Outcomes of patients loss to follow -up after antiretroviral therapy - - PDF document

1 Outcomes of patients’ loss to follow-up after antiretroviral therapy initiation in north- eastern South Africa Julie Ambia1, Chodziwadziwa Kabudula2, David Etoori1, F. Xavier Gomez-Olive2, Ryan Wagner2, Kathleen Kahn2, Georges Reniers1

1London School of Hygiene and Tropical Medicine 2MRC/Wits Agincourt, Johannesburg, South Africa

Introduction High rates of loss to follow-up (LTFU) have been reported in many HIV care and treatment programmes in sub-Saharan Africa. Perception of good health, use of alternative medicine, stigma, treatment fatigue, lack of knowledge, transport fare and competing demands for time are some of the factors that contribute to these high rates of disengagement [1]. However, some patients could have switched facilities, and still be adhering to ART [2, 3]. It is important to better understand the vital and treatment status of the LTFU patients both for improving the delivery of HIV services and for estimating the impact of ART on HIV associated mortality. For several years now, the SPECTRUM model has been used by UNAIDS for making most national and subnational estimates of HIV epidemic trends. ART coverage and the survival

• f patients on ART are key inputs on SPECTRUM that influence various components in the

model, including mortality, fertility and onwards transmission. Its inputs need to be carefully calibrated with empirical evidence [4]. Data on the survival of people living with HIV (PLHIV) on ART often come from clinical cohorts, and it was quickly understood that high LTFU rates could not be ignored. Several strategies have been employed to obtain information on vital status of patients LTFU, including the review of obituaries in newspapers and record linkage with Civil Registration and Vital Statistics databases; both with the intent to identify patients who deceased after being LTFU [5, 6]. The IeDEA Network of East Africa has resorted to tracing a random sample of patients LTFU patients by phone or home visits, which has the added advantage that data can be used for estimating undocumented transfers to other facilities providing treatment [3, 7, 8]. In this study, we use data from Agincourt Health Demographic Surveillance System (HDSS) [9] that was individually linked with clinic visit records of chronic care patients at eight

2 Primary Health Care facilities. The vital status (also migration status) of any patient LTFU can thus be ascertained via the HDSS. Moreover, the link with the Agincourt HDSS will allow us to identify patients who are LTFU at one facility and are subsequently registered at another facility. Study setting and population The AHDSS study area covers 475 km2 in Bushbuckridge, Mpumalanga province, north- eastern South Africa. As of 2014 the population was approximately 115,000 people living in 17,000 households spread in 31 villages [10]. Overall, 47.9% of the population are men and 52.1% are women [11]. Three fifths of the men aged between 30 and 54 years and 25% of the women aged between 25 and 49 years are regarded as labour migrant [12]. These labour migrants work in farms, mines and manufacturing industries outside the study site, returning home whenever possible. One third of the population are former Mozambican refugees who moved into the area during the civil war that began in 1977 and ended in 1992. Data sources Since 30th April 2014, records of HIV patients from eight facilities (7 government facilities and 1 public/private partnership) have been linked with their Agincourt HDSS records. Currently, clinic visit information on test dates and results of HIV diagnosis, CD4 counts, and viral loads have been linked with individual Agincourt HDSS records. Operational definitions A patient was considered LTFU if they did not return for more than 90 days after their scheduled appointment. Time on ART was measured from the date ART was initiated to the earliest of the date of either LTFU, transferring out to another facility, death or end of analysis period (12th August 2017). Time lost to follow-up was measured from the date of LTFU to the earliest date re-engagement, out-migration, death or end of analysis period (12th August 2017). Duplicate records were checked in the data set to identify those who had transferred to another health facility (both self and documented transfers). Migration was considered if an individual was reported to have left the Agincourt HDSS study area. Re- engagement was defined as the resumption of clinic visits in the same clinic or another clinic.

3 Statistical methods With LTFU defined as more than 90 days late for a scheduled appointment, we used competing risks survival analysis to estimate the cumulative incidence of (i) LTFU, (ii) transferring out to another facility and (iii) death after initiating ART. The complement of these three probabilities represents the proportion of patients that are still living in the Agincourt HDSS and taking ART. We used competing risks survival analysis to also estimate the cumulative incidence of (i) re-engagement, (ii) death and (iii) migration out of Agincourt HDSS after LTFU. The complement of these three probabilities represents the proportion of patients that are LTFU but still alive, living in the study site, and not on treatment elsewhere. We used clinical data from the eight facilities to calculate naïve estimates of mortality among PLHIV who start ART, and to adjust the mortality estimates after incorporating LTFU

• utcomes from linked datasets.

We used Cox regression to estimate cause-specific hazard ratios for LTFU. To model the hazard of LTFU after starting ART, patients were right censored at the time of death or transferring out. Covariates with evidence of association were included in the adjusted

• model. Schoenfeld residuals were used to evaluate proportional hazards assumption.

4 Results Table 1. Characteristics of patients initiating ART in the Agincourt sub-district in Mpumalanga Province, South Africa

Patient characteristic Total n=1,138 % Patient category Men 242 21.2 Women (non-pregnant) 704 61.9 Women (pregnant or breastfeeding) 192 16.9 Nationality South African 752 66.0 Other nationalities 386 34.0 Age at ART/ care initiation (in years) <1-19 61 5.4 20-29 319 28.0 30-39 356 31.3 40-49 214 18.8 ≥50 187 16.5 Missing 1 Year of ART/care initiation 2014 132 11.6 2015 415 36.5 2016 548 48.1 2017 43 3.8 Baseline CD4 count <350 cells/µL 722 65.1 351-500 cells/µL 244 22.0 >501 cells/µL 143 12.9 Missing 29 WHO stage at ART initiation WHO I/II 1,008 91.1 WHO III/IV 99 8.9 Missing 31 Tuberculosis at ART initiation Yes 29 2.6 No 1,092 97.4 Missing 17

Baseline characteristics There were more women (78.8%) than men (21.2%) who initiated ART. In total, 66% were South Africans and 78.1% were between 20 and 49 years old. Median age was 39.4 years [IQR: 31.4-47.2 years] among men, 36.1 years [IQR: 28.5-47.1 years] among non-pregnant women and 27.4 years [IQR: 23.3-32.2 years] among pregnant women. The highest number

• f patients initiated ART in 2016 (48.1%) compared to 2015 (36.5%) and 2014 (11.6%).

Overall, 91.1% of the patients were in WHO stage I/II and 2.6% were co-infected with TB at ART initiation. The median follow-up time for these patients on treatment was 13.9 months

5 [IQR: 0.2-44.5 months]. The 1,138 patients have contributed a total of 1,369 person-years on ART. Outcomes of patients initiating ART Cumulative incidence of LTFU (n=456) was 25.8% (23.1%-28.5%) at one year and 37.0% (33.6%-40.4%) at two years following ART initiation (see Figure 1). The cumulative incidence of transferring out (n=12) was 0.9% (0.4%-2.5%) at one year after initiation of

• ART. The cumulative incidence of death (n=27) was 1.57% (0.93%-2.48%) at one year after

initiation of ART and 3.2% (2.04%-4.64%) at two years. Overall, 25 of 27 (93%) patients died within three months of their last clinic visit, median number of days was 37 [IQR: 20-58 days]. Figure 1. Cumulative incidence curves of outcomes of patients after ART initiation Outcomes of patients LTFU Among those LTFU, cumulative incidence of re-engaging (n=103) which was 24.4% (19.6%- 29.6%) at one year after being LTFU and 42.4% (33.9%-50.6%) at two years (see Figure 2). Cumulative incidence of out-migration (n=77) was 13.4% (9.6%-17.9%) at one year after being LTFU and 32.0% (24.0%-40.4%) at two years. Only one patient died and two patients never returned after initiating ART.

.1 .2 .3 .4 .5 .6 1 2 3 4 Years since initiating ART Lost to follow-up Dead TransferOut

Outcomes of patients initiating ART

6 Figure 2. Cumulative incidence curves of outcomes of patients LTFU Risk factors associated with loss to follow-up among ART patients Factors associated with LTFU among ART patients are presented in Table 2. In crude and adjusted analysis, being pregnant was associated with a higher risk of LTFU (HR: 2.30; 95% CI 1.80-2.93, AHR: 1.78; 95% CI 1.38-2.30). Being a man was also associated with increased hazard of LTFU in both crude and adjusted analyses (HR: 1.48; 95% CI 1.15-1.90, AHR: 1.55; 95% CI 1.20-2.00). Younger age was associated with higher incidence of LTFU, and those above age 50 had lower LTFU. Other individual attributes such as nationality or the WHO staging were not associated with LTFU. Evaluation of Schoenfeld residuals suggested no violation of the proportional hazards assumption for any of the variables included in the analyses.

.1 .2 .3 .4 .5 .6 1 2 3 4 Years lost to follow-up Reengaged Dead Migrated

Outcomes of patients LTFU

7 Table 2. Factors associated with loss to follow-up among ART patients

Crude HR (95% CI) P value Adjusted HRa (95% CI) P value Patient category Women (non-pregnant) 1 1 Pregnant women 2.29 (1.80-2.92) 0.001 1.77 (1.38-2.29) 0.001 Men 1.47 (1.14-1.89) 0.003 1.55 (1.20-2.00) 0.001 Nationality South African 1 Mozambican 0.92 (0.74-1.145) 0.464 Age at ART initiation <1-19 years 2.14 (1.47-3.10) 0.001 2.18 (1.49-3.17) 0.001 20-29 1.64 (1.28-2.11) 0.001 1.60 (1.24-2.06) 0.001 30-39 1 1 40-49 0.94 (0.68-1.28) 0.679 0.99 (0.72-1.36) 0.960 ≥50 years 0.47 (0.31-0.70) 0.001 0.55 (0.36-0.82) 0.004 Year of ART initiation 2014 0.68 (0.49-0.94) 0.018 0.68 (0.49-0.94) 0.020 2015 0.75 (0.60-0.94) 0.012 0.76 (0.60-0.96) 0.020 2016 1 1 2017

• 1.000
• 1.000

WHO stage at ART initiation WHO I/II 1 WHO III/IV 1.02 (0.71-1.47) 0.913

a>90 days late

No of subjects – 1,056 No of failures – 379 Time at risk – 1368.9

Conclusion The use of Agincourt HDSS data linked with individual clinical records provided a unique

• pportunity to investigate definitive outcomes of ART patients LTFU in the Agincourt sub-

district in South Africa. Most patients initiated ART at early stages having WHO stage I or II. Overall, 26% of the patients were LTFU within one-year and 24% of those patients had re- engaged at the same (or another) clinic within one year. As reported by other tracing studies conducted in Malawi [13] and South Africa [14], we also found that mortality was highest within three months of the last clinic visit. This may be a reflection of missed opportunities by health care workers to provide individualised care to the sickest patients during their scheduled appointments [3] to improve their outcomes. Thus, there is need to include a simple screening tool in ART programs to identify these patients and provide quality care. In this study, men, pregnant women, and younger age were at an increased risk of LTFU. These patients may benefit from active tracing and enhanced counselling. For instance,

8 these patients should be called after 24 hours if they have missed their scheduled appointment, home visits should be done if they have not returned to the clinic after one week.

9 References

• 1. Loeliger KB, Niccolai LM, Mtungwa LN, Moll A, Shenoi SV. Antiretroviral therapy initiation

and adherence in rural South Africa: community health workers' perspectives on barriers and

• facilitators. AIDS Care. 2016;28(8):982-93.
• 2. Wilkinson LS, Skordis-Worrall J, Ajose O, Ford N. Self-transfer and mortality amongst

adults lost to follow-up in ART programmes in low- and middle-income countries: systematic review and meta-analysis. Trop Med Int Health. 2015;20(3):365-79.

• 3. Geng EH, Odeny TA, Lyamuya RE, Nakiwogga-Muwanga A, Diero L, Bwana M, et al.

Estimation of Mortality among HIV-infected people on antiretroviral therapy treatment in east Africa: a sampling based approach in an observational, multisite, cohort study. Lancet HIV. 2015;2(3):e107-e16.

• 4. Mahy M, Lewden C, Brinkhof MW, Dabis F, Tassie JM, Souteyrand Y, et al. Derivation of

parameters used in Spectrum for eligibility for antiretroviral therapy and survival on antiretroviral therapy. Sex Transm Infect. 2010;86 Suppl 2:ii28-34.

• 5. Anglaret X, Toure S, Gourvellec G, Tchehy A, Zio L, Zaho M, et al. Impact of vital status

investigation procedures on estimates of survival in cohorts of HIV-infected patients from Sub-Saharan Africa. J Acquir Immune Defic Syndr. 2004;35(3):320-3.

• 6. Cornell M, Lessells R, Fox MP, Garone DB, Giddy J, Fenner L, et al. Mortality among

adults transferred and lost to follow-up from antiretroviral therapy programmes in South Africa: a multicenter cohort study. J Acquir Immune Defic Syndr. 2014;67(2):e67-75.

• 7. Geng EH, Odeny TA, Lyamuya R, Nakiwogga-Muwanga A, Diero L, Bwana M, et al.

Retention in Care and Patient-Reported Reasons for Undocumented Transfer or Stopping Care Among HIV-Infected Patients on Antiretroviral Therapy in Eastern Africa: Application of a Sampling-Based Approach. Clin Infect Dis. 2016;62(7):935-44.

• 8. Geng EH, Glidden DV, Bwana MB, Musinguzi N, Emenyonu N, Muyindike W, et al.

Retention in care and connection to care among HIV-infected patients on antiretroviral therapy in Africa: estimation via a sampling-based approach. PLoS One. 2011;6(7):e21797.

• 9. Kahn K, Collinson MA, Gomez-Olive FX, Mokoena O, Twine R, Mee P, et al. Profile:

Agincourt health and socio-demographic surveillance system. Int J Epidemiol. 2012;41(4):988-1001.

• 10. Kabudula CW, Houle B, Collinson MA, Kahn K, Tollman S, Clark S. Assessing Changes

in Household Socioeconomic Status in Rural South Africa, 2001–2013: A Distributional Analysis Using Household Asset Indicators. Social Indicators Research. 2016:1-27.

• 11. Agincourt HDSS fact sheet 2014. 2014.

10 12. Tollman SM, Kahn K, Sartorius B, Collinson MA, Clark SJ, Garenne ML. Implications

• f mortality transition for primary health care in rural South Africa: a population-based

surveillance study. Lancet. 2008;372(9642):893-901. 13. Yu JK, Chen SC, Wang KY, Chang CS, Makombe SD, Schouten EJ, et al. True

• utcomes for patients on antiretroviral therapy who are "lost to follow-up" in Malawi. Bull

World Health Organ. 2007;85(7):550-4. 14. Van Cutsem G, Ford N, Hildebrand K, Goemaere E, Mathee S, Abrahams M, et al. Correcting for Mortality Among Patients Lost to Follow Up on Antiretroviral Therapy in South Africa: A Cohort Analysis. PLoS One. 2011;6(2).