JMIR Publications

ABSTRACT

Background:

Individuals with complex, chronic diseases are now living longer, making reproductive health an important topic to address in the health care setting. Self-respondent surveys are a feasible way to collect past contraceptive use and pregnancy history to assess contraceptive safety and effectiveness. Because sensitive topics, such as contraception and pregnancy outcomes can vary across social groups or cultures, piloting questions and evaluating survey administration procedures in the target population is necessary for precise and reliable responses prior to wide distribution.

Objective:

To develop a precise and reliable survey instrument and related procedures among individuals with cystic fibrosis (CF) regarding contraceptive use and obstetrical history.

Methods:

We piloted and tested electronic questions related to contraceptive use and pregnancy history among 50 participants with and without cystic fibrosis ages 18-45 years using a 3-tier process. Findings from each tier informed changes to the questionnaire before testing in the subsequent tier. Tier 1 used cognitive pre-testing to assess question understanding and need for memory prompts. In tier 2, we employed electronic test-retest self- and interviewer-administered approaches to assess question reliability, evaluate response missingness, and determine confidence between two types of survey administration approaches. In tier 3, we tested the questionnaire for clarity, time to complete, and whether additional prompts were necessary.

Results:

In tier 1, respondents suggested improvements to the electronic survey questions and to the written and visual prompts for better recall regarding past contraceptive use. In tier 2, the test-retest reliability between self- and interviewer-administrative procedures of “ever use” contraceptive method questions were similar, with percent absolute agreement ranging between 84%-100%. When the survey was self-administered, the percentage of missing responses was higher and respondent confidence about month and/or year when contraceptive methods were used was lower. Most respondents reported they preferred the self-administered survey because it was more convenient and faster to complete.

Conclusions:

Our three-tier process to pilot electronic survey questions related to contraceptive and obstetrical history in our complex disease population helped us tailor content and format to achieve more complete and reliable responses. Our final electronic survey included improved written and pictorial prompts. We also included a 10%-respondent test-retest self- and interviewer-administered quality control component to better inform imputation procedures of missing data. Clinical Trial: This qualitative study was observational and thus did not require clinical trial registration per ICMJE (International Committee of Medical Journal Editors) guidelines.