The Impact of War on the HIV Care Continiuum in Tigray, Ethopia: A Multi-Methods Approach

Background: Armed conflicts represent devastating threats to HIV care continuity in subSaharan Africa, where the dual burden of conflict and HIV infection remains disproportionately high. The Tigray War (November 2020–November 2022) caused catastrophic devastation to northern Ethiopia’s health system, resulting in massive excess mortality, near-total infrastructure destruction, extended antiretroviral medication stockouts, severe workforce reductions, and prolonged humanitarian access blockades. Despite this crisis and the region's substantial HIV burden, significant knowledge gaps persist. Prior research has not systematically examined conflict impacts across the full HIV care cascade; from testing and diagnosis through treatment continuity, co-morbidity management, and preventive services. No study has employed robust quasi-experimental designs to establish counterfactual trajectories for conflict-affected HIV programmes. The effects on integrated services including tuberculosis care, preventive therapy, and maternal - child health remain unquantified. Patient perspectives on service quality during post-conflict recovery have not been assessed. These gaps provided the rationale for a comprehensive multi-methods investigation in this conflictaffected setting. Aims: The aims of this thesis are threefold. First, the project synthesised global evidence on armed conflicts and HIV treatment outcomes in sub-Saharan Africa. Second, it quantified the Tigray War’s impact on HIV care utilisation, treatment cascade outcomes, co-morbidity management, preventive therapy, and maternal-child health services. Third, it examined patient satisfaction with pharmaceutical, laboratory, and clinical HIV services as a proxy for perceived quality in post-conflict Mekelle. Methods: This thesis-by-publication design integrated nine manuscripts employing complementary multi-methods approaches. Paper 1 comprised a systematic review and metaanalysis examining conflict impacts on HIV treatment outcomes across sub-Saharan Africa from evidence published in MEDLINE, PubMed, CINAHL, and Scopus dated January 1, 2002, to December 30, 2022. The review utilised Agency for Healthcare Research and Quality criteria for quality assessment, Joanna Briggs Institute standardised instruments for data extraction, and Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines to report results. Papers 2–6 utilised secondary electronic medical record data of the entire HIV care and treatment cascade from seven health facilities in Mekelle City spanning pre-war, war, and post-war periods over two decades: Paper 2 examined HIV testing, linkage, and treatment initiation using interrupted time-series analysis with segmented regression; Paper 3 assessed retention, viral suppression, mortality, and reengagement using interrupted timeseries and Cox proportional hazards survival analyses; Paper 4 investigated tuberculosis care cascade outcomes using ARIMA forecasting, segmented regression, and survival analysis; Paper 5 quantified preventive therapy impacts using interrupted time-series analysis with generalised least squares estimation; and Paper 6 assessed mother-to-child transmission, infant outcomes, and cervical screening using Firth’s penalised logistic regression, generalised estimating equations, and mixed-effects modelling. Papers 7–9 employed primary crosssectional surveys assessing patient satisfaction with pharmaceutical, laboratory, and clinical services in the HIV clinic respectively, analysed using structural equation modelling. Data were analysed using R software. Results: The systematic review (Paper 1) identified 16 studies revealing loss to follow-up proportions of 5.4% to 43.5%, pooled virologic non-suppression of 30% (95% CI 0.26–0.33), adherence rates exceeding 88%, and mortality rates of 4.2% to 13%. Females had higher odds of loss to follow-up than males (OR 1.51, 95% CI 1.05–2.17). However, the evidence base showed methodological limitations and insufficient depth of coverage. Analyses of HIV service utilisation (Paper 2) demonstrated substantial war-related disruptions: HIV diagnoses declined by 29% (IRR 0.71, 95% CI 0.58 to 0.86; p<0.001), linkage to care by 32% (IRR 0.68, 95% CI 0.53 to 0.86; p=0.002), and ART initiation by 30% (IRR 0.70, 95% CI 0.54 to 0.91; p=0.008). Women experienced 43% reductions in diagnoses (IRR 0.57, 95% CI 0.46 to 0.72) compared with 8% among men. Youth aged 0 to 24 years faced 41% declines (IRR 0.59, 95% CI 0.43 to 0.81), and NGO-managed facilities suffered 79% reductions (IRR 0.21, 95% CI 0.13 to 0.36) without post-war recovery. Treatment continuity outcomes (Paper 3) revealed catastrophic collapse among 20,259 patients (60.3% female): loss to follow-up accelerated during conflict (β=+1.40 percentage points/quarter, 95% CI 0.81 to 1.99; p<0.0001), with hazard ratios from 5.94 to 13.35 across Cox, IPCW, and Fine-Gray models. Retention eroded (β=−1.19; p=0.052) and viral load testing coverage fell (β=−3.58; p<0.001) before partial post-conflict recovery. Documented mortality was likely under-ascertained (328 probable vs 66 recorded deaths). Re-engagement was only 4.30% (95% CI 3.80 to 4.80). Tuberculosis care (Paper 4) was substantially disrupted among 4,984 ART patients, with 433 TB episodes recorded across pre-war, COVID-19, war, and post-war periods. TB diagnosis rates rose from 7.62% pre-war to 11.65% post-war, while treatment initiation remained at or above 91% throughout. In Cox models, TB hazard within 365 days of ART initiation was elevated during war (HR 2.75, 95% CI 1.96 to 3.87) and post-war (HR 2.43, 95% CI 1.76 to 3.36) versus pre-war. Effect modification by facility level (p<0.001) and ownership (p=0.001) was significant; sex and age were not. Preventive therapy delivery (Paper 5) showed profound sustained collapse across 77,499 cotrimoxazole and 10,193 TPT initiations. War-onset IRRs were 0.013 (95% CI 0.010 to 0.016) for cotrimoxazole and 0.067 (95% CI 0.014 to 0.315) for TPT, reductions of 98.7% and 93.3% (both p<0.001). Counterfactuals imply 8,852 cotrimoxazole and 3,095 TPT initiations missed during the war alone, with similar post-war deficits and no recovery. Post-conflict maternal-child services (Paper 6) showed mixed outcomes among 405 HIVexposed infants: MTCT rate was 2.72% (95% CI 1.36 to 4.81%), declining from 5.56% in 2022 to 0% in early 2025 (p=0.0015), with all transmissions in primary-level facilities. Exclusive breastfeeding prevalence was 90.9% (95% CI 89.3 to 92.4%), and programme retention reached 99.8% with no losses to follow-up. Growth failure was rare among exclusively breastfed infants (0.48%) but markedly higher among replacement-fed infants (26.2%). Cervical cancer screening achieved a 98.6% offer rate but only 76.9% completion. Patient satisfaction surveys revealed moderate perceived quality across services. Pharmaceutical services (Paper 7) showed 62.9% overall satisfaction, with gaps in medication information and drug problem resolution; Provider Communication loaded most strongly (β=0.975), and appointment interval was the only demographic predictor (AOR=4.15, 95% CI 1.91 to 9.04). Laboratory services (Paper 8) showed 64.2% satisfaction with Timeliness loading most strongly (β=0.995). Clinical services (Paper 9) showed 62.4% satisfaction, with all four domains strongly loading (β=0.920 to 0.985, all p<0.001). Clients at tertiary-level (OR 0.50) and public facilities (OR 0.64, both p<0.001) had lower satisfaction odds, indicating persistent structural deficiencies. Conclusions: This thesis demonstrates that the Tigray War caused catastrophic, multifaceted, and sustained collapse of the HIV care continuum. HIV testing, care linkage, and treatment initiation declined substantially; loss to follow-up increased severalfold; viral suppression halved; mortality increased dramatically; and tuberculosis and preventive therapy services were largely dismantled. Recovery remains incomplete more than two years post-ceasefire, with minimal patient reengagement and satisfaction well below national benchmarks. Impacts were inequitably distributed, with women, youth, and NGO facility clients bearing disproportionate burdens. Given that HIV prevalence has approximately doubled in post-war Tigray, regaining epidemic control will require service delivery at substantially expanded capacity to address both ongoing needs and accumulated burden. Similarly, the collapse of tuberculosis screening and treatment likely resulted in substantial undiagnosed and untreated TB cases, while prolonged treatment interruptions and suboptimal viral suppression may have accelerated the emergence of antiretroviral drug resistance. The overall satisfaction of pharmaceutical, laboratory or clinical services are below Ethiopia's national benchmark of 80%. These findings highlight the need for conflict-resilient health systems, HIV care prioritisation within humanitarian responses, and post-conflict investment sufficient to address the accumulated disease burden.