Proportional Hazards Model on Attrition and its Predictors in Community Antiretroviral Refill Groups among ART Users in Eastern Ethiopia

Background

The HIV epidemic remains a global public health threat, with an estimated 37 million people living with HIV worldwide, 24.4 million in Sub-Saharan African countries, and Ethiopia, approximately 60% of PLHIV know their status.^1,2 Of the 37 million people estimated to be living with HIV globally, approximately 62% are on life-saving antiretroviral therapy³ and 53% are virally suppressed. ⁴ The prevalence of antiretroviral treatment(ART) failure was 15.9%. ⁵ ART plays a crucial role in managing HIV by reducing viral load, morbidity, mortality, and transmission. ⁶ There is presently no cure for HIV. However, it can be manageable with life-long antiretroviral therapy by patient-centered approaches to ART delivery, a strategy termed differentiated service delivery models (DSDM) therapy. ³ The community-based ART model is one of the DSDM strategies for stable ART clients, who aim to provide a more client-centered approach to HIV care and treatment. ⁷ In sub-Saharan Africa, ⁸ two community ART delivery models exist: group models managed by clients and individual models overseen by health facilities. In the group models, PLHIV self-organizes groups and represents each other to collect ART refills. The use of these two models has been associated with high retention and favorable clinical outcomes however, community Art refill groups (CAGs) are more recommendable.^4,9,10 CAGs are more recommendable for stable ART patients who live in the same geographic area and have agreed to join the group. Each group will consist of 6 to 12 individuals who will pick up their ART refill outside of a medical institution.^11,12

Though, Ethiopia has implemented decentralization of services and task shifting of ART delivery service through community ART group refill approach and the model is a critical step in treating people living with HIV, there is limited evidence about the attrition of ART clients in this model, particularly in the study area. Therefore, this study aimed to assess attrition and its predictors in community antiretroviral ART refill groups among ART users in Eastern Ethiopia.

Methods

Sample Size and Sampling Procedure

The sample size was determined by using a single population proportion formula with 95% confidence level, 5% margin of error, and 50% (rules of thumb) since there is no previous study was conducted in the study area. With 10% contingency, the final sample size was 423.

Based on the city administration health department report, approximately 2213 PLHIV were enrolled in CAGs in six health facilities. Due to the different numbers of clients in each health facility, the proportional allocation was done based on the medical records between January 01, 2021 and December 31, 2023. Finally, from the selected cards in each facility, a simple random sampling technique (lottery methods) was used (Figure 1).

OPEN IN VIEWER

Figure 1.

Schematic presentation of the sampling procedure for CAG model attrition among ART users in public health facilities in Eastern Ethiopia, 2023.

Results

In the study, a total of 423 respondents with a mean age of 40.90 (SD ± 8.12) years were included. The majority of the respondents, 319 (75.11% had community recruitment levels. Two-thirds, 285(67.38%) respondents had ART regimen change (Table 1).

OPEN IN VIEWER

Table 1.

Demographic, Clinical, and Personal Characteristics of Respondents in CAGs among ART Users in Public Health Facilities of Eastern Ethiopia, 2023.

Variables	Categories	Frequency	Percentage (%)
Age (years)	18–24	154	36.41
	25–67	260	61.47
	≥68	9	2.13
Sex	Male	105	24.82
	female	318	75.18
Marital status	married	144	34.04
	unmarried	279	65.96
Educational status	Can’t read and write	57	13.48
	Can to read & write	55	13.00
	1–8 grade	114	26.95
	9–12 grade	148	34.99
	College and above	49	11.58
Previous residency	Rural	155	36.64
	Urban	268	63.36
Recruitment level	Community	319	75.11
	Health facility	105	24.59
Distance from health facility	<5km	160	37.83
	≥5 km	263	62.17
Group size	<3	12	3.31
	3–12	396	93.62
	>12	13	3.07
Viral load outcome	Suppressed	104	24.59
	Unsuppressed	279	65.96
	Missed*	40	9.46
WHO clinical stage	Stage I	100	23.64
	Stage II	101	23.88
	Stage III	129	30.49
	Stage V	93	21.99
ART regimen change	No	138	32.62
	Yes	285	67.38
History of LTFU	No	134	31.68
	yes	289	68.32
Duration on ART (years)	≥ 5	104	24.59
	< 5	319	75.41

Key: Missed = viral load not detected by test used.

Incidence of Attrition and its Predictors

In this study, the incidence of attrition was 6.63 (95% CI: 5.78, 7.48) per 100 person-years. The median duration of months in CAGs from the start till the end of the follow-up period was 9 months (IQR = 24).

The study indicated that the recruitment level from health facilities was 39% times less likely to undergo attrition as compared to the community recruitment level [AHR = 0.61, 95% CI, 0.42, 0.90]. Respondents with no history of LTFU were 59% times less likely to undergo attrition as compared to their counterparts [AHR = 0.41, 95% CI, 0.09, 0.54]. The risk of attrition among those with stage IV was 1.13 times more likely as compared to those on WHO stage I [AHR = 1.13, 95% CI, 1.04, 3.76] (Table 2).

OPEN IN VIEWER

Table 2.

Factors Associated with Attrition of CAGs among ART Users in Public Health Facilities at Dessie, Eastern Ethiopia, 2023.

		Attrition in CAGs		CHR [95% CI]	AHR [95% CI]
Variables	Category	Yes	No
Age (in years)	18–24	81	73	1.00	1.00
	25–67	149	11	0.08 [0.32, 1.89]	0.13 [0.32, 1.89]
	>68	7	2	2.10 [0.34, 3.83]	0.64 [0.31, 4.72]
Marital status	Unmarried	150	129	1.00	1.00
	Married	82	62	0.87 [0.40, 4.90]	0.11 [0.31, 3.92]
Recruitment level	Community	192	127	1.00	1.00
	Health facility	40	64	2.41 [0.80, 7.90] **	0.61 [0.42, 0.90] *
History of LTFU	Yes	71	63	1.00	1.00
	No	161	128	0.89 [0.12, 0.90] **	0.41 [0.09, 0.54] *
ART regimen change	No	74	64	1.00	1.00
	Yes	158	127	0.92 [0.39, 0.97] **	0.44 [0.28, 1.93]
Viral load outcome	Suppressed	60	44	1.00	1.00
	Unsuppressed	143	136	1.29 [1.10, 3.95] **	1.11 [0.22, 3.55]
	Missed	29	11	0.51 [0.10, 4.95]	0.11 [0.02, 2.65]
WHO clinical stage	Stage I	68	32	1.00	1.00
	Stage II	55	46	1.77 [1.05, 2.23] **	1.13 [0.94, 4.76]
	Stage III	59	70	2.52 [0.85, 7.23]	2.13 [0.64, 4.76]
	Stage V	50	43	1.82 [1.05, 5.23] **	1.13 [1.04, 3.76] *

Key: ** = P < 0.25, * = p-value <0.05, CHR = Crude Hazard Ratio, AHR: Adjusted Hazard Ratio, CI = Confidence Interval &1.00= Reference.

Discussion

The declining retention rates observed over time highlight the challenges in maintaining long-term adherence to treatment in community ART groups. ¹⁵ Our study showed that the incidence of attrition was 6.63 per 100 person-years This finding is supported by previous research highlighting the challenges of sustaining retention in care among ART users in community settings. ¹⁴ Because this study highlights the importance of addressing aggressiveness in community ART groups, the need for interventions to improve retention in these groups is further emphasized.

Recruitment level was found to be a significant predictor of attrition, with a lower risk observed among health facilities recruited. This finding suggests that individuals recruited from health facilities may have better access to care and support, leading to improved retention rates in community ART groups. Additionally, the Ministry of Health rolled out a plan for the scale-up of CAGs in the community to offer psychosocial support and they are comfortable rather than recruited from the health facilities. ¹⁶

Respondents with no history of LTFU were found to have a significantly lower likelihood of attrition compared to those with a history of LTFU. This finding suggests that individuals who have previously not experienced loss to follow-up may have a heightened awareness of the importance of adherence and engagement in care, leading to improved retention rates. ¹⁷

Furthermore, the study identified that individuals in the advanced WHO clinical stage (Stage IV) had a higher risk of attrition compared to those in the early stages. This finding highlights the need for targeted interventions and support for individuals with advanced disease stages to ensure their continued engagement in care. ¹⁸ Based on the study findings, it is recommended that healthcare programs and policymakers prioritize interventions targeted at the identified risk factors on improving CAGs, updating CAG recruiting levels, enhancing virologic monitoring, and providing continuous support and counseling. ¹⁹

Conclusion

Our study findings underscore the importance of targeted interventions to enhance long-term adherence to treatment. By addressing specific risk factors such as recruitment level, history of LTFU, and clinical stage, healthcare programs can design and implement interventions that promote sustained engagement in care for individuals living with HIV. These interventions may include strengthening recruitment processes, addressing loss to follow-up, tailoring support for advanced disease stages, and implementing continuous monitoring and evaluation.