S1 Data -

Background
Treat-all strategies improved patient outcomes, despite higher rates of loss to follow-up compared to the pre-treat era. Patients in Ethiopia experienced a higher rate of LTFU during the treat-all strategy period; however, studies did not identify contributing factors in comparison with previous strategies. This study aimed to assess the incidence and predictors of loss to follow-up before and after the start of the treat-all strategy among adults on anti-retroviral therapy in public health facilities in Hawassa City, Ethiopia.
Methods
An institution-based retrospective follow-up study was conducted among 1190 randomly selected adults on antiretroviral therapy in public health facilities in Hawassa City. Using the Open Data Kit (ODK), data were collected from medical records and exported to Stata version 16 and R 4.2.1 for analysis. A Grays test and cumulative incidence curve were used to compare the cumulative incidence function of loss to follow-up. Bivariable and multivariable competing risk regression were fitted to identify predictors of LTFU and variables with a p-value <0.05 were considered significant.
Results
The cumulative incidence of lost-to-follow-up was 4.92(3.84,6.3) and 8.67(7.26,10.3) per 100 person-years (PY) in pre-treat all and treat all cohorts, respectively. The cumulative incidence of mortality was 5.86(4.67,7.35) and 3(2.26,4.12) per 100 PY in pre-treat and treat all cohorts, respectively. Fair/poor adherence (aSHR:5.17; (95% CI 1.97, 13.51), underweight (aSHR:2.13; 95% CI: 1.15–3.93) and WHO stage III/IV (aSHR:2.69; 95% CI: 1.27, 5.71) were predictors of loss up in pre—treat all, whereas fair/poor adherence (aSHR = 2.07; 95% CI: 1.18, 3.68), underweight (aSHR:1.71; 95% CI: 1.13, 2.56), and CD4 cell >350 cell/m3 (aSHR: 1.67; 95% CI: 1.05, 2.65) predicts of loss up in treat all cohorts.
Conclusion
This study demonstrated that the incidence of loss to follow-up was considerably higher in the treat-all period as compared to the pre-treat-all era. Poor medication compliance, underweight, and a CD4 level >350 cells/m3 contributed to the higher rate of LTFU in the treat-all strategy. Targeted interventions, such as nutritional support and strengthening medication adherence counseling, should be implemented to maintain treatment retention and reduce antiretroviral therapy dropout rates.

研究背景
全治疗（treat-all）策略虽较治疗前时代的失访（Loss to Follow-up, LTFU）率更高，但可改善患者转归。埃塞俄比亚的患者在全治疗策略实施期间的失访率更高，但现有研究未结合既往治疗策略分析其潜在影响因素。本研究旨在评估埃塞俄比亚哈瓦萨市公立卫生机构接受抗逆转录病毒疗法（anti-retroviral therapy, ART）的成人患者，在全治疗策略启动前后的失访发生率及其预测因素。

研究方法
本研究为基于机构的回顾性队列随访研究，纳入哈瓦萨市公立卫生机构中1190名经随机选取的接受抗逆转录病毒疗法的成人患者。研究采用开放数据套件（Open Data Kit, ODK）收集医疗记录数据，并将数据导出至Stata 16版及R 4.2.1软件进行分析。采用格雷检验（Grays test）及累积发病率曲线比较失访的累积发病率函数；拟合双变量及多变量竞争风险回归模型以识别失访的预测因素，以P值<0.05作为差异具有统计学意义的标准。

研究结果
全治疗前队列与全治疗队列的失访累积发病率分别为每100人年（person-years, PY）4.92（95%置信区间：3.84, 6.30）与8.67（95%置信区间：7.26, 10.30）。全治疗前队列与全治疗队列的死亡率累积发病率分别为每100人年5.86（95%置信区间：4.67, 7.35）与3.00（95%置信区间：2.26, 4.12）。在全治疗前队列中，依从性一般/较差（调整后的亚分布风险比aSHR:5.17; 95%置信区间：1.97, 13.51）、体重不足（aSHR:2.13; 95%置信区间：1.15–3.93）以及WHO分期III/IV期（aSHR:2.69; 95%置信区间：1.27, 5.71）为失访的独立预测因素；而在全治疗队列中，依从性一般/较差（aSHR=2.07; 95%置信区间：1.18, 3.68）、体重不足（aSHR:1.71; 95%置信区间：1.13, 2.56）以及CD4细胞计数>350个/立方米（aSHR:1.67; 95%置信区间：1.05, 2.65）为失访的独立预测因素。

研究结论
本研究证实，相较于全治疗前时代，全治疗策略实施期间的失访发生率显著升高。药物依从性差、体重不足以及CD4细胞计数>350个/立方米是全治疗策略下失访率升高的关键影响因素。临床应实施针对性干预措施，例如营养支持及强化药物依从性咨询，以维持患者的治疗留存率并降低抗逆转录病毒疗法的脱落率。