Part 1:
1. Choose a paper describing the development or validation of a measure of relevance in health disparities research (please give the full citation and/or upload the paper if that's possible).
Peek, M. E., Nunez-Smith, M., Drum, M., & Lewis, T. T. (2011). Adapting the everyday discrimination scale to medical settings: reliability and validity testing in a sample of African American patients. Ethnicity & disease, 21(4), 502-9.
2. What was the definition of the construct?
The construct in this study was the Discrimination in Medical Settings (DMS) scale, which was modified from the Everyday Discrimination Scale in order to assess discrimination faced by patients in medical settings.
3. How did the authors provide evidence on the validity of the measure? Could you think of additional approaches to validating the measure?
The construct was validated in 74 African American patients at an academic medical center in Chicago. Convergent validity was assessed through the following Spearman correlations: between the DMS scale and the Krieger Experiences of discrimination (EOD) scale (which is another measure of racial discrimination that has been validated in multi-ethnic populations); between the DMS scale and the Center for Epidemiologic Studies Depression Scale (CES-D) (which is a screening tool for clinical depression); and between the DMS scale and the African American Trust in Health Care scale (which is a measure of overall mistrust of the health care system). To assess test discriminant validity, the Spearman correlation between the DMS scale and the Modified Marlowe-Crowne Social Desirability Scale (MC-Form C) was used. The following evidence on the validity of the DMS scale were provided: DMS scale was significantly correlated with EOD scale (r=0.51, p<0.01) and with the African American Trust in Health Care scale (r=0.27, p=0.02). DMS scale was inversely associated with the Social Desirability Scale (r=0.18, p=0.13). In addition to depression, the authors could also have checked for correlation of DMS scale with other negative health outcomes, such as psychological distress, anxiety, or smoking, in order to further validate the measure.
4. How did the authors provide evidence on the reliability of the measure? Could you think of additional approaches to evaluating the reliability of the measure?
The reliability was assessed through internal consistency (through Cronbach’s alpha in both the original sample and the retest sample) and test-retest reliability (by Spearman rank correlations between baseline scores and at two month follow-up). The following evidence on the reliability of the DMS scale were provided: The DMS scale had a Cronbach’s alpha of .89 in the original sample and .85 in the retest sample. The test-retest reliability was .58 (P<.001). I cannot think of any additional approaches to evaluate the reliability.
5. Describe the implications of a lack of measurement validity or reliability for future research applications.
If there was a lack of measurement validity or reliability in this study, it would not be wise for future studies to use the DMS scale among African Americans. This is because when an instrument is not valid and not reliable, we cannot successfully measure our “predictor variable”, since we would not know if what has been measured by the scale is indeed the correct predictor. In addition, it should be kept in mind that although this study showed the DMS scale to be valid and reliable for African Americans, the results from this study cannot be applied to other racial/ethnic groups. The DMS scale would have to be validated separately in those other groups. This is because various other factors, such as gender roles, cultural norms and language differences, unrelated to the construct can systematically influence the item responses on the scale.
Part 2:
1. Find a paper describing a health disparity (please give the full citation or, even better, upload the paper so everyone else can look at it too)
Brydsten, A., Hammarström, A., & San Sebastian, M. (2018). Health inequalities between employed and unemployed in northern Sweden: a decomposition analysis of social determinants for mental health. International journal for equity in health, 17(1), 59. doi:10.1186/s12939-018-0773-5
2. Summarize the construct and
measurement of the dimension of disparity (e.g., racial inequalities?, SES
inequalities?) and the outcome measured (e.g., self-rated health).
The study was conducted on the Health on Equal Terms survey, which is a Swedish public health survey conducted every four years, where an effective sample of 10,407 individuals was taken from those who responded. The dimension of disparity that was analyzed was employment status. The construct used to measure this was self-reported employment (employed or unemployed) on the Health on Equal Terms survey. The health outcome assessed was psychological distress, as measured by the twelve-item version of the General Health Questionnaire (GHQ-12). In addition to the exposure and outcome variables described, other social determinants of health inequalities that were grouped into four dimensions: socioeconomic status, economic resources, social network and trust in institutional systems, were also measured. Each of these four dimensions were measured using relevant variables. For instance, SES was measured by education, civil status and occupational class; economic resources was measured by income, cash margin and financial strain, and so on. These social determinants were entered in decomposition analysis to help explain the difference in mental health outcomes between the employed vs unemployed groups.
3. What is the evidence for the validity and reliability of the measures?
While there was no explicit measurement of validity and reliability of the measures in this study, since the investigators were using data from national level surveys to obtain measures of employment and the various social determinants, and because this survey is conducted on a regular basis, we can have some confidence in validity and reliability of the measures. In addition, it is mentioned that the twelve-item General Health Questionnaire (used to measure the outcome) is a validated scale (i.e. it was validated in previous studies in the Swedish population).
4. What is the reference category used for the disparity measure (ie, who is the comparison group)? Why does this reference category make sense (or not) for this research question?
The reference category is employed individuals. Since this study is comparing the psychological distress between adults who are unemployed and employed, this seems to be an appropriate reference category.
5. How is the disparity quantified or measured? Is this an absolute or relative measure or are both provided? Describe which type of measure you would prefer for this research area, or, if both, why.
The mental health disparity (i.e. disparity in psychological distress between the employed and unemployed) was quantified using the twelve-item General Health Questionnaire, which is a widely used screening instrument for non-psychotic mental illnesses. Patients are asked questions about their symptoms and behaviour in the last few weeks and each response is reported in a 4-point scale based on severity. The items are then scored into an index with a range of 0-12. In the analysis phase, the difference in psychological distress was estimated, between the unemployed group and the employed group and was presented as log odds, which is a relative measure. In my opinion, this is a good measure in mental health research, where the outcomes in many cases cannot be quantified in absolute terms.
Part 3:
1. Read someone else's response to part 1 above (identifying a construct) and comment, specifically noting whether you can see any additional implications of measurement quality for future research or whether you agree with those noted by your classmate.
I agree with Nicholas’ suggestion that it would be better to have the content validity of the diabetes health literacy scale to be measured by patients who have very good diabetes control and have good health literacy (as deemed by their providers), rather than by experts on health literacy and diabetes educators. This is because our “target population” are patients with diabetes. As such, our construct should be validated in this population for more optimal validation results as opposed to being validated by experts.