Situational analysis of varying models of adherence support and loss to follow up rates; findings from 27 treatment facilities in eight resource limited countries

Authors


Corresponding Author Martine Etienne, Institute of Human Virology of the University of Maryland School of Medicine, Baltimore MD, USA. E-mail: metienne@ihv.umaryland.edu

Summary

Objectives  Large-scale provision of ART in the absence of viral load monitoring, resistance testing, and limited second-line treatment options places adherence support as a vital therapeutic intervention. We aimed to compare patient loss to follow up rates with the degree of adherence support through a retrospective review of patients enrolled in the AIDSRelief program between August 2004 and June 2005.

Methods  Loss to follow up data were analysed and programs were categorised into one of four tiered levels of adherence support models: Tier I, II, III, and IV which increase from lowest to highest support. Bivariate and t-test analyses were used to test for significant differences between the models.

Results  13,391 patients at 27 treatment facilities from six African and two Caribbean countries began antiretroviral therapy within the first year of the AIDSRelief program. The mean loss to follow up within the first year was 7.5%. Eight facilities were Tier I, three (Tier II), nine (Tier III), and seven (Tier IV). Facilities in Tier I had a loss to follow up rate of 14%, Tier II (10%), Tier III (5%), and Tier IV (1%). The proportion of loss to follow up for Tier I and Tier III were significantly different from each other (P < 0.02), as were Tier I and Tier IV (P < 0.006). There were differences between Tier II and Tier IV (< 0.009) as well as Tier III and Tier IV (P < 0.017).

Conclusion  These data strongly support the use of proactive adherence support programs, beyond routine patient counselling and defaulter tracking to support the’public health approach’to ART.

Introduction

In 2003 there were an estimated 100 000 patients on antiretroviral therapy in sub-Saharan Africa and only 50 programs delivering antiretroviral therapy (WHO/UNAIDS 2006). By 2006 there were about 5000 clinics estimated to be delivering ART to close to 1 million patients (WHO/UNAIDS 2006). The success of these programs in providing access to treatment has been enormous. According to the 2008 report released by UNAIDS/WHO/UNICEF (2008a,b) only 31% of those requiring ART have been treated.

There is an unquestionable impact on AIDS related death and morbidity expected from antiretroviral therapy, yet the loss to follow up rates and mortality of those started on ART are less than desirable (Braitstein et al. 2006). Some programs have reported loss to follow up rates of 25–54% with the median time on treatment of 40 weeks–24 months (Lawn et al. 2005; Severe et al. 2005; Wools-Kaloustian & Kimaiyo 2006; Ferradini & Leakhena 2007; Rosen et al. 2007; Bwirire et al. 2008; Martin et al. 2008) These figures are similar to those reported in clinical cohorts in the United States and Europe, which showed that up to 50% of patients may no longer be on first-line therapy by 18 months (Lucas et al. 1999; Amoroso et al. 2004).

Determinants that negatively impact durable viral suppression including co-morbidities, drug toxicity, resistance, and requirements of strict adherence are amplified in resource limited settings both through the advanced stage at presentation for treatment of most patients and the current choice of antiretroviral regimens in most of these countries. Programs face these challenges without the luxury of access to viral load assays and genotypes to guide clinical decision making. These additional challenges place significant clinical importance on the patient’s ability to adhere to medications and the programs’ ability to support patients through therapy especially during the period of treatment initiation and immunological reconstitution (Gross et al. 2001; Ferradini et al. 2006).

Ensuring adherence to medical therapy and care thus becomes a vital therapeutic intervention which programs can use to support their patients. Despite this intuitive solution there are few described structured adherence interventions to guide global scale-up efforts. The success of increasing access to antiretroviral therapy leads to the urgent need to identify effective and generalizable adherence support models that can lead to durable viral suppression of the initial treatment regimen, reduce clinical reliance on viral load monitoring, and in turn help prevent the problems associated with widespread drug resistance.

In terms of cost effectiveness and health systems factors facilities that are able to remove patient co-payments for antiretroviral treatment as well as treatment for opportunistic infections lose fewer patients (Bisson & Stringer 2009). Facilities that have been able to train providers and care workers in strategies on preventing and/or reducing patient attrition, as well as providing meals and patient transport funds for patient visits have also seen a reduction in patient attrition (Bisson & Stringer 2009). The patient provider relationship also greatly affects patient adherence (Hofer et al. 2004).

Loss to follow up has been linked to fatal patient outcomes (Etard et al. 2006; Lanoy et al. 2006). In fact, many patients lost to follow-up are found dead when defaulter tracking is implemented after the fact (Schneider et al. 2007; Variava 2007). A study by Losina et al. (2009) highlights the importance of not only scaling-up ART care but strengthening the quality of the care provided. We aimed to evaluate the impact that sequentially more supportive adherence programs had on the 12 month loss to follow up rates of patients initiating ART across a variety of resource limited clinics and to define the optimal adherence support strategies which lead to reproducible success in these patients.

Methods

Since its inception in August of 2004, the AIDSRelief program, a President’s Emergency Plan for AIDSRelief (PEPFAR) grantee, has facilitated the care and treatment of over 300 000 patients in approximately 250 faith-based, private, and public hospitals and clinics in eight countries (Kenya, Rwanda, Uganda, Tanzania, Zambia, Nigeria, Haiti, and Guyana). Through the provision of antiretroviral therapy, laboratory testing, drugs for opportunistic infections, and a wide variety of clinical and community based training, AIDSRelief has been able to support rapid scale up of care and treatment in these countries. The scale up has been rapid yet cautious in an effort to maintain the quality of the medical care provided. This retrospective analysis is limited to the first year of the AIDSRelief program; August 2004 through June 2005 in which there were a total of 34 sites and 16 611 patients who started antiretroviral therapy. Included were sites that began their antiretroviral program between August 2004 and June 2005 and also had complete data on loss to follow up information. Facilities that had incomplete loss to follow up data were excluded. Seven facilities were excluded from the analysis due to missing data. The total population analysed included 27 facilities with 13 391 patients who started antiretroviral therapy between August 2004 and June 2005.

Loss to follow up data were aggregated from the Centers for Disease Control (CDC) quarterly reports of ART treated patients. Treatment facilities are required to track all patients on ART that fail to attend clinic for a previously scheduled appointment or fail to attend within a certain time period. When a patient misses a visit, this is recorded in a log form. At the end of each clinic day, the treatment facility reviews the log and patients who missed their appointments would be followed up with a home visit or a telephone call the following day. A patient is considered lost to follow up if s/he has not been to clinic or picked up drugs for at least 3 months (CDC 2006). All of the AIDSRelief facilities in this analysis actively attempted to contact patients with missed visits either through home visits or cell phone calls. Therefore, in addition to tracking clinic appointments, if a patient cannot be contacted for 3 months through home visit or community tracking, this patient is designated lost to follow up. These data are recorded in individual patient charts and in clinic logs. In this analysis, loss to follow up does not include patients that are known to have died, transferred to another treatment facility, or have stopped ART either by patient preference or provider initiative.

Treatment facilities were characterised according to the adherence model used to support patients starting ART. Each facility was categorised in one of four ‘Support Tiers’ based on the level of adherence support given to the patient.

 Tier 1Tier2Tier3Tier4
Adherence CounsellingXXXX
Structured Treatment Preparation XXX
Community Home Visits  XX
Supportive Supervision by Community Nurse   X
  • Tier I: This model provided one-on-one adherence counselling only, prior to starting a patient on ART.

  • Tier II: This model included adherence counselling as in Tier I and a structured HIV education program for preparing patients for therapy. The facility had a number of treatment preparation sessions that covered essential topics that were mandatory for patients and their families to attend prior to starting antiretroviral. These sessions were designed with guidance from the AIDSRelief technical trainers and all patient education was supported by the local facility staff.

  • Tier III: In this model, the facility provided the adherence counselling, the structured treatment preparation sessions, and included an intensive community based follow up program for patients starting antiretroviral therapy. Home visits are conducted by volunteers and/or paramedical staff trained using existing countries’ Ministry of Health community tools and supplemented by the AIDSRelief Community Adherence Support Curriculum tools to identify the different antiretroviral medications, side effects. Staff encourage continued adherence to antiretroviral and clinic appointments, publicise secondary prevention messages and strategies aiming to reduce high risky behaviours, promoting healthy nutrition and disease prevention. The cadre of workers in this level are usually infected or affected by HIV. These volunteers receive incentives such as small stipends, home visiting materials, refresher trainings, bicycles, lunches and other motivational effects. Some programs employed volunteers to become ‘adherence counsellor staff’ on part time employment bases. Patient information is transferred from the field to the hub clinic via home visit reports, phone calls, and integrated meetings.

  • Tier IV: Includes all the components from Tier I–Tier III. In addition, a community nurse is dedicated to work in the field with the community volunteers. This community nurse holds a supervisory position, both coordinating and medically supporting the work of the community health volunteers and other non-medical staff. The community nurse provides planned home visits, as well as immediate triage care for patients in the home based on volunteer or clinic provider concern. Patient information is collected via home visit reports as in Tier III, but the community nurse is the direct recipient of this information in the field prior to the hub clinic, allowing for a higher level of home based support and communication with the clinic providers.

Data analysis

Bivariate and t-test analyses were used to test for differences between the models. Tukey’s HSD test for the difference between mean number of providers, total number of persons with ART, median CD4 at baseline, CD4 at 6 months, site population characteristics (stable or migratory) and accessibility to transportation at different levels of tiers were performed using anova.

Results

Within the first year of the AIDSRelief program 13 391 patients at 27 treatment facilities from Guyana, Haiti, Kenya, Nigeria, Rwanda, Tanzania, Uganda, and Zambia began antiretroviral therapy. The demographics of the AIDSRelief facilities are dynamic, reaching very rural and marginalised patients (Table 1). Several of the facilities are private non-for-profit faith-based health centers providing antiretroviral therapy in a catchment area of 10 000–100 000 people. Many of these facilities are also the sole source of care within their communities. Some patients travel long distances to reach a health center.

Table 1.   Affiliation of faith based facilities and type of area served
Affiliationn%
 Catholic1970
 Protestant829
Total27100
Areas servedn%
 Rural1659
 Urban829
 Sub-urban311
Total27100

The mean loss to follow up within the first year of AIDSRelief programs was 7.5%. Eight facilities were Tier I, three were Tier II, nine were Tier III, and seven were Tier IV. Facilities in Tier I had a cumulative loss to follow up of 14%vs. facilities in Tier II (10%), Tier III (5%), and Tier IV (1%) (Figure 1). Although the proportion of loss to follow up in Tier II was lower than Tier I, there were no statistical differences between the two tiers (P < 0.301) (Table 2). The proportion of loss to follow up in Tier III was significantly lower (P < 0.02) than in Tier I (Table 3). The proportion of loss to follow up in Tier IV was lower than Tier I (P < 0.006) (Table 4). Although the proportion of loss to follow up in Tier III was lower than Tier II, there were no statistical differences between the two tiers (P < 0.083) (Table 5). Tier IV loss to follow up was also lower than Tier II (P < 0.009) and Tier III (< 0.017) (Tables 6 and 7).

Figure 1.

 Percent loss to follow-up within 12 months on ART (Guyana, Haiti, Nigeria, Rwanda, Tanzania, Kenya, Uganda, and Zambia).

Table 2.   T-test of differences in proportions between Tier I and Tier II
TiersProportions (%)95% CIP value
Tier I140.040 to 0.2310.300
Tier II10−0.110 to 0.302
Table 3.   T-test of differences in proportions between Tier I and Tier III
TiersProportions (%)95% CIP value
Tier I140.040–0.2310.020
Tier III50.014–0.075
Table 4.   T-test of differences in proportions between Tier I and Tier IV
TiersProportions (%)95% CIP value
Tier I140.040 to 0.2310.006
Tier IV1−0.003 to 0.020
Table 5.   T-test of differences in proportions between Tier II and Tier III
TiersProportions (%)95% CIP value
Tier II100.110–0.3020.083
Tier III50.014–0.075
Table 6.   T-test of differences in proportions between Tier II and Tier IV
TiersProportions (%)95% CIP value
Tier II10−0.110 to 0.3020.009
Tier IV1−0.003 to 0.020
Table 7.   T-test of differences in proportions between Tier III and Tier IV
TiersProportions (%)95% CIP value
Tier III50.014 to 0.0750.017
Tier IV1−0.003 to 0.020

Tukey’s HSD test results show no significant differences of mean of number of providers, total number of persons with ART, median CD4 at baseline, CD4 at 6 months, site population charactictics (stable or migratory) at different levels of tiers. However, access to transportation in Tier I is significantly (P < 0.05) different from Tier II. In addition, access to transportation in Tier II is also significantly (P < 0.05) different from Tier III.

Discussion

AIDSRelief provided antiretroviral therapy to 34 facilities within the first year of the program in six countries across Africa and two countries in the Caribbean. This analysis reviews 27 of these facilities. The mean loss to follow up within the first year of AIDSRelief programs was 7.5%, compared to 16% (Brinkoff, et al. 2008) and up to 56% (Rosen et al. 2007) in other international programs. It is clear that retroactively following up of patients does not offer the most supportive model for care and treatment. The inability to provide highly supportive care and treatment to patients and their families, leads to poor patient outcomes as well as serious public health implications (Bisson et al. 2008).

Home visits markedly improved patient’s retention in care. These data support the use of proactive adherence support programs, beyond routine patient counselling and defaulter tracking to support the ‘public health approach’ to providing antiretroviral therapy. Programs with at least a defined structured treatment preparation plan for patients starting antiretroviral therapy will significantly lower the likelihood of high loss to follow up rates. The addition of the community nurse presented a statistically significant difference between Tier IV and Tier III as well as Tier IV and Tier I. This demonstrates the importance of continuous follow-up in the community as a key to successful treatment outcomes. A comprehensive ART adherence program is able to support patients before, during and continuously in the community. Programs that may be limited in financial and staffing resources can still develop strong support programs that promote adherence and patient support.

Limitations include the exclusion of several sites because of lack of necessary data for this specific analysis. Excluded facilities had similar characteristics as included facilities and were not significantly different. It is important to note that these data are abstracted quarterly from patient charts. Additionally, some sites already had active home-based care programs and community links and providing antiretroviral therapy became an addition to these programs. Finally, few sites had greater capacity for treating patients on antiretroviral therapy than other sites with limited experience. We strongly believe these data have important public health implications.

Although there were nearly 3 million people receiving ART at the end of 2007 (UNAIDS/WHO/UNICEF 2008a,b), it is clear that access does not equal outcome. The loss to follow up rate is one of the leading causes of attrition in many HIV/AIDS programs providing ART, with rates as high as 56% (Rosen et al. 2007). With such high rates of loss to follow up, HIV/AIDS programs should be required to develop operational methods and strategies of treatment and care delivery necessary to ensure high levels of success, limiting the development and spread of resistant virus in the community. We must establish models of treatment delivery that provide highly supportive therapy to match our expected treatment outcomes (Farmer et al. 2001; Walton et al. 2004). Provision of home-based adherence support with the use of community health care workers will ensure optimal treatment outcomes (Mukherjee et al. 2006; Koenig et al. 2008). The ability to identify and aggressively implement effective adherence support models that promote successful outcomes will contribute to the goal of not only providing access to ART but achieving quality care and treatment.

Acknowledgements

The AIDSRelief Year One Clinics: Kamokchya Christian Community Center, Nile Treatment Center, Nsambya Home based Care, Nsambya MTCT, Nsambya Private Clinic, Bethelehem Treatment Center, Bushenyi Kabwohe, Bushenyi Katungu, Chikuni Hospital, Chogoria Hospital, CRESCO Hospital, Comboni Hospital, Consolata Hospital, Kabarole Hospital, Kalongo Hospital, Kasanga Hospital, Katondwe Hospital, Kendu Bay Hospital, Kijabe AIC Hospital, Kikuyu Hospital, Kololo Treatment Center, Macha Hospital, Mombasa CBHC, Mtendere Hospital, Mwea Hospital, Nazareth Hospital, St. Monica Hospital, Tumu Tumu Hospital, Villa Maria Hospital, Virika Hospital, Wusakile Hospital. University of Maryland School of Medicine Institute of Human Virology: Dr Anthony Amoroso, Dr Robert Redfield, Dr Mian Hossain.

Conflicts of interest

The authors have declared that they have no conflicts of interest.

Ancillary