Outcome of Antiretroviral Therapy and Its Predictors Among Clients Enrolled to Community Differentiated ART Service Delivery Model in Amhara Region, Ethiopia, 2024; Multi-Centered Retrospective Follow up Study

Abstract

Background

Community-based ART service delivery brings HIV care closer to clinically stable clients and may improve outcomes while reducing burdens on patients and health facilities. However, evidence from low income setting remains limited.

Methods

Institution based retrospective follow-up study was conducted from 422 adult ART users selected by simple random sampling. Data were analyzed using STATA 15. Multivariable Cox proportional hazards regression identified predictors of treatment outcomes.

Results

The cumulative incidence of good ART treatment outcome was 81.99% (95% CI: 77.98–85.54%). Median survival time was 11 months (IQR: 5–12). Missing ART doses reduced survival by 54% (AHR=0.46; 95% CI: 0.22–0.95). Good retention in care (AHR=1.58), tuberculosis prophylaxis (AHR=1.38), and adequate drug supply (AHR=1.39) significantly improved survival.

Conclusion

Treatment outcomes were suboptimal. Strengthening retention, tuberculosis prophylaxis, drug supply, and adherence support could improve survival.

Introduction

Antiretroviral therapy (ART) is a medical treatment that involves the use of a combination of antiretroviral drugs to suppress the replication of the human immunodeficiency virus (HIV) in the body. ¹ Good anti-retroviral therapy outcome refers to antiretroviral therapy and other related combined cares success in achieving its intended goal of suppressing the replication of the human immunodeficiency virus and improving the health and well-being of individuals living with HIV.^2,3 Effective antiretroviral therapy (ART) treatment outcomes are critically important for both individual patients and the broader community. ⁴ From a community perspective, good ART treatment outcomes contribute significantly to public health by reducing the transmission of the virus. ⁵ Viral load suppression is a key indicator of ART effectiveness and is widely used to assess the success of HIV treatment programs. ⁶ However, viral load suppression should be interpreted within the context of viral load coverage (VLC), as suppression estimates only represent individuals with documented viral load results. ⁷ A high viral load suppression rate may overestimate program performance if a substantial proportion of people receiving ART have not received viral load testing. ⁶ Therefore, adequate viral load coverage is essential to ensure that viral suppression reflects the true treatment outcomes of the entire ART population. ⁶

A community-based differentiated ART service delivery model is an approach to HIV care that tailors service provision to the specific needs and circumstances of individuals living with HIV or those at increased risk of HIV infection. ⁸ It’s a client-centered approach that aims to: Improve client outcomes, ensure the health system runs efficiently, and refocus resources to those who need them most. ⁸ Since 2015, the World Health Organization has endorsed differentiated service delivery (DSD) as a strategic approach for HIV treatment delivery.^8,9 Some principles of DSD include: Equitable access, Efficient use of resources, Safe, high quality of care for patients, Robust and standardized set of outcome measures and evaluation processes. ¹⁰ One example of a differentiated service delivery (DSD) model is the adherence club approach, in which small groups of patients convene either at health facilities or designated community sites, where nurses or community health workers provide clinical monitoring, psychosocial support, and antiretroviral medication refills. ¹¹

According to a study conducted in Sub-Saharan Africa, there were a dearth of high-quality and quantitative evidence regarding the treatment outcomes of community differentiated ART service delivery models for HIV treatment in the region, but it does indicate that viral suppression and retention in care are roughly showed better outcomes. ¹² A study conducted in Tanzania found that A community differentiated art service delivery model is a promising entry point to improve ART treatment outcomes and strategies are needed to harmonize the comprehensiveness and quality of services across facilities providing HIV services. ¹³ Similarly, studies conducted in Rwanda and Uganda showed that retention rates of individuals living with HIV and on ART high after the application of a community differentiated art service delivery model. ¹⁴

There are different factors that alter the outcome of ART using a community differentiated art service delivery model, which are socio-demographic characteristics such as age, gender, level of education, residence and as well as clinical predictors like viral load, opportunistic infections, treatment regimen, ARV drug adherence, mental health, duration on ART, retention and drug reaction.^13,15 Other factors that may influence treatment outcome include health system predictors like, type of health facilities, ARV drug supply, access to viral load laboratory, distance to health facilities and healthcare services. ¹⁶

In order to reduce facility visits and maintain treatment continuity, the World Health Organization (WHO 2016) advises that people living with HIV who are stable on antiretroviral therapy (ART) be provided drug refills for a maximum of three to six months through differentiated ART service delivery models.^17,18 As a result, community differentiated ART service delivery model (CAGs) help patients to get access to medications by cutting down the time, and money needed for frequent clinic visits. Plus, by promoting peer support at the community level, CAG also help patients form a social network and lessen the perception of stigma and all these in turn helps to improve ART treatment outcomes of clients enrolled to the program.^19,20

Despite all efforts, progress against HIV remains fragile in many countries and profoundly inadequate among priority populations globally. ²¹ In spite of the availability of excellent medication, AIDS-related causes of death claimed about 700,000 lives in 2019. Refocusing the HIV response on how to provide life-saving assistance to everyone in need across all nations and communities is imperative. ²² HIV testing and starting antiretroviral medication (ART) are important measures for persons living with HIV (PLHIV), but it is difficult to keep PLHIV on treatment over the long term. ¹⁴

Ethiopia has adopted the community differentiated ART service delivery platform on late October 2018 and in 2020 it scaled up to Amhara Region. The model is new in its kind and there is a scarcity of study in this area in the country. Therefore, this study aimed to assess antiretroviral therapy treatment outcomes and identify their predictors among clients enrolled in a community differentiated ART service delivery model in the Amhara Region.

Methods

Study Settings

This retrospective follow-up study was conducted in 17 randomly selected health facilities providing community differentiated ART services out of a total of 55 such facilities in the Amhara Region. Geographically, the Amhara National Regional State (ANRS) is situated between 9°21′ to 14°00′ North latitude and 36°20′ to 40°20′ East longitude. The region covers an estimated area of 154,708.96 km² and is administratively divided into 22 zones. The regional state capital is Bahir Dar which is the seat of the Regional Government of Amhara. Based on the central statistical agency (CSA) 2023 population projection, Amhara Region population is about 23,216,000 of which 11,634,000 are males and 11,582,000 are females. ²³ Currently Amhara region has 55 health facilities which provide Community Differentiated ART Service delivery. At the time of the study, approximately 6,487 clients were receiving antiretroviral therapy either through health extension worker-led community ART groups (HEP-CAG) or peer-led community ART groups (PCAG) across 55 health facilities in the Amhara Region.

Study Design and Period

An institution-based retrospective follow-up study design was conducted from May 6 to May 31, 2024.

Population

For this retrospective follow-up study, all clients living with HIV who were enrolled in community differentiated ART service delivery models in the Amhara Region from October 1, 2020, to May 5, 2024, were considered eligible. Clients with missed, incomplete, or insufficient medical records were excluded from the study.

Eligibility Criteria

Inclusion Criteria

All stable clients living with HIV and enrolled to community differentiated ART service delivery model from October 1/2020 up to May 05/2024 in Amhara Region.

Exclusion Criteria

ART clients who have been enrolled to community differentiated ART service delivery model with missed and incomplete files and documents.

Sample Size, Technique and Procedure

The sample size for this retrospective follow-up study was determined using the double population proportion formula based on significant predictors from previous studies conducted in Southern Ethiopia and South Africa. Opportunistic infection was selected as it provided the maximum sample size, with 51.5% outcome among exposed, 36.8% among non-exposed groups, and an AHR of 1.4. Using Epi Info version 7.2.6, a 1:1 allocation ratio, 95% confidence level, 80% power, and 10% contingency for incomplete records, the final sample size was 422 ART patient records^24,25(Table 1).

Table 1.

Sample Size Calculation to Assess ART Treatment Outcome and Its Predictors Among Clients Enrolled to Community Differentiated ART Service Delivery Model in Amhara Region, North West Ethiopia

Significant predictors	Outcome in unexposed group	Confidence level	Study power	AHR	Ratio	Sample size	Sample size with 10 %	Reference
Opportunistic Infection	36.8 %	95 %	80 %	1.4	1	384	422	​
Types of health facilities	66 %	95 %	80 %	1.3	1	164	180	​

For this retrospective follow-up study, eligible ART client records were proportionally allocated among 17 randomly selected health facilities in the Amhara region based on the number of clients enrolled in community differentiated ART service delivery models from October 1, 2020, to May 31, 2024. A list of eligible clients enrolled in the community ART delivery model was prepared from existing registration records, and client records were assigned serial numbers. Using a computer-generated simple random sampling technique, follow-up records were selected through unique ART numbers until the required sample size of 422 client records was achieved (Table 2).

Table 2.

Schematic Presentation of Sampling Procedure for the Study to Assess ART Treatment Outcome and Its Predictors Among Clients Enrolled to Community DSDM, in Amhara Region Public Health Facilities, Ethiopia, 2024

S/N	Randomly selected health facilities In amhara region	Total clients in community DSDM	Proportional allocation	Computer generated simple random sampling method	Total sample size
1	Dessie comprehensive specialized hospital	482	51	​	422
2	Dessie Health Center	239	25
3	Gondar University comprehensive specialized hospital	437	47
4	Gondar health center	116	13
5	Kombolcha 03 Health Center	327	35
6	Debre Birhan comprehensive specialized hospital	242	26
7	Shewa Robit Health Center	161	17
8	Debre Birhan 04 Health Center	126	13
9	Finote Selam general hospital	200	21
10	Felege Hiwot comprehensive specialized hospital	283	30
11	Bahir Dar Han Health Center	124	13
12	Debre Tabor comprehensive specialized hospital	221	24
13	Abay Health Center	108	12
14	Woldia comprehensive specialized hospital	332	35
15	Debre Markos comprehensive specialized hospital	267	29
16	Bichena Health Center	172	18
17	Debre Markos Health Center	120	13
Total		3957	422

Operational Definitions

ART Treatment Outcome: it involves assessing the impact of ART on various aspects of clients’ health, such as viral suppression rates, adherence to treatment, retention in care and opportunistic infection.^14,26

Good ART Treatment outcome: All clients who are retained in care, no opportunistic infection, adhered to treatment and virally suppressed during the follow up time was considered as having good ART treatment outcome. Unless they assigned as having poor ART Treatment Outcome. ⁷

Viral suppression: If the client has a viral load that is undetectable, equal to or less than 50 copies/ml during the time covered by the study. ²⁷

Opportunistic Infection: Patient who have additional illness other than HIV/AIDS during the study period. ²⁸

Good Retention in care: In this study, individuals who remained active in care from the time of enrollment into the community differentiated ART service delivery model through the completion of a one-year follow-up period were classified as having good retention in care. ²⁹

Good ART adherence: Taking at least 95% of prescribed ARV drugs per specified period. ⁷

Distance from Health Facility: According to the World Health Organization, areas located within a two-hour walking distance or within 10 km of a primary health care unit are considered accessible. ³⁰

Event: - Good ART Treatment Outcome During the Follow up Period

Censored: - All ART treatment outcomes other than events (shift to other facility-based DSD models, transfer out to other health facilities, interruption in treatment, ART treatment outcomes after the follow up period and Death).

Follow up time: -From the time of enrollment to community differentiated ART service delivery model until either an event or censorship occurs with a maximum of one year follow up time.

Data Collection Procedures

A structured checklist was developed in English that covers the socio-demographic characteristics, clinical factors and health system related factors based on a review of relevant literatures.^24,31,32 Unique ART numbers/patient record numbers were used as identifiers to link patients with their respective medical records, while patients’ addresses were obtained from ART health facilities. The paper-based data collection tool was digitized using Kobo Collect, which improved data quality at the point of entry and reduced the time required for data processing and entry. Then, the checklist was administered by the trained data collectors to obtain information by reviewing clients’ documentations like ART cards, viral load results reports, ART registers, community differentiated service delivery Register and other relevant documents from sampled ART health facilities.

Data Quality Assurance and Control

To ensure data quality, eight degree-holding health officers were recruited as data collectors, and a one-day training session was provided to enhance their understanding of the data collection tool prior to data collection. The checklist was pretested on 5% of the total sample size, and necessary revisions were made accordingly to improve the internal validity of the study. The data collector and the investigator were met virtually to discuss issues and resolve questionnaire problems. Moreover, during the data collection, there has been close mentoring, coaching and remote supervision virtually amid in the ongoing conflict. Finally, the collected data was reviewed cleaned, and edited before data analysis.

Data Processing and Analysis

Data were entered into KoboCollect and exported to Stata software version 15 for analysis. The outcome variable for each participant was dichotomized into good or poor ART outcomes, and incidence rates were computed for the entire follow-up period. Multicollinearity among independent variables was assessed using Spearman’s rank correlation test for categorical predictors. Descriptive statistics, including frequency, proportion, interquartile range, cumulative incidence, and person-time incidence rates, were performed. The proportional hazards assumption was evaluated both graphically using log-minus-log survival plots and statistically using a global test. Model adequacy was assessed using Cox-Snell residuals and the log-likelihood ratio test.

Candidate variables for multivariable Cox proportional hazards regression were selected based on bivariable Cox regression analysis. Variables with a p-value ≤ 0.25 in bivariable analysis and those satisfying the proportional hazards assumption were included in the multivariable model. The final Cox proportional hazards regression analysis was conducted at a 95% confidence level, and statistical significance was declared at p < 0.05.

Results

Sociodemographic Characteristics

A total of 422 clients enrolled in community ART groups participated in the study, yielding a 100% response rate. Of these, 75.36% (n = 318) were female. The mean age of participants was 43 years (SD ± 10 years). The majority of patients, 84.60% (n = 357), were urban residents. Regarding to educational level, 39.34% (166) patients did not get formal education. Among all participates, 34.83% (147) were unemployed in their occupational status (Table 3).

Table 3.

Socio-Demographic Characteristics of Treatment Outcome and Its Predictors Among Clients Enrolled to Community Differentiated ART Service Delivery Model in Amhara Region, North West Ethiopia, 2024. (N=422)

Variables	Categories	Frequency	Percent
Sex	Female	318	75.36
	Male	104	24.64
Residence	Rural	65	15.40
	Urban	357	84.60
Educational status	No formal education	166	39.34
	Primary	124	29.38
	Secondary	76	18.01
	College and above	56	13.27
Occupational status	Unemployed	147	34.83
	Employed	70	16.59
	Merchant	75	17.77
	Farmer	130	30.81
Marital status	Single	55	13.03
	Married	213	50.47
	Divorced	109	25.83
	Widowed	45	10.66

Clinical Related Characteristics

From all (422) participated clients, 81.28% (343) clients, their viral load status were suppressed at the beginning of the study. Most of the participants, 83.18% (351) patients were taken one dose medicine per day. Majority of participants, 85.55% (361) did not miss their prescribed drugs with 12 months duration. Above half of the participants, 76.54% (323) drug adherence condition was good. Above half of ART users, 84.36% (356) retention in care status was good. Majority of respondents, 71.09 % (300) received tuberculosis prophylaxis at the beginning of this study. Nearly all participants 98.58 % (416) were on 1J ARV drug regimen. The median duration time patients stay on ART was 10 years with interquartile range of 4 up to 18 years(Table 4).

Table 4.

Base Line Clinical Related Characteristics of Treatment Outcome and Its Predictors Among Clients Enrolled to Community Differentiated ART Service Delivery Model in Amhara Region, North West Ethiopia, 2024. (N=422)

Variables	Categories	Frequency	Percent
Viral load status	Low level viraemia	79	18.72
	Virally suppressed	343	81.28
Opportunistic infection	No	349	82.70
	Yes	73	17.30
Frequency of ART drug per day	Daily	351	83.18
	Two time	71	16.82
Missed ART drugs	No	361	85.55
	Yes	61	14.45
ART drugs adherence	Poor	46	10.90
	Fair	53	12.56
	Good	323	76.54
Drug regiment	1e	2	0.47
	1i	4	0.95
	1j	416	98.58
Drug reactions	No	369	87.44
	Yes	53	12.56
Types of CAG enrollment	PCAD	269	63.74
	HEP-CAG	153	36.26
Retention in care status	Poor	66	15.64
	Good	356	84.36
Body mass index	Under weight	64	15.17
	Normal	339	80.33
	Over weight	19	4.50
TB prophylaxis treatment	No	122	28.91
	Yes	300	71.09

Note. 1e = TDF + 3TC + EFV (Tenofovir + Lamivudine + Efavirenz); 1i = TDF + 3TC + DTG (Tenofovir + Lamivudine + Dolutegravir); 1j = AZT + 3TC + DTG (Zidovudine + Lamivudine + Dolutegravir).

Health Facilities Characteristics

In this study, 44.08% (186) ART users follow up, care and treatment center are at specialized hospitals. Majority of patients, 83.41% (352) had good access of drug supply. Majority of patients, 77.73% (328) had good access of laboratory services. The average distance of health facilities for patient was 6.81km with the range of 1-16 km (Table 5).

Table 5.

Health Facilities Characteristics of Treatment Outcome and Its Predictors Among Clients Enrolled to Community Differentiated ART Service Delivery Model in Amhara Region, North West Ethiopia, 2024. (N=422)

Variables	Categories	Frequency	Percent
Type of health facilities	Health centers	103	24.41
	Primary hospitals	85	2.14
	General hospitals	48	11.37
	Specialized hospitals	186	44.08
Drug supply status	Poor	70	16.59
	Good	352	83.41
Laboratory access	No	94	22.27
	Yes	328	77.73

ART Treatment Outcomes

The cumulative incidence and incidence rate of ART treatment outcomes were 10 per 100 person-days of observation (95% CI: 9–11 per 100 person-days) and 81.99% (95% CI: 77.98%–85.54%), respectively. The total follow-up time contributed by all participants was 3,415 person-months, with a median follow-up duration of 11 months (range: 10–11 months) (Figure 1).

Figure 1.

Outcome status of patients on ART for treatment outcome and its predictors among clients enrolled to community differentiated ART service delivery model in Amhara Region, North West Ethiopia, 2024. (N=422)

Survival Time of ART Treatment Outcomes

Clients on ART had an overall median survival period of eleven months, with survival interquartile ranging from five to twelve months. 346 patients, or 81.99 percent, experienced the incident during the research period. 18.01% (76) of the remaining clients were not shown the event; these were censored because of death (0.95% (4)), lost to follow-up (5.92%) (25), transferred to another medical facility (4.74% (20), and the event was not developed during the follow-up period (until a 12-month follow-up), 6.39% (27). 50% of patients took eleven months to exhibit the event, according to the overall Kaplan Meier survival curve (good outcome) (Figure 2).

Figure 2.

Over all Kaplan Meier survival probability curve for treatment outcome and its predictors among clients enrolled to community differentiated ART service delivery model in Amhara Region, North West Ethiopia, 2024. (N=422)

Proportional Hazards Assumption

The validity of a fitted Cox proportional hazards model and the interpretation of its results depend on the proportional hazards assumption. The proportional hazards ratio is assumed to remain constant throughout time by the CPHM. The most popular techniques for evaluating this assumption are the statistical method of the global goodness of fit test and the graphical way of visualizing the log minus log curve. Plotting log minus log survivor curves against survival time for categorical predictors and determining whether the curves are fairly parallel or cross one another was the graphical method employed in this work to verify proportional hazards assumptions. For instance, log rank test = <0.001 for drug supply (Figure 3).

Figure 3.

Log-log graphical test for treatment outcome and its predictors among clients enrolled to community differentiated ART service delivery model in Amhara Region, North West Ethiopia, 2024. (N=422)

Each predictor’s graph was created, and those that satisfied the criteria—if they also met the statistical method—were taken into account for the final model. These included TB prevention, opportunistic infection, missed ART medication, ART medication adherence, CAG enrolment, retention in care status, drug supply status, laboratory access, and the distance to the medical facility. Graphical assessment of the proportional hazards assumption in Cox regression is often considered subjective and may vary in interpretation; therefore, it should be complemented by formal statistical tests. Among the available methods, Schoenfeld’s global goodness-of-fit test is widely recommended as a robust approach. In this study, the proportional hazards assumption was assessed using the global test, and the result indicated that the assumption was not violated (p = 0.24), confirming that the proportional hazards assumption was satisfied.

Model Adequacy Test

The multivariable Cox proportional hazards regression model was fitted, and model adequacy was assessed graphically using Cox-Snell residuals. The Cox-Snell residuals were compared with the Nelson–Aalen cumulative hazard function. A 45-degree reference line on the Nelson–Aalen plot indicates an adequate model fit. In this study, the observed Cox-Snell residual curve closely followed the Nelson–Aalen cumulative hazard function, suggesting that the model provided a reasonable fit to the data, despite minor deviations from the reference line (Figure 4).

Figure 4.

Cox Snell residual goodness of model test of Cox proportional hazards regression analysis for treatment outcome and its predictors among clients enrolled to community differentiated ART service delivery model in Amhara Region, North West Ethiopia, 2024

Graphical (Cox Snell residual) of model check method is subjective; therefore, there is need of support it through objective model fitting method/technique that is the likelihood ratio test (dequacy test). The results of the likelihood ratio test were χ² = 1836.92 (p < 0.001), indicating that the Cox proportional hazards model provided a statistically significant improvement in model fit and was appropriate for the data.

Predictors of ART Treatment Outcomes

The bi-variable CPHRM contained twenty-two possible factors that were believed to be predictors of the ART treatment result (the patient’s survival status while on ART). Sex, age, residence, body mass index, TB prophylaxis, health facility type, drug supply status, laboratory access, marital status, opportunist infection, viral load, frequency of ART drug intake, missed ART drug, adherence to ART, duration on ART, ART drug regimen, drug reaction, CAG enrolment, retention in care status, body mass index, and distance from the health facility were among these. At the 0.25 level of significance, ten predictors on the bi-variable CPHRM were significant. These included TB prevention, CAG enrolment, retention care status, opportunistic infection, missed ART medication, ART medication adherence, viral load, medication supply status, laboratory access, and distance from the medical facility.

Both statistical and graphical assessments confirmed that the proportional hazards assumption was satisfied for all ten predictors. Accordingly, a multivariable Cox proportional hazards regression model was fitted including these variables. In the final model, four predictors were found to be statistically significant at p < 0.05: missed ART medication, retention in care, TB prophylaxis, and drug supply status.

Clients who missed ART medication had a 54% lower hazard of favorable survival compared with those who did not miss their medication (AHR = 0.46; 95% CI: 0.22–0.95). Participants with good retention in care had a 58% higher hazard of survival compared to those with poor retention (AHR = 1.58; 95% CI: 1.11–2.26). Similarly, those who received TB prophylaxis had a higher survival hazard (AHR = 1.38; 95% CI: 1.02–1.85), indicating approximately a 1.4-fold increase compared with those who did not receive prophylaxis. In addition, participants with good drug supply access had a 39% higher hazard of survival compared to those with poor access (AHR = 1.39; 95% CI: 1.07–1.81) (Table 6).

Table 6.

Bi-Variable and Multi-Variable Cox Proportional Hazards Regression Analysis for Treatment Outcome and Its Predictors Among Clients Enrolled to Community Differentiated ART Service Delivery Model in Amhara Region, North West Ethiopia, 2024. (N=422)

Variables	Categories	ART Rx outcomes		CHR (95%CI)	AHR (95% CI)	P-value
		Event	Censored
Viral load	Low level viraemia	33	46	1	1	​
	suppressed	313	30	1.97(1.40-2.76)	0.95(0.66-1.37)	0.789
Opportunist infections	No	309	40	1	1	​
	Yes	37	36	0.56(0.37-0.68)	0.91(0.69-1.22)	0.541
Missed ART drug	No	333	28	1	1	​
	Yes	13	48	0.23(0.13-0.39)	0.46(0.22-0.95)	0.035*
Drug adherence	Poor	23	23	1	1	​
	Fair	32	21	1.17(0.73-1.86)	1.00(0.66-1.52)	0.989
	Good	291	32	1.70(1.17-2.47)	1.06(0.75-1.50)	0.730
Type of CAG enrollment	PCAD	212	57	1	1	​
	HEP-CAG	134	19	1.18(0.99-1.40)	1.16(0.97-1.39)	0.105
Retention in care status	Poor	22	44	1	1	​
	Good	324	32	3.14(2.14-4.59)	1.58(1.11-2.26)	0.011*
TB prophylaxis	No	52	70	1	1	​
	Yes	294	6	2.18(1.67-2.85)	1.38(1.02-1.85)	0.034*
Drug supply status	Poor	37	33	1	1	​
	Good	309	43	2.01(1.51-2.68)	1.39(1.07-1.81)	0.012*
Laboratory access	No	40	54	1	1	​
	Yes	306	22	2.01(1.46-2.77)	1.05(0.73-1.51)	0.787
Distance of health facility	​			1.00(0.98-1.03)	0.99(0.97-1.02)	0.620

Discussion

The cumulative incidence and incidence rate of ART treatment outcomes observed in this study were 81.99% (95% CI 77.98% - 85.54%) and 10 per 100 person-days of observation (95% CI 9/100 – 11/100), respectively. These findings are comparable to, yet slightly higher than, the outcomes reported in similar studies. For instance, a study conducted in Kenya reported a cumulative incidence of 79% with an incidence rate of 8 per 100 person-days over a 12-month follow-up period. ³³ The Kenyan study employed a similar sampling method, enrolling patients from various socio-economic backgrounds and using inclusion criteria that selected patients who were enrolled to community differentiated ART service delivery models.

In another study from Malawi, the cumulative incidence was reported to be 83%, with an incidence rate of 9 per 100 person-days over a follow-up period of 10 months. ³⁴ This study also included a broad socio-economic demographic and similar study design to the current research. The slight difference in incidence rates might be due to variations in healthcare infrastructure, patient support systems and health care workers. Malawi’s study had a slightly shorter follow-up period, which aligns closely with the median follow-up time in the current study, potentially contributing to the similarity in outcomes.

Conversely, a study conducted in Uganda reported a lower cumulative incidence of 75% and an incidence rate of 7 per 100 person-days over a 14-month follow-up period. ³⁵ This lower incidence could be due to differences in socio-economic status and healthcare accessibility among the study population. The Ugandan study employed a more stringent inclusion criterion, focusing on patients with advanced HIV infection, which might have contributed to lower overall treatment success rates. Furthermore, the longer follow-up period allowed for more variability in patient adherence and retention, likely impacting the cumulative incidence and incidence rate.

In addition, a study conducted in South Africa reported a cumulative incidence of 82% and an incidence rate of 10 per 100 person-days over a 12-month follow-up period. ¹² The inclusion criteria of the South African study, which enrolled patients who had been on ART for at least six months, together with its emphasis on urban populations with relatively better access to health services, may partly account for the similarity in findings with the present study. In addition, the socio-economic profile of participants comprising predominantly low to middle-income individuals was comparable to that of the current study population, which may further explain the observed consistency in results. The overall median survival time of patient on ART was eleven months with the survival interquartile ranges of five up to twelve months. These findings are consistent with several studies in similar settings. For instance, a study conducted in Malawi reported a median survival time of twelve months, with 80% of patients has showed viral suppression within the first follow up year. ¹² This study also noted similar reasons for patient censoring, including a 4.7% transfer rate to other facilities and a 6% rate of loss to follow-up, closely aligning with our observations.

Moreover, a research conducted in Kenya has found a median survival time of ten months, with an 82% event rate. ³³ The reasons for censoring were comparable, including a 5% transfer rate and a 5.5% loss to follow-up rate. These similarities reinforce the robustness of our findings and the effectiveness of ART programs across different settings.

A study in Uganda further supports our results, reporting a median survival time of eleven months with an 81% event rate. ³⁵ This study also identified similar patterns in patient attrition, with 4.8% of patients transferring to other facilities and 6.2% lost to follow-up, highlighting common challenges in ART retention and continuity of care across different regions.

Moreover, research in South Africa found a median survival time of eleven months with a slightly higher event rate of 83%. ³⁶ The censoring proportions in the South African study due to death (1%), loss to follow-up (5.9%), and transfer to other facilities (4.5%) were comparable to those observed in the present study. This similarity enhances the comparability of findings across settings and underscores the importance of strengthening follow-up systems, as well as continuous monitoring and support, to reduce attrition and improve retention in ART programs. The study found that the survival rate of ART patients who missed their medication was lower as compared to those who consistently adhered to their ART regimen. This finding is consistent with several other studies on adult patients enrolled in community ART delivery models. ³⁷ The similarity might be due to similar study design and used similar inclusion criteria. Most patients missed their drug to get relief from drug side effect. Because of drug side effects can have a significant impact on the physiological changes. The drug reactions can range from mild side effects to severe and life-threatening complications, and they may vary depending on the specific drug or drug combination used. These reactions can lead to decreased adherence, treatment interruptions, and even discontinuation of therapy, ultimately that compromising the effectiveness of ART care and treatment services.^38,39 In study in Kenya observed a significant reduction in survival rates among non-adherent patients, reporting that missing ART doses increased the risk of mortality by approximately 50%. ³³ Similarly, research in Malawi found that non-adherence due to drug side effects was a primary factor contributing to higher mortality rates, with non-adherent patients experiencing a 55% reduction in survival compared to those who maintained consistent medication intake. ³³ Moreover, a study in South Africa reported that patients who missed ART doses had a 53% lower survival rate, aligning closely with the findings of this study. ¹² These studies employed similar designs, following adult patients in community ART programs over comparable periods and using consistent inclusion criteria.

The study found that the survival rate of ART patients with good retention in care was higher as compared to those with poor retention. This finding aligns with several other studies that emphasize the critical role of retention in care for improving ART treatment outcomes. A study conducted in Uganda reported that patients with consistent retention in ART programs had significantly higher survival rates, similar to the current study’s findings ( ⁴⁰ ). The similarity can be attributed to the comparable study designs and inclusion criteria, which focused on following adult patients over 12 months to assess the impact of retention on survival or good ART treatment outcomes. Both studies underscore the importance of continuous care and regular follow-up, which help ensure adherence to ART and timely management of any complications that arise. Contrastingly, some studies have reported different impact of retention on survival. A study in South Africa found that patients with good retention were more likely to survive than those with poor retention. ⁴¹ This higher disparity might be due to differences in healthcare infrastructure, the implementation of retention strategies, level of development and wealth index. South Africa’s ART programs often include more robust support systems, such as community health workers and adherence clubs, which may enhance patient retention more effectively. Furthermore, socio-economic factors and access to healthcare services vary between regions, potentially influencing the degree of retention’s impact on survival. These contrasting findings highlight the need to tailor retention strategies to specific contexts to optimize ART outcomes. ⁴²

The survival of ART patients who have completed tuberculosis prophylaxis was higher as compared to patients who did not take tuberculosis prophylaxis. This finding aligns with a study conducted in Uganda, which reported similar outcomes. ⁴³ The similarity between these studies may be attributed to their comparable sample sizes and inclusion criteria and consistent baseline characteristics among participants. Furthermore, a study) in South Africa reported a similar increase in survival rates, with ART patients who received TB prophylaxis showed higher survival rate compared to those who did not. ⁴⁴ These studies collectively underscore the beneficial role of TB prophylaxis in improving the survival of ART patients. The association between TB preventive therapy (TPT) and ART treatment outcomes has been extensively explored in previous researches. Similarly, studies in Ethiopia and Brazil have shown significant reductions in TB incidence and improved survival rates among ART patients receiving TB prophylaxis. ⁴⁵ These consistent findings across different geographical locations highlight the critical role of TB prophylaxis in enhancing the effectiveness of ART. Contrary to the finding that ART patients who completed tuberculosis prophylaxis, some studies report no significant difference in survival outcomes. For instance, a study conducted in India found that TB prophylaxis did not significantly improve survival rates among HIV patients on ART, possibly due to high rates of drug resistance and suboptimal adherence to TB prophylaxis. ⁴⁶ Similarly, a study which was done in Botswana observed no significant survival benefit, attributing the disparity to variations in healthcare infrastructure and differences in the implementation and monitoring of TB prophylaxis programs. ⁴⁷

The survival of ART patients who had good access of drug supply was higher as compared to patients who had poor access of drug supply. This study finding is supported by recent studies, for instance, a study conducted in Ethiopia found that consistent access to ART medication significantly improved survival rates, with patients having regular drug supplies showing a 1.45 times higher survival rate. ⁴⁸ This study emphasized the importance of reliable drug supply chains in maintaining continuous treatment and preventing interruptions. This similarity might be due to similar study designs and community-based ART service delivery models. Besides, a study conducted in South Africa reinforced these findings, showing that patients with stable and uninterrupted ART drug supplies had a markedly higher survival rate, at 1.42 times greater than those with irregular supplies. ⁴⁹ This study highlighted the critical role of drug availability in enhancing adherence to ART regimens and reducing the risk of developing drug resistance. The consistency in these findings across different regions underscores the importance of robust supply chains and community-based delivery models in improving ART outcomes. ⁵⁰

Conclusion

This study’s lower-than-expected prevalence of ART treatment results highlights important areas for HIV care and treatment services provided through community-based diversified service delivery platforms. Predictors of ART patient survival were found in the current investigation, including drug supply status, retention in care, missing ART medications, and tuberculosis prophylaxis. Improvements in these areas may result in higher survival rates for ART patients, as these characteristics were proven to have a substantial impact on patient outcomes. While retention in care highlights the necessity of ongoing patient involvement and follow-up, missed doses of ART medications underscore the significance of adherence support programs. The necessity for integrated care strategies that address co-infections is demonstrated by the contribution of tuberculosis prevention to increased survival rates. Moreover, ensuring a stable and consistent drug supply is crucial for maintaining uninterrupted treatment and preventing the development of drug resistance.

Appendix.

Abbreviations

AHR

Adjusted Hazard Ratio

AIDS

Acquired Immune Deficiency Virus

ANRS

Amhara National Regional State

ART

Anti-retroviral therapy

ARV

Anti-retroviral

CAD

Community ART Distribution

CAG

Community ART Group

CCLAD

Community Client Led Antiretroviral Distribution

CDSDM

Community Differentiated ART Service Delivery Model

CI

Confidence Interval

CSA

Central Statistical Agency

DSD

Differentiated Service Delivery

EPHI

Ethiopian Public Health Institute

HEPCAG

Health Extension Professionals Led Community ART Group

HIV

Human Immunodeficiency Virus

LTFU

Loss to Follow-Up

MMD

Multi-Month Dispensing

PCAD

Peer Led Community ART Distribution (PCAD)

PEPFAR

US President’s Emergency Plan for AIDS Relief

LHIV

People Living with HIV

UNAIDS

United Nations Program on Acquired Immune Deficiency

USAID

United States’ Agency for International Development

VLS

Viral Load Suppression

WHO

World Health Organization.

Ethical Considerations

Ethical approval was obtained from the Ethical Institutional Review Board (IRB) of GAMBY Medical and Business College. A letter of cooperation (protocol number 03/21358) was issued on March 26, 2024, to the health facilities in the Amhara Region. In addition, the Ethical Review Committee (ERC) of the respective health facilities waived the requirement for informed consent, given the retrospective nature of the study based on medical chart review of people living with HIV enrolled in a community-differentiated ART service delivery model. By not including names or medical records as identifying information in the data extraction form and limiting the usage of the data to the proposed study, the confidentiality of the information was maintained throughout the investigation. The Helsinki Declaration was followed in this investigation.

Consent to participate

Informed consent was waived by the Ethical Institutional Review Board (IRB) of the participating health institutions in the Amhara Region, as the study involved a retrospective review of medical charts of people living with HIV enrolled in a community differentiated ART service delivery model, and informed consent was therefore deemed not applicable given the nature of the study.

Data Availability Statement

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.*