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).
Sorhaindo AM, Karver TS, Karver JG, Garcia SG “Constructing a Validated Scale to Measure Community-Level Abortion Stigma in Mexico” Contraception (2016) doi10.1016/ j.contraception.01.013 [Epub ahead of print]
2. What was the definition of the construct?
The construct was a 33-question survey assessing abortion stigma among Mexican nationals, with the thought that high abortion stigma leads to women seeking unsafe abortion and increasing their morbidity/mortality. There are 5 sub-scales to the survey: autonomy, discrimination, guilt/shame, religion and secrecy. For each of the 33 statements, the respondents were asked whether they agreed or disagreed using a 5-point Likert scale.
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 developed initially through an in-depth qualitative study in 5 states and Mexico city an found the most prevalent themes to be: social pressure to have children results in stigma in those seeking abortion, women who are known to have had an abortion are criticized by people in their communities, and deep Catholic values make women feel as if they have to pay for a sin by seeking and having an abortion. They then developed a preliminary list of 50 statements related to abortion stigma. Conducting cognitive interviews with 18 men and women of varying ages and levels of education tested face validity and they were asked, “In your own words, what do you think this phrase means?” They subsequently dropped 17 of the 50 preliminary statements to create the final 33 questions based on these cognitive interviews. Then confirmatory and exploratory factor analysis with reverse coding and sub-scale development was done. Those questions with factor loadings equal or greater than 0.3 were retained for each sub-scale. A Cronbach’s alpha was then determined for scale reliability (which was 0.92), and evaluation of the CA with certain items dropped from the questionnaire was tested and dropped if it changed the CA by 0.03 points or more. This ended in 32 or the 33 questions being retained in the final questionnaire.
I would consider additionally evaluating content validity in this construct by having experts assess the clinical credibility of the questionnaire in assessing our outcome to ensure that we are truly assessing what was intended to be measured. I would have also considered using formal linguistics tests to ensure readability of the survey during the cognitive interviewing and piloting phase. Additionally, in the future I would consider using this scale in other populations outside of Mexico to establish whether this can be used more widely.
4. Describe the implications of a lack of measurement validity or reliability for future research applications.
By accurately and correctly measuring levels of abortion stigma in women seeking or considering abortion, we will be able to improve our work on destigmatizing abortion with these patients which could potentially have significant implications on reducing unsafe abortion and maternal morbidity/mortality. Additionally, there is hope that accurately measuring stigma and correlating its effects on unsafe abortion may affect policy changes in countries with high abortion restrictions to encourage relaxation of those policies.
Part 2
1. Find a paper describing a health disparity (please give the full citation or upload the paper)
Alosaimi et al “Measure of Maternal Socioeconomic Status in Yemen and Association with Maternal and Child Health Outcomes.” Matern Child Health J (2016) 20:386-97
2. Summarize the construct and measurement of the dimension of disparity (e.g. race, SES) and the outcome measures (e.g. self-rated health)
The construct was SES indices that included data on ownership of vehicle (motorcycles, automobile, farm wagon etc), ownership of agricultural land, ownership of livestock, household access to basic service such as drinking water and electricity, characteristics of place of dwelling and outerwall material (concrete, wood, etc) and household head level of education. The outcome measures were indices of maternal and child heath status: maternal mortality, spontaneous abortion, stillbirth, neonatal and infant mortality.
3. What is the evidence for the validity and reliability of the measures?
A factor analysis using principal component analysis was done based on assigned scores and presence or absence of items from a list of household assets used in previous studies. They then used a factor loading cutoff of 0.4.
4. What is the reference category used for the disparity measure? Why does this measure make sense (or not) for this research question?
The reference categories were the lowest tertile of the wealth, educational and housing quality indices. It appears from their data that the majority of the sample was in the 2nd or 3rd tertile of the wealth index, with the least amount of people in the lowest tertile. I think it would have made more sense to have the 3rd tertile be the reference group and compare the lower tertiles to this group to examine outcome association. I believe what the authors are most interested in is if being in the lowest tertile of each of the SES indices increases the risk of adverse maternal and child health outcomes, so in that setting the highest tertiles should be the reference group and comparisons should be made to that group.
5. How is the disparity quantified? Is this an absolute or relative measure or are both provided?
Both absolute and relative measures of the disparities among SES index tertiles are provided in this study.
6. Describe which type of measure you would prefer for this research area, or , if both, why.
I think providing both in this type of study is preferred. Absolute measures are helpful to estimate the public health impact of these SES indices on maternal child health and potentially improve any public policies on these measures. Relative measures I think can make interpretation of the comparative risks easier to understand in a clinical sense.