Comprehensive Critique

Wenyang Cai

Professor Rita DeBonis

May 6, 2016

 

A research critique is an objective assessment of a study’s strengths and limitations. A critical reading of a research article involves a careful appraisal of the researcher’s major conceptual and methodological decisions. Critiques usually conclude with the reviewer’s summary of the study’s merits, recommendations regarding the value of the evidence and suggestions about improving the study or the report. In the critique, the problem statement and purpose, method and research design,  population and sample, date collection and measurement, procedures, results, data analysis, findings, and summary assessment of the study by Jenerette, Brewer, Edwards, Mishel &Gil (2014), will be examined.

Problem Statement and Purpose

The problem statement in the study is well structured, and indicates for most of the problem statements, components that include: Problem identification, background scope of the problem, consequences of the problem and proposed solution. It clearly concludes what needs to be changed. (Polit & Beck, 2014).  It provides the background that stigmas often exist among young adults with sickle cell disease (SCD) when they seek care for pain. The scope of the problem specifically points to young adults with SCD and this population remains consistent throughout the study. The study reports that there is a stigma towards young adults with SCD, preventing them from seeking adequate pain management. Therefore, the consequence of the problem indicates that when such stigma exists, they would not be able to seek proper pain management and would suffer from pain crisis. The study proposes an intervention; the use of care seeking intervention (CSI) to decrease health-related stigma in young adults with SCD. The problem statement is strong, with most of the components well illustrated.

The purpose statement establishes the general direction of the inquiry and captures, usually in one or two sentences, the study’s substance. It is usually easy to identify a purpose statement because the word purpose is explicitly stated and can be identified by words “aim, goal, or objective” (Polit & Beck, 2014).  The purpose of the pilot study was to test an intervention to decrease health-related stigma during care seeking (Jenerette et al., 2014). The aim of the pilot study was to obtain preliminary estimates of the efficacy of a self-care management intervention to lower self-perceived health-related stigma for young adults with SCD (Jenerette et al., 2014). The purpose of the study is concisely and clearly stated.  It is easily identified by the words “purpose and aim”.  In a quantitative study, a statement of purpose identifies the key study variables and their possible interrelationships, as well as the population of interest (Polit & Beck, 2014).  In the purpose statement, the independent variable is the self-care management intervention and the dependent variable is self-perceived health-related stigma towards young adults with SCD. The relationship between these two variables is clearly identified in the purpose statement and is clearly stated.  The self-care management is the intervention used to lower the self-perceived health-related stigma towards young adults with SCD (Jenerette et al., 2014).

The purpose of research is significant in the nursing practice. Sickle cells in blood vessels lead to vaso-occlusion that can be painful and damaging to tissues and organs. Morbidities associated with SCD affect adolescents’ participation in academic, extracurricular, and everyday activities. SCD negatively affects transition from pediatric to adult health care and independent living (Abel et al., 2015). Advances in SCD treatment have proven significance in decreased pain and prolonged survival (Jenerette et al., 2014). Other studies also show that SCD has a significant association with having an overt or silent stroke, or cerebrovascular accident (CVA) (Abel et al., 2015). These evidences indicate that stigma preventing young adults with SCD from seeking health care visits during painful crises could negatively affect individual health in many aspects.

Research Design

In experimental research, researchers actively introduce an intervention or treatment, most often to address therapy questions. While in non-experimental research researchers collect data without intervening or introducing treatments (Polit & Beck, 2014). The pilot study used a prospective, longitudinal design where participants were randomized to either the Care-Seeking Intervention (CSI) group or an attention control group (Jenerette et al., 2014). In order to estimate the efficacy of a self-care management intervention to lower self-perceived health-related stigma for young adults with SCD, experimental measure must be applied. In the study, the researcher adopted the experimental research, which is appropriate for illustrating the research problem, purpose and research questions.

CSI contains two sessions. Session 1 focused on describing possible advantages of seeking care for a pain crisis earlier before the crisis fully evolves, and session 2 focused on assertive communication skills with videos titled “Speak Up to Get the Care You Need” that depicted possible patient-provider interactions featured the most positive outcomes were the result of the patient seeking care early and using the SBAR communication technique (Jenerette et al., 2014). The two sessions in CSI were the first comparison in the study used in order to make a cogent interpretation of the findings.  Attention control groups were used to be consistent with the two sessions of the CSI group, with one focused on review of childhood memories to high school and the other focused on the participants’ lives as young adult living with SCD. (Jenerette et al., 2014). The research control provided validity to the method used to measure variables. In addition, a couple of other comparisons were also made in different parts of the study measures session. In the part of measuring the communication skills, the measurement of communication skills use a communication score assessing scale. Indirect construct validity has been supported by using within- and between-sample comparison to support trends in patient-provider communication found in existing literature (Jenerette et al., 2014). In the part of assessing Health-related stigma, the scale of measurement has three subscales: General Public, Physicians, and Family. The total score of Chronic Pain Stigma Scale is obtained by summing the mean score of the three subscales after required reverse coding. (Jenerette et al., 2014). This comparison of scores from different aspects comparison made the findings more convincible.  These different comparisons made in the study method were ample to enhance interpretability of the finding.

Longitudinal designs involve collecting data multiple times over an extended period. Such designs are useful for studying changes over time and for establishing the sequencing of phenomena. (Polit & Beck, 2014). In the study, the researcher indicated that the pilot study used a prospective longitudinal design where participants were randomized to either the CSI group or an attention control group. Data were collected form August 2010 to September 2012 (Jenerette et al., 2014). Longitudinal design is appropriate for the research because the time period is consistent with the criteria of longitudinal design, it studied changes overtime from August 2010 to September 2012, and data was collected multiple times, with changes being studied by participants completing two post-baseline sessions (Time 1) then followed by each group completing two follow-up sessions for data collection (Time 2 and Time 3) (Jenerette et al., 2014). However, there are defects in this longitudinal design-study. Polit & Beck (2014) states that “A serious challenge in longitudinal studies is the loss of participants (attrition) over time. Attrition is problematic because those who drop out of the study usually differ in important respects from those who continue to participate resulting in potential biases, the risk of faulty inferences, and concerns about the generalizability of the findings”. The study here did encounter this challenge because in the study it indicated that for the participants in CSI Session I, the number n for completed Time 1, completed session I, completed time 2 was 48, 39, and 29 respectively. In CSI Session Ⅱ, n for completed Time 1, completed CSI session Ⅱ and completed Time 3 was 48, 35, and 27 respectively. Same situation in control group, n for completed 1, completed Control session I and completed Time 2 was 42, 34, and 28 respectively. In CSI Session Ⅱ, n for completed Time 1, completed control session Ⅱ and completed Time 3 was 42, 33, and 27 respectively. 40% (36/90) of the sample did not complete the study.  These groups of data illustrate that loss of attrition over time did occurs in the study, this resulting potential biases and increase risk of faulty inferences, which makes the findings less convincing.

The study is not internally valid but is partial externally valid. Internal validity is the extent to which it can be inferred that the independent variable is truly causing the outcome (Polit & Beck, 2014). Inconsistently with the study hypothesis, in change from baseline session, young adults with SCD in CSI group had lower GLM scores than young adults in control group in all dimensions. In CSI group, the p value is p=.3872, p=.1663 in Health-related stigma total scale time 2, time 3 respectively, and p=.2274, p=.0876 in Health-related stigma of doctors subscales time 2, time 3 respectively.  In control group, p=.0133, and p=.0232 in the health-related stigma total scale time 2, time 3 respectively. P=.1040, p=.1280 in Health-related stigma doctors subscale respectively. These findings were contrary to the proposed purpose of the study-to decrease health-related stigma with the CSI. (Jenerette et al., 2014), which makes internal validity not valid.

External validity concerns inferences about whether relationships found for study participants might hold true for different people, conditions, and setting (Polit & Beck, 2014). The sample of participants were obtained through a convenience sample of 90 young adults with SCD that were recruited from the adult outpatient sickle cell clinic affiliated with the Comprehensive Sickle Cell Program at the University of North Carolina at Chapel Hill. This is a representation of the population. In addition, the theory of Self-Care Management for Sickle Cell Disease could be used to guide a similar study in an adolescent population of individuals with SCD as CSI skills may be useful as they begin to interact with health care providers in the adult health care system (Jenerette et al., 2014). However, it may not be able to generalize to populations other than adolescents, or generalize to other conditions and settings.  Therefore, it is considered to be partial externally valid.

Efforts were made to reduce threat to internal validity due to Attrition. The minimization of attrition was attempted in the study by including monetary incentives, parking passes, and participant reminders through telephone calls and text messaging (Jenerette et al., 2014).  However, despite these attempts, 40% (36/90) of the sample did not complete the study (Jenerette et al., 2014). This was problematic as participants that drop out of the study differ in important aspects from those who continue to participate, resulting in potential biases, possibly lead to faulty inferences and make findings internally invalid.

In conclusion, the study design is a quantitative research. It is an experimental design that involves researchers actively introducing an intervention, the Care Seeking Intervention, to participants in order to test the relationship between the independent variable; CSI intervention and dependent variable; efficacy to lower self-perceived health-related stigma among young adults with SCD. It is a longitudinal design that involve collecting data multiple times over an extended period by collecting data continue via Time 2 and Time 3 follow up sessions.

Populations and Samples

Quantitative researchers need to know what characteristic the study participants should possess, and clarify the group to whom study results can be generalized. A population is all the individuals or objects with common, defining characteristics (Polit & Beck, 2014).  The population in the study was identified as “Young adults with sickle cell disease (SCD)”. The sample is 90 young adults with SCD from the outpatient sickle cell clinic affiliated with the Comprehensive Sickle Cell Program at the University of North Carolina at Chapel Hill. The inclusion criteria were a diagnosis of SCD, the ability to read, write, and understand english and the age rage of 18 to 35 years. Individuals were screened by clinic staff and were excluded if they had any known cognitive impairment that would limit their ability to participate in study activities (Jenerette et al., 2014). This sample gave sufficient detail to the sample description.

The study utilized nonprobability sampling. According to Polit & Beck (2014), nonprobability sampling is defined as researchers select elements by nonrandom methods in which every element usually does not have a chance to be included.  The samples chosen were not random since the adults from the outpatient sickle cell clinic affiliated with the Comprehensive Sickle Cell Program at the University of North Carolina at Chapel Hill were available for the study. A convenience sample of 90 young adults with SCD was recruited from there (Jenerette et al., 2014). This sampling is considered to be convenience sampling, which entails selecting the most conveniently available people as participants (Polit & Beck, 2014). Quantitative researchers want samples that allow them to generalize their results to a broader population, and that have adequate power for statistical conculation validity (Polit & Beck, 2014).  The study is externally valid; it can be generalized to adolescent population of individuals with SCD as CSI skills may be useful as they begin to interact with health care providers in the adult health care system. Therefore, this sample selection is appropriate as it allows them to generalize their results to a broader population (Jenerette et al., 2014). A representative sample is one whose characteristics closely resemble those of populations. Sampling bias is the systematic overrepresentation or underrepresentation of some segment of the population in terms of key characteristics (Polit & Beck, 2014).  Inclusion criteria for the study sampling were a diagnosis of SCS, the ability to read, write, and understand English, and the age 18 to 35 years. Individuals were screened by clinic staff and were excluded if they had any known cognitive impairment that would limit their ability to participate in study activities (Jenerette et al., 2014). These criteria used for sample selection minimized the sample biases. However, the sample was not representative of the defined population.  According to Polit & Beck (2014), “nonprobability samples are rarely representative of the population, when every element in the population does not have a chance of being selected”.

Power analysis was not used to estimate sample size needs in the study.  Researchers can estimate how large their samples should be for testing their research hypotheses through power analysis. When using power analysis, researchers must estimate how large the group difference will be (Polit & Beck, 2014). In the study, no evidence indicated that power analysis was performed to determine the adequate sample size. In quantitative studies, researchers can use a power analysis to estimate sample size needs. Large samples are preferable because they enhance statistical conclusion validity and tend to be more representative, but even large samples do not guarantee representativeness (Polit & Beck, 2014).  In the study, the sample size only consists of 90 participants, and 40% (36/90) of the sample did not complete the study (Jenerette et al., 2014), which means only 54 participants were in the sample. This indicated that the sample size was too small to be considered as adequate.

Date Collection and Measurement

Self-Reports method was used in the study to measure the relationship between the variables. According to Polit & Beck (2014), “Structured self-report methods are used when researchers know in advance exactly what they need to know and can frame appropriate questions to obtain the needed information” (p.184). While observation is used for “participants cannot be asked questions or cannot be expected to provide reliable answers, such as infants, children or people whose communication skills are impaired” (p.189). In the study, the researchers clearly defined that the sample’s demographic, communication skills and Health-related stigma are the dimensions that they want to measure. Therefore, the use of Self-Reports is the best method for measuring the key variables.

The study used combination instruments of Self-Report, which include questionnaire for demographic measurement, interviews for communication skills and Health-related stigma measurements, and scales incorporated in interviews of communication skills and health-related stigma (Jenerette et al., 2014). For the Demographic questionnaire, it “requested information to describe the sample by age, sex, education and questions specific to SCD, such as number of crises per year that require hospitalization and the SCD genotype.” (Jenerette et al., 2014). These reviewed the questionnaire content and clearly stated what were aimed to be measured.  For the interviews of Communication skills and Health-related stigma measurement, Likert-type scale was used, the total score and the subscale scores were obtained by summing responses, with higher responses indicating higher levels of communication skills (Jenerette et al., 2014). Both Communication skills and Health-related stigma were measured at different time intervals, including baseline T1 and follow T2 and T3 post intervention or control sessions. (Jenerette et al., 2014). Each of the instrument has been given adequate description of its method and purpose. The combination of the instruments yielded adequate evidences from different aspects and provide comparison of different scales to support the study purpose.

In general, validity is described as the ability of the instrument to measure what it is supposed to measure and reliability is the instrument’s ability to consistently and accurately measure the concept under study (Coughlan et al., 2007). One of the aspect of reliability is internal consistency and it is evaluated by calculating coefficient alpha (Cronbach’s alpha) (Polit & Beck 2014).  The statistic of the study indicated that for the communication skills section, the total scale and internal consistency reliabilities have been reported to range from 0.76 to 0 .92. At the baseline the Cronbach’s alpha reliability coefficients for the total score and subscales were .66, .81, .89, .90 and .07 respectively.  For the Health-related stigma section, Cronbach’s alpha reliability coefficients for the total score and subscales were .84, .70, .69, and .81. These numbers are all within the normal range of values for the reliability index (from .00 to +1.00). (Jenerette et al., 2014). These data indicated that report provide evidences that the data collection methods yielded data that were high on reliability. In addition, study indicated that indirect construct validity has been supported by using within- and between-sample comparison to support trends in patient-provider communication found in existing literature (Jenerette et al., 2014). This use of comparison between within- and between-sample was to ensure that the study was measuring what it is supposed to measure, which is the definition of validity.

Procedures

According to Polit & Beck (2014), “Experiments involve an intervention researcher manipulates the independent variable by introducing an intervention; control (including the use of a control group that is not given the intervention and represents the comparative counterfactual); and randomization or random assignment (with participants allocated to experimental and control groups at random to make the groups comparable at the outset) (p.173). The study clearly states that eligible participants received a letter and were approached for enrollment by the researcher of research assistant in the clinic. Written informed consent was provided, follow with a baseline questionnaire packet that included demographic form and measurement scale of pain, communication and health-related stigma were given. “After these were completed, participants were randomized into either CSI or the attention control group with both groups completing two post-balance sessions to participate in either the CSI or the attention control activity. Then, all subjects completed the outcomes measure, health-related stigma, at two follow-up measurement occasions related to a care-seeking experience that varied by subject.” (Jenerette et al., 2014). The above statement adequately describes that the study adopted an experimental design, which involves use CSI as intervention, control group as to be used for comparison, and randomization as to be used for assigns participants to a control or to the CSI group on a random basis. All characteristics of true experiments have been met, therefore the intervention is adequately stated and properly implemented. “Intervention fidelity is to take steps to monitor that an intervention is faithfully delivered in accordance with its plan and that the intended treatment was actually received.” (Polit & Beck, 2014, p.164).  The study did not provide evidence to indicate that efforts has been made to monitor or enhance intervention fidelity. In addition, evidence shows that 40% of participants quit the study during the Time 2 and Time 3 follow-up secessions (Jenerette et al., 2014), but no actions were taken to insure intervention to be faithfully delivered in accordance with the plan.

For quantitative researchers, powerful tools for minimizing bias include randomness-having certain features of the study established by chance rather than by design or personal preference, and blinding-used to avoid biases stemming from participants’ or research agents’ awareness of study hypotheses or research status. (Polit & Beck, 2014).  As stated in the study, “After completion of the baseline data forms, participants were randomized into either CSI or the attention control group with both groups completing two post-baseline sessions to participate in either the CSI or the attention control activity” (Jenerette et al., 2014), we can see that randomness was used to minimize bias. However, no evidences indicate that blinding was used for the study, neither for the participants’ awareness of the study nor for the research result interpreters. Study shows that some professional efficacy measurement models, such as generalized linear model (GLM) for longitudinal and other correlated outcome data analyzing, ANCOVA model for measurements of score change form baseline in the health-related stigma and doctor subscale, and SAS software for all analyses performed, were used for data collection, measurement and interpretation (Jenerette et al., 2014). All these professional data collection and measurement tools required professional trained staffs to perform. These evidences indicate that staffs who collected data were appropriately trained.

Results

The study did point out that the primary objective was to compare the efficacy of the CSI and attention control interventions in decreasing overall health-related stigma (Jenerette et al., 2014). In the result section, the participants of the sample were described based on sex, employment status, and family relationship status. However, these background characteristics of the sample was not sufficiently described, as it shows no significant differences between CSI and control except with employment where 47.6% of CSI compared with 25% of the control group reported employment (Jenerette et al., 2014).  Study categories include outcomes means by time comparisons between control group and CSI group from time 1 to time2, change in study outcome mean comparison by time1 and time 2, and baseline characteristics by completion status of young adult with SCD in a randomized study of CSI versus attention control group. The mean values were presented in the charts, as well as the SD and P value (Jenerette et al., 2014).  It can be said that the appropriate statistical methods were used as the variables were measured in three sufficient categories with values clearly stated.

Two analytical methods, the GLM model and the ANCOVA model, were used for the analysis. Overall, both methods yielded similar outcomes, and supported the results of the primary analysis conducted through the use of the GLM analysis (Jenerette et al., 2014). The used of combination methods for analysis that yielded the similar result contribute to the study’s reliability, which gave more support to the result’s accuracy and thus made the result more convincible. The methods of analysis are appropriate for the study.

There are four tables in the study that clearly outlines the findings of the study. Titles were carefully labeled column headings, they were baseline characteristics of sample of CSI group versus attention control group, study outcomes mean by time, change in study outcome mean comparison by time, baseline characteristics by completion status of the sample of CSI group versus attention control group (Jenerette et al., 2014). Which were closely adhere to the contents that were discussed in the efficacy result, and were clearly presented to summarize large amounts of statistical information.

Data Analysis

The method of analysis is appropriate for the study. There is no chance for confounding variables to interfere with the study because of both of the CSI group’s outcome and the control group’s outcome has been measured. In compare with the CSI group, the control group’s outcome was measured in order to determine other extraneous variables that may affect the dependent variable would not interfere the outcome of the independent variable of the CSI group. Also, the inclusion criteria for this study sampling was “a diagnosis of SCS, the ability to read, write, and understand English, and the age 18 to 35 years. Individuals were screened by clinic staff and were excluded if they had any known cognitive impairment that would limit their ability to participate in study activities” (Jenerette et al., 2014), which also helps to eliminate confounding variables.

It is clearly stated that the hypotheses were not supported in the study. The efficacy results section presents the probability of the findings. The result of the sensitivity analysis for both GLM analysis and ANCOVA analysis were similar, with significant finding (p<0.05) of baseline health-related stigma total score for both groups, doctor’s subscale of both groups, and care-seeking pain level. In the discussion section, it indicated that subjects in the CSI group reported higher levels of health-related stigma than the attention control group. This was contrary to the proposed hypotheses of the study, which was to decrease health-related stigma with the CSI (Jenerette et al., 2014).

Findings

The information about statistical significance were presented in the study. The demographic comparison between the experiment group and the control group, was demonstrated by using the SD value and percentage of respondents.  Based on these statistics, the conclusion of the demographic and disease characteristics of study subjects were comparable between the CSI and the attention group can be drawn, and it indicated that slightly higher proportion of SCD subjects completed the intervention in the attention control group (64.3%; 27/42) compared with the CSI group (56.3%; 27/48). (Jenerette et al., 2014). In addition, the efficacy results analysis of the GLM model and ANCOVA model show that subjects in the CSI group reported higher levels of health-related stigma than the attention control group. The failure of the CSI intervention was identified as lack of support that may be due to many young adults have already had negative encounter with health care providers when seeking care for SCD pain, which make it more difficult for them to change behaviors. (Jenerette et al., 2014). Based on this finding, the statistical significance was able to be applied to the generalizability of the result, which was identified as that the theory of Self-Care Management for Sickle Cell Disease could be used to guide a similar study in an adolescent population of individuals with SCD as CSI skills may be useful as they begin to interact with health care provider in the adult health care system. (Jenerette et al., 2014).

Information about effect size and probability statistics were presented. The study pointed out the limitation of the probability statistics significance, which was a relatively high proportion of subjects have withdrawn early in the study, and this made a low probability statistics power. This limitation produces threat to statistical conclusion validity. However, this limitation contributes to improvement of statistical significance in the future. As this is a pilot study, the lessons learned would guide the design and conduct of future studies to prevent this problem, which is that future studies need to identifying factors related to missed study activities and it is an important step in designing tailored studies that include strategies to retain subjects. (Jenerette et al., 2014).

Summary Assessment

Despite the limitations, the study findings appear to be valid. In general, validity is described as the ability of the instrument to measure what it is supposed to measure (Coughlan, Cronin & Ryan, 2007). Validity concerns the quality of evidence about the relationship between the independent variable and the dependent variable (Polit & Beck ,2014). The study struggled to use multiple comparison, which included the comparison of the two sessions of CSI group, the subscales comparison of Health-related stigma, and the attention control group versus the CSI group (Jenerette et al., 2014), to ensure that these comparison of studies’ independent variable is the truly the causes of the dependent variable.

Study has pointed out that the result of the study is useful for health care providers, as they can become aware of the health-related stigma towards SCD patients. In addition, the patient should learn to use the SBAR technique, which is frequently used between health care   providers, this would provide the patient and health care providers with common language and may lead to more positive provider’s perception of the credibility of individuals with SCD and thus treat the patient more promptly and appropriately (Jenerette et al., 2014).  As nurses, one of our roles is to be an educator. We can teach patients with SCD the SBAR communication technique, therefore the patient can use it to seek care for pain management more effectively as health-related stigmatization is reduced.  According to Clochesy et al., empowerment of patients to improve communication with healthcare provider is an important component in ensuing self-management of chronic conditions (2015). This article discusses the perspective of patients on effective communication strategies with healthcare encounter and the proposal on communication tool that will enhance the communication with providers of healthcare. Based on these evidences we can see that helping SCD patients with the SBAR communication technique can lead to stigmatization reduction and thus improve treatment seeking effectiveness. This can lead to big progress in nursing practice.

 

References

Abel, R. A., Cho, E., Chadwick-Mansker, K. R., D’Souza, N., Housten, A. J., & King, A. A. (2015). Transition Needs of Adolescents With Sickle Cell Disease. American Journal of Occupational Therapy, 69(2), 1-5 5p. doi:10.5014/ajot.2015.013730

CLOCHESY, J. M., DOLANSKY, M. A., HICKMAN JR., R. L., GITTNER, L. S., & Hickman, R. J. (2015). ENHANCING COMMUNICATION BETWEEN PATIENTS AND HEALTHCARE PROVIDERS: SBAR3.

Coughlan, M., Cronin, P., & Ryan, F. (2007). Step-by-step guide to critiquing research. Part 1: quantitative research. British Journal of Nursing, 16(11), 658-663 6p.

Jenerette, C. M., Brewer, C. A., Edwards, L. J., Mishel, M. H., & Gil, K. M. (2014). An Intervention to Decrease Stigma in Young Adults with Sickle Cell Disease. Western Journal Of Nursing Research, 36(5), 599-619 21p. doi:10.1177/0193945913512724

Polit, D., & Beck, C. (2014).  Essentials of Nursing Research:  Appraising Evidence for Nursing Practice.  Philadelphia. Wolters Kluwer Lippincott Williams & Wilkins.