Collecting health narratives at scale: A multi-study evaluation of speech-to-text in health surveys
- 1University of Zurich: Universitat Zurich
Abstract
Patient narratives can reveal aspects of health that standardized measures miss, yet collecting open-ended responses at population scale remains difficult. We evaluated whether speech-to-text (STT) could enable richer responses in online health surveys without the resource demands of interviews. Across four surveys involving people with post-COVID-19 condition, healthy adults, people who engage in sex work, and older adults, participants could answer open-ended questions by typing or using STT. We compared response length, lexical diversity, stop-word proportion, and total unique content words, alongside STT uptake and participant feedback. Quantitative comparisons were made using a mixed-effects model including modality (STT vs typed) as fixed effect and a random intercept per participant. STT uptake varied markedly across populations (0.7-80.0%). Spoken responses were estimated to contain more characters than typed responses in two studies, by 245.0 and 891.8 characters on average (both p<0.001), and contained more unique content words, despite lower lexical diversity and higher stop-word proportions. Participants generally valued STT for its convenience and spontaneity, but uptake was constrained by context, privacy, interface design and individual preferences. STT may enable population-scale collection of richer health narratives, provided implementation is tailored to the study population and research context.
Related
- Assessing Person-Specific Symptom Networks From Brief Narratives: Development and Initial Evaluation of the Narrative-to-Network (N2N) Approach
- The Online Health Safety Gap: Consensus Alignment Does Not Imply Safety in Peer-to-Peer Health Narratives.
- The Medicine Quality Monitoring Globe: A public health intelligence system for collecting and curating reports of substandard and falsified medical products from publicly available Internet news media
- Cooking location and child health outcomes in Rwanda: a pooled analysis of two demographic and health surveys
- Digital development and health in China: a serial pathway model through internet use and physical exercise in repeated cross-sectional surveys
BibTeX
@misc{baumer2026collecting,
title = {{Collecting health narratives at scale: A multi-study evaluation of speech-to-text in health surveys}},
author = {A. M. Baumer and P. Naoumis and J. Fabian and S. Strukova and M. Wolf and T. Ballouz and A. Farnham and V. X. C. Hofmann and G. Spitale and V. Dellwo and A. Glaessel and M. A. Puhan and V. von Wyl},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.23.26363819},
url = {https://www.medrxiv.org/content/10.64898/2026.09.23.26363819v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Collecting health narratives at scale: A multi-study evaluation of speech-to-text in health surveys
AU - A. M. Baumer
AU - P. Naoumis
AU - J. Fabian
AU - S. Strukova
AU - M. Wolf
AU - T. Ballouz
AU - A. Farnham
AU - V. X. C. Hofmann
AU - G. Spitale
AU - V. Dellwo
AU - A. Glaessel
AU - M. A. Puhan
AU - V. von Wyl
PY - 2026
DA - 2026/09/23
PB - medRxiv
DO - 10.64898/2026.09.23.26363819
UR - https://www.medrxiv.org/content/10.64898/2026.09.23.26363819v1
AB - Patient narratives can reveal aspects of health that standardized measures miss, yet collecting open-ended responses at population scale remains difficult. We evaluated whether speech-to-text (STT) could enable richer responses in online health surveys without the resource demands of interviews. Across four surveys involving people with post-COVID-19 condition, healthy adults, people who engage in sex work, and older adults, participants could answer open-ended questions by typing or using STT. We compared response length, lexical diversity, stop-word proportion, and total unique content words, alongside STT uptake and participant feedback. Quantitative comparisons were made using a mixed-effects model including modality (STT vs typed) as fixed effect and a random intercept per participant. STT uptake varied markedly across populations (0.7-80.0%). Spoken responses were estimated to contain more characters than typed responses in two studies, by 245.0 and 891.8 characters on average (both p<0.001), and contained more unique content words, despite lower lexical diversity and higher stop-word proportions. Participants generally valued STT for its convenience and spontaneity, but uptake was constrained by context, privacy, interface design and individual preferences. STT may enable population-scale collection of richer health narratives, provided implementation is tailored to the study population and research context.
ER -