SEARCH

SEARCH BY CITATION

Keywords:

  • antiretroviral therapy;
  • loss to follow-up

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Objectives  To determine the proportion of patients returning to antiretroviral treatment (ART) and factors associated with their return in a resource-limited setting.

Methods  Between September 2006 and March 2009, at two ART-providing facilities in Lilongwe, Malawi, we identified patients who had missed clinic appointments by more than 3 weeks and therefore would have run out of antiretroviral drugs. We traced these individuals, documented reasons for missed appointments and, where appropriate, arranged another ART clinic appointment.

Results  Between April 2006 and March 2009, 2653 patients on ART had missed 3098 scheduled appointments. We successfully traced 85%, of whom 30% had died. Of the 1580 patients found alive, 25% had transferred to another ART clinic, 21% had collected drugs from other sources, 11% had treatment gaps; 40% had stopped taking drugs, 1% had not started taking drugs despite collecting them and 2% refused to be interviewed. Of the 1158 LTFU patients who had not died, transferred out or declined to be interviewed, 89% promised to return to their ART clinic and 74% actually did. The probability of returning to the clinic was significantly associated with being women, aged over 39 at ART initiation and having either treatment gaps or uninterrupted therapy. The B2C project reduced the proportion of patients finally classified as LTFU by 59%.

Conclusion  Early active follow-up of LTFU patients resulted in marked improvement in known patient outcomes and improved retention in the treatment programme.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Since the advent of antiretroviral therapy (ART) in sub-Saharan Africa (SSA), many countries with high HIV prevalence have been scaling up ART programmes. In 2005, an estimated 25.8 million people lived with HIV, and only 810 000 HIV-positive people had been started on ART in SSA (WHO 2006). By the end of 2008, 3 million started ART in SSA (UNAIDS 2009). Despite this initial success, the scale-up process towards universal access to ART faces a number of challenges. More than half of the patients enrolled in ART programmes in the SSA were lost to follow-up (LTFU) from treatment within 2 years of ART initiation (Rosen et al. 2007; Bisson et al. 2008). In Malawi, 223 437 patients had been registered for ART by December 2008, of whom 24 409 (11%) were LTFU (HIV Unit 2008). A large proportion of patients classified as LTFU have died (Yu et al. 2007); however, in cases where death is not the reason for loss to follow-up, there are concerns about drug resistance and treatment failure (Adam et al. 2003). Treatment interruptions reduce much of the immunological benefit of ART and increase AIDS-related mortality and morbidity (Hogg et al. 2002; Adam et al. 2003). Improving ART programme retention remains a critical challenge to programme success.

Programme attrition can be reduced through active tracking of patients on ART who miss visits and by developing a strategy for their return to care (Rosen et al. 2007). We therefore piloted an intervention designed to improve long-term retention in treatment (the ‘Back-To-Care (B2C)’ project). In this article, we highlight the design of the programme and determine the proportion of patients LTFU and the factors associated with successful return to the ART clinic.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Background

Data used for the study came from two ART facilities in Lilongwe, Malawi, one at Kamuzu Central Hospital (Lighthouse, the largest public sector ART provider in the Central region of Malawi) and the other at the district hospital (Martin Preuss Centre (MPC)). By April 2009, Lighthouse and MPC had 5334 patients on ART and 3959 patients alive and on ART, respectively. Both Lighthouse and MPC use an electronic data system (EDS) for routine collection of ART data; Lighthouse since January 2005 and MPC since December 2006. The B2C project was introduced at Lighthouse in July 2006 and at MPC in September 2007, aiming to improve long-term patient retention on ART.

Data collection

At the clinic first visit, patient and guardian demographics, phone numbers and residential addresses were entered in the EDS. On initiation of ART, a schematic map with landmarks indicating patient residence was drawn on a locator form. Patients were seen 2 weeks after ART initiation and then routinely every month for clinical assessment and antiretroviral (ARV) drug dispensing. At each clinic visit, the treatment outcome of the patient was updated in the EDS, and the remaining pill counts were entered separately for each type of ARV drugs, including tablets brought to the clinic and tablets left at home. The next appointment date was calculated electronically based on regimen, prescribed schedule and number of tablets newly dispensed plus tablets remaining at present visit. Visit intervals were extended to 2 months if patients had completed 6 months of ART without complications. Outcomes for the overall ART programme were classified as ‘Alive and on ART’, ‘Died’, ‘LTFU’ (no visit within 2 months of running out of ARVs)’, ‘Transfer out’ and ‘Stopped ART’.

For the purposes of the B2C project, so as to identify patients more promptly, ‘LTFU’ was defined as missing a scheduled ART clinic appointment by 3 weeks or more. As a patient could appear on the tracing list more than once during the study period, we defined an ‘LTFU case’ as a single episode of a missed appointment. The EDS produced a weekly list of overdue patients. One receptionist and 3 trained health extension workers were responsible full-time for verifying the status of patients on the list to rule out data errors and for tracing confirmed LTFU patients by phone or home visit. The B2C staff contacted the patient/guardian to ascertain the patient’s true ART outcome status. Status was recorded as either dead, transfer to other ART facility, alive and on ART (uninterrupted therapy or treatment gaps), stopped ART (stopped or never started ART), tracing rejected, refused comment and LTFU. ‘Uninterrupted therapy’ meant that the patient was still taking ARVs in spite of not coming to the ART clinic, and ‘treatment gaps’ referred to a situation where a patient took none or fewer than the prescribed drugs before the interview date. ‘Transfers’ were classified as either an ‘official transfer’ if transfer-out notes were available in the patient’s health passport but not at the ART facility, or as ‘self-transfer’ if the patient had arranged the transfer independently. ‘Tracing rejected’ referred to patients that could not be traced because they lived outside the clinics’ catchment areas. All patients found alive who had not transferred to other ART facilities were asked to return to the clinic (defined as ‘Patients expected back’) and among those agreeing to return a specific appointment was scheduled with the clinic. Two further call back attempts were made for patients who did not realize these appointments before they were classified as ‘LTFU’. Patients reporting back to the clinic were defined as ‘return-to-care’. Tracing outcomes were recorded on a standardized form.

Statistical analysis

Data were entered into an MS Access 2003 database and analysed using STATA 10.0. Descriptive analysis was carried out to determine the rate of loss to follow-up and describe the study population, followed by univariable analysis of the association between explanatory variables and the outcome variable (‘return-to-care’). Only records for patients who had uninterrupted therapy, treatment gaps, stopped or had never started ART were used in the univariable and multivariable analyses because these were the patients expected back at the clinic.

Logistic regression was then used to identify patient characteristics that were independently associated with the outcome of return-to-care. Explanatory variables were included in the logistic models if they demonstrated significant association with ‘return-to-care’, or if they confounded association involving other explanatory variables. Separate multivariable analyses were run for men and women to investigate whether factors associated with return-to-care differed between them. The level of significance was set at  0.05, and 95% confidence intervals (CI) were used throughout.

Ethical approval

Informed consent for tracing in the event of missed appointments was obtained from patients during clinic registration as a component of routine practice. The Malawi National Health Science Research Committee provides general oversight and approval for the collection and use of routine programmatic data for monitoring and evaluation, as was the case with this study. The study was also approved by the Ethical Advisory Group (EAG) of The International Union Against Tuberculosis and Lung Disease, Paris, France (The Union).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Patient cohort

Between April 2006 and April 2009, 13 981 patients accessed ART at Lighthouse and MPC. Of these, 8095 (58%) were women and 777 (6%) were children aged 14 or less. The median age at ART initiation was 34; 5707 (41%) were aged between 30 and 40. At ART initiation, 3144 (23%) were in WHO clinical stage 1 or 2 with CD4 cell count below 250/mm3, 8117 (59%) were in stage 3, 2529 (18%) in stage 4 and in 191 (1%) the reason for starting ART was not available.

Back-to-Care (B2C)

Between April 2006 and April 2009, 2653 patients had missed 3098 scheduled appointments by 3 weeks or more. Most of these patients (2320, 87%) had missed one appointment, 251 (9%) had missed two appointments and 82 (3%) had missed 3 or more appointments. The rate of loss to follow-up was 9.3 per 100 person-years. Among the B2C cohort, 1453 (55%) were women and 163 (6%) were children aged 14 or less. The median age at ART initiation was 33; 999 (38%) were aged between 30 and 40. Reason for starting ART was documented in 2597 (98%) cases; 614 (24%) of these started in WHO clinical stage 4, 1539 (59%) in stage 3 and 444 (17%) in stage 1 or 2 with a CD4 count <250/mm3. Most patients (1702, 64%) had been taking ARVs for <6 months. Patients with missed appointments were more commonly men, patients who started ART in WHO clinical stage 4 and younger (<40 years) than patients who did not appear on the B2C list (Table 1).

Table 1.   Characteristics of patients who accessed ART at Lighthouse and MPC between April 2006 and April 2009 (N = 13 981)
Patient characteristicsNon-B2C patientsB2C patientsP-value
n = 11 291 (%)n = 2 690 (%)
Sex
 Male4676 (41)1210 (44)0.001
 Female6615 (59)1480 (55)
Age at ART initiation
 <15612 (5)165 (6)<0.001
 15–397302 (65)1826 (68)
 ≥403377 (30)699 (26)
Reason for starting ART
 WHO clinical stage 1 and 22689 (24)455 (17)<0.001
 WHO clinical stage 36559 (59)1558 (59)
 WHO clinical stage 41909 (17)620 (24)

The B2C profile for all patients LTFU is presented in Figure 1. Among the 3098 cases identified as LTFU, 845 (27%) could not be traced either because of incorrect addresses on the locator form (815 (96%)) or because they lived outside the clinics’ catchment area (30 (4%)). Of the total LTFU patients, 2253 (73%) were successfully traced. The most common reason for missed appointments was death (673 cases, 30%). Of the 1580 LTFU patients who were found alive, 312 (20%) had officially transferred to another ART clinic, which was confirmed by transfer-out notes in the patient’s health passport while no record of the transfer had been kept at the clinic.; 83 (5%) had self-transferred to another ART clinic; 335 (21%) were on uninterrupted therapy because they had collected drugs from alternative sources (friends, relatives or unlicensed vendors); 179 (11%) had treatment gaps; 630 (40%) had stopped taking drugs; 14 (1%) had not started taking drugs despite collecting them; and 27 (2%) refused to be interviewed.

image

Figure 1.  Profile of all patients identified as lost to follow-up by the B2C program.

Download figure to PowerPoint

Of the 1158 LTFU patients who had not died, transferred to another clinic or declined to be interviewed, 1030 (89%) promised to return to their ART clinic and 764 (74%) actually did.

In univariable analyses, age, duration on ART and tracing outcome were associated with return-to-care (Table 2); sex and WHO stage at initiation were not. Patients who initiated treatment at age >39 or <15 had a higher probability of returning to care than those aged 15–39 (age ≥ 40: OR 2.33, 95% CI 1.55–3.50, age < 15: OR 2.65, 95% CI 1.39–5.03). The probability of returning to care was also higher in patients who had been on ART for more than 6 months (OR 2.20, 95% CI 1.60–3.00). Patients with treatment gaps had a higher probability of returning to care than patients who had stopped or never started ART (OR 10.1, 95% CI 4.59–22.1). Patients in WHO clinical stage 4 at ART initiation had a higher probability of returning to care than patients in WHO clinical stage 3 (OR 1.55, 95% CI 1.01–2.35). Multivariable analyses showed no significant association between WHO stage at ART initiation and returning to care. Men were less likely to return to care than women (adjusted OR 0.56, 95% CI 0.39–0.80). Patients aged 40 and older had a higher probability of returning to care than those aged 15–39 (adjusted OR 2.35, 95% CI 1.51–3.64). Patients with treatment gaps were more likely to return to care than those who had stopped/never started (adjusted OR 8.69, 95% CI 3.92–19.29), and patients who had uninterrupted therapy were more likely to return to care compared to those who had stopped/never started (adjusted OR 3.11, 95% CI 2.01–4.83). Factors associated with returning to care were similar between men and women except for duration on ART and WHO clinical stage 4. Being in WHO clinical stage 4 and on ART for more than 6 months was significantly associated with returning to care among women (WHO stage 4 OR 2.14 95% CI 1.07–4.25; ≥6 months OR 2.10 95% CI 1.29–3.43) (Table 3).

Table 2.   Distribution of patient characteristics and factors associated with return-to-care (n = 812)*
Patient characteristicsnReturn to care (n)OR (crude)OR (adjusted)†
  1. *Patients who were wanted back identified from their last default case.

  2. †adjusted for sex, age, WHO clinical stage, duration on ART and tracing outcome.

  3. ‡Nine patients do not have WHO clinic stage.

Sex
 Male3632490.74 (0.55–1.01)0.56 (0.39–0.80)
 Female44933511
Age at ART initiation
 <1575632.65 (1.39–5.03)1.94 (0.95–3.97)
 15–3954035911
 ≥401971622.33 (1.55–3.50)2.35 (1.51–3.64)
WHO stage at ART initiation‡
 WHO clinical stage 1 and 21541030.81 (0.55–1.19)0.91 (0.59–1.40)
 WHO clinical stage 347934211
 WHO clinical stage 41701351.55 (1.01–2.35)1.54 (0.98–2.42)
Duration on ART
 ≤6 months40025611
 >6 months4123282.20 (1.60–3.00)1.78 (1.25–2.51)
Tracing outcome
 ARV gaps11510810.1 (4.59–22.1)8.69 (3.92–19.29)
 Uninterrupted therapy2211883.72 (2.46–5.62)3.11 (2.01–4.83)
 Stopped/Never started ART47628811
Table 3.   Factors associated with return to care by gender
Patient characteristicsCrude OR (95% CI)Adjusted OR (95% CI)*
Female (n = 449)Male(n = 363)Female (n = 449)Male(n = 363)
  1. *adjusted for age, WHO clinical stage, duration on ART and tracing outcome.

Age at ART initiation
 <151.78 (0.76–4.17)4.59 (1.72–12.28)1.41 (0.54–3.71)2.68 (0.93–7.69)
 15–391111
 ≥402.49 (1.22–5.05)2.87 (1.71–4.83)2.34 (1.12–4.88)2.36 (1.36–4.11)
WHO stage at ART initiation
 WHO clinical stage 1 and 20.80 (0.49–1.31)0.61 (0.30–1.26)1.10 (0.65–1.88)0.68 (0.31–1.49)
 WHO clinical stage 31111
 WHO clinical stage 41.90 (0.99–3.67)1.33 (0.76–2.32)2.14 (1.07–4.25)1.17 (0.64–2.15)
Duration on ART
 ≤6 months1111
 >6 months2.53 (1.62–3.94)1.93 (1.23–3.02)2.10 (1.29–3.43)1.50 (0.91–2.48)
Tracing outcome
 ARV gaps7.43 (2.61–21.13)14.57 (4.41–48.11)6.87 (2.38–19.82)11.2 (3.33–37.66)
 Uninterrupted therapy3.58 (2.00–6.41)3.98 (2.21–7.17)3.19 (1.71–5.92)3.18 (1.70–5.97)
 Stopped/never started ART1111

The B2C project reduced the proportion of patients finally classified as ‘LTFU’ by 59%. Without the intervention, 3245 (23%) of 13 960 registered would have been classified as loss to follow-up while, the B2C programme reduced this number to 1344 (10%) (Figure 2).

image

Figure 2.  Comparison of current ART outcomes at Lighthouse and MPC with and without the B2C programme.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

This is one of the largest and most detailed analyses of loss to follow-up of patients on ART in SSA. Our Back-to-Care (B2C) programme resulted in a marked increase in ascertainment of ART patient outcomes and identified reasons for missing appointments that are rarely considered by the ART programme and that have not been previously quantified. While death was the most common reason for missing scheduled visits, we found that 21% of cases had transferred to other clinics and that 51% of LTFU patients found alive returned for continuation of care, therefore markedly improving overall retention in the treatment programme. Patients most likely to return to care through this programme were women, those who had been on treatment for more than 6 months, patients aged 40 or older at ART initiation and patients with treatment gaps compared to those who had stopped/never started ART.

As in other studies that evaluated reasons for loss to follow-up (Rosen et al. 2007; Yu et al. 2007), we found that mortality was a common reason for failure to return to the clinic. The Malawi ART programme defines ‘loss to follow-up’ as failure to return to the clinic within 2 months of running out of ARVs, and active follow-up is usually only initiated after a minimum of this period. This strategy primarily serves to ascertain patient survival status for programme monitoring. The B2C programme at Lighthouse and MPC initiated active follow-up when patients were 3 weeks overdue for their appointment to bring patients back into care and to prevent prolonged treatment interruptions and adverse long-term outcomes. ART retention rates improved considerably after implementing early active follow-up at Lighthouse and MPC, and the proportion of untraceable patients fell. It is likely that Malawi’s national ART programme outcomes could be improved using a similar strategy. We have demonstrated that this intervention can be successfully implemented within the constraints of the public health services in Malawi.

Given that Malawi’s ART programme has seen a rapid expansion and roll-out to new clinics, we saw a large number of both official- and self-transfers to other ART clinics. While it is not surprising that patients would choose clinics close by, the fact that the majority of these transfers was not documented resulted in additional work for the B2C team. Our findings suggest that transfers may be under-represented or that duplicate counting of patients may be more common than the current estimate of 10% in the national cohort analyses (HIV Unit 2008). We found that many of the patients LTFU collected their drugs from alternative sources, including sharing with family members and/or friends. This is undoubtedly another consequence of increased access to ART. It is unclear how this sharing behaviour adversely affects patient outcomes or whether this is a warning sign of potential abuse of drugs.

For every 100 patients on ART, we had to follow-up about 9 LTFU patients per year, and about 19% of our patients missed at least one appointment during the 3-year study period in spite of providing routine pre-treatment counselling and six monthly counselling education sessions. We did not collect data on the underlying reasons for missing appointments, and it will be important to conduct further qualitative studies to understand and address these problems. Despite the pre-treatment counselling sessions about the benefits and implication of ART, some patients did not start treatment although they had collected their drugs; it may be that the complexity of the information given during counselling daunted some patients.

Our results showed an independent association between age and returning to care. Older patients (≥40 years of age) were more likely to return to the clinic compared to younger patients aged between 15 and 39. Older patients may have more settled lifestyles into which the management of ARVs could be more easily incorporated. In agreement with Vreeman et al. (2009), we found that children were more likely to return to the clinic, possibly because of the practice of only initiating therapy if they had a stable guardian. Duration on ART and WHO clinical stage 4 at ART initiation was associated with returning to the clinic among women only, similar to other studies (Mocroft et al. 2001). Patients who had ARV gaps and uninterrupted therapy were also more likely to return to the clinic. While we have no data on motivation and barriers to adhere to appointments, it is possible that these groups, while understanding of the importance of ART adherence, could source their drugs from friends, relatives or unlicensed vendors and thus did not see the need to keep their scheduled appointments. Moreover, patients with uninterrupted therapy were less likely to return to care than patients with ARV gaps because they might have more stable drug sources than those with ARV gaps. An additional limitation is lack of wide range of baseline characteristics of patients (e.g. education), thus limiting analysis of factors associated with the probability of returning to the clinic.

The B2C project reduced the proportion of patients finally classified as ‘LTFU’ by 59%. WHO has established early warning indicators for the emergence of drug resistance that include maintaining loss-to-follow-up rates of <20% (Bennett et al. 2008). In the absence of the B2C programme, our clinics would not meet these criteria. However, with the programme, our loss-to-follow-up rate fell to 10%, suggesting that the B2C strategy may help to reduce the risk of HIV drug resistance.

Despite our intensive efforts, 26% of the LTFU patients could not be traced because of incorrect addresses or insufficient information on patient residence. While it is possible that some patients would have given incorrect addresses, there is generally high mobility of people in urban areas, and patients’ details were not routinely checked or updated at clinic visits. About a quarter of the successfully traced patients had transferred out to new ART clinics but had no documents at their previous ART clinic. To avoid tracing patients who had officially transferred out, transfer information should have been documented clearly in patient files and registers by clinic staff at the time of transferring out. The B2C programme maintained relatively few staff, but the required physical tracing for many clients likely increased overall programme costs, and strategies to minimize unnecessary tracing are critical considerations in evaluating the cost-effectiveness of such programmes.

In conclusion, our study shows that early active follow-up of patients can improve retention in treatment and programme outcomes. ART programmes should invest into obtaining accurate, complete and up-to-date patients’ addresses to reduce the numbers of patients LTFU after tracing. At national level, there is a need to review the methods for monitoring ART clinic transfers to minimize double-counting as the current procedures are likely to misclassify a considerable proportion of patients who transfer out as LTFU.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

We are grateful to all the staff at Lighthouse and Martin Preuss Center who worked with the B2C team to collect data. We thank numerous donors supporting the Lighthouse and the Malawi national ART program. HT is supported as an operational research fellow by the International Union Against Tuberculosis and Lung Disease, Paris, France.

Conflicts of interest

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

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

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References
  • Adam BD, Maticka-Tyndale E & Cohen JJ (2003) Adherence practices among people living with HIV. AIDS Care 15, 263274.
  • Bennett DE, Bertagnolio S, Sutherland D & Gilks CF (2008) The World Health Organization’s Global Strategy for Prevention and Assessment of HIV Drug Resistance. WHO, Geneva.
  • Bisson GP, Gaolathe T, Gross R et al. (2008) Overestimates of survival after HAART: implications for global scale-up efforts. PLoS ONE 3, e1725.
  • HIV Unit, MoH (2008) ARV Therapy in Malawi – up to 31st December 2008. Quarterly Report. MoH, Lilongwe.
  • Hogg RS, Heath K, Bangsberg D et al. (2002) Intermittent use of triple-combination therapy is predictive of mortality at baseline and after 1 year of follow-up. AIDS 16, 10511058.
  • Mocroft A, Youle M, Moore A et al. (2001) Reasons for modification and discontinuation of antiretrovirals: results from a single treatment centre. AIDS 15, 185194.
  • Rosen S, Fox MP & Gill CJ (2007) Patient retention in antiretroviral therapy programs in sub-Saharan Africa: a systematic review. PLoS Med 4, e298.
  • UNAIDS (2009) Report on the Global AIDS Epidemic. UNAIDS, Geneva.
  • Vreeman RC, Nyandiko WM, Sang E et al. (2009) Impact of the Kenya post-election crisis on clinic attendance and medication adherence for HIV-infected children in western Kenya. Confl Health 3, 5.
  • WHO (2006) Progress on Global Access to HIV Antiviral Therapy: Report on ‘3 by 5’ and Beyond WHO, Geneva.
  • Yu JK, Chen SC, Wang KY et al. (2007) True outcomes for patients on antiretroviral therapy who are “lost to follow-up” in Malawi. Bulletin of the World Health Organization 85, 550554.