Quantitative Appraisal of a Psychoeducation Research Study
In the literature review of a research process, it is important to establish the type of study to be done in order to have a good foundation and apply appropriate steps. Qualitative Research is focused on patient patterns in a particular population which is useful to identify a definitive intervention (Terry, 2018, p. 74). The article by McGillion, et al., (2008) is a qualitative study to evaluate the effect of a Chronic Angina Self Management Program (CASMP) on patients with Chronic Stable Angina (CSA). This paper will address the research purpose, question, hypothesis, sample size, study design, the validity of the measurement variables, and reliability of the results and what the research means for clinical practice.
Understanding the purpose of a study helps fill in the gap of why the study is relevant. A brief background on why the study is needed helps to identify what is missing and leads to the study purpose (Melnyk & Fineout-Overholt, 2015, p. 88). The study by McGillion, et al., (2008) was done because the authors saw the need to implement a Chronic Angina Self Management Program (CASMP) to help patients with chronic stable angina. Current interventions did not adequately address the “multidimensional ischemic and persistent pain problem” (McGillion, et al., 2008, p. 127). The goal was for patients with chronic stable angina to have an improved health-related quality of life (HRQL).
The Research Question
Selecting the research question for this study using PICOT unabbreviated as Population, Intervention, Comparison, Outcome and Time (Melnyk & Fineout-Overholt, 2015, p. 29) shows that the population will be made up of individuals with Chronic Stable Angina (CSA), Intervention is the implementation of CASMP, Comparison are individuals who do not get the CASMP intervention, Outcome is to see an improvement in the HRQL, and time is three months. So, the research question is: Will the implementation of CASMP help patients with chronic stable angina (CSA) have an improved health-related quality of life, by being able to better manage pain at three months compared to patients who will not receive CASMP?
A hypothesis is defined as a prediction about the relationships between two variables (Dependent and independent) in a study (Melnyk & Fineout-Overholt, 2015, p. 452). In the article titled “Randomized controlled trial of a psychoeducational program for the self-management of chronic cardiac pain” the authors hypothesized that with patients who have CSA (dependent variable), the implementation of CASMP (independent variable) will help them have an improved HRQL and better manage pain in three months.
Sample Size and Type of Sampling Used
In addressing the sample size, the external validity should be considered as the study results should be applicable to a broader population. A power analysis should be done to determine the sample size and minimize findings based on chance (Melnyk & Fineout-Overholt, 2015, p. 89). Although the samples may not perfectly represent the identified population, there can be reasonable estimates (Melnyk & Fineout-Overholt, 2015, p. 465). Type of sampling for this study was Random Sampling and a priori power analysis was done. The study was conducted in Canada over an 18-month period, participants had to be able to speak, read and understand English. Recruitment was from three cardiac outpatient programs in three university teaching hospitals. The sample size was 130 participants who were randomized to CASMP or three-month wait list of usual care. A total of 117 completed the study (McGillion, et al., 2008, p. 127). The authors specified a 10-point difference in SF-36 scores as being clinically significant and the sample size tested was set to test for this difference. Larger standard deviation (SD) was reported for the two scales of SF-36 which required estimated sample sizes beyond the allowable timeframe for the study so, it was expected that there may be an inadequate power analysis to detect a meaningful change to the two SF-36 scales. The program used to compute the sample size estimate was the nQuery Advisor 4.0.
Selecting an appropriate study design is crucial as it forms the foundation of the study, to adequately test the study hypothesis. The study design used in the article was a randomized controlled trial which is the best design to study cause and effect, providing strong evidence to improve clinical practice (Melnyk & Fineout-Overholt, 2015, pp. 452-453). Participants were randomly selected to either the six-week intervention group or the three-month waitlist control group. The authors controlled the randomization process centrally using a university-based tamper proof computerized system(McGillion, et al., 2008, p. 128). This was an experimental study as the research question or hypothesis is focused on testing the effects of the intervention of CASMP on patient’s outcome. A good quantitative design should adequately test the hypothesis, lack bias, control extraneous variables and be able to detect statistically significant findings (Melnyk ; Fineout-Overholt, 2015, p. 452).
There are two variables to consider in a study; independent and dependent. “The independent variable leads to the effect produced in the dependent variable” (Terry, 2018, p. 24). The study variables in the article are the use of CASMP and the HRQL outcome on CSA patients. HRQL outcome on CSL patients is the dependent variable, while the use of CASMP is the independent variable.
Reliability of the Measurements of Major Variables
According to Melnyk and Fineout-Overholt, 2015 (p. 89), a measurement tool is accurate when it measures what it is supposed to measure, is stable over time and made up of various items that consistently analyze the same construct. Study findings are deemed accurate when measures of internal consistency are at least 0.70 (p. 470).
A baseline questionnaire was used to obtain the sociodemographic information, angina and all clinically related characteristics. Braden’s Evidence-based Self-help Model of Learned Response to Chronic Illness Experience guided the author’s selection of trial outcomes (McGillion, et al., 2008, p. 130).
HRQL was measured using the Medical Outcome Study 36-Item Short Form (SF-36). This form is noted to be a comprehensive instrument that captures functional status that includes behavioral and dysfunction, distress and wellbeing and also, self-evaluation of general health status. Eight subscales that represent overall quality of life were used which are “physical functioning (PF), role limitations due to physical problems (RP), social functioning (SF), bodily pain (BP), mental health (MH), role limitations due to emotional problems (RE), vitality (VT), and general health perception (GH)” (McGillion, et al., 2008, p. 130).
A reliability estimate was carried out for all eight SF-36 subscales and evidence showed that some potential exists for the SF-36 to be insensitive to changes in the angina class so the authors used another instrument specific to the disease, the Seattle Angina Questionnaire (SAQ) in their evaluation of HRQL. The SAQ has 19 items that can quantify five important domains of CAD: Angina, physical limitations, pain stability frequency, satisfaction of treatment and perception of the disease. It is noted that this instrument has been used in a number of similar studies and proofs to be a valid measurement tool. Internal consistency reliabilities were noted to be 0.85 for PL, 0.71 AF, 0.73 TS, and 0.68 DP.
Self-Efficacy was measured with a modified version of the 11-item pain and Resourcefulness was measured by Rosen-baum’s Self-Control Schedule (SCS). Both had an internal consistency of 0.94 and 0.80 respectively.
In preparation to analyze data, it is important to assess the study for completeness and determine what strategies to use in cases of missing data (Melnyk & Fineout-Overholt, 2015, p. 452). If more than 30% of data is missing from a questionnaire, in most cases it is eliminated. If less than 30% of data is missing, it is acceptable to use the mean for missing items. Researchers should be aware of common statistical errors such as focusing on the p-value, which is when the researcher chooses the test based on the statistical significance, selective reporting of outcomes with significant findings and incomplete data.
According to the McGillion, et al., (2008) the data were analyzed based on the intention-to-treat principle. The Chi-squared analysis for discrete level data and the Student t-test for continuous level data were used for the equivalence of groups on baseline demographic characteristics and pretest scores (p. 131). To prevent Type 1 error (A false positive that shows a difference exist between interventions when in actuality no difference exists) on the SF-36 and SAQ data, the authors conducted a multivariance analysis of variance (MANOVA) before the analysis of covariance (ANCOVA). To ensure that the effect of the interventions is apparent, the authors chose a change score approach so that the differences in change scores between the control and treatment groups would be available to the reader (pp. 131-132).
Reliability and Validity of the Results
The validity of the study results is based on if the results were obtained through solid scientific methods by eliminating bias. Bias such as selection bias, which refers to the selection process, measurement bias which refers to systematic errors or errors from data collection, recall bias regarding memory recall and informational bias with regards to recording different information from what was given in interviews or patient records (Melnyk & Fineout-Overholt, 2015, pp. 94-95). Since quantitative studies depend on statistical data the numerical results from the study needs to be analyzed to access for reliability.
A well-rounded method was used to minimize bias and random error to include priori power analysis, randomization that was centrally controlled, blinding of the data collectors, intention-to-treat analysis and an examination for the possibility of an interventional cohort effect. Also, the use of a wait-list control condition and diligent follow up ensured adequate follow-up and minimal loss (McGillion, et al., 2008, p. 136). Lastly, an external auditor verified the treatment integrity. It was noted that because the participants could not be blinded, performance bias could not be ruled out.
The authors mentioned that the participant’s enhancement of physical activity may account for the slight improvement in the SF-36 result on physical conditioning and general health. Also, because the sample size had participants with differing angina sub-class, some of the SF-36 subscales may have been insensitive to improvements in angina induced disability. As noted above, performance bias could not be ruled out as it was impossible to blind participants in a social-based study. Sample size bias was also considered and a recommendation to use a larger sample size in future studies (McGillion, et al., 2008, pp. 136-137).
How the Study Fits with Previous Research in that Area
Prior research hasn’t been done on psychoeducational-based trials for CSA using SF-36 or SAQ in comparison to HRQL. However, the authors did note that studies have used other means to evaluate HRQL. Four Psychoeducation trials were reported that showed significant symptom improvement, including duration, frequency, and cardiac pain. Lefort et al.’s CPSMP trial is a study that was focused on a different population however used the SF-36 to assess how psychoeducation affects chronic pain. In comparison to the study on the use of CASMP, Lefort et al. study showed that their CPSMP program significantly improved SF-36 bodily pain, vitality, physical functioning and mental health for individuals with noncancerous pain (McGillion, et al., 2008, p. 135).
Impact on Clinical Practice
According to Melnyk and Fineout-Overholt (2015), the reason for reviewing research and its relevance in evidence-based decision making is to improve clinical practice and outcome. This study is relevant as it provides evidence that shows the effectiveness of CASMP in improving physical functioning, angina pain and frequency, and self-efficacy to manage angina in patients who have CSA in three months. It is noted that there is a risk of bias in the sample size and questions on its applicability to a larger participant group. So, further research is recommended on a larger sample size.