Physician preferences for prescribing Chinese patent medicines during pregnancy: a discrete choice experiment

Prescribing Chinese patent medicines (CPMs) during pregnancy requires physicians to balance fetal safety, expected maternal benefit, patient preference, affordability, and limited pregnancy-specific evidence. Quantitative evidence on how these factors are traded off remains limited.

To quantify physicians' preferences for prescribing CPMs during pregnancy, estimate the DCE-derived relative importance of key prescribing attributes, and explore preference patterns across physician subgroups.

An unlabeled, forced-choice discrete choice experiment was conducted among physicians from a multi-tier medical consortium comprising 10 healthcare institutions in Anhui Province, China. Of 237 submitted questionnaires, 198 were retained for analysis, corresponding to a questionnaire validity rate of 83.54%. Seven attributes were included: fetal safety impact, treatment stage, patient willingness, prior prescribing experience with the specific CPMs, route of administration, patient affordability, and drug efficacy. A mixed logit model was used to estimate preference parameters and DCE-derived relative importance.

Compared with significant fetal impact, profiles indicating no fetal impact (β = 3.074, 95% CI 2.587 to 3.560, P < 0.001) or minor fetal impact (β = 1.954, 95% CI 1.632 to 2.276, P < 0.001) had higher utility. Patient willingness (β = 0.604, 95% CI 0.423 to 0.786, P < 0.001), topical administration (β = 0.361, 95% CI 0.143 to 0.580, P = 0.001), and affordability (β = 0.177, 95% CI 0.100 to 0.258, P < 0.001) were positively associated with profile selection. Early-pregnancy treatment (β = -0.200, 95% CI -0.397 to -0.004, P = 0.046) and uncertain efficacy (β = -0.316, 95% CI -0.526 to -0.107, P = 0.003) were negatively associated with selection. Fetal safety impact accounted for 71.07% of the total DCE-derived relative importance based on attribute-level utility ranges, followed by patient willingness and route of administration. Subgroup analyses were exploratory.

Fetal safety was the most influential attribute in physicians' choices between hypothetical CPMs profiles during pregnancy. Patient willingness, administration route, affordability, treatment stage, and efficacy also contributed to choices, whereas prior prescribing experience with the specific CPMs was not significantly associated with selection. The findings may inform evidence summaries, decision-support tools, and structured communication, but effects on actual prescribing and clinical outcomes require prospective evaluation.