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).
Baer TE, Scherer EA, Fleegler EW, Hassan A. Food Insecurity and the Burden of Health-Related Social Problems in an Urban Youth Population. The Journal of adolescent health: official publication of the Society for Adolescent Medicine 2015;57:601-7.
2. What was the definition of the construct?
The authors used the USDA’s definition of Food insecurity, “uncertainty of having, or inability to acquire, enough food to meet the requirements of all members of a household because of financial or resource constraint.” The study objectives were to look at the prevalence and severity of food insecurity in young adults, to look for associations between degree of food insecurity and other health-related social needs, and to look at the sensitivity and specificity of a 2-item food insecurity tool developed in a pediatric clinic (in caregivers of children under 3 years old) in a young adult urban primary care clinic, against the age appropriate USDA-Food Security Survey (FSS), which is considered the gold standard for FI screening.
3.How did the authors provide evidence on the validity of the measure? Could you think of additional approaches to validating the measure?
The tool being tested, a two-item food insecurity screening tool developed by Hager et al. reads: Q1: “Within the past 12 months we worried that our food would run out before we got money to buy more” and Q2: “Within the past 12 months the food we bought just did not last and we did not have money to get more.” Answers options for both questions are: “often true,” “sometimes true,” or “never true.” Participants are considered food insecure if they answer “often true” or “sometimes true” to either question, and considered food secure if they answered “never true” to both questions.
As noted above, participants were screened for food insecurity using the age-appropriate USDA Food Security Survey. If patients aged 18-25 years identified as parents, they completed the 18-item FSS; non-parents aged 18-25 years completed the 10-item Adult FSS; and participants aged 12-17 years completed the 9-item FSS for Youths. Survey answers were scored per USDA algorithms, being broken down into categories of high food security, marginal food security, low and very low food security.
The authors assessed validity of the two-item screening tool by calculating sensitivity, specificity, negative predictive value and positive predictive value of the two-item tool compared to the age appropriate USDA-FSS.
Not having the level of statistical detail as the Krieger article, the authors could have developed and tested a validation model to determine if the food insecurity questions were all evaluating the same underlying constructs. They could have specified structural equation models exploring the structure of the different measures of food in security and fit the structural equation models by the comparative fit index and root mean square error of approximation, as they did in Krieger (methods for which I honestly do not fully grasp).
Alternatively, instead of comparing the 2-item screener to the USDA-FSS, they could have conducted qualitative interviews with patients to ask additional questions about food security, to assess food security status and ability of the 2-item screener to uncover food insecurity.
Hager ER, Quigg AM, Black MM, et al. Development and validity of a 2-item screen to identify families at risk for food insecurity. Pediatrics 2010;126:e26-32.
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 authors provided evidence of reliability by testing the tool, previously validated in caregivers of children 0-3 years of age, in an adolescent population. The authors did not use any specific analytic approach to reliability, such as Cronbach’s alpha or test-re-test correlations, as they did in Kriegers, which they could have done.
Aside from testing the 2-item screening in adolescents, additional populations (both other demographics of youth and non-youths) should be done.
5. Describe the implications of a lack of measurement validity or reliability for future research applications.
Without measurement of validity or reliability, it is difficult, if not impossible, to know the implications of the study, and how it might apply to other study populations. Not knowing if the tool is valid or reliable in other patient populations is challenging for clinics wanting to screen for food insecurity but not knowing what tool to use. The Hunger Vital Signs (the name for the 2-item screener) is the most widely adopted/recommended, but still largely un-validated in diverse patient groups—and not yet validated in adults. No food insecurity screening tool has been validated for use in non-pediatric patients within the health care setting.
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)
Lau M, Lin H, Flores G. Race/Ethnic Disparities in Health and Health Care among U.S. Adolescents. Health Serv Res. 2012 Oct;47(5):2031-59
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 authors focus on race and ethnic disparities as defined by parent reported race/ethnicity on the 2003 National Survey of Children’s Health (NSCH). The outcome of interest was broad, looking at 40 health and health care variables over three domains of parent-reported medical and dental health status, access to care and use of medical and dental services (all parent-reported). Focused on parents of adolescents aged 10-17 years old. Race/ethnicity was broken into: White, African American, Latino, Asian/Pacific Islander, American Indian/Alaskan Native, and multiracial (if parent selected more than one race).
3. What is the evidence for the validity and reliability of the measures?
There was limited evidence for validity and reliability. The authors controlled for sociodemographic characteristics, to control for confounding, and performed bivariate and logistic regression analyses. The authors utilized previously reported on NSCH analysis techniques, providing support for their analyses.
Being a cross-sectional analyses of parent-reported information, there was no way for the authors to verify any reported information, which is a limitation of the study.
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 was white race in bivariate analyses to look at disparities in health and health care in the study. White is the typical reference category, often being the largest single race/ethnicity in many studies. White is also typically associated with better health and health care access, thus, using white as the reference to detecting disparities with other races makes sense for this research question.
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.
This article measured disparities using relative measures, reporting on proportions of respondents by race/ethnicity and for each health or health care variable of interest, and odds ratios in the regression modeling. As the Harper/Lynch article notes, absolute and relative measurements can tell very different stories—I therefore think both should be reported, or at least available in supplemental data to evaluate if there are different trends in absolute vs. relative measures. The Harper/Lynch article also specifically recommends using Between-Group Variance for absolute disparity summaries when looking at non-ordinal variables, such as race/ethnicity, and general entropy class measures for relative disparities—both of which I would love to learn more about!