Background
The rapid expansion of antiretroviral treatment (ART) is one of the most remarkable achievements in public health history. ART was provided to eight million people by the end of 2011, which is a 20-fold increase since 2003 [
1]. In 2011, for the first time, a majority (54%) of people eligible for ART in low- and middle-income countries were receiving the treatment [
1].
ART is a life-long intervention that requires a robust framework to adequately monitor and evaluate processes, outcomes and long-term impact, not only at individual patient level but also at health facility and program levels. Patient retention in care is one of the crucial indicators of the success of ART programs [
2‐
7], mainly because high levels of patient retention in care are related to improved adherence to ART, slow progression to AIDS, and increased survival. Moreover, patients who are not retained due to loss to follow-up are likely to develop a high viral load which is associated with an increased risk of infecting other people [
8‐
10]. Hence, countries face the dual challenge of managing and sustaining growing cohorts of patients on ART, in addition to the need for increasing access to ART for the patients who still do not have access to it.
Since the inception of large-scale expansion of ART, ART programs in Africa had retained about 60% of their patients at the end of two years on ART by 2007 [
2]. Loss to follow-up was the major cause of attrition, followed by death. Data on the proportion of patients retained on ART over time continue to show that most patient attrition occurs within the first year and that retention rates tend to stabilize thereafter [
11]. In 2009, the average global retention rate at 12 months was 82%. It dropped to 77% at 24 months and remained stable at 75% and 74.5% at 36 and 48 months, respectively [
3]. These figures are consistent with those from an updated meta-analysis of 39 cohorts from sub-Saharan Africa in 2011 [
4]. These findings indicate that retention in care remains to be a challenge for ART programs though it is improving over time [
3,
4].
Many ART programs have therefore been striving hard to identify and implement appropriate strategies to optimize their retention levels [
1]. In addition, it has been identified that levels of retention vary widely across health facilities and programs [
3,
5]. Hence, health facilities and programs that have achieved higher levels of retention can serve as models for future improvements.
The objectives of this study were to estimate levels of retention in care in nine health facilities in Ethiopia, explain the variability in levels of retention in care across these health facilities, and develop a framework, which will potentially serve as a model, for improving retention. We hypothesized that health facilities with higher and improving retention in care were implementing a number of interventions that positively impact retention.
Methods
The antiretroviral treatment program in Ethiopia
A number of initiatives, including resource mobilization, cost reduction, public-private partnerships, and the public health approach, have been undertaken to expand access to ART in Ethiopia [
6,
10,
12]. As a result, ART services have been decentralized and are available in both health centres and hospitals [62]. By mid-2011, more than 333,400 patients were ever started on ART, and 247,800 patients were alive and taking ART. Retention in care has been identified as a real challenge for the ART program in the country; and, hence, a lot of initiatives, including the case management, the peer education and the expert patient programs, have been implemented to improve it [
5,
13]. These different initiatives have been implemented first as pilot projects since 2007/8, and scaled up in 2009/10. Currently, the case management program is a national initiative launched to be implemented in all health facilities, though its implementation might vary across health facilities.
Study design, data collection and analysis
A retrospective cohort study was conducted in 2009 to determine the outcomes of the ART services in 55 health facilities which were selected using a multi-stage random sampling from all regions in the country [
12]. Among these 55 health facilities, nine health facilities (one tertiary hospital (HP), two general HPs, two urban health centers (HCs), and four rural HCs), with quite variable levels of cumulative retention in care in 2007/8, were selected purposively and conveniently based on their extreme levels of retention and logistics feasibility, respectively, for an in-depth analysis to identify the reasons for the variability in retention in care across these health facilities.
A mixed-methods study, based on the
positive deviance approach[
14], was conducted in 2011/2012. The
positive deviance approach is based on the assumption that the solution to a problem can be found within by identifying and learning from organizations and individuals who do their job better than others. The study was conducted in such a way that the rates of retention in care in the nine health facilities were compared to identify health facilities with higher levels of retention in care. Then the package of interventions implemented in the health facilities with higher or improving levels of retention in care, in comparison with interventions in the health facilities with relatively lower levels of retention in care, were explored and compared.
Quantitative data
‘Current retention’ in care was the primary outcome we used to compare health facilities for their levels of performance.
‘Current retention’ in care is defined as the retention rate in a specific ‘
calendar’ year among patients who were on ART sometime during the
“calendar”. The rates of the ‘
current retention’ in care in the nine health facilities were estimated using the tools developed recently to measure retention in care in ART programs [
15]. In estimating the
‘current retention’ in care, patients who were lost to follow-up sometime before the ‘
calendar’ year, but traced back and restarted on ART during the ‘
calendar’ year, were included in both the denominator and the numerator.
The rates of retention in care in these health facilities were then compared against a reference to identify health facilities with relatively higher and lower levels of retention in care. FH hospital (HP) and WT health center (HC) were used as references for HPs and HCs, respectively, because of their relatively higher levels of retention in care in 2007/8. The odds of retention in care were then calculated using Epi Info-3.5.1 to check for the significance of the difference in the rates of retention in care in health facilities against the reference health facility. Trends in retention in care were also developed to check if health facilities were improving their levels of retention in care over time. Data were collected from patient registers and individual patient files.
Qualitative data
Data collection and sampling: Key informant interviews were conducted with service providers to understand the different interventions implemented by the health facilities with better or improving retention in care or by the community-based organizations linked to them. A total of 72 key informants were included in the study until we reached information saturation. The interviewees were clinicians (one to two from each health facility), adherence counselors (one to two from each health facility), case managers (one to two from each health facility), adherence supporters (one to two from each health facility) and community-based service providers (one to two from each community-based organization providing care and support services). The interviewees were purposively selected as key informants since they were thought to have the potential to provide rich, relevant and diverse information pertinent to retention in care and treatment. The interviews were conducted in local language and tape recorded after consent was received. A question guide, focusing on retention in care, was developed and used to facilitate the interview. The guide includes questions related to challenges for and benefits of retention in care, approaches for improving retention in care and interventions implemented to improve retention in care. The guide also asks for the date when these interventions were started to be implemented in the health facility. The key informant interviews were conducted concurrently with the quantitative data collection in such a way that the interviewers and interviewees were blinded to the ‘current retention’ levels of the health facilities under investigation.
A focus group discussion (FGD), with 12 ART mentors to the nine health facilities with better or improving retention in care, was conducted to identify the different interventions implemented by these health facilities in order to improve retention in care. The FGD participants were purposively selected on the basis of their experience in the field and thought to provide rich, relevant and diverse information. A question guide was used to facilitate the discussion. The guide includes questions related to challenges for and benefits of retention in care, theoretical approaches and practical interventions for improving retention in care. The discussion was conducted in local language. It was conducted for one hour and 45 minutes in one of the health facilities with better retention in care in Addis Ababa. It was tape recorded after consent was received from the participants.
The operational definitions of the different variables used for the study are presented in Table
1.
Table 1
Operational definitions of the variables related to retention in care
Retention
| All patients who are not registered as deceased or LTFU for any reason | Number of patients alive and on ART | Number of patients alive and on ART plus death plus LTFU |
Loss to follow-up
| Patients who miss scheduled visits to the clinic for more than three months after the last visit | Not applicable (NA) | NA |
Transfer out
| It refers to the official transfer of the patient to another clinic | NA | NA |
Transfer in
| It refers to the official transfer of the patient from another clinic | NA | NA |
Cumulative retention
| The total retention by the end of the calendar among patients ever started on ART | Number of patients alive and on ART by the end of the calendar | The total number of patients ever started on ART |
Current retention
| The retention rate during a specific “calendar” among patients who were on ART sometime during the “calendar”
| Number of patients alive and on ART by the end of the “calendar”
| Number of patients alive and on ART by the end of the calendar plus number of patients who died plus LTFU during the “calendar”
|
Calendar
| The time during which the level of the “current retention” is estimated | NA | NA |
The researchers were all health professionals with experience in HIV/AIDS program management and qualitative and mixed-methods research. The FGD was facilitated by an experienced moderator, the first author, and attended by an observer who took notes.
Data analysis: A concurrent constant comparison was conducted on the field notes and transcripts of the records in line with the question guides used during the interviews and FGD. NVivo version 9 was used to support the qualitative data analysis.
Ethics statement
This study was approved by the ethical clearance committee of the Ethiopian Health and Nutrition Research Institute. We obtained informed verbal consent from study participants for both conducting and recording the interview. The verbal consent was tape recorded in local languages. We have also got a letter of support from the Federal authorities to collect patient data from the health facilities.
Discussion
We found that the baseline characteristics of the patients (CD4-cells count, median age and gender) did not vary significantly across the health facilities (Table
2). However, the level of retention in care was variable across these health facilities: DT HP had the least retention in care (OR = 0.46 (0.35, 0.60), P-value = 0.000) among HPs, and BR HC and NM HC had the least retention in care (OR = 0.44 (0.28, 0.70), p-value = 0.000) among HCs in 2009/2010 (Tables
3). We also found that health facilities which had poor retention in care in 2005/6 were able to catch up with health facilities with better retention in care in 2009/10 (Figure
1). Retention in care dropped between 2005/6 and 2007/8; on the contrary, it had improved between 2007/8 and 2009/10. The variability in levels of retention in care among health facilities was less in 2009/10 than in 2006/7 (Figure
1).
In the earlier phases of the ART scale up in Ethiopia, there was a lot of attention for increasing access to ART. However, there was little attention for retaining patients in care. As a result, there was a fast growing problem of attrition of patients from the ART program in 2005/6-2007/8. It took time before it was recognized that retention in care was a real challenge for the ART program. Later, cognizant of the challenge, a lot of initiatives were implemented to improve retention in care. A “case management program” was thus introduced systematically as a pilot project in very few health facilities in 2007/8. It was afterwards scaled up in a number of health facilities, and decided to be a national program to improve retention in care in the country. A number of health facilities, which were not included in the pilot project, with poor retention in care were able to catch up with health facilities with better retention in care. This was possible as a result of diffusion of best practices through different management practices such as supportive supervisions, review meetings and experience sharing visits among health facilities [
6,
10,
15].
There is a lot of evidence that poor retention in care in resource limited countries is due to factors related to health systems, community and individual patient [
16]. In a previous study we found that lack of trust in the services, distance and transport cost, nutrition, opting for alternative traditional medicines, stigma, feeling well, and lack of or inadequate family and community support mechanisms are the main reasons contributing for poor retention in care [
5]. These reasons are also described in many studies in developing countries [
16].
Our qualitative study identified interventions implemented by health facilities and the community-based organizations to address these barriers for retention in care. Health facilities with better and improving retention in care were found to implement comprehensive packages of interventions. We categorized these interventions into four themes: (1) retention in care promoting activities by the health facility, (2) retention in care promoting activities by the community-based organizations, (3) the coordination of the retention in care promoting activities by the case manager(s), and (4) patient information systems managed by the data clerks(s). These comprehensive packages of interventions were identified to be priorities in high-performing health facilities while they were either low priorities or virtually lacking in low-performing health facilities (Table
4). Based on these themes and sub-themes that emerged from the interviews and FGDs, a framework was developed (Figure
2). The framework consists of four themes presented above and discussed below one by one.
Retention in care promoting activities by the health facility level interventions include: ensuring continuity of care (including consultations, medicines, laboratories, and others); provision of care and support services (including transport, nutrition and other related services); coordination of care within and outside the health facilities; preparedness of health care teams for the needs of patients (including clinical, communication, counseling and related skills); support for patient self-management; implementing models of care that facilitate task shifting and “multi-disciplinary team” approaches (involvement of less qualified health workers and community members); provision of adherence counseling; implementation of defaulter tracing activities; and linkage and coordination with community-based organizations. Health facilities which had high priority and focus on such and related interventions were said to have patients who are more informed, motivated and likely to adhere than the patients in health facilities where these interventions are either not priority interventions or not there at all. Moreover, these health facilities were able to identify patients at risk of poor adherence and/or retention, initiate earlier tracing of patients lost to follow-up.
Retention in care promoting activities by the community level interventions include: presence of community-based organizations which work on awareness creation and stigma reduction; mobilization and coordination of community resources; provision of complementary services like counseling, care and support; presence of family-and peer-support mechanisms; and, coordination of the care of patients with health facilities and other community-based organizations. Such kinds of services are either rarely implemented or not available around the health facilities with relatively low level of retention in care.
Patient information system was also found to be one of the building blocks for improving retention in care in health facilities with better or improving retention in care. Health facilities and community-based organizations have not only patient information and monitoring systems but also the culture of sharing and coordinating the information of patients in their catchment areas. Both health facilities and the community-based organizations have patient information and monitoring systems that enable them to identify patients at risk of poor adherence and/or retention, and take appropriate measures accordingly. The data clerk is at the center of the patient information systems.
In addition to the services and the patient information systems in place, the coordination of the care of patients was also found to be a key building block to improve retention in care. The “case manager(s)” in these health facilities are at the center of coordination of the care of patients. The “case managers” coordinate the patient care given by both health facilities and community-based organizations. Moreover, the “case managers” participate in the “multi-disciplinary team” meetings and “catchment-area” meetings. HIV/AIDS case management is a mode of service delivery for chronic illnesses such as HIV/AIDS, and involves health facilities, community-based organizations, faith-based organizations, governmental and nongovernmental organizations and other community resources. The case management program utilizes a “multi-disciplinary team” approach and a network model around its catchment [
17].
Our findings are in line with the findings in other studies which highlight the need for comprehensive packages of interventions to improve retention in care [
18]. These interventions were started to be implemented in other chronic diseases such as diabetes and mental illness when a lot of evidence was generated that patients with chronic diseases need services which go beyond health facilities and are delivered at both home and community levels [
19,
20]. However, health systems in developing countries are basically designed more for acute problems than chronic problems [
21]. Moreover, service delivery models in developing countries are labour-intensive and very much relying on physicians, in spite of the lack of highly qualified health workers in these countries [
22‐
24]. It is therefore important that health systems in these countries adapt their health service organisation and delivery in line with the health systems realities of the countries and the life-long needs of chronic patients: delivery models which require less doctor-time and allow rational redistribution of tasks, and respond to the life-long needs of patients [
22,
25‐
29].
Moreover, care providers are confronted with transitions (epidemiologic and technologic) that affect the patient-provider relationship with the need to redirect certain care relations towards a more horizontal partnership [
30]. The framework in Figure
2 was developed to address the needs of patients with lifelong treatment, the health systems realities of low-income countries, and in line with the chronic care model for patients with chronic illnesses [
31,
32].
This study has both strengths and weaknesses. The first strength of the study is that it is a mixed methods study that aimed to identify health facilities with relatively better and less retention in care and explore how health facilities with better retention in care were able to achieve that level of retention compared to those health facilities which were not able to do that. This facilitates the design of practical models of care that improve retention in care. The second strength of the study is that it included all tiers of health facilities providing ART including tertiary hospitals, general hospitals and health centers. This can give more robust information than a study that includes only one health facility or health facilities from a limited tier of the health facilities. The third strength of the study is that a framework for improving retention is developed based on the themes that emerged from the interviews and FGDs. The first limitation of this study is that it does not estimate the cost-effectiveness of the interventions implemented by health facilities (and community-based organizations) with better retention in care compared to health facilities with less retention. The second limitation of the study is that it cannot give an estimate of the relative contribution of the different interventions implemented by the health facilities with better and improving retention in care. The third limitation of the study is that the design is not able to assess cause and effect relationship, and there might be other explanatory factors that could not be controlled or accounted for.
This study has both theoretical and practical relevance. The theoretical relevance is that it adds to the body of knowledge for interventions to improve retention in care by developing an evidence-based framework structuring the activities to improve patient retention in a resource-limited setting. The practical relevance of the study is that it is addressing the real challenge of many ART programs which are striving hard to manage and sustain them towards universal access to care and treatment services. Hence, the findings from this study will help policy makers, program managers and implementers to design and implement interventions towards better retention in care and improved patient outcomes.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
YA: conceived the study, coordinated and participated in the data collection, conducted the data analysis and interpretation, developed the first draft, and revised subsequent drafts. LL: commented on successive drafts. EW: commented on successive drafts. FR: commented on successive drafts. KR: commented on successive drafts. WVD: advised on the conception of the study idea, data analysis and interpretation, commented on successive drafts. All authors approved the final version for submission.