Ethical Article Analysis

Application of AnalysisBy Elissa Torres Essential Questions

• What are the essential elements in evaluating prior research? • How does the analysis of quantitative versus qualitative studies differ? • How are results communicated from data collection and analysis?

Introduction The use of statistics and statistical analysis is part of the clinical practitioner’s role. This may appear in different ways from reviewing existing clinical research to participating in a study. There are some critical questions when understanding statistics and the role of clinician in health care:

• Why is it important to keep up-to-date on clinical research? • Why is it important for health care facilities to conduct ongoing studies? • What type of studies are important?

Previous chapters focused on understanding the elements of statistics and research, including how to select and conduct hypothesis testing based upon the type of data collected. This chapter focuses on the application of prior information to understand information written in prior research studies and set up statistical tests and interpret results both statistically and clinically.

Academic Research Study Extraction In the evaluation of research articles, it is important that key areas can be identified for interpretation and understanding. In the review of both qualitative and quantitative research, it can be daunting to extract the relevant information to determine the primary goals and outcomes of the study. For clinical studies, this also means addressing the epidemiology.

The simplest way to extract relevant information is to first start with those key areas.

1. Topic: What is the broad topic research area/title? 2. Problem statement: What is the problem that the research is attempting to address? In

many studies, authors identify a lack of research in a specific area or population. 3. Purpose statement: Why did the author complete the study? In some studies, this often

appears in a sentence containing the phrase, “the focus of this study … ” 4. Research questions: What specific questions does the author need to address? In many

articles, this is not explicitly written but can be derived. 5. Hypothesis, variables, or phenomena: What are the variables the author has identified to

address the research goal (quantitative)? How is the phenomena described that the author seeks to better understand (qualitative)?

6. Sample and location: What was the sample used, and where did the study take place? 7. Methodology: Was the research quantitative or qualitative? Did the author provide any

more details, such as quantitative correlational or qualitative case study?

8. Data collection: How did the author approach data collection? For example, did the author use surveys, interviews, or clinical studies?

9. Data analysis: What approach did the author use to analyze the data? Did the author mention statistical tests? What type of statistical data was provided? What type of information is provided with qualitative studies?

10. Results: What were the results of the study? Did the author find anything significant? Did the study address epidemiology?

These 10 questions for article evaluation are useful to perform a quick review of the study’s key elements; however, it is important to start the process by first reading the full article. The format in which information is displayed in Table 5.1 can be used as a template to organize information found for each of these article elements. In some studies, information can be easily located in the abstract and in clearly organized sections; however, this is not always the case. Table 5.1

Quantitative Article Evaluation

Article Citation Aljohani, A. H., Alrubyyi, M. A., Alharbi, A. B., Alomair, A. M., Alomair, A. A., Aldossari, N. A., & ... Tallab, O. M. (2018). The relation between diabetes type II and anemia. The Egyptian Journal of Hospital Medicine, 70(4), 526. doi:10.12816/0043795

Point Description

Broad Topic Area/Title

The Relation Between Diabetes Type II and Anemia

Problem Statement

“There is consequently a need for more studies on the incidence and prevalence of anemia among patients with diabetes mainly those with renal malfunction” (p. 527).

Purpose Statement

“This study consequently purposed to determine the pervasiveness of anemia due to renal insufficiency among patients with type 2 diabetes” (p. 526, 527).

Research Questions

Is there a relationship between patients with anemia and patients with type II diabetes?

Define Variables/ Hypotheses

Categorical variable: Gender Continuous variables: Age, Hb, Ferritin, MCV, TIBC, FBG, Erythroietin, eGFR, Urea, Na, K, CA, and HbA1c (found on pages 528 and 529)

Sample 50 participants Case group: 25 participants with diabetes (8 male/17 female) Control group: 25 participants without diabetes (7 male/18 female (p. 528)

Methodology Quantitative, case-control study (p. 527)

How was Data Collected?

Medical records for the patients were examined from physical examinations (p. 528)

How was Data Analyzed?

SPSS; descriptive statistics for categorical; summary statistics, independent t-test; and ANOVA test; Pearson correlation for Hb and HG for both male and female (p. 528)

What Were the Results?

The study indicated the following were statistically significant (low p-values) between the case group and control group. Hb Male and Hb Female Ferritin Male and Ferritin Female MCV TIBC Of the biochemical parameters, the following were significant: FBG, Erthropoietin, eGFR, Urea, K, C1, Ca, HbA1c Creatinine was not significant In the correlation test, HB and HG (female) was significant, but HB and HG (male) was not significant. (pp. 528-529) Clinical implications: The study did find a higher occurrence of anemia in patients with diabetes (87.5% males, 82.3% female). The study also concluded that the presence of anemia may increase the likelihood of poorly controlled diabetes (p. 529).

Check for Understanding

1. Would there be any additional evaluation of the article?

2. Did the researchers appear to follow ethical guidelines?

3. What were the assumptions and limitations of the study? Table 5.2

Qualitative Article Evaluation

Article Citation Jangland, E., Nyberg, B., & Yngman-Uhlin, P. (2017). It's a matter of patient safety: Understanding challenges in everyday clinical practice for achieving good care on the surgical ward - a qualitative study. Scandinavian Journal of Caring Sciences, 31(2), 323- 331. doi:10.1111/scs.12350

Point Description

Broad Topic Area/Title

Identify the challenges and barriers linked to quality care and patient safety in the surgical ward.

Problem Statement

“Identify the challenges and barriers linked to quality of care and patient safety in the surgical ward” (p. 324). Study addresses gap where there were only a few studies that looked at both the nurses’ and leaders’ perspective.

Purpose Statement

“The aim of this study was to explore, from the perspectives of care leaders, the situations and processes that support or hinder good and safe care on the surgical ward” (p. 324).

Research Questions

What are the perspectives of leaders on the processes that support good quality care in the surgical ward? What are the perspectives of leaders on processes that hinder good quality care in the surgical ward? How do leaders’ experiences inform improvement in clinical practice?

Describe Phenomena

Categorical variable: Gender Continuous variables: Age, Hb, Ferritin, MCV, TIBC, FBG, Erythroietin, eGFR, Urea, Na, K, CA, and HbA1c (found on pages 528 and 529)

Sample “10 leaders in surgery departments (four department leaders and six nursing managers) from 1 university hospital and 2 county hospitals in different regions in Sweden” (pp. 324- 325).

Methodology Qualitative-descriptive design

How was Data Collected?

Repeated reflective interviews using semistructured interviews

How was Data Analyzed?

Systematic text condensation

What Were the Results?

Study identified four major themes (pp. 326-328):

1. Constant demands for increased efficiency and production 2. Continual nursing turnover and loss of competence 3. A traditional hierarchical culture 4. Vague goals and responsibilities in the development of nursing care

Clinical implications: Based upon the study, which has limitations as it was performed in one country (Sweden), organizational changes are required to ensure higher levels of competence of staff and resources available to surgical ward nurses to ensure higher quality care (p. 330).

The two evaluations above provide a roadmap for reviewing prior research. Much of the research completed in the clinical setting may not be as comprehensive; however, it is important to understand the process. In a clinical setting, there may be opportunities to reduce cycle time, increase quality, or participate in studies that influence health outcomes. Understanding the process, knowing how to evaluate the data, and communicating the results enables contribution to the organization.

Application of Statistics to Scenario A medical office has noticed an increase in patient dissatisfaction and as well as an increase in usage of urgent care facility services rather than seeing their primary care physicians (PCPs). To increase understanding of the patient perception, the office surveyed the patients and received 81 responses. The survey includes a total of eight questions. The first five questions capture satisfaction and urgent care utilization responses, and the last three questions capture data on education, gender, and age group.

• Q1: You meet with your Primary Care Physician greater than one time per year. Responses Strongly Disagree to Strongly Agree.

• Q2: You spend more than 10 minutes with your Primary Care Physician discussing health concerns. Responses Strongly Disagree to Strongly Agree.

• Q3: You are more likely to go to urgent care versus your Primary Care Physician. Responses Strongly Disagree to Strongly Agree.

• Q4: What is the number of times you went to urgent care in the past 12 months? Numerical response requested.

• Q5: Rate your overall satisfaction with the medical office. Responses Strongly Disagree to Strongly Agree.

• Q6: What is the highest level of education you completed? • Q7: What is your gender? • Q8: What is your age?

To review the responses from the data collected in the scenario, click on the button below. Scenario Data

Table 5.3

Patient Dissatisfaction Application Scenario

Point Description

Broad Topic Area/Title Understand the relationship between patient satisfaction and usage of services at urgent care facilities.

Problem Statement Recent indicator identified lower patient satisfaction and higher incidence of using services at urgent care facilities. There is a need to understand the perception of patient satisfaction for the XYZ medical office and decrease usage of urgent care.

Research Questions What is the patient perception of satisfaction with the medical office? Do patients use urgent care as an alternative to the primary care physician (PCP)? Is there a relationship between patient satisfaction and usage of urgent care facilities?

Hypothesis H10: There is no relationship between the perception for number of visits and perception of time spent with PCP. H1A: There is a relationship between the perception for number of visits and perception of time spent with PCP. H20: There is no relationship between the perception for number of visits and the likelihood to go to urgent care. H2A: There is a relationship between the perception for number of visits and the likelihood to go to urgent care. H30: There is no relationship between the perception for number of visits and the overall satisfaction.

H3A: There is a relationship between the perception for number of visits and the overall satisfaction. H40: There is no relationship between the perception time spent with PCP and likelihood to go to urgent care. H4A: There is a relationship between the perception of time spent with PCP and likelihood to go to urgent care. H50: There is no relationship between the perception of time spent with PCP and overall satisfaction. H5A: There is a relationship between the perception of time spent with PCP and overall satisfaction. H50: There is no relationship between the number of visits to urgent care in past 12 months and overall satisfaction. H5A: There is no relationship between the number of times went to urgent care in past 12 months and overall satisfaction.

Describe Phenomena (qualitative) or Define Variables/ Hypotheses (quantitative)

Nominal: education, gender, age group Ordinal: Survey Questions 1-3 and 5 Continuous: Survey Question 4: Number of visits to urgent care in last 12 months

Sample 80 patients from XYZ medical office

How is Data Being Collected? Sent electronic survey to 300 patients, and received 80 responses.

How Will Data be Analyzed Descriptive statistics

Correlation analysis

What Were the Results? Statistical relationships were identified. The null hypothesis would be rejected and the alternative hypothesis would be accepted in all cases. From a practical perspective, while the results indicated higher scores for the likelihood to go to urgent care versus the PCP, the actual descriptive statistics for urgent care visits do not support this.

Communicating Results The data can be sorted for communication based upon summary and descriptive statistics for some of the variables prior to the hypothesis tests. As an example, to describe the sample respondents by age group and gender, the data can be converted in Excel to percentages (see Table 5.4). These percentages can be written out or included in a table. Table 5.4

Converting Frequency to Percentage Example

Age Group Female Percent Female Male Percent Male Total Percent Total by Age Group

< 20 9 18.0% 2 6.7% 11 13.8%

20-25 7 14.0% 4 13.3% 11 13.8%

23-31 10 20.0% 5 16.7% 15 18.8%

32-37 8 16.0% 4 13.3% 12 15.0%

38-43 6 12.0% 4 13.3% 10 12.5%

> 44 10 20.0% 11 36.7% 21 26.3%

Total 50 30 80

Even though the responses to the survey questions were ordinal as they were translated from Strongly Disagree (1) to Strongly Agree (5), with larger samples, responses can be treated as continuous. Frequently, the three most common forms of descriptive statistics are displayed in a chart. These include the mean, median, and standard deviation (see Table 5.5). Table 5.5

Example of Descriptive Statistics

Question n M Mdn SD

Q1 80 1.93 2.00 1.11

Q2 80 2.15 2.00 1.29

Q3 80 3.31 4.00 1.41

Q4 80 1.41 1.00 1.37

Q5 80 3.13 3.00 1.31

Beyond addressing some information on descriptive statistics, the hypothesis tests need to be addressed. Prior to conducting statistical testing, the data needs to be assessed for normality. When assessing for normality, a statistical program, such as SPSS, determines if the data meets the conditions of a normal distribution. Often, when data is derived from survey data responses with ranges from strongly disagree to strongly agree, the data is not normally distributed unless the samples are very large. In this case, the sample received was only 80. Table 5.6 displays the normality tests for the variables that will be tested. Because the sample size is lower, the Shapiro-Wilk results should be used. The Kolmogorov-Smirnov test is most applicable for samples of more than 2,000 data points. Based upon a 0.05 level of significance, a researcher would reject the null hypothesis, which stated that the data was normally distributed. Table 5.6

Test for Normality

Tests of Normality

Kolmogoroz-Smirnova Shapiro-Wilk

Statistic df Sig. Statistic df Sig.

Q1 .247 80 .000 .771 80 .000

Q2 .250 80 .000 .810 80 .000

Q3 .237 80 .000 .866 80 .000

Q4 .256 80 .000 .801 80 .000

Q5 .211 80 .000 .895 80 .000

a. Lilliefors Significance Correction

Because the test results identified that the data was not normally distributed, a nonparametric test would be used to conduct the hypothesis testing for correlation. The correlation test to use in this scenario is the Spearman Rho test. If the data was normally distributed, the commonly used Pearson Product Moment test would be used. Table 5.7 demonstrates the SPSS output for the Spearman Rho correlation test between survey Questions 1 and 2. Correlation coefficients are reviewed on a scale of -1 to +1. The relationship is stronger if the calculated coefficient is closer to either -1 or +1. In this case, there is a strong relationship between meeting with the PCP more than one time per year and spending more than 10 minutes with the PCP discussing health concerns. Another statistic to review in the output is the p value. If the p-value is less than the level of significance identified in the study, the null hypothesis would be rejected and the alternative hypothesis would be accepted. Table 5.7

Test for Correlation Q1&Q2

Spearman’s rho Q1 Correlation Coefficient 1.000 .777**

Sig. (2-tailed) . .000

N 80 80

Q2 Correlation Coefficient .777** 1.000

Sig. (2-tailed) .000 .

N 80 80

Correlation coefficients are reviewed on a scale of -1 to +1. The relationship is stronger if the calculated coefficient is closer to either -1 or +1. If the correlation coefficient is positive, then the two variables are moved upward in the same direction. If the statistic is negative, then one variable increases as the other variable results decrease (Levine, Krehbiel, Berenson, 2013). In this case, there is a strong relationship between meeting with PCP more than one time per year and spending more than 10 minutes with the PCP discussing health concerns. Another statistic to review in the output is the p-value. If the p-value is less than the level of significance identified in the study, the null hypothesis would be rejected and the alternative hypothesis would be accepted. Table 5.8 displays the remaining correlation coefficients depicted in the table as r and the corresponding p-values for the test. Table 5.8

Correlation tests from Example

Variable n r’s p-value

Q1&Q2 80 .777 .000*

Q1&Q3 80 .566 .000*

Q1&Q5 80 -.313 .005*

Q2&Q3 80 .419 .000*

Q2&Q5 80 -.348 .002*

Q4&Q5 80 -.212 .060*

Table 5.8 demonstrates that there is a statistical correlation between all variables tested at a 0.05 level significance except Q4 (number of times visited urgent care in the last 12 months) and Q5 (overall satisfaction with the medical office). The data output requires analysis to the original hypothesis questions in the study.

Reflective Summary This chapter reviewed the application of statistics to research, how to identify data, select the appropriate tests, and apply this to data sets. The chapter also explored how to review articles or studies for the key elements for understanding. This understanding was further applied to a practical scenario including analysis of data collected. The statistical and practical analysis of results for communication are essential in the roles of a clinician and the tools learned in this course provided the framework for increased understanding.

Key Terms Hypothesis: A testable statement of a relationship; an epidemiologic hypothesis is the relationship is between the exposure (person, time, and/or place) and the occurrence of a disease or condition.

M: Table notation for statistical mean of data array.

Mdn: Table notation for statistical median of data array.

N: Table notation representing the sample size.

P values: The probability that there is enough evidence to make conclusions resulting from the data collected in the study.

r: Table notation representing the coefficient of correlation.

SD: Table notation representing the standard deviation of the data array.

Variable: A data item such as characteristics, numbers, properties, or quantities that can be measured or counted. The value of the data item can vary or be manipulated from one entity to another. There are three different types of variables—dependent, independent, and extraneous.

References Aljohani, A. H., Alrubyyi, M. A., Alharbi, A. B., Alomair, A. M., Alomair, A. A., Aldossari, N. A.,

& ... Tallab, O. M. (2018). The relation between diabetes type II and anemia. The Egyptian Journal of Hospital Medicine, 70(4), 526. doi:10.12816/0043795

Levine, D. M., Krehbiel, T. C., & Berenson, M. L. (2013). Business statistics: A first course (6th ed.). Upper Saddle River, NJ: Pearson.

Jangland, E., Nyberg, B., & Yngman-Uhlin, P. (2017). It's a matter of patient safety: Understanding challenges in everyday clinical practice for achieving good care on the surgical ward - a qualitative study. Scandinavian Journal of Caring Sciences, 31(2), 323-331. doi:10.1111/scs.12350