JMIR Publications

ABSTRACT

Background:

Short-video platforms have become a primary channel through which the Chinese public encounters health information, and thousands of hospital-affiliated physicians now maintain creator accounts. Yet almost all existing evidence comes from small content audits or audience surveys; whether professional credentials, such as academic rank and institutional prestige, predict reach on these platforms has never been tested at national scale.

Objective:

We aimed to quantify how audience is distributed across Chinese physician accounts and to test whether institutional authority is associated with follower reach once city tier, clinical specialty, and platform portfolio are taken into account.

Methods:

We conducted a cross-sectional audit of 3,880 physician accounts affiliated with 959 hospitals in 177 Chinese cities, covering Douyin, Kuaishou and Xiaohongshu (snapshot: July 2026). Account-level attributes were linked to professional title, hospital tier, specialty, city and facility size. A content sub-study coded 36,645 videos from 3,615 accounts into 12 formats grouped into 4 paradigms. Concentration was measured with the Gini coefficient (bootstrap 95% CI), associations with Spearman rank correlation and Kruskal-Wallis tests, and multivariable associations with a log-linear model using cluster-robust standard errors at hospital level.

Results:

Median follower count was 145,000 (IQR 80,000-340,000). Audience was highly concentrated (Gini 0.621, 95% CI 0.604-0.637): the top 10% of accounts held 51.9% of all followers, the top 1% held 15.5%, and the bottom half held 11.9%. Professional title was only weakly associated with reach (Spearman rho=0.062, P<.001; Kruskal-Wallis epsilon-squared=0.035) and the gradient was non-monotonic: median followers were 157,000 for associate chief physicians, 124,500 for chief physicians and 68,000 for attending physicians. Hospital tier was not associated with reach (rho=-0.011, P=.47; Kruskal-Wallis P=.58). The full model explained 6.3% of the variance in log followers; each additional platform was associated with 25% more followers (P<.001) and practicing in a Tier-1 city with 22% more (P<.001). Content was dominated by a single format: 69.8% of videos were talking-head lectures, whereas clinical scenes and medical documentaries accounted for 11.6% each.

Conclusions:

On Chinese short-video platforms, institutional authority operates as an entry ticket rather than as an amplifier: it qualifies a physician to be followed but explains almost none of the variation in how many followers they attract. Audience allocation appears to be driven by platform-native performance and portfolio strategy. Hospitals that wish to build online presence should invest in account portfolios and format diversification rather than assume that seniority or institutional prestige will translate into reach.