Delivering Integrated Community Case Management During Conflict: Service Delivery, Medicine Availability and Routine Data Quality in Yemen

Abstract Background Integrated community case management (iCCM) depends on more than the presence of trained community health workers (CHWs); it also requires reliable commodities, functioning referral pathways and trustworthy routine information. This study examined service delivery, referral, medicine readiness and routine-data quality as interconnected components of a conflict-affected community child-health platform in Yemen. Methods We conducted a retrospective descriptive repeated-measures analysis of aggregate weekly service and medicine reports submitted by 33 programme personnel (30 CHWs and three community health supervisors) in four programme districts (Al Khawkhah and Hays, Al Hudaydah governorate; Khanfar, Abyan governorate; and Al Azarq, Al Dhala'e governorate) from November 2025 to July 2026. Data were cleaned before aggregation to CHW-month level. Desk-based data-quality checks assessed identifiers, duplicates, referral-case consistency and inventory balance; source-register or physical-stock verification was not undertaken. We summarised monthly trends, district referral rates and descriptive, hypothesis-generating comparisons of referral rates according to reported commodity zero-balance status. Results The cleaned datasets contained 999 weekly service reports, 1,008 weekly medicine reports and 256 observed CHW-months. The 33 programme personnel recorded 9,708 illness episodes: respiratory tract infection (RTI) 44.4%, fever 34.3% and diarrhoea 21.3%. There were 690 recorded referrals (7.1%); referral rates ranged from 2.9% in Hays to 20.9% in Al Azarq and increased in the final study months. At least one available weekly report recorded a zero remaining balance in 207/256 CHW-months (80.9%) for zinc, 105/256 (41.0%) for amoxicillin and 89/256 (34.8%) for oral rehydration salts (ORS). Zinc zero-balance occurrence remained high in every study month. Referral counts exceeded recorded case counts in at least one illness-age-sex stratum in 142/999 weekly service reports (14.2%). Conclusions Community-based iCCM recorded substantial service activity in conflict-affected Yemen, but frequent reported zero medicine balances, marked referral variation and material routine-data inconsistencies constrain interpretation of programme performance and may limit treatment readiness. The findings support integrated attention to commodity management, referral continuity, supportive supervision and routine-data quality, while emphasising that a reported zero balance is not equivalent to a verified or month-long stock-out.