Private healthcare facilities’ participation in Uganda’s viral hemorrhagic fever surveillance program: contributions, selective case capture, and underutilized early-warning potential

Background Private healthcare facilities are important initial points of contact, especially for patients with undifferentiated febrile illnesses in mixed health systems. However, their contribution to viral hemorrhagic fever surveillance remains to be quantified, particularly in outbreak-prone regions. We assessed their contribution to Uganda’s national surveillance system and examined variation by case mix, geography, and outbreak period. Methods We conducted a retrospective cross-sectional analysis of all suspected VHF specimen submissions recorded in Uganda’s national Result Dispatch System from 1 January 2023 to 14 August 2026, spanning the 2025 Sudan virus disease and 2026 Ebola (Bundibugyo virus, BDBV) outbreaks. Ownership was classified against the national health facility master list. Pre-specified outcomes were test positivity, recorded bleeding, and alive-at-reporting status. Adjusted analyses used generalized linear mixed models with a district-level random intercept, pre-specified sensitivity analyses, and a test of effect modification by Kampala. Results Of 7,929 submissions, 504 (6.4%) were private. Private healthcare facilities contributed 21 of 149 VHF-positive submissions and 10 of 32 Ebola-positive submissions. During May–June 2026, they contributed 212 of 1,230 submissions and 10 of 20 Ebola-positive submissions; Ebola positivity was 4.72% versus 0.98% in public submissions (post-hoc comparison; crude odds ratio 4.99, p = 0.0007). Private ownership was associated with overall VHF positivity (adjusted odds ratio [aOR] 3.44, 95% confidence interval [CI] 1.95–6.05), recorded bleeding (4.97, 3.86–6.41), and alive-at-reporting status among known-status records (5.43, 4.09–7.21). Positivity attenuated after reporting-status adjustment (2.40, 1.34–4.28); intervals included 1 in alive-only and pre-2026 analyses. Bleeding remained associated after adjustment for symptom recording (3.85, 2.98–4.97). The alive-at-reporting association was stronger in Kampala but remained elevated elsewhere. All 16 initially significant symptom comparisons remained significant after false discovery rate adjustment. Conclusions Private healthcare facilities contributed a small share of submissions but a disproportionate share of detections, recorded VHF-compatible bleeding far more often, and far more often reported patients still alive, unlike the post-mortem case-finding dominating public submissions in the capital. This is consistent with an underused early-warning role and supports targeted, differentiated integration; establishing any effect on timeliness requires prospective study. Trial registration Clinical trial number: not applicable.