Correlates of completion rate and question comprehension for a multi-ethnicity online study of ovulation and menstruation health
Williams, Anna Sophia
MetadataShow full item record
BACKGROUND: Polycystic ovary syndrome (PCOS), a diagnosis of exclusion, is considered the most common endocrinopathy in females of reproductive age. Current cohorts of individuals with PCOS exist but are limited, specifically in racial/ethnic diversity, due to the complexity of multiple choice survey questions and difficulty to obtain data. METHODS: After examining existing cohorts and their questionnaires, an online survey was designed specifically to assess PCOS characteristics in a diverse population. The survey was field tested for cognitive understanding and usability in a diverse population, with respect to race/ethnicity and education. After cognitive testing, the survey was launched online and participants were recruited to complete the survey. Survey completion rates by level of education, race/ethnicity, and birthplace were established by determining the proportion of participants in each subgroup to complete the entire survey. We also looked at the average percentage of the survey participants in each subgroup completed before discontinuing the survey. Univariate regression analyses were performed to evaluate these results. Question comprehension was determined by evaluating all survey questions by length and complexity. The survey question asking participants to report their typical menstrual cycle length (MCL) was selected to assess participants’ question comprehension. Responses that fell outside of the normal range for MCL were examined and those that may have been inaccurate were isolated. Level of education and birthplace reported by participants who may have responded to the MCL question inaccurately were noted and assessed collectively. RESULTS: Two-hundred and forty-eight participants, aged 18-53 years, began the survey between August 9, 2017 and October 23, 2017. Of these participants, 71.7% identified as White; 11.7% as Hispanic, Latina, or of Spanish Origin; 12.6% as Black or African American; 13.2% as other races/ethnicities; and 9.2% as more than one race/ethnicity. Of the participants with some high school education, a high school diploma, or a GED, 74% completed the entire survey; of those with some college education or a 2-year degree, 81% completed the entire survey; and of those with a 4-year college degree or more, 90% completed the entire survey. On average, participants with a high school education completed 86% of the survey before discontinuing; those with some college education completed 91% of the survey before discontinuing; and those with a college degree or more completed 95% of the survey before discontinuing. Of the participants who identified as White, 92% completed the entire survey; of those who identified as Hispanic, Latina, or of Spanish Origin, 68% completed the entire survey; and of those who identified as Black or African American, 70% completed the entire survey, with statistical significance (p = 0.002). On average, participants who identified as White completed 96% of the survey before discontinuing; those who identified as Hispanic, Latina, or of Spanish Origin completed 82% of the survey before discontinuing; and those who identified as Black or African American completed 86% of the survey before discontinuing, with statistical significance (p = 0.002). Of the participants who were born in the United States, 92% completed the entire survey, and of the participants who were not born in the United States, 66% completed the entire survey, with statistical significance (p < 0.0001). On average, participants who were born in the United States completed 96% of the survey before discontinuing and participants who were not born in the United States completed 81% of the survey before discontinuing, with statistical significance (p < 0.0001). When asked to report their typical MCL, 66 participants (28.1%) responded with irregular answers and 20 (8.5%) may have responded inaccurately. Of these 20 participants, 18 (90%) completed at least some amount of college and 16 (80%) were born in the United States. DISCUSSION: This pilot study succeeds in establishing a cohort that is more racially/ethnically diverse than existing cohorts of individuals with PCOS. In order to gather more information on the ways in which PCOS manifests across different groups, future recruitment efforts ought to be more targeted towards individuals belonging to different racial/ethnic groups and socioeconomic statuses. Additionally, the baseline survey and future modules can be made more accessible to diverse groups through language translation. Furthermore, the addition of a menstrual cycle tracker component to the survey may increase the accuracy of information on participants’ menstrual cycle patterns.