guinea into Liberia and Sierra Leone. Cases were reported in

Monrovia, Liberia in 2014. This is unique since in the past most EVD outbreaks have

occurred in remote settings (Gatherer, 2014). The spread of the epidemic into an urban

center presented new challenges to managing the disease in a dense population (Gatherer,

2014).

Τhis particular epidemic was the largest EVD epidemic because of the number of

cases reported as well as because of the large geographic area that was affected (Gomes,

Piontti, Rossi, Chao, Longini, Halloran, & Vespignani, 2014). Though response was

rapid, the disease continued to spread rapidly throughout Western Africa (Gomes, et al.,

2014). After months of efforts by the World Health Organization and others to contain

the virus and set up EVD Treatment Centers, the epidemic was declared over in

December 2016 (WHO, 2018). A total of over 28,000 cases were confirmed and many

more suspected (WHO, 2018).

The 2014-2016 EVD Outbreak in Monrovia, Liberia

As of the 2008 census, Monrovia, Liberia contained over a fourth of the entire

population of the country of Liberia or around 970,000 individuals (LIS-GIS, 2008). As

the capital city of Liberia, Monrovia has served for decades as the Liberian hub of

commerce and trade (LIS-GIS, 2008). However, at times Monrovia has had set backs. In

the 1990s a violent civil war took place in Liberia leaving thousands of orphans and

damaging nearly every infrastructure in Monrovia (Huband, 2013). Though the 2014

EVD outbreak occurred years later, Monrovia was still recovering from the war (Huband,

2013). As a result, in many respects the city was considerably vulnerable when the

disease hit.

In a dense population setting like Monrovia, disease has the opportunity to spread

more rapidly even when sound infrastructure is in place (Eisenstein, 2016). Individuals

are forced into closer proximity due to volume which may be increased even more so in

cities with limited resources (Eisenstein, 2016). Diseases like EVD spread quickly when

urban centers lack adequate housing and sanitation (Eisenstein, 2016). Though the city

has made great developmental strides over recent years, there are still areas in Monrovia

that could be improved (Eisenstein, 2016). EVD entering this city created a challenge for

those attempting to contain the virus. It is important to establish prevention techniques in

case this disease or others spread into the city again before the city has completely

rebuilt.

Possible Prevention Techniques- The Role of the Workplace

There are certain universal precautions that may be taken to lower the risk of

contracting EVD. During an EVD outbreak, individuals who work in healthcare or who

prepare bodies for funerals should use PPE or proper protective equipment (Suwantarat &

Apisarnthanarak, 2015). For EVD, this should include facial shields, gloves, boots, and

multiple layers of protective clothing. Individuals who work outside of healthcare should

get loved ones suspected of having the disease to a hospital immediately and not try to

devise their own PPE at home (Suwantarat & Apisarnthanarak, 2015).

During an outbreak, everyone should adhere to locally mandated curfews and

quarantine measures, report suspected cases, and practice good hygiene (Pandey et al.,

2014). Good hygiene includes the use of bleach to clean possibly contaminated items and

the use of soap and water or hand sanitizer after being in public (Gatherer, 2014).

Additionally, no one should consume or handle bush meat (Rizkalla, Blanco-Silva, &

Gruver, 2007).

The WHO (2015) investigated best practices that could be used to prevent the

spread of EVD through improved hygiene and infection control. It was recommended

that these practices be shared with healthcare workers to reduce the spread of the disease

among this population (WHO, 2015). These recommendations include steps that anyone

can take to reduce the spread of disease. This information therefore would also be

beneficial to share in many other professions. It is one hope of this research that an

educational campaign may develop to improve workplace hygiene. This paper by WHO

lays out the procedure for improving workplace hygiene.

Improving workplace hygiene would be a logical step to reduce EVD outbreaks

since according to the CIA, in 2014, the unemployment rate in Liberia was only 2.8%

(CIA, 2018). Though much of the country still falls beneath the poverty line, the majority

of its citizens are employed to some degree. Employers could reach their employees

through workplace educational programs and thus also reach a large portion of the

population (CDC, 2016). In the SARS outbreak and during the 1918 flu outbreak,

workplaces, dormitories, schools, and other confined groups were targeted by public

health workers (CDC, 2016). Hygiene programs in many instances worked to reduce

some of the illness being transferred in these confined groups (CDC, 2016).

In order to conduct a successful workplace education program, it is wise to

consider that certain groups within this population may be more or less at risk.

Individuals may be more or less at risk based on their age, gender, occupation, or

education. This assumption is supported by research done by Glynn (2015). Glynn found

in his research that there exists variability in EVD virus infection rates among different

age groups and sexes though the reason for this trend is unknown.

In addition to certain ages and sexes being more at risk, a situation report

published by WHO (2016) noted that healthcare workers were more at risk for catching

EVD. While this group may be more educated about EVD, they are still being exposed

more than the general population (WHO, 2016). It is possible that other occupations are

more at risk as well. The CDC notes that those who interact with dead bodies for example

may also be more at risk (2018).

Decreasing the spread of disease in the workplace not only saves lives, it saves

incomes. Arbogast, Moore-Schilts, Jarvis, Harpster-Hagen, Hughes, and Parker (2016)

found that hand hygiene education amongst employees significantly reduced various

types of illnesses and absenteeism in the workplace. In an epidemic, this will stabilize the

national economy by ensuring that vital businesses continue to stay open, make money,

and serve the citizens (Adegun, 2014). This also ensures that individuals are healthy

enough to keep working and thus providing financially for their families (Adegun, 2014).

Definitions

ETC or EVD treatment centers. EVD Treatment Centers abbreviated as ETCs on

most CDC and WHO maps are hospitals, clinics, or aid stations that have the ability to

adequately treat and diagnose EVD through lab techniques (CDC, 2018).

EVD or EVD. EVD commonly abbreviated as EVD is a hemorrhagic fever that

caused an epidemic in 2014-2015 throughout Western Africa (CDC, 2018).

EVD Fatality Rates. The number of individuals who died as a result of EVD in

Liberia has been calculated and will be used in this study (CDC, 2018).

Hemorrhagic fever. Hemorrhagic fevers are a unique type of illness which can

break down capillaries causing both internal and external hemorrhaging (CDC, 2018).

The subsequent loss of blood and bodily fluids can easily lead to death without medical

intervention (CDC, 2018). Though others exist, the only hemorrhagic fever being studied

in this research is the EVD virus (CDC, 2018).

PPE (Personal Protective Equipment). Personal protective equipment sometimes

abbreviated as PPE includes protective clothing which is used as standard protocol when

dealing with EVD patients (OSHA, 2018). EVD treatment centers would have access to

this equipment (OSHA, 2018).

Preventative Hygiene. For EVD, preventative hygiene practices may include

washing hands, using hand sanitizer, and the use of bleach (CDC, 2018).

Protective factors. Protective factors is a term that is used in public health to refer

to factors that may protect an individual from becoming ill, in this case with EVD (CDC,

2018). One example of a protective factor may have been access to bleach prior to the

outbreak (CDC, 2018).

Remission. Remission often refers to someone who has been cured from an illness

or is no longer demonstrating symptoms of an illness (CDC, 2018). EVD cannot be

cured, but through supportive therapy it is possible for someone to survive EVD (CDC,

2018). Something unique about EVD is that during remission, EVD patients do not have

EVD symptoms, but they may still be contagious for a period of time (CDC, 2018).

Individuals may also still have long-term health issues associated with EVD after the

disease goes into remission (CDC, 2018).

Supportive therapy. There is no known cure for EVD illness so the only way to

treat EVD symptoms is through supportive therapy such as fluids and pain medication

(CDC, 2018). What supportive therapy is used can differ based on the illness so it is

important to note what can be done for EVD treatment (CDC, 2018).

Viral load. Viral load refers to the amount of virus in a given amount of fluid

(CDC, 2018). The viral load necessary to contract EVD is very small (CDC, 2018).

Assumptions

This study was based on the following four assumptions:

1. The survey used was developed through a combined effort of MIT and the

Liberian government. It was assumed that the survey was the same one

published on the Humanitarian Data Exchange and MIT websites. It was also

assumed that the survey was distributed as reported by the Humanitarian Data

Exchange.

2. Assumptions have been made for fatality rates used in some references

throughout this work. Some cases of EVD may not have been reported due to

a lack of immediate relatives, misdiagnosis, or fear of reporting to the

government. It was subsequently assumed that these cases were not the

majority of cases reported and were ubiquitous in occurrence throughout the

country. Because it was assumed that these cases occurred in small proportion

and evenly throughout the country, it was also assumed that they do not create

bias in the data.

3. It was assumed that the doctors and treatment centers reporting EVD cases

diagnosed the disease accurately and provided the correct data to the World

Health Organization.

Scope and Delimitations

This study was based on data collected by Liberian nonprofit organizations and

the Liberian government during the 2014-2016 EVD epidemic. Specifically, these data

refer to Monrovia, the capital city of Liberia. It was a cross-sectional survey that

attempted to get a comparable response rate in each section of the city. The expected

response rate was based on population number and did not discriminate based on age,

gender, occupation, or religion. These data created certain delimitations.

1. The data used were part of a survey and therefore did not include a control

group.

2. This study constitutes an analysis of secondary data and therefore I did not

have any contact with patients or any control over data collection.

3. This study was delimited by the information that was collected by doctors in

the field.

4. The time of this study delimits the study to only the 2014-2015 outbreak in

Liberia and not anywhere else or during any other timeframe.

Limitations

This research has five limitations due to the methods used to collect data.

1. Data may be missing because some respondents could have refused to

complete the survey. The investigators did assume that they would not get a

100% response rate when they began the survey (Humanitarian Data

Exchange, 2017). The surveyors divided the city into districts and attempted

to get a comparable response rate per the population in each district to give an

adequate overview of the city as a whole without one district being over- or

underrepresented (Humanitarian Data Exchange, 2017). It is believed that this

was accomplished, based on the data and based on the investigators’

description of the data (Humanitarian Data Exchange, 2017).

2. This survey was reliant on individuals in the field to accurately collect and

report the data. Individual error may have occurred, but making the survey

short, simple, and electronic may have alleviated some of this concern

(Humanitarian Data Exchange, 2017). Investigators were also told to report

issues (Humanitarian Data Exchange, 2017).

3. The data collection occurred over several days, so it is possible that data

collection changed during that time frame.

4. As cross-sectional research, results will show correlation and not necessarily

causation (HHS, 2007).

5. As a cross-sectional survey, a single moment in time is being explored (HHS,

2007). While both populations and knowledge can be dynamic, this survey

nevertheless offers an important view of how much people knew about EVD

in the early stages of the epidemic—when public health measures are most

important (CDC, 2018).

Significance

During the 2014-2016 outbreak in Liberia, nearly 5,000 people died with over

10,000 confirmed total cases (CDC, 2018). It is imperative to the sustained progress of

Liberia to understand how much information the working class in Monrovia received

about EVD. This knowledge can inform social change by preventing similar rates of

morbidity and mortality in a future epidemic. These individuals are, in a sense, a captive

audience with their employers. Results might encourage employers to offer more EVD

education to ensure that workers stay healthy. If workers are as uneducated as the rest of

the population, then this may reveal a gap in prevention strategies that could be easily

resolved through workplace initiatives.

Social change implications include not only saving lives, but the development of

better workplace health campaigns that could prevent thousands of workers from having

to stay home. This, in turn, would protect the economy which suffered during the

epidemic. In Monrovia, 20% of all businesses closed completely during the epidemic

(Bowles, Hjort, Melvin, & Werker, 2015). Several issues occurred when businesses

closed: an economic decline, increased poverty, and over 75% of the population reported

food shortages (CIA, 2018). Education to prevent EVD could reduce the number of

individuals who become ill, thus allowing businesses to remain open and serve the

public.

Summary and Transition

Though EVD has existed for decades, the most recent outbreak in West Africa

was the most devastating, with over 11,000 confirmed EVD virus fatalities (CDC, 2017).

This quantitative study used secondary data collected by the Liberian Government,

organized by MIT, and published by the World Bank to further understand what

sociodemographic and socioeconomic factors among working adults may be associated

with EVD fatality in Liberia. The theory of reasoned action/ planned behavior was used

to further understand what roles society, communities, relationships, and individuals

played in the spread of EVD. These associations were analyzed using regression

techniques and SPSS software.

The study’s methodology and the analysis plan are described in the Section 2.

Section 2: Research Design and Data Collection

Introduction

In the previous section, I offered an overview of the current literature relevant to

EVD outbreaks in Monrovia, Liberia. A survey was conducted in Liberia during the early

phases of the most recent outbreak of EVD in Liberia in 2014 (Humanitarian Data

Exchange, 2017). The survey asked individuals about preventative hygiene and the

spread of EVD (Humanitarian Data Exchange, 2017). Through this survey, a gap in the

literature was addressed. Specifically, I used these data to examine whether or not there

are differences by sociodemographic groups in Monrovia, Liberia in the uptake of these

preventative hygiene techniques and general understanding about EVD at the beginning

of the EVD outbreak.

In this section, I will provide an outline of the research design and data collection

methodology. This will involve the data collection process, design rationale, and

instrumentation. I will also address possible ethical concerns and threats to validity. I will

explain how these concerns will be addressed, where applicable. Finally, I will

summarize what will be done to analyze the data in the next section.

Research Design and Rationale

The purpose of the 2014 primary study was to perform a population based cross-

sectional survey that explored EVD knowledge among individuals living in Monrovia,

Liberia during the 2014-2016 EVD outbreak (Humanitarian Data Exchange, 2017). A

cross-sectional survey was chosen by the groups in charge of the initial project because

they wanted a baseline understanding of what individuals living in the region understood

about EVD (Morse, 2015). I sought to use the data collected by the survey on hygiene

and basic EVD knowledge among working adults to determine whether there is a

difference in understanding among various sociodemographic groups. In this study, I

used a combination of univariate, bivariate, and multivariable analysis techniques to

determine which predictors have a significant effect on outcome and which did not.

Some challenges with the data did exist. Cross-sectional surveys do not

investigate change over time and thus can predict only association and not causation

(HHS, 2007). Another limitation of this study was the possibility of social desirability

bias (Lavrakas, 2008). Individuals may have lied on the survey to appear as though they

were using better hygiene practices than they really were (Lavrakas, 2008). The survey

attempted to reassure respondents that all responses were anonymous and that the

importance of honesty was imperative (Morse, 2015). This eliminated some of the

possible respondent bias.

Methodology

Study Population

Monrovia is on the Western Atlantic coast of the African nation, Liberia (CIA,

2018). Monrovia is the capital city of Liberia and by far the largest city in the country

(CIA, 2018). The population of Monrovia was estimated to be 970,000, according to the

most recent census conducted in 2008 (LIS-GIS). In 2003, a civil war left Monrovia

decimated (Bastian, 2014). According to the World Bank, in 2008, prior to the outbreak,

there were only 50 doctors in the entire country (World Bank, 2019). In 2014, there were

still too few doctors to meet the needs of the population, and thus, during the outbreak,

many hospitals were forced to close: many of the health workers had died (Bastian,

2014). By the time of the survey, efforts to stem the spread of EVD in Monrovia took on

two forms: public health initiatives, with a focus on prevention, and outside aid (Bastian,

2014). Specifically, groups like the World Health Organization and Red Cross began

setting up mobile ETRs or EVD treatment centers (Bastian, 2014). Additionally, these

same groups used ad campaigns and community forums to spread the word about the

epidemic (Bastian, 2014).

The primary survey attempted to gain an accurate representation of the entire

city’s population through a three-tiered sampling technique (Morse, 2015). Of the

households surveyed, a total of 1,572 individuals responded to the survey (Morse,

Grépin, Blair, & Tsai, 2016). Other than being an adult, there were no restrictions on

respondents (Morse, 2015). Respondents were therefore a mixture of genders, ages,

education levels, religions, and economic classes (Humanitarian Data Exchange, 2017).

Homes where the only adult present had an active case of EVD were excluded from the

data collection to protect the surveyors and because of the assumption that those seriously

ill would not be able to answer the survey (Morse et al., 2016).

Sampling and Sampling Procedures

Sampling of the population took place in three unique stages in Monrovia,

Liberia (Morse, 2015). Monrovia is divided into wards which are large clusters of

communities similar to the American concept of counties. From these wards, 77

communities were chosen based on their population size so representation would be

proportional throughout each ward (Morse, 2015). For example, a larger ward may have

ten communities chosen for sampling relevant to the ward’s population size while a

smaller ward may only have one community chosen for sampling (Morse, 2015). From

each community, 20 households were randomly selected to participate in the study

(Morse, 2015). From each household, only one adult (over the age of 18) respondent was

allowed to answer the survey (Morse, 2015). If an individual was not home, another

home in the sampling section was randomly chosen (Morse, et al., 2016). Overall, the

participation rate was 95% (Morse, et al., 2016).

Power Analysis

From the survey, only specific questions will be used to address the research

questions. Using all of the available sample taken from the 1572 participants ensures

adequate power and also improves validity. A priori power analysis was conducted to

determine sample size (G* Power calculator Universitaet Duesseldorf, 2010). While there

are no similar studies in this population group, a medium effect size (OR = 1.5) was

selected for regression analysis (Chen et al., 2010). The estimated sample size was 417

participants to achieve satisfactory statistical power (>0.95). Also, after the completion of

the study, a post hoc power analysis was also conducted to confirm that adequate power

was achieved.

Data Collection and Management

The MIT Lab of Governance designed and commissioned the project now

considered part of the EVD Trust of Communication and Cooperation (Morse, 2015).

MIT used Parley to conduct the door to door surveys in Monrovia, Liberia (Morse, 2015).

Parley Liberia is a non-profit organization that functions outside of the Liberian

government and works on primarily social services projects (Bloomberg, 2019).

Surveyors used handheld devices with Pendragon Software to conduct the surveys

(Pendragon, 2019).

Pendragon Software allows users to access data that is being collected remotely

via a cloud (2019). This is why the principle lab at MIT chose this software. Access to

the cloud is through a secure, remotely encrypted ODBC connection (Pendragon, 2019).

This prevents the data from being tampered with and protects the data from being altered

by anyone outside of administrators with access to the cloud (Pendragon, 2019).

Data were reviewed at least once daily by the principle investigators to make sure

the surveys were being done completely and accurately (Morse et al., 2016).

Furthermore, principle investigators checked the timestamp on the surveys and made sure

GPS was enabled on the hand-held devices (Morse, et al., 2016). The investigators then

checked the time and GPS coordinates frequently to make sure that the individuals in the

field doing the surveys were going to the correct locations (Morse et al., 2016).

Individuals who conducted the surveys went through extensive training and none of the

surveyors reported any negative experiences (Morse, 2015).

Data Accessibility and Permissions

MIT oversaw the development and implementation of the EVD response survey

(Morse, 2015). These results were then shared with the Humanitarian Data Exchange in

an attempt to share information with other organizations who may respond to the EVD

crisis (2017). The Humanitarian Data Exchange is an open data sharing platform (2017).

The website was developed by the United Nations and is managed in Hague

(Humanitarian Data Exchange, 2017).

Instrumentation and Operationalization

In order to decrease survey collection error, surveys were done electronically on

hand-held devices (Morse, 2015). The surveyors used Pendragon Survey Software to ask

respondents the questions (Morse, 2015). The survey lasted on average 45 minutes and

included questions about food security, economic security, government response, EVD

knowledge, and demographic characteristics (Humanitarian Data Exchange, 2017). Data

were secured by not including personally identifiable information per the regulations

established by the organizations collecting the data (Humanitarian Data Exchange, 2017).

Operationalization of Variables

All the predictor variables in this study were nominal (Table 2). The dependent

variables are dichotomous with yes or no categories.

Table 2

Operationalization of Variables

Dependent Variable Coding Predictors Coding

RQ1: Use of

preventive hygiene

techniques (use of

chlorine, use of hand

sanitizer, not shaking

hands) by employed

adults.

No = 0

Yes = 1

Don’t know = 97

Don’t want to answer

this specific question:

Education 0-None

1-Some ABC

2-Completed ABC

3-Some junior high

4-Completed junior

high

5-Some high school

6-Completed high

school

7-Some university

8-Completed

university

88-Other

RQ2: Understanding

of the basic nature of

the EVD (common

symptoms and how

it’s spread) by

employed adults.

No = 0

Yes = 1

Don’t know = 97

Age 18-34 = 1

35-54 = 2

55-70 = 3

70+ = 4

Do not wish to report

age: 5

Gender Male = 0

Female = 1

Other = 2

Occupation 0-None

Professional/technical

2-Clerical

3-Sales and services

4-Skilled manual

5-Unskilled manual

6-Domestic service

7-Agriculture

88-Other

Data were downloaded from the Humanitarian Data Exchange as an Excel

worksheet. The Humanitarian Data Exchange is an open platform for sharing data that

may be helpful to other organizations during a crisis (2017). Data analysis of the Excel

worksheet was accomplished using the IBM Statistical Package for Social Sciences also

known as SPSS (IBM, 2019). Data were cleaned prior to publication for any outliers or

duplicate cases by the investigators (Morse, 2015). According to the website, there is a

review committee that ensures submissions were collected ethically and do not disclose

personally identifiable information (Humanitarian Data Exchange, 2017).

Data analysis methods were chosen based on the type of data collected

(quantitative). Descriptive statistics were used to examine each of the two dependent

variables against all of the independent variables (age, gender, education, and

occupation). Univariate analysis included frequency displayed as percentage of each of

the demographic response categories. Further, multivariable analysis (binomial

regression) was applied to test the association between the sociodemographic predictors

and the two outcome variables. A result was considered statistically significant with a p ≤

.05.

Threats to Validity

Door to door surveys can offer a unique insight into the thoughts and problems

being faced by a population at a given moment in time (Morse, 2015). However, cross

sectional surveys can also have a negative impact on validity (HHS, 2007). Limitations of

this cross-sectional survey include the missing data of the homes skipped for active EVD

cases and possible inaccurate reporting. Limitations of cross-sectional surveys in general

is that they only allow for a glimpse of an issue over a distinct period of time (HHS,

2007). This means that only association and not causation can be determined (HHS,

2007).

Surveyors were told to randomly pick 20 homes within each designated zone but

to skip any home with someone actively infected with EVD (Humanitarian Data

Exchange, 2017). If someone was not home, surveyors went to the next home and so on

as long as they remained in the pre-designated sample zone (Morse, 2015). As a result,

there was an over 95% response rate (Morse, 2015). The surveys collected were

completely finished so there was no incomplete or missing data within the surveys

analyzed (Humanitarian Data Exchange, 2017). Accurate reporting was confirmed as

much as possible by the investigators by making the surveys electronic, monitoring

survey data entry multiple times a day, ensuring the data collected correlated with the

sampling plan geographically by having GPS on the survey devices, and by frequently

communicating with the surveyors who also underwent extensive training to prevent

error (Morse et al., 2016).

External Validity

One common threat to external validity includes selection bias (Alexander, Lopes,

Ricchetti-Mastersson, & Yeatts, 2013). In this research, the primary investigators ensured

that the population surveyed was a proportional representation of the population of each

community (Morse, 2015). Additionally, they ensured that the homes chosen for surveys

were random with the one exception of homes where the only adult had an active case of

EVD were avoided (Morse et al., 2016). This was a necessary step to protect the

surveyors (Morse et al., 2016). Additionally, individuals with EVD were often unable to

answer a survey because they were so ill.

External validity was also limited because Liberia has had unique issues

historically that made the country even more unstable entering into this crisis (Bastian,

2014). A civil war which ended in 2003 saw the death of over 250, 000 Liberian citizens

(CIA, 2018). The war wreaked havoc on the country’s economy and health infrastructure,

of which neither had fully recovered by 2014 (Bastian, 2014). The civil war and declining

economy may have therefore limited the ability to generalize this study to other countries

in some ways. Economically though, many other sub-Saharan African nations have a

similar GDP and poor health infrastructure (CIA, 2018). Additionally, Monrovia may

face another epidemic or other urban areas within Liberia may face an epidemic of

similar nature and thus the benefits of this research are still important (WHO, 2018).

Internal Validity

Instrumentation was the primary internal validity concern entering into the

primary data collection. To alleviate user error, an electronic version was made (Morse,

2015). No surveyors reported experiencing difficulty using the program (Morse, et al.,

2016). Investigators monitored data entry remotely through the day to make sure the data

was being collected correctly (Morse et al., 2016).

Furthermore, there was no evidence of any historical events that could have

affected internal validity during the time frame samples were collected, though it took a

month to get all of the over 1,000 surveys completed (Morse et al., 2016). It is possible

that over a month an individual could learn more about EVD and preventive hygiene.

Therefore, hypothetically, in December someone may answer the questions differently

compared to if they had been surveyed later in January. Overall, though, a month was the

shortest feasible amount of time possible to collect the data and each community was

completed on the same day to ensure consistency at least among that cluster of

respondents (Morse et al., 2016).

Ethical Procedures

For the primary study, the group that originally conducted the research at MIT

obtained approval from the IRB at MIT (Morse et al., 2016). In Liberia, approval for

conducting the survey was obtained through the Peacebuilding Office at the Ministry of

Internal Affairs (Morse et al., 2016). An informed consent was obtained from everyone

who participated in the survey (Morse et al., 2016). The Humanitarian Data Exchange

where the data are publicly published also has a rigorous ethics process prior to allowing

a source to publish data on the site (2017). As a part of the United Nations Secretariat,

The Humanitarian Data Exchange ensures that personally identifiable information is

never published (2017). The site has an internal review process to make sure such

sensitive data is never shared (Humanitarian Data Exchange, 2017). Additionally, the site

has a way for viewers and publishers to report data that is suspected to be unethical

(Humanitarian Data Exchange, 2017). For the secondary data study, the Walden

University IRB approval was obtained prior to analyzing the data.

Ethical Considerations

One ethical consideration was privacy of the population being surveyed.

Individuals were promised anonymity in the survey (Morse, 2015). In order to ensure

this, the surveyors did not collect names or identifiable information (Humanitarian Data

Exchange, 2017). Instead respondents were given a number to symbolize their response

(Humanitarian Data Exchange, 2017). Omitting names prevents the data collected from

being traced to an individual and eliminates concerns about identity protection.

Summary and Transition

In this section, I outlined my research design and rationale. I explored the

instrumentalization used to collect the primary data and what will be done to analyze the

data in this project. I overviewed the operationalization of variables that I will be using to

perform my statistical analysis and then offered a plan for that analysis using univariate

and multivariate analysis techniques. In Section 3, I will report and display the results of

my SPSS analyses.

Section 3: Presentation of the Results and Findings

The purpose of this study was to determine if there was a significant difference in

EVD knowledge and prevention among different demographic groups of employed adults

in Monrovia, Liberia, during the 2014-2016 EVD epidemic.Age, education level, gender,

occupation, religion, and ward were chosen from the data source for further investigation

as possible predictors of EVD understanding and preventative practices.

The study’s research questions were as follows:

1. Is there an association between socioeconomic and demographic factors

(education, age, gender, ward, religion, and occupation) and frequency of

employed adults who report the use of preventative hygiene practices (use of

hand sanitizer and bleach) in Monrovia, Liberia?

2. Is there an association between socioeconomic and demographic factors

(education, age, gender, ward, religion, and occupation) and frequency of

employed adults who report understanding the basic nature of the EVD

(common symptoms and how it is spread) in Monrovia, Liberia?

This section includes a description of the data that were collected for analysis, and

how the data were cleaned from a larger survey, leaving only the statistics necessary for

this research’s purpose. From the final data set, secondary analysis was conducted using

SPSS v. 25 software to produce descriptive statistics, chi-square analyses, and

multivariable analyses. Finally, this section will be summarized and a transition to section

4 will be provided.

Description of Data Abstracted for Analysis

The primary data were derived from a survey conducted by MIT in 2015, which

was then published on the public data forum, the Humanitarian Data Exchange in 2017.

The survey was collected in Monrovia, but not by MIT researchers. Instead, they hired an

agency in Monrovia to use door to door electronic surveys that uploaded results to a

cloud managed remotely in real time by the MIT group. MIT made efforts (seen in

Section 2 under “Sampling and Sampling Procedures”) to ensure that the survey

questions were written to be widely inclusive of the current EVD situation in the city.

The primary researchers also mapped the city in a way that would ensure equitable

survey representation of the citizens in the various wards, and they thoroughly trained

those collecting the survey data to decrease field-related errors. This yielded a 95%

overall survey participation rate.

Survey restrictions for participation were minimal and included the following:

respondents had to be over the age of 18, the only representative of their household

taking the survey, and had to be asymptomatic for the surveyor’s protection. Those

surveyed were asked their age, gender, occupation, ward of residence, religion, and

education level in addition to EVD-related questions. For more detail on data collection

methods, see Section 2.

Data Preparation

Data analysis began after approval from the Walden University Institution Review

Board (Approval No. 12-13-19-0608067). These data had already been deidentified by

the primary data collectors. To begin analysis though, the data had to be further refined.

The survey utilized in this study included multiple questions that covered the

current state of EVD progression in Monrovia at the time the survey was conducted.

Many of the questions and answers were not necessary for inclusion in this research such

as political and personal viewpoints. Certain assessment questions and categories were

therefore excluded in order to narrow the focus of the project to address the research

questions specifically. Other individual survey questions were grouped together into one

question so more meaningful analysis could be performed.

The following survey questions were combined for the category, “EVD

knowledge”:

• Understanding of symptoms/ signs of EVD

• Awareness of how EVD is transmitted

• General knowledge about the virus

The following survey questions were combined into the category, “EVD

prevention”:

• Use of a bucket with bleach before entering the home

• Use of hand sanitizer

According to the inclusion criteria, the number of cases in the final dataset was

1334. The data analysis was performed by doing descriptive statistics of all variables of

the dataset, chi- square analysis between each predictor and the dependent variables, and

multivariable analysis.

Descriptive Statistics

First, descriptive statistics were conducted on both the dependent and independent

variables. Age was converted to categorical from numerical, and thus all data was

categorical in nature. The descriptive statistics table 3 below shows the number and

percent of employed adults per variable category for 1334 collected survey responses.

Table 3

Descriptive Statistics of the EVD Monrovia Study Sample (N = 1334)

Variable Employed Adults

N %

Ward

New Kru Town 103 7.8

Congo Town 46 3.5

Paynesville 414 31.3

Gardnersnesville 93 7.0

New George 81 6.1

Bardensville 35 2.6

Caldwell 58 4.4

Logan Town 50 3.8

Clara Town 72 5.4

West Point 79 6.0

Sonwein 53 4.0

Slipway 66 5.0

Sinkor 53 4.0

Lakpazee 55 4.2

Old Road 64 4.8

Gender

Male 625 46.9

Female 709 53.1

Education

None 158 12.0

Some completed-abc 130 9.9

Some completed-junior

high

210 15.9

Some completed-high

school

559 42.4

Some completed-

university

260 19.7

Religion

Christian 1193 89.4

Muslim 133 10.0

Other 8 6.0

Age

18-34 639 47.9

35-54 570 42.7

>54 125 9.4

Occupation

Professional-clerical 225 16.8

Manual 411 30.6

Sales-services 538 40.1

Other 168 12.5

Bucket Use

No 299 22.4

Yes

1035 77.6

Sanitizer Use

No 752 56.4

Yes 582 43.6

How you get EVD

Aware of how you get

EVD

1230 91.7

Do not know how you get

112 8.3

Symptoms/ signs

Aware of symptoms/

signs

1271 94.7

Do not know symptoms/

signs

71 5.3

Knowledge

Low knowledge 911 67.9

High knowledge 431 32.1

Since the data were categorical, it was necessary to display frequency (labeled N)

of the individual survey responses and valid percent of the responses as well. Valid

percent figures will be designated with a % after them in the following breakdown of

results. All variables had 1334 responses with no missing data included.

Notable results of the descriptive statistics included that most participants had a

high school or college degree, were Christian, lived in Paynesville, and worked in sales-

services or manual labor. Additionally, there were more individuals in the 18- 54 age

categories.

Several facts may contribute to the data seen above. For age, a civil war in the

early 2000s had a high death tole (Bastian, 2014). Many in the 18-34 category would

have been too young to fight. The next age groups would have likely seen the most

fatalities.

For ward, Paynesville had the most residents. Paynesville is a large suburb, larger

than Monrovia city itself, that expands from a busy market area. Markets are essential to

the Liberian economy (CIA, 2018). This combined with the large geographic area of

Paynesville contribute to its higher population. Individuals from this survey primarily

work in sales- service and manual labor positions. According to the CIA World Fact

Book, mining and agriculture are the primary occupations in Liberia which would be

classified as manual positions (2018). Within agriculture, some who sell their goods may

have selected that on the survey response as well. Markets are common in Liberia where

agricultural products, manmade goods, and meats are sold (CIA, 2018).

Within the gender category, the frequency was slightly skewed towards females

who accounted for 53.1% of the population. The CIA World Fact Book says that

maternal mortality is a serious issue in Liberia, and that female genital cutting is putting

female lives at risk (2018). While these facts would suggest there should be fewer women

than men, the Fact Book also notes these issues are more prevalent among tribes and in

rural settings (CIA, 2018). This study was set in Monrovia, an urban setting.

In the religion category, 1193 reported to be Christian (89.4%), 133 Muslim

(10%), and 8 other (6.0%). In cities like Monrovia, tribal cultures are becoming scarcer

and are often only seen in outlying areas of Liberia. Traditional burial practices are a

high-risk activity for EVD transmission (CDC, 2018).

Within the research question related to preventative hygiene, respondents were

more likely to use buckets. However, slightly less people used hand sanitizer than those

who did not.

Within the research question pertaining to overall understanding of EVD, most

respondents were aware of how you get EVD, the signs and symptoms of EVD, and

overall knowledge.

Bivariate Analysis

This phase of analysis utilized the chi-square test and Cramer’s V measure for

effect. Chi-square tests are used with categorical data sets to determine if two variables in

the same population are related. Chi-square tests examine independence between the

observed and expected data, and for this research, the tests were held to a p value at 0.05.

Cramer’s V was included when an association was found through the chi-square tests to

demonstrate the strength of the association.

This research examined five dependent variables and six independent variables.

To make the data more user friendly, all dependent variables were combined under each

independent variable creating 6 total tables instead of several. All independent variables

and dependent variables were not further combined into one table because that table

would appear too congested and thus would not be easy to read. The following tables

show the combined chi-square results with Cramer’s V included.

Table 4 displays the results for the association between age, preventative hygiene

practices, and understanding EVD.

Table 4

Bivariate analysis (chi-square test) between age and preventative hygiene practices and

understanding of EVD (N = 1326).

Age

Dependent

Variables

18-34 years

N (%)

35-54 years

N (%)

>54 years

N (%)

Total X

P Cramer’s V

Total 636 565

125 1326

Bucket Use 4.351 .114 -

No 156 (52.5) 120 (40.4) 21 (7.1) 297

Yes 480 (46.6) 445 (43.2) 104(10.1) 1029

Sanitizer

Use

1.653 .438 -

No 364 (48.7) 308 (41.2) 75 (10.0) 747

Yes 272 (47.0) 257 (44.4) 50 (8.6) 579

Total 639 570 125 1334

How you

get EVD

.809 .667 -

Aware of

how you

get EVD

588 (48.1) 523 (42.8) 112 (9.2) 1223

Do not

know how

you get it

51 (45.9) 47 (42.3) 13 (11.7) 111

EVD

Symptoms/

Signs

1.004 .605 -

Aware of

symptoms/

signs

607 (48.1) 540 (42.8) 116 (9.2) 1263

Do not

know

symptoms/

signs

32 (45.1) 30 (42.3) 9 (12.7) 71

EVD

Knowledge

.733 .693 -

Low

knowledge

430 (47.5) 387 (42.7) 89 (9.8) 906

High

knowledge

209 (48.8) 183 (42.8) 36 (8.4) 428

Among those who did not use a bucket with bleach, 52.5% were 18-34 years old,

40.4% were 35-54 years old, and 7.1% older than 54 years. A sentence can't start with a

number, unless it’s spelled out 46.6% of 18-34 years old, 43.2% of 35-54 years old, and

10.1% of greater than 54 years old used buckets with bleach to clean. According to the

chi-square test, with a X

value of 4.351 and p value of .114, there was no significant

association found between age and bucket use.

Among hand sanitizer use, 48.7% of 18-34 years old, 41.2% of 35-54 year olds,

and 10% of those older than 54 years did not use hand sanitizer. Forty-seven percent of

18-34 year-olds, 44.4% of 35-54 year-olds, and 8.6% of greater than 54 years old used

hand sanitizer. Using the chi-square test to examine the relationship between sanitizer use

and age, with a X

value of 1.653 and p value of .438, there was no significant difference.

For the age category, 48.1% of 18- 34 year-olds, 42.8% of 35-54 year-olds, and

9.2% of > 54 year-olds, were aware of how you get EVD. For the next question, 45.9% of

the 18-34 age category, 42.3% of 35-54 year-olds, and 11.7% of > 54 year-olds were

unaware of how you get EVD. According to the chi-square of association between how

you get EVD and age, with a X

value of .809 and p value of .667, there was no

significant difference.

For those who understood the signs and symptoms of EVD, 48.1% of 18-34 year-

olds, 42.8% of 35-54 year-olds, and 9.2% of > 54 years were aware of the signs and

symptoms. Next, for 18-34 year-olds 45.1%, 42.3% of 35-54 year-olds, and 12.7% of >

54 year-olds were unaware. According to the chi-square test, with a X

value of 1.004 and

p value of .605,, there was no significant difference.

Results for the last dependent variable tested for the independent variable age in

bivariate analysis, showed that 47.5% of 18-34 year-olds, 42.7% of 35-54 year-olds, and

9.8% of > 54 year-olds, have a low level of knowledge about EVD. The next result

showed, 48.8% of the 18-34 age category, 42.8% of 35-54 year-olds, and 8.4% of > 54

year-olds, have high knowledge of EVD. According to the chi-square test, these

differences were not significant with a X

value of .773 and p value of .693.

Table 5 displays the results for the association between education and

preventative hygiene practices, and understanding EVD.

Table 5

Bivariate analysis (chi-square test) between education and preventative hygiene

practices and understanding of EVD (N = 1326).

Education

Dependent

Variables

0 None

N (%)

Some

completed

ABC N (%)

Some

completed

junior high N

(%)

Some

completed

high school N

(%)

Some

completed

university N

(%)

Total X

P Cramer’s V

Total 157 130

209 553 260 1309

Bucket Use 24.395 .000 .137

No 37

(12.6)

46 (15.7) 54 (18.4) 119 (40.6) 37 (12.6) 293

Yes 120

(9.2)

84 (8.3) 155 (11.8) 434 (42.7) 223 (17.0) 1016

Sanitizer Use 68.484 .000 .229

No 111

(15.0)

99 (13.4) 133 (18.0) 291 (39.4) 105 (14.2) 739

Yes 46 (8.1) 31 (5.4) 76 (13.3) 262 (46.0) 155 (27.2) 570

Total 158 130 210 559 260 1317

How you get

EVD

55.704 .000 .206

Aware of how

you get EVD

135

(11.2)

104 (8.6) 187 (15.5) 526 (43.5) 257 (21.3)

1209

Do not know

how you get it

(21.3)

26 (24.1) 23 (21.3) 33 (30.6) 3 (2.8) 108

EVD

Symptoms/

Signs

49.912 .000 .195

Aware of

symptoms/

signs

138

(11.0)

114 (9.1) 196 (15.7) 542 (43.4) 259 (20.7) 1249

Do not know

symptoms/

signs

(29.4)

16 (23.5) 14 (20.6) 17 (25.0) 1 (1.5) 68

EVD

Knowledge

85.911 .000 .255

Low

knowledge

124

(13.8)

107 (11.9) 171 (19.0) 372 (41.4) 125 (13.9) 899 .

High

knowledge

34 (8.1) 23 (5.5) 39 (9.3) 187 (44.7) 135 (32.3) 418

The next independent variable explored was education. This was broken into no

education, some completed ABC (elementary school), some completed junior high, some

completed high school, and some completed university. According to the results, 12.6%

of those not using buckets had no education, 15.7% ABC, 18.4% junior high, 40.6% high

school, and 12.6% university. On the other hand, 9.2% of those using buckets had no

education, 8.3% ABC, 11.8% junior high, 42.7% high school, and 17% university chose

to use buckets. The chi-square result showed a X

value of 24.395 and p< .0001. With p

below .05 and Cramer’s V at .137 (small effect size) there was a significant association

between bucket use and education level thus rejecting the null hypothesis and accepting

the alternative. This association is shown through the statistics to be an increase in bucket

use as education levels increase, at least until the high school level.

Regarding hand sanitizer, 15% of those not using it had no education, 13.4%

ABC, 18% junior high, 39.4% high school, and 14.2% university. On the contrary, 8.1%

of those using hand sanitizer had no education, 5.4% ABC, 13.3% junior high, 46% high

school, and 27.2% university. The chi-square result showed a X

value of 64.484 5 and p<

.0001. With p value below .05 and Cramer’s V at .229 (small effect size) there was a

significant association between hand sanitizer use and education level thus rejecting the

null hypothesis and accepting the alternative. This association is shown through the

statistics to be an increase in hand sanitizer use as education levels increase.

For participants having individual awareness about contracting EVD, 11.2% had

no education, 8.6% ABC, 15.5% junior high, 43.5% high school, and 21.3% university.

Those unaware included 21.3% with no education, 24.1% ABC, 21.3% junior high,

30.6% high school, and 2.8% university. There was an increase in awareness of how you

get EVD as education levels increase. The chi-square test showed a X

value of 55.704

and p<.0001. Cramer’s V = .195 (small effect size) thus interpreted that this increase in

awareness with education level is significant, and the null hypothesis can be rejected.

The next dependent variable examined with education was awareness of EVD

signs and symptoms; 11% of those of no education, 9.1% of ABC, 15.7% of junior high,

43.4% of high school, and 20.7% of university were aware of the signs and symptoms.

The dispersal of percentages among those unaware of the symptoms and sign of EVD per

education group were as follows: 29.4% with no education, 23.5% of ABC, 20.6% of

junior high, 25% of high school, and 1.5% of university. This showed an increase in

awareness with some education levels and a significant difference demonstrated by chi-

square results that included a X

value = 49.912, p value<.0001, and a Cramer’s V =

.195 (small effect size). The null hypothesis can be rejected in favor of the alternative

hypothesis that there is a significant difference among education levels and EVD signs

and symptoms.

For overall knowledge of EVD, 13.8% with no education, 11.9% of ABC, 19% of

junior high, 41.4% of high school and 13.9% of university had low knowledge. For

overall high knowledge of EVD among education groups the results were as follows:

8.1% with no education, 5.5% ABC, 9.3% junior high, 44.7% high school, and 32.3%

university. X

= 85.911, p<.0001, and Cramer’s V = .255 (small effect size) for these

tests therefore showing a significant difference between the two categories. Knowledge

increased with education levels which allow rejection of the null hypothesis.

Table 6 displays the results for the association between gender and preventative

hygiene practices, and understanding EVD.

Table 6

Bivariate analysis (chi-square test) between gender and preventative hygiene practices

and understanding of EVD (N = 1326).

Gender

Dependent

Variables

Male

N (%)

Female N

(%)

Total X

P Cramer’s

Total 622 704 1326

Bucket Use 3.090 .079 -

No 126

(42.4)

171 (57.6) 297

Yes 496

(48.2)

533 (51.8) 1029

Sanitizer

Use

16.313 .000 .111

No 314

(42.0)

433 (58) 747 9

Yes 308

(53.2)

271 (46.8) 579

Total 625 709 1334

How you

get EVD

14.254 .000 .103

Aware of

how you get

EVD

592

(48.4)

631 (51.6)

1223

Do not

know how

you get it

(29.7)

78 (70.3) 111

EVD

Symptoms/

Signs

8.986 .003 .082

Aware of

symptoms/

signs

604

(47.8)

659 (52.2) 1263

Do not

know

symptoms/

signs

(29.6)

50 (70.4) 71

EVD

Knowledge

26.113 .000 .140

Low

knowledge

381

(42.1)

525 (57.9) 906 .

High

knowledge

244

(57.0)

184 (43.0) 428

The third independent variable examined for association was gender. The male

population had 42.4% of respondents and 57.6% of females did not use buckets. Further,

48.2% did use buckets. The chi-square results yielded a X

value of 3.090 and p value of

.079. The value.079 is higher than the threshold for significance of .05 thus the null

hypothesis is accepted.

For the hand sanitizer category, 42% of males and 58% of females did not use

hand sanitizer. 53.2% of males did use hand sanitizer while 46.8% of females did. The

chi-square test for independence resulted in a X

= 16.313, a p<.0001, and a Cramer’s V

= .111 (small effect size). There was a significant difference (more males used hand

sanitizer than females) thus the null hypothesis is rejected.

Next, 48.4% of males and 51.6% of females were aware of how you get EVD.

However, 29.7% of males and 70.3% of females were not aware of how you get EVD.

Chi-square testing revealed an X

of 14.254, a p value<.0001, and Cramer’s V = .103.

Males were more likely to know how you get EVD than women and this result was

significant.

For the signs and symptoms variable, 47.8% of males and 52.2% of females were

aware of the signs and symptoms of EVD, but fewer males (29.6%) and 70.4% of

females did not know the signs and symptoms of EVD. The chi-square test gave a X

8.986, a p = .003. Additionally, Cramer’s V = .082 (small effect size). These results

showed that males were aware of the signs and symptoms of EVD significantly more

than women.

Finally, 42.1% of males and 57.9% of females had low knowledge of EVD while

57% of males and 43% of females had high knowledge of EVD. Males knew more than

women in this category, and chi-square testing revealed this difference was significant

= 26.113, p<.0001 and Cramer’s V = .140).

Table 7 displays the results for the association between occupation and

preventative hygiene practices, and understanding EVD.

Table 7

Bivariate analysis (chi-square test) between occupation and preventative hygiene

practices and understanding of EVD (N = 1326).

Occupation

Dependent

Variables

Professional-clerical

N (%)

Manual N (%) Sales-services

(%)

Other N

(%)

Total X

P Cramer’s V

Total 225 407

535 167 1334

Bucket Use 14.917 .002 .106

No 32 (10.7) 84 (28.1) 141 (47.2) 42 (14.0) 299

Yes 193 (18.6) 323 (31.2) 394 (38.1) 125

(12.1)

1035

Sanitizer Use 26.515 .000 .141

No 97 (12.9) 229 (30.5) 312 (41.5) 114

(15.2)

752

Yes 128 (22.0) 178 (30.6) 223 (38.3) 53 (9.1) 582

Total 225 411 538 168 1342

How you get

EVD

26.175 .000 .140

Aware of how

you get EVD

216 (17.6) 386 (31.4) 489 (39.8) 139

(11.3)

1230

Do not know

how you get it

9 (8.0) 25 (22.3) 49 (43.8) 29 (25.9) 112

EVD Symptoms/

Signs

23.766 .000 .133

Aware of

symptoms/ signs

222 (17.5) 395 (31.1) 506 (39.8) 148

(11.6)

1271

Do not know

symptoms/ signs

3 (4.2) 16 (22.5) 32 (45.1) 20 (28.2) 71

EVD Knowledge 9.600 .022 .085

Low knowledge 133 (14.6) 287 (31.5) 373 (40.9) 118

(13.0)

911 .

High knowledge 92 (21.3) 124 (28.8) 165 (38.3) 50 (11.6) 431

Occupation was the next independent variable analyzed. For bucket use, 10.7% of

professional-clerical, 28.1% of manual professions, 47.2% of sales services, and 14% of

other did not use buckets. 18.6% of professional-clerical, 31.2% of manual, 38.1% of

sales-services, and 12.1% of other used buckets. The chi-square test yielded a X

14.917, p = .002, and Cramer’s V = .106. There is a significant association between

bucket use and occupation. The participants who were more likely to use buckets

included the professional and manual job groups.

Next, 12.9% of professional-clerical, 30.5% of manual, 41.5% of sales services,

and 15.2% of other jobs did not use hand sanitizer. 22% of professional clerical, 30.6% of

manual, 38.3% of sales services and 9.1% of other did use hand sanitizer. Significantly

more professional clerical workers and manual workers used hand sanitizer (X

= 26.515,

a p value<.0001, and a Cramer’s V = .141-small effect size).

For those aware of how you get EVD, 17.6% of professional clerical, 31.4% of

manual, 39.8% of sales services, and 11.3% of others fell into the awareness category.

The lack of awareness category included 8.0% of professional clerical, 22.3% of manual

laborers, 43.8% of sales and services, and 25.9% of other. Chi-square analysis yielded a

of 26.175, a p< .0001, and a Cramer’s V = .140 (small effect size). A significant

higher amount of higher awareness was found among professional clerical workers and

manual workers.

For the next variable, 17.5% of professional clerical, 31.1% of manual, 39.8% of

sales services, and 11.6% of other reported knowing the signs and symptoms of EVD.

4.2% of professional clerical, 22.5% of manual, 45.1% of sales services, and 28.2% of

other did not know any of the signs or symptoms of EVD. Chi-square showed a

significant with a difference between professional clerical and manual having higher

awareness than sales service or other (X

= 23.766, p <.0001, and Cramer’s V = .133-

small effect size).

Finally, 14.6% of professional clerical, 31.5% of manual, 40.9% of sales services,

and 13% of others had low knowledge of EVD. 21.3%. of professional clerical, 28.8% of

manual, 38.3% of sales services, and 11.6% of other had high knowledge of EVD. This

too was significant as found by chi-square testing, although with a very small effect size

= 9.6, p = .022, and Cramer’s V = .085).

Table 8 displays the results for the association between religion and preventative

hygiene practices, and understanding EVD.

Table 8

Bivariate analysis (chi-square test) between religion and preventative hygiene practices

and understanding of EVD (N = 1326).

Religion

Dependent

Variables

Christian N

(%)

Muslim N (%) Other N (%) Total X

P Cramer’s V

Total 1186 132

8 1326

Bucket Use .345 .842 -

No 268 (90.2) 27 (9.1) 2 (0.7) 297

Yes 918 (89.2) 105 (10.2) 6 (0.6) 1029

Sanitizer Use 1.307 .520 -

No 669 (89.6) 72 (9.6) 6 (0.8) 747

Yes 517 (89.3) 60 (10.4) 2 (0.3) 579

Total 1193 133 8 1334

How you get

EVD

1.112 .573 -

Aware of how

you get EVD

1095 (89.5) 120 (9.8) 8 (0.7) 1223

Do not know

how you get it

98 (88.3) 13 (11.7) 0 (0.0) 111

EVD Symptoms/

Signs

.455 .797 -

Aware of

symptoms/ signs

1129 (89.4) 126 (10.0) 8 (0.6) 1263

Do not know

symptoms/ signs

64 (90.1) 7 (9.9) 0 (0.0) 71

EVD Knowledge 4.131 .127 -

Low knowledge 805 (88.9) 93 (10.3) 8 (0.9) 906 .

High knowledge 388 (90.7) 40 (9.3) 0 (0.0) 428

Religion was also examined in this research as a possible predictor variable. For

the individuals who did not use buckets, 90.2% were Christians did not use buckets, 9.1%

Muslims, and 0.7% of other religions. Similarly, individuals who used buckets were

89.2% Christians, 10.2% Muslims, and 0.6% of other religions. Thus, there was no

significant difference in the association between religion and bucket use (X

= .345 and

p = .842).

There was also no significant difference between religion and use of hand

sanitizer (X

of 1.307 and a p of .520); 89.6% of Christians did not use hand sanitizer,

9.6% of Muslims did not, and neither did 0.8% of other religions. However, 89.3% of

Christians used hand sanitizer, 10.4% of Muslims used hand sanitizer, and 0.3% of others

used hand sanitizer.

Further, the participants who were aware of how you get EVD were 89.5%

Christians, 9.8% Muslims, and 0.7% of others. 88.3% Christians, 11.7% Muslims, and

0% of other religions. There was no significant difference between religion and

awareness of how you get EVD (X

of 1.112 and a p = .573).

Within the signs and symptoms category, 89.4% Christians were aware of the

signs and symptoms, 10% Muslims and 0.6% of others. Among those not aware of the

signs and symptoms, 90.1% were Christians and 9.9% were Muslims. Thus, there was no

significant association between religion and awareness of EVD signs and symptoms (X

= .455 and p value = .797).

For overall knowledge of EVD, 88.9% of Christians, 10.3% of Muslims, and

0.9% of other religions had low knowledge. 90.7% of Christians, 9.3% of Muslims, and

0% of others had high knowledge, but these results were not significantly different (X

4.131 and p = .127).

Table 9 displays the results for the association between ward and preventative

hygiene practices, and understanding EVD.

Table 9

Bivariate analysis (chi-square test) between ward and preventative hygiene practices and

understanding of EVD (N = 1326).

Ward

Dependent

Variables

Z100-New

KruTown

N (%)

Z1000-

Congo

Town

N (%)

Z1100-

Paynesville

N (%)

Z1200-

Gardnersnesville

N (%)

Total X

P Cramer’s V

Total 103 44 412 92

Bucket Use

No 18 (6.1) 2 (0.7) 101 (34.2) 27 (9.2)

Yes 85 (8.3) 42 (4.1) 311 (30.5) 65 (6.4)

Sanitizer Use

No 54 (7.3) 20 (2.7) 238 (32.1) 47 (6.3)

Yes 49 (8.6) 24 (4.2) 174 (30.4) 45 (7.9)

Total 103 46 414 93

How you get

EVD

Aware of how

you get EVD

93 (7.7) 44 (3.6) 394 (32.6) 87 (7.2)

Do not know

how you get it

10 (8.9) 2 (1.8) 20 (17.9) 6 (5.4)

EVD Symptoms/

Signs

Aware of

symptoms/ signs

97 (7.8) 44 (3.5) 396 (31.7) 90 (7.2)

Do not know

symptoms/ signs

6 (8.5) 2 (2.8) 18 (25.4) 3 (4.2)

EVD Knowledge

Low knowledge 76 (8.5) 31 (3.5) 269 (30.1) 52 (5.8) .

High knowledge 27 (6.3) 15 (3.5) 145 (34.0) 41 (9.6)

Ward

Dependent

Variables

Z1300-

New

George

N (%)

Z1400-

Bardnesville

N (%)

Z1600-Caldwell

N (%)

Z200-Logan

Town

N (%)

Total X

P Cramer’s V

Total 80 35

58 50

Bucket Use

No 15 (5.1) 15 (5.1) 23 (7.8) 4 (1.4)

Yes 65 (6.4) 20 (2.0) 35 (3.4) 46 (4.5)

Sanitizer Use

No 43 (5.8) 24 (3.2) 41 (5.5) 25 (3.4)

Yes 37 (6.5) 11 (1.9) 17 (3.0) 25 (4.4)

Total 81 35 58 50

How you get

EVD

Aware of how

you get EVD

72 (6.0) 33 (2.7) 50 (4.1) 40 (3.3)

Do not know

how you get it

9 (8.0) 2 (1.8) 8 (7.1) 10 (8.9)

EVD Symptoms/

Signs

Aware of

symptoms/ signs

78 (6.2) 33 (2.6) 54 (4.3) 44 (3.5)

Do not know

symptoms/ signs

3 (4.2) 2 (2.8) 4 (5.6) 6 (8.5)

EVD Knowledge

Low knowledge 52 (5.8) 22 (2.5) 49 (5.5) 32 (3.6)

High knowledge 29 (6.8) 13 (3.0) 9 (2.1) 18 (4.2)

Ward

Dependent

Variables

Z300-

Clara

Town =

N (%)

Z400-West Point

= 11

N (%)

Z500-Soniwein

= 12

N (%)

Z600-Slipway =

N (%)

Total X

P Cramer’s V

Total 70 79

53 66

Bucket Use

No 5 (1.7) 3 (1.0) 15 (5.1) 26 (8.8)

Yes 65 (6.4) 76 (7.5) 38 (3.7) 40 (3.9)

Sanitizer Use

No 40 (5.4) 47 (6.3) 30 (4.0) 43 (5.8)

Yes 30 (5.2) 32 (5.6) 23 (4.0) 23 (4.0)

Total 72 79 53 66

How you get

EVD

Aware of how

you get EVD

64 (5.3) 69 (5.7) 49 (4.0) 62 (5.1)

Do not know

how you get it

8 (7.1) 10 (8.9) 4 (3.6) 4 (3.6)

EVD Symptoms/

Signs

Aware of

symptoms/ signs

67 (5.4) 77 (6.2) 50 (4.0) 63 (5.0)

Do not know

symptoms/ signs

5 (7.0) 2 (2.8) 3 (4.2) 3 (4.2)

EVD Knowledge

Low knowledge 55 (6.1) 60 (6.7) 44 (4.9) 41 (4.6)

High knowledge 17 (4.0) 19 (4.4) 9 (2.1) 25 (5.9)

Ward

Dependent

Variables

Z700-

Sinkor =

N (%)

Z800-Lakpazee

= 15

N (%)

Z900-Old Road

= 16

N (%)

Total X

P Cramer’s V

Total 53 55 64 1314

Bucket Use 78.647 .000 .245

No 8 (2.7) 15 (5.1) 18 (6.1) 295

Yes 45 (4.4) 40 (3.9) 46 (4.5) 1019

Sanitizer Use 26.098 .025 .141

No 18 (2.4) 34 (4.6) 37 (5.0) 741

Yes 35 (6.1) 21 (3.7) 27 (4.7) 573

Total 53 55 64 1322

How you get

EVD

30.228 .007 .151

Aware of how

you get EVD

51 (4.2) 47 (3.9) 55 (4.5) 1210

Do not know

how you get it

2 (1.8) 8 (7.1) 9 (8.0) 112

EVD Symptoms/

Signs

15.466 .348 -

Aware of

symptoms/ signs

51 (4.1) 48 (3.8) 59 (4.7) 1251

Do not know

symptoms/ signs

2 (2.8) 7 (9.9) 5 (7.0) 71

EVD Knowledge 33.622 .002 .159

Low knowledge 29 (3.2) 39 (4.4) 44 (4.9) 895 .

High knowledge 24 (5.6) 16 (3.7) 20 (4.7) 427

Finally, the wards, how the city of Monrovia is divided into communities, was

tested for association with the dependent variables. First, 6.1% New Kru Town, 0.7%

Congo Town, 34.2% Paynesville, 9.2% Gardnernesville, 5.1% New George, 5.1%

Bardnesville, 7.8% Caldwell, 1.4% Logan Town, 1.7% Clara Town, 1% West Point,

5.1% Soniwein, 8.8% Slipway, 2.7% Sinkor, 5.1% Lakpazee, and 6.1% Old Road did

not use buckets. 8.3% New Kru Town, 4.1% Congo Town, 30.5% Paynesville, 6.4%

Gardnernesville, 6.4% New George, 2% Bardnesville, 3.4% Caldwell, 4.5% Logan

Town, 6.4 % Clara Town, 7.5% West Point, 3.7% Soniwein, 3.9% Slipway, 4.4% Sinkor,

3.9% Lakpazee, and 4.5% Old Road did use buckets. The chi-square test resulted in X

78.647 and a p<.0001, and Cramer’s V = .245 (small effect size). There is a statistically

significant difference between ward and bucket use. Specifically, New Kru Town, Congo

Town, New George, Logan Town, Clara Town, West Point, and Sinkor all used buckets

at higher rates per population size within dependent variable inquiry.

The dependent variable tested next was hand sanitizer use. For this variable, 7.3%

New Kru Town, 2.7% Congo Town, 32.1% Paynesville, 6.3% Gardnernesville, 5.8%

New George, 3.2% Bardnesville, 5.5% Caldwell, 3.4% Logan Town, 5.4% Clara Town,

6.3% West Point, 4% Soniwein, 5.8% Slipway, 2.4% Sinkor, 4.6% Lakpazee, and 5%

Old Road did not use hand sanitizer. 8.6% New Kru Town, 4.2% Congo Town, 30.4%

Paynesville, 7.9% Gardnernesville, 6.5%% New George, 1.9% Bardnesville, 3%

Caldwell, 4.4% Logan Town, 5.2% Clara Town, 5.6% West Point, 4% Soniwein, 4%

Slipway, 6.1% Sinkor, 3.7% Lakpazee, and 4.7% Old Road did use hand sanitizer. The

chi-square test results were X

= 26.098 and p = .025, and Cramer’s V was .141 (small

effect size). There was an association found between hand sanitizer use and ward.

Specifically, New Kru Town, Congo Town, Paynesville, New George, Logan Town, and

Sinkor all used hand sanitizer at higher rates per population size within dependent

variable inquiry.

The next variable showed that 7.7% New Kru Town, 3.6% Congo Town, 32.6%

Paynesville, 7.2% Gardnernesville, 6% New George, 2.7% Bardnesville, 4.1% Caldwell,

3.3% Logan Town, 5.3% Clara Town, 5.7% West Point, 4% Soniwein, 5.1% Slipway,

4.2% Sinkor, 3.9% Lakpazee, and 4.5% Old Road reported being aware of how you get

EVD. 8.9% New Kru Town, 1.8% Congo Town, 17.9% Paynesville, 5.4%

Gardnernesville, 8% New George, 1.8% Bardnesville, 7.1% Caldwell, 8.9% Logan

Town, 7.1% Clara Town, 8.9% West Point, 3.6% Soniwein, 3.6% Slipway, 1.8%

Sinkor, 9.9% Lakpazee, and 8% Old Road reported not being aware of how you get

EVD. The chi-square test yielded a X

= 30.228, a p = .007, and Cramer’s V = .151

(small effect size). Thus, there was a significant association found between EVD

knowledge and ward. Specifically, Congo Town, Paynesville, Gardenrnesville,

Bardnesville, Soniwein, Slipway, and Sinkor had greater awareness about contracting

EVD compared to the other wards.

The following percentages per ward show how respondents to the survey reported

awareness of the signs and symptoms of EVD: 7.8% New Kru Town, 3.5% Congo Town,

31.7% Paynesville, 7.2% Gardnernesville, 6.2% New George, 2.6% Bardnesville, 4.3%

Caldwell, 3.5% Logan Town, 5.4% Clara Town, 6.2% West Point, 4% Soniwein, 5%

Slipway, 4.1% Sinkor, 3.8% Lakpazee, and 4.7% Old Road. 8.5% New Kru Town, 2.8%

Congo Town, 25.4% Paynesville, 4.2% Gardnernesville, 4.2% New George, 2.8%

Bardnesville, 5.6% Caldwell, 8.5% Logan Town, 7% Clara Town, 2.8% West Point,

4.2% Soniwein, 4.2% Slipway, 2.8% Sinkor, 9.9% Lakpazee, and 7% Old Road do not

know the symptoms/ signs of EVD. These differences were not statistically significant

= 15.466 and a p = .348).

Lastly, 8.5% New Kru Town, 3.5% Congo Town, 30.1% Paynesville, 5.8%

Gardnernesville, 5.8% New George, 2.5% Bardnesville, 5.5% Caldwell, 3.6% Logan

Town, 6.1% Clara Town, 6.7% West Point, 4.9% Soniwein, 4.6% Slipway, 3.2% Sinkor,

4.4% Lakpazee, and 4.9% Old Road had low knowledge about EVD. 6.3% New Kru

Town, 3.5% Congo Town, 34% Paynesville, 9.6% Gardnernesville, 6.8% New George,

3% Bardnesville, 2.1% Caldwell, 4.2% Logan Town, 4% Clara Town, 4.4% West Point,

2.1% Soniwein, 5.9% Slipway, 5.6% Sinkor, 3.7% Lakpazee, and 4.7% Old Road had

high knowledge. The chi-square test for independence yielded an X

of 33.622 and p

value of .002, and the Cramer’s V = .159 (small effect size). Therefore, the difference is

significant, and the null hypothesis is rejected. The differences were most notable with

Paynesville, Gardnernesville, New George, Bardnesville, Logan Town, Slipway, and

Sinkor having more knowledge.

Multivariable Analysis

Binomial logistic regression was used for multivariable analysis. Binomial

logistic regression was chosen because it can show the association between multiple

independent variables and a dependent variable. Predictors occupation, education, and

gender were analyzed with each dependent variable from the research questions. Table 10

answers research question 2: Is there an association between socioeconomic and

demographic factors (education, age, gender, and occupation) and frequency of employed

adults who report understanding the basic nature of the EVD (common symptoms and

how it is spread) in Monrovia, Liberia?

Table 10.

Multivariable analysis for the dependent variable “How you get EVD” with predictors

occupation, education, and gender.

The Hosmer and Lemeshow test showed the results adequately fit the data at p =

.869. Additionally, according to the Nagelkerke R

, a 13% variation in the outcome is

explained by this model.

Occupation was a significant predictor of understanding how you get EVD (p =

0.001). The interpretation of the results are as follows: individuals of manual (OR = .541,

95% CI = .233-1.256), sales services (OR = .277, 95% CI = .152-.504), or other

occupation (OR = .529, 95% CI = .312-.899 ) are less likely to understand how you get

EVD, compared to professional-clerical participants.

With p < .0001, education is also a significant predictor of how you get EVD. All

those with formal education were more likely to know more about how you get EVD than

those with no formal education. Compared to no education in descending order, junior

high had the greatest level of understanding (OR = 18.259, 95% CI = 5.087-65.547)

followed by ABC (OR = 14.5, 95% CI = 4.066-51.712), high school (OR = 10.786, 95%

CI = 3.052-38.116), and university (OR = 5.662, 95% CI = 1.678-19.107). In summary,

compared to no education, those with ABC were 14.5 times more likely to understand

how you get EVD, those who finished junior high were 18.259 times more likely to know

how you get EVD, those who completed high school were 10.786 times more likely than

those with no education to know how you get EVD, and those who finished university

were 5.662 times more likely to know how you get EVD compared to those with no

formal education.

For gender, the p value is .141 which is above the 5% significance level. Gender

is therefore, not a significant predictor of understanding how you get EVD.

Table 11 partially answers research question 2: Is there an association between

socioeconomic and demographic factors (education, age, gender, and occupation) and

frequency of employed adults who report understanding the basic nature of the EVD

(common symptoms and how it is spread) in Monrovia, Liberia? This table shows the

association between independent variables occupation, education, and gender and the

dependent variable EVD knowledge.

Table 11

Multivariable analysis for the dependent variable “Knowledge” with predictors

occupation, education, and gender.

The Hosmer and Lemeshow test showed the results adequately fit the data at p =

.270. Additionally, according to the Nagelkerke R

, a 9.7% variation in the outcome is

explained by this model.

Occupation with a p value of .391 was not a significant predictor of EVD

knowledge. With a p<.0001, education is a significant predictor of EVD knowledge.

Compared to no education in descending order, university had the highest level of

knowledge (OR = .452, 95% CI = .326-.626) followed by ABC (OR = .266, 95% CI =

.164-.432), junior high (OR = .212, 95% CI = .123-.367), and high school (OR = .210,

95% CI = .133-.331).

For EVD knowledge among genders the p value is .009 thus showing a

significant difference between males and females. Examination shows that with males

used as the reference gender and an odds ratio of 1.417 (95% CI = 1.092-1.840), females

are more likely to have greater knowledge about EVD.

Table 12 shows the association between sanitizer use and predictors occupation,

education, and gender. This table corresponds with the first research question: Is there an

association between socioeconomic and demographic factors (education, age, gender, and

occupation) and frequency of employed adults who report the use of preventative hygiene

practices (use of hand sanitizer and bleach) in Monrovia, Liberia?

Table 12

Multivariable analysis for the dependent variable “Sanitizer” with predictors

occupation, education, and gender

The Hosmer and Lemeshow test showed the results adequately fit the data at p =

.737. Further, according to the Nagelkerke R

, 8.1% variation in the outcome is explained

by this model.

Occupation was a significant predictor of hand sanitizer use among this

population with a p value = .019. individuals with occupations besides professional-

clerical were less likely to use hand sanitizer. Those in manual jobs were 1.827 less likely

to use hand sanitizer (95% CI = 1.165-2.865). Individuals in sales-services used hand

sanitizer 1.759 times less (95% CI = 1.184-2.613), and those in the other category used

hand sanitizer 1.435 times less (95% CI = .976-2.130).

Education is a significant predictor of hand sanitizer use with p value<.0001.

Increased hand sanitizer use was seen among those with formal education. The use of

hand sanitizer can be seen as follows among education levels: university (OR = .648,

95% CI = .470-.892), high school (OR = .285-.641), ABC (OR = .315, 95% CI = .200-

.497), and junior high (OR = .255, 95% CI = .153-.423).

For gender, p value = .212. Gender is not a significant predictor of hand sanitizer

use.

The Table 13 demonstrates the association between occupation, education, and

gender with the dependent variable signs and symptoms. Signs and symptoms was a

component of the second research question looking at overall EVD knowledge among the

study population.

Table 13

Multivariable analysis for the dependent variable “Signs and Symptoms” with predictors

occupation, education, and gender.

The Hosmer and Lemeshow test showed the results adequately fit the data at p =

.498. Further, according to the Nagelkerke R

, a 14.2% variation in the outcome is

explained by this model.

For occupation, the p value = .004. Occupation is thus a significant predictor of

understanding EVD signs and symptoms. Individuals working in professional or clerical

fields were more likely to understand the signs and symptoms of EVD. This is

demonstrated through the odds ratios of the other professions: other (OR = .474, 95% CI

= .252-.893), sales-service (OR = .287, 95% CI = .141-.581), and manual (OR = .279,

95% CI = .076-1.018).

Education is also a significant predictor in this category with p value< .0001. All

those with formal education were more likely to know more about EVD signs and

symptoms. Specifically, those with ABC education were 31.398 times more likely to

know the signs and symptoms (95% CI = 3.990-247.098), junior high 25.362 times more

likely (95% CI = 3.155-203.898), high school 15.685 times (95% CI = 1.971-124.815),

and university 7.347 times (95% CI = .952-56.734).

For gender, the p value is .424 which is above the 5% significance level.

Therefore, gender is not a significant predictor of EVD sign and symptom recognition.

Bucket use was the second part of the first research question which looked at

preventative hygiene use in the study population. Sown below is bucket use among

working adults of the following demographics: education, occupation, and gender.

Table 14

Multivariable analysis for the dependent variable “Bucket Use” with predictors

occupation, education, and gender

The Hosmer and Lemeshow test showed the results adequately fit the data at p =

.995. Further, according to the Nagelkerke R

, 3.7% variation in the outcome is explained

by this model.

Occupation was a significant predictor of bucket use among this population with a

p value = .049. Individuals with occupations besides professional-clerical were less

likely to use hand sanitizer. The interpretation of the results are as follows: individuals of

manual (OR = 1.432, 95% CI = .829-2.472) sales services (OR = 1.316, 95% CI = .853-

2.030) or other occupation (OR = .899, 95% CI = .595-1.358) are less likely to

understand how you get EVD, compared to professional-clerical participants.

Education was a significant predictor of bucket use with p value = .005. All

education levels beyond no education had increased bucket use; in decreasing order

bucket use was as follows: university (OR = .652, 95% CI = .427-.998), ABC (OR =

.601, 95% CI = ..348-.1.037), high school (OR = .536, 95% CI = .326-.883), and

junior high (OR = .353, 95% CI = .205-.610).

For gender, the p value is is .583, thus gender is not a significant predictor of

bucket use.

Summary and Transition

The results of the descriptive statistics, chi-square analysis, and binomial logistic

regression were presented in Section 3. These results explored the association between

independent variables (age, occupation, education, ward, and religion) and preventative

hygiene measures including the use of hand sanitizer and buckets with bleach. The

second research question explored the association of these independent and EVD

understanding as measured by overall knowledge, knowledge of the signs and symptoms,

and awareness of how you get EVD.

Chi-square analysis revealed that age, ward and religion were not significant

predictors of EVD understanding and preventive hygiene and were eliminated as

independent variables for multivariable analysis. Based on BLR analysis, education was a

significant predictor of EVD understanding for all criteria/ categories, so the first null

hypothesis is rejected. Education is a significant predictor of preventative hygiene use

among both criteria defined in the first hypothesis, so the null hypothesis is rejected.

Gender was not a significant predictor of knowledge pertaining to how you get EVD and

signs and symptoms but was for overall knowledge with females knowing more than

males. Gender was not a predictor for hand sanitizer use or bucket use. Occupation was a

significant predictor for how you get EVD and signs and symptoms but not general EVD

knowledge. For hypothesis two, occupation was a significant predictor of hand sanitizer

use but not of bucket use.

Section 4 is next and will cover a more detailed interpretation of these findings.

Also, in Section 4 the limitations, recommendations for future study, and the use of this

research to promote social change will be discussed.

Section 4: Application to Professional Practice and Implications for Social Change

Healthcare and public health professionals are becoming increasingly concerned

as EVD outbreaks over the last decade have shown rising mortality rates (CDC, 2017).

The EVD outbreak of 2014-2016 caused over 11,000 deaths throughout Africa with

Liberia experiencing the highest morbidity and mortality (CDC, 2017). Spread into urban

centers like Monrovia, Liberia is partially responsible for this new trend (CDC, 2017). A

cross-sectional survey conducted by MIT in Monrovia during the epidemic was used to

explore the nature of EVD in Liberia in more detail.

Using the survey, this study explored the impact of sociodemographic and

socioeconomic factors on the use of bleach and hand sanitizer and on general EVD

knowledge among employed adult Monrovians, by using bivariate (chi-square) and

multivariable analysis (binomial logistic regression). The survey was conducted in all 15

of Monrovia’s wards. Seventy-seven communities were chosen within the wards, based

on population size, and then homes were chosen randomly by the surveyors. Data were

collected electronically and uploaded in real time to a cloud, where researchers could

oversee its accuracy and address concerns.

A better understanding of these associations could lead to more effective

preventative programming among employed adults in future outbreaks. In Section 4,

these results will be discussed within the context of the literature, and recommendations

for practice and future research, as well as social change implications, will be provided.

Interpretation of the Findings

Key findings of the study

According to the results of the study, occupation and education were significant

predictors of hand sanitizer use, but gender was not. People working in professional-

clerical jobs were more likely to use hand sanitizer compared to other occupations. Those

who finished university were most likely to use hand sanitizer followed by those who

finished high school, ABC, junior high, and no school.

Occupation and education were also significant predictors of bucket use. Those in

professional-clerical positions were most likely to use buckets with bleach as were those

with formal education. Bucket use decreased among levels of education in the following

order: university, ABC, high school, junior high, and none. Therefore, occupation and

education are significant predictors of use of preventative hygiene practices.

In addition, I explored the association between EVD knowledge and

sociodemographic factors. The first variable, “How you get EVD,” was analyzed with

predictors occupation, education, and gender. Occupation and education, but not gender,

were found to be significant predictors. Within occupation, those working in manual,

sales-services, or “other” occupations were less likely to understand how you get EVD

compared to individuals working in professional-clerical occupations. For education,

those who finished junior high had the greatest level of knowledge followed by ABC,

high school, university, and no formal education.

The second variable, “Knowledge,” encompassed general knowledge questions

about EVD. Occupation was not a significant predictor of EVD knowledge, but education

and gender were. Knowledge about EVD decreased in the following order among the

categories: university, ABC, junior high, high school, and none. Females were more

likely to know about EVD.

Occupation and education, but not gender, were significant predictors of

understanding EVD signs and symptoms. Those working in clerical-professional fields

had the highest amount of knowledge, and those with formal education were more likely

to know the signs and symptoms. For education, knowledge decreased in the following

order: ABC, junior high, high school, and university.

Finally, occupation and education were predictors for understanding how you get

EVD and understanding of EVD signs and symptoms. For EVD knowledge, education

and gender were also significant predictors.

Findings in Relation to the Literature

Research has shown that certain age groups, especially the older, are more at risk

for dying or developing complications from EVD (Glynn, 2015). Older individuals are

also more likely than the rest of the Liberian population to be impoverished or lack

healthcare (Maharaj, 2012). These facts make age an important demographic to consider

when examining the 2014-2016 EVD epidemic. Results from this study showed that age

was not a significant predictor of EVD knowledge or preventative hygiene. This indicates

that despite lack of adequate healthcare or finances, older are still able to practice basic

EVD hygiene measures and understand general EVD knowledge.

In this study, religion was not a significant predictor of EVD knowledge or

preventative hygiene. One of the highest risk activities for contracting EVD is any

contact with bodily fluids including through Liberian burial practices (Pandey, Atkins,

Medlock, Wenzel, Townsend, Childs, & Galvani, 2014). Though burial ceremonies

themselves may differ among religions, the Liberian burial ceremonies that increase

exposure may be more cultural compared to religious as indicated by literature (CDC,

2018). Use of PPE while burying an EVD positive body and use of professional burial

services should continue to be promoted indiscriminately among all religions studied:

Christians, Muslims, tribal, etc. (Suwantarat & Apisarnthanarak, 2015).

Ward was not considered a significant predictor of EVD knowledge or

preventative hygiene use among employed Monrovians, probably because the significant

predictors found in this study (occupation, education, and gender) were more or less

evenly distributed among the communities in Monrovia. Any investment into educational

campaigns proposed in professional practice should thus be done throughout Monrovia

and not be focused in one specific community.

Professional-clerical occupations were found to be significantly more aware of

EVD signs and symptoms and how EVD is spread and were more likely to use

preventative hygiene compared to other professions. This field would include healthcare

workers who are known to be more at risk for contracting EVD (Suwantarat &

Apisarnthanarak, 2015). Literature revealed a lack of resources including medicine and

PPE likely led to increased infection rates among this field (Gee & Skovdal, 2017).

However, these results show that more basic hygiene practices like hand sanitizer and

bleach were being used. It is also logical that healthcare workers would have more

knowledge about EVD signs and symptoms and how EVD is spread compared to other

professions because of their training (Oladimeji, et al., 2015). Though this same study by

Oladimeji, et al. found that often the quality of training varied (2015). Other professional

occupations may be at a similar advantage because of trainings and access to hand

sanitizer or more financial means to purchase preventative hygiene measures (Doocy, &

Burnham, 2006).

Other occupations may put some workers more at risk, such as miners who are

exposed to bats which are known carriers of EVD (CDC, 2018). Hunters and those who

sell bush meat in the markets are also more at risk for contracting EVD (CDC, 2018). It is

therefore concerning that these professions are included in the group of occupations who

were less likely to know how EVD is spread, the signs and symptoms of EVD, and were

less likely to use preventative hygiene. Educational efforts or aide for hygiene items

would be a valuable initiative among these groups (Doocy & Burnham, 2006).

For the variable education, those with formal education (ABC-university)

understood more about EVD and practiced more preventative EVD hygiene measures

compared to those with no formal education. Public health-based education programs

have been proven to be one way to mitigate the spread of EVD (CDC, 2018). Varying

levels of EVD knowledge and hygiene practice among grade levels cannot be clearly

determined, but based on these results and current literature, education on EVD

throughout school would likely be beneficial (CDC, 2018).

Several studies have shown differences in male vs. female hygiene practices and

public health knowledge (Judah et al., 2009). This research revealed women had more

general EVD knowledge compared to men, but the two genders were comparable when

looking at preventative hygiene measures, knowledge of signs and symptoms, and

knowledge of how EVD is spread.

Women are traditionally the caregivers for those who are sick in the home as well

as for extended family (Suwantarat & Apisarnthanarak, 2015). As the epidemic

progressed, mistrust of the government and hospitals led to even more home care

(Suwantarat & Apisarnthanarak, 2015). Women were considered more at risk for

contracting EVD for this reason, and because they were more likely to prepare bodies for

burial (Suwantarat & Apisarnthanarak, 2015). Based on this literature, females in the

study population would have been most likely to have personally cared for an EVD

relative or to have prepared a body for burial (Suwantarat & Apisarnthanarak, 2015).

Thus, women are seeing and experiencing a side of the EVD epidemic statistically more

than men which may be giving them more insight into generally what happens during

EVD progression (Suwantarat & Apisarnthanarak, 2015). Those who care for the sick are

also more likely to take precautions (if possible), which could include seeking additional

educational (Gatherer, 2014).

Interpretation of Findings in the Context of the Theoretical Framework

The Theory of Reasoned Action/ Planned Behavior examines the relationship

between attitudes and behaviors (Bensley & Brookins-Fisher, 2003). For this study, the

behaviors being analyzed were the use of preventative hygiene and learning/remembering

knowledge about EVD. Contracting EVD is the potential negative outcome of not taking

these measures. This analysis assumes that people want to prevent EVD infection. There

are several reasons individuals would be motivated to prevent EVD infection including

working to provide for themselves and their families, serving as caregivers, and because

they witnessed the personal and economic effects from previous epidemics (Suwantarat

& Apisarnthanarak, 2015).

Attitudes about these behaviors can be influenced by several internal and external

factors including beliefs about what outcome the behavior will yield and how valuable

that outcome may be (Bensley & Brookins-Fisher, 2003). In schools, consistent education

may be used to reinforce the belief that learning and remembering EVD facts and using

preventative hygiene will result in not getting infected, and further that not getting

infected will prevent serious illness and possibly death (CDC, 2018). Kids will be more

likely to practice preventative hygiene and retain knowledge if they believe that

remembering the signs and symptoms of EVD, knowing about EVD, and knowing how

to identify EVD prevent infection.

This research showed that individuals working in professional-clerical positions

were more likely to know about EVD signs and symptoms and how EVD is spread and to

practice EVD preventative hygiene. A person in healthcare may believe the hygiene

measures discussed prevent EVD because they have lived experience in addition to

training (CDC, 2016). This belief will lead to those in healthcare being more likely to

practice these behaviors (Bensley & Brookins-Fisher, 2003). If training about

preventative hygiene and the need to have EVD knowledge to reduce transmission is less

among other professions or if access to preventative hygiene is less, than beliefs about

their effectiveness in disease prevention may be less as well (Bensley & Brookins-Fisher,

2003).

Women serve an important role in Liberian society by functioning as caregivers,

mothers, teachers, cooks, etc. (Suwantarat & Apisarnthanarak, 2015). They are essential

to the running of the household and family in the Liberian culture (Suwantarat &

Apisarnthanarak, 2015). Females were identified as having more EVD knowledge than

men in this study population. Within the context of the TRA/ PB, women believe that

having more EVD knowledge is likely to help prevent EVD infection (Bensley &

Brookins-Fisher, 2003). It is unknown whether this is because of lived experience

because they remembered the trauma from previous EVD outbreaks or if it is done

purposefully to prevent the spread to their families after caring for a sick loved one and to

protect themselves as a crucial part of the home.

Social norms also play an important role in behavioral decision making (Bensley

& Brookins-Fisher, 2003). With this study, variations within different education levels

offer an opportunity to look at social norms in more detail. Children will be more likely

to do a desired behavior if a child believes his or her parents, teachers, and peers want

them to learn more about EVD and want them to practice preventative hygiene. Parents

and teachers should show children how proud they are when they practice and remember

what they have learned.

People want approval from their peers, not just throughout school, but as adults

too (Bensley & Brookins-Fisher, 2003). This may offer insight into the occupations

where disparities were seen in EVD knowledge and preventative hygiene practices. If an

individual sees his coworkers taking more precautions or believes his coworkers approve

or want him/her to adopt these behaviors than they are more likely to do so themselves.

Workers were motivated to get back to work after the previous EVD epidemic caused

significant job losses and economic hardship. EVD infection prevention can prevent this

(CDC, 2018). Education could reinforce this as an additional benefit of EVD prevention

beyond preventing serious illness (CDC, 2018). If coworkers and bosses are motivated in

some way to practice these skills, then slowly everyone in the workplace will get on

board. Establishing social norms may be done through training or purchasing hand

sanitizer and bleach for employees as is done for healthcare workers, the population with

the most knowledge of EVD and best prevention practices.

Based on cultural norms, women believe they must serve the role as the caretaker

(Suwantarat & Apisarnthanarak, 2015). As such, women likely believe their family wants

them to take precautions to know about EVD so they can identify the ill and to protect

themselves. However, these theories do not account for other cultural or demographic

motivators. Ward, for example, was a potential predictor used in this study, but it was not

found to be significant (Bensley & Brookins-Fisher, 2003). On the other hand, cultural

factors like traditional burials or hunting bushmeat are considered to be risk factors for

infection (CDC, 2018).

Other external and internal barriers that cannot be overcome may exist as well,

such as financial barriers to purchase hygiene measures, personality barriers (some

people may not be influenced by social norms or have attitudes that are unchangeable by

reason), some people may be afraid, some may be unable to get a professional-clerical

job, and some people may lack access to education (Bensley & Brookins-Fisher, 2003).

Extending education and media campaigns to all literacy levels, even those without

formal education, to all occupations, and possible government subsidizing may be

options. Campaigns could also address mistrust of the government and fear of the disease

so people can begin to feel in control and make more reasoned decisions.

Limitations

There are some limitations to the secondary data set which was chosen for this

research. The primary investigators mitigated many potential issues with the data such as

using a non-governmental agency to do the surveying, collecting data electronically for

monitoring and accuracy, and mapping out the data collection strategically so an accurate

representation of the population would occur. However, regardless of these measures

certain challenges exist.

First, the survey was cross sectional meaning that the data collected can only

demonstrate association and not necessarily causation (HHS, 2007). Cross sectional

surveys like this one also only refer to a brief time period. Thus, this study may not be an

accurate representation of the preventative hygiene practices used throughout the entire

epidemic or the knowledge of EVD among the population for those two years. However,

the data were collected at the early stages of the epidemic in Liberia and still can offer

insights into the questions being posed in this research.

Additionally, there is always the possibility of field error despite safeguards.

Investigators were told to report any problems, and respondents were given anonymity.

Further, keeping the survey simple and electronic hopefully reduced incorrect responses.

Assignement introduction Demographic Characteristics and Response Preparedness

Students also viewed