Yes, no, maybe so: the importance of cognitive interviewing to enhance structured surveys on respectful maternity care in northern India.

Quantitative survey findings are important in measuring health-related phenomena, including on sensitive topics such as respectful maternity care (RMC). But how well do survey results truly capture respondent experiences and opinions? Quantitative tool development and piloting often involve translating questions from other settings and assessing the mechanics of implementation, which fails to deeply explore how respondents understand survey questions and response options. To address this gap, we conducted cognitive interviews on survey questions (n = 88) adapted from validated RMC instruments used in Ethiopia, Kenya and elsewhere in India. Cognitive interviews with rural women (n = 21) in Madhya Pradesh, India involved asking the respondent the survey question, recording her response, then interviewing her about what the question and response options meant to her. We analysed the interviews to revise the tool and identify question failures, which we grouped into six areas: issues with sequencing, length and sensitivity; problematic response options; inappropriate vocabulary; temporal and spatial confusion; accessing different cognitive domains; and failure to resonate with the respondent's worldview and reality. Although women tended to provide initial answers to the survey questions, cognitive interviews revealed widespread mismatch between respondent interpretation and question intent. Likert scale response options were generally incomprehensible and questions involving hypothetical scenarios could be interpreted in unexpected ways. Many key terms and concepts from the international RMC literature did not translate well and showed low resonance with respondents, including consent and being involved in decisions about one's care. This study highlights the threat to data quality and the validity of findings when translating quantitative surveys between languages and cultures and showcases the value of cognitive interviews in identifying question failures. While survey tool revision can address many of these issues, further critical discussion is needed on the use of standardized questions to assess the same domains across contexts.