Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized © International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Internet: www.worldbank.org; Telephone: 202 473 1000 This work is a product of the staff of The World Bank with external contributions. Note that The World Bank does not necessarily own each component of the content included in this work. The World Bank therefore does not warrant that the use of the content contained in the work will not infringe on the rights of third parties. The risk of claims resulting from such infringement rests solely with you. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or other partner institutions or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 Unported licence (CC BY 3.0) http://creativecommons.org/licences/by/3.0. Under the Creative Commons Attribution license, you are free to copy, distribute and adapt this work, including for commercial purposes, under the following conditions: Attribution – Please cite the work as follows: Evaluation of the National Adherence Guidelines for Chronic Diseases in South Africa Using Routinely Collected Data. First Enrollment Report. Washington DC: World Bank. License: Creative Commons Attribution CC BY 3.0 Translations – If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in its translation. All queries on rights and licenses should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington DC, 20433, USA; fax: 202-522-2625; email: pubrights@worldbank.org. Evaluation of the National Adherence Guidelines for Chronic Diseases in South Africa Using Routinely Collected Data FIRST ENROLLMENT REPORT 31 OCTOBER 2016 World Bank co-authors: Marelize Görgens and Nicole Fraser-Hurt This page is for collation purposes iv TABLE OF CONTENTS 4.1 General inclusion/Exclusion criteria ............................................................................... 3 4.2 Cohort specific inclusion/exclusion criteria ................................................................ 3 6.1 Methods for identifying eligible individuals ................................................................. 6 6.2 Intervention Sites .................................................................................................................... 6 6.3 Control sites ............................................................................................................................... 7 6.4 Cohort Specific Enrollment .................................................................................................. 7 6.5 Enrollment process/Identification of cohorts .......................................................... 12 8.1 Timing of cohort initiation ................................................................................................ 15 8.2 Enrollment by cohort .......................................................................................................... 15 8.3 Enrollment over time .......................................................................................................... 17 8.4 Enrollment by facility/district ........................................................................................ 19 8.5 Ineligible Patients ................................................................................................................. 20 v FIGURES 1 Screening and enrollment by cohort over time compared to the target total through October 6th (lines for HIV cohorts have been updated to reflect 20% increased enrollment target) .........................................................................18 2 Enrollment over time by province and clinic ..........................................................................................19 3 Enrollment by District and Study Cohort ..................................................................................................20 4 Gantt Chart of Timeline for Project..............................................................................................................28 TABLES 1 Population data (facility headcount and total active patients) at each facility and total numbers eligible by intervention (I) and control (C) for each intervention ........................................................................................................................................... 5 2 Assumptions and sample sizes for each cohort ......................................................................................14 3 Timing of cohort initiation by site and cohort ........................................................................................15 4 Enrollment by cohort as of October 6th ......................................................................................................16 5 Ineligible subjects by cohort and reason for exclusion .......................................................................20 6 Baseline characteristics of the cohorts ......................................................................................................22 7 Adherence clubs...................................................................................................................................................22 8 Decentralized medicine delivery ..................................................................................................................23 9 Enhanced adherence counseling ..................................................................................................................23 10 Tracing and retention in care .........................................................................................................................24 11 TB, hypertension, and diabetes .....................................................................................................................25 12 Study evaluation outcomes for protocol 1 ................................................................................................26 13 Short-term outcomes (ART initiation within 30 days) for those eligible for FTIC cohort ............................................................................................................................................................27 vi ACRONYMS ART Antiretroviral treatment ARV Antiretroviral CBO Community based organization CRF Electronic case report form DMD Decentralized medicine delivery EAC Enhanced adherence counseling FSW Female sex worker FTIC Fast track initiation counseling HIV Human Immunodeficiency Virus NCD Non-communicable disease NGO Non-Governmental Organization KP Key Population MSM Men who have sex with men NDoH National Department of Health NHLS National Health Laboratory Services PHC Primary health care TB Tuberculosis TBHD TB, hypertension and diabetes TRIC Tracing and retention in care TROA Total remaining on ART vii This page is for collation purposes ACKNOWLEDGEMENTS This report was initially drafted by Matthew Fox of Boston University and Sophie Pascoe of the Health Economics and Epidemiology Office with guidance and input from Nicole Fraser-Hurt and Marelize Görgens of the World Bank. Data management and analysis was conducted by Matthew Fox, Sophie Pascoe and Amy Huber. Nicole Fraser-Hurt provided substantive comments which led to revision of the report. ix This page is for collation purposes EXECUTIVE SUMMARY This report describes enrollment into the cohorts for protocol 1 for the Evaluation of the National Adherence Guidelines for Chronic Diseases in South Africa Using Routinely Collected Data. The study is evaluating short-term and long-term effects of five interventions being implemented by the National Department of Health (NDoH) in South Africa to improve adherence to HIV care and chronic disease care in general: Fast track initiation counselling, decentralized medicine delivery, adherence clubs, early patient tracing and enhanced adherence counselling. The study uses a randomized evaluation design to compare sites where the intervention was rolled out with sites where it was not. After completing data enhancement activities at each of the 24 sites (12 intervention and 12 control sites in 4 provinces in South Africa), the teams began enrollment into 7 study cohorts (5 HIV intervention evaluation cohorts and 2 TB, hypertension and diabetes (TBHD) observational cohorts). Enrollment began on 20 June 2016 and is on track to complete by end of October 2016 for the HIV cohorts while enrollment for the TBHD cohorts is ongoing (with expected completion by December 2016). The team faced several challenges to completing enrollment, including delays in the rollout of the interventions within the sites and deviations from the interventions described in the guidelines, inability to access registers of patients who received the interventions, variation in the ways the interventions were implemented, implementation of decentralized medicine delivery in control sites and changes to the interventions caused by a concurrent NDoH strategy to “decant” stable patients from HIV care sites. The team adapted to these conditions by working with the World Bank and NDoH co- principal investigators and developing strategies to ensure data collection could commence. These included changing the timing of enrollment, working with NDoH and the implementing partners to ensure rollout of the interventions and adapting the analytic approach to deal with the changes to the counterfactual created by the rollout of decentralized medicine delivery in control sites. Once enrollment began it was fairly smooth for all cohorts with the exception of adherence clubs and early tracing which took longer for sites to rollout. Enrollment was uneven in time and by province, but eventually all sites were on track to complete enrolling the appropriate number of subjects. Enrollment is now completing for the HIV cohorts and passive follow up has now begun for patients through monitoring data collection systems (TIER.net and NHLS) and patient files for study outcomes as work shifts to focus on protocol 2 where qualitative data will be collected to supplement the quantitative analyses that will result from protocol 1. xi This page is for collation purposes BACKGROUND For antiretroviral therapy (ART) for HIV and treatments for other chronic diseases to be effective, patients must remain in care for longer periods of time, initiate treatment as early as allowed under prevailing guidelines, consistently achieve high levels of adherence to their treatment regimen and, as a result, exhibit low and stable monitoring test results and/or treatment completion. In the case of HIV, treatment is lifelong and requires consistent, nearly complete adherence to sustain an undetectable viral load. Numerous studies and reviews have indicated that retention in care and adherence to ART in South Africa are sub-optimal and pose a serious threat to the long-term success of the national HIV response. Although there is less evidence on hand, these same problems almost certainly also pertain to tuberculosis (TB), for which treatment completion and cure rates do not approach global targets and to non-communicable diseases (NCDs), for which almost no treatment adherence data are available. To address this challenge, in 2014 the NDOH developed the “National Adherence Guidelines for Chronic Diseases (HIV, TB and NCDs).” The guidelines address the provision of a minimum package of interventions to increase linkage to care, retention in care, and adherence to treatment. The minimum package of interventions includes five interventions that are being evaluated under this study: 1) fast track initiation counseling (FTIC); 2) enhanced adherence counseling for unstable patients (EAC); 3) adherence clubs (AC); 4) decentralized medicine delivery (DMD); and 5) early tracing of all missed appointments to improve retention in care (TRIC). The study is a matched cluster randomized study in 24 clinics, 12 of which receive early implementation of the minimum package and 12 delay implementation and serve as control sites. Clinics were matched on clinic characteristics: total remaining on ART, clinic size, setting, location and viral suppression. STUDY AIM AND OBJECTIVES The overall aims of this study are to assess the impact of a subset of the National Adherence Guidelines’ minimum package of interventions on HIV patients’ treatment outcomes at public sector clinics; estimate the costs of the interventions; and describe the cascade of care for TB, hypertension, and diabetes at these same clinics. The study has 8 specific aims which are detailed in the two study protocols entitled “Evaluation of the National Department of Health's National Adherence Guidelines for Chronic Diseases in South Africa Using Routinely Collected Data” and “Process Evaluation of the National Department of Health's National Adherence Guidelines for Chronic Diseases in South Africa”. The objectives are: 1 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA ► Among HIV-infected patients newly eligible for antiretroviral therapy, evaluate the impact of Fast Track Treatment Initiation Counselling on initiation and viral suppression. ► Among HIV-infected patients who are stable on antiretroviral therapy, evaluate the impact of Adherence Clubs on ART adherence and viral suppression. ► Among HIV-infected patients who are stable on antiretroviral therapy, evaluate the impact of Decentralized Medicine delivery on ART adherence and viral suppression. ► Among HIV-infected patients who have poor adherence to antiretroviral therapy, evaluate the impact of Enhanced Adherence Counselling on treatment adherence and viral suppression. ► Among HIV-infected patients in antiretroviral therapy programs who miss a scheduled appointment by 5 days or more, evaluate the impact of Early Patient Tracing on retention in care. ► For each clinic included in the study, evaluate the overall impact of the Adherence Guidelines on patient outcomes. ► Estimate the incremental and total cost of each of the interventions listed above compared to standard of care. ► Describe the current status of the cascade of care and adherence to treatment for tuberculosis, hypertension, and diabetes, for the purpose of tailoring the minimum package of interventions to these conditions in the national rollout. PURPOSE OF THE REPORT The purpose of this report is to describe the process and current progress with enrollment for protocol 1 which is on track to be completed for the HIV cohorts by end of October and is ongoing for the TB, Hypertension and Diabetes cohorts. As part of this report we will describe the data sources used to collect baseline information on the cohorts and give an overview of the study sites based on the data collected at the sites. We will then review the eligibility criteria for each of the cohorts and describe the methodology used to identify eligible subjects for each cohort. Finally, we will review the enrollment within each cohort and describe the cohort in terms of basic clinical and demographic characteristics of each. 2 FIRST ENROLLMENT REPORT ELIGIBILITY CRITERIA For all cohorts, we set up general inclusion criteria designed to focus the study on non-pregnant adults. All cohorts followed these criteria, including the TBHD cohorts. Below we describe the general inclusion and exclusion criteria which applied to all cohorts within the study. 4.1 General inclusion/Exclusion criteria Inclusion criteria ► ≥ 18 years old ► Meet the inclusion criteria for one or more of the cohorts Exclusion criteria ► Not resident in the facility’s catchment area ► Recorded intention to transfer care to a different facility within 12 months ► Pregnant and eligible for PMTCT 4.2 Cohort specific inclusion/exclusion criteria Within each cohort, specific inclusion/exclusion criteria were set to allow enrollment of appropriate subjects. The enrollment criteria followed the December 2014 national guidelines for HIV care and ART and November 2015 National Adherence Guidelines for Chronic Disease (HIV, TB and NCDs) with some updates based on the August 2016 version of the Guidelines. Below we describe inclusion/exclusion criteria applied to each cohort. Cohort 1: Fast track treatment initiation inclusion criteria (Patients newly eligible for ART) ► Determined to be eligible to start ART under prevailing national guidelines (CD4 count < 500, WHO Stage 3 or 4 condition, no TB or cryptococcal meningitis) Cohorts 2 and 3: Adherence clubs or decentralized medicine delivery (Patients stable on ART) ► On same ART treatment regimen for at least 12 months ► Most recent viral load taken in past 3 months ► Two consecutive viral loads undetectable (<400 copies/ml3) 3 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Cohort 4: Enhanced adherence counseling (Patients on ART with poor adherence) ► On first line ART for at least 3 months ► Poor adherence as indicated by an elevated viral load (>400 copies/ml3) Cohort 5: Early patient tracing (Patients lost from ART programs) ► Initiated ART ► Fail to return for a scheduled appointment within 5 to 90 days from their appointment date Cohort 6: TBHD cohort (Tuberculosis, hypertension, and diabetes patients) ► Screened for or diagnosed with TB or hypertension within the last month, or diabetes within last 6 months ► Does not have existing TB, hypertension (diagnosed more than 1 month ago) or diabetes (diagnosed more than 6 months ago) In the section below on exclusions we describe the reasons, for each cohort, that patients did not meet the eligibility criteria. OVERVIEW OF STUDY SITES The study is being conducted at 24 primary health care clinics (PHCs) in South Africa. In consultation with the NDoH, the study team chose six clinics from one district each in Gauteng, KwaZulu Natal, Limpopo, and North West Provinces giving a total of 24 sites. These provinces were chosen because in most cases they are high HIV burden provinces with high burden districts and high volume clinics. Each site was required to meet the following criteria: 1) It had to be a high volume site as defined by having its Total Remaining on ART (TROA) being above 1000 patients; 2) It could not already be a National Health Insurance pilot site; 3) It had to be generating computerized TIER.net Phase 2 (all HIV patient data captured electronically including back capture of historical data) which is essential for the evaluation outcomes; and 4) It could not be participating in any other adherence-related studies or pilots. Sites were chosen so that they could be matched with another site within a district roughly on TROA, proportion of patients virally suppressed, setting (rural/urban/formal/informal) and location (sites near each other). After choosing the 12 pairs of sites (24 total), NDoH, World Bank, Boston University and HE2RO randomly allocated one site in each pair to be an intervention site and one pair to be a control site. Table 1 below describes the study populations at each facility using data from TIER.net and DHIS on patients who were active in care (i.e. had visited the clinic within the last four months) and who 4 FIRST ENROLLMENT REPORT would have been eligible for enrollment by virtue of having met our eligibility criteria. Catchment population and PHC headcount are from DHIS while the eligibility data come from TIER.net data. Table 1 Population data (facility headcount and total active patients) at each facility and total numbers eligible by intervention (I) and control (C) for each intervention 2015 Total PHC Total active FTIC AC/DMD catchment headcount, on ART on eligible eligible EAC eligible TRIC eligible population monthly 2015 30 June 2016 June 1-30 30 June 2016 June 1-30 30 June2016 Facility average N N N (%)* N (%) N (%) EKURHULENI I: Motsamai Clinic 12681 2701 1563 43 486 (31%) 53 (3%) 86 (6%) C: Tamaho Clinic 18686 3980 1741 18 366 (21%) 8 (0%) 374 (21%) I: Phola Park CHC 70047 14920 3434 17 1422 (41%) 97 (3%) 259 (8%) C: Ramokonopi CHC 76257 16242 3502 63 999 (29%) 146 (4%) 190 (5%) I: Khumalo Clinic 26255 5592 2587 22 808 (31%) 66 (3%) 185 (7%) C: Zonkizizwe 1 Clinic 21856 4655 1614 35 528 (33%) 55 (3%) 121 (7%) EKURHULENI TOTAL 225782 48090 14441 198 4609 (32%) 425 (3%) 1215 (8%) MOPANI I: Grace Mugodeni CHC 22533 5 212 2061 15 982 (48%) 31 (2%) 158 (8%) C: Motupa Clinic 17263 4 511 1708 23 546 (32%) 17 (1%) 204 (12%) I: Giyani CHC 25982 7 149 2184 30 621 (28%) 60 (3%) 363 (17%) C: Dzumeri Clinic 23644 5 753 1597 10 598 (37%) 16 (1%) 253 (16%) I: Tzaneen Clinic 17258 4265 1851 19 700 (38%) 2 (0%) 334 (18%) C: Nkowankowa CHC 22629 4 881 1132 10 233 (21%) 6 (1%) 97 (9%) MOPANI TOTAL 129309 31771 10533 107 3680 (35%) 132 (1%) 1409 (13%) BOJANALA I: Letlhabile CHC 69554 9825 3855 43 1729 (45%) 31 (1%) 487 (13%) C: Wonderkop Clinic 21236 3200 1848 49 875 (47%) 28 (2%) 272 (15%) I: Hebron Clinic 32352 4301 1714 29 870 (51%) 34 (2%) 130 (8%) C: Majakaneng Clinic 21539 3259 1431 24 751 (52%) 12 (1%) 136 (10%) I: Tlhabane CHC 79290 15130 5202 86 1330 (26%) 50 (1%) 667 (13%) C: Bafokeng CHC 62149 10810 3113 37 1252 (40%) 36 (1%) 520 (17%) BOJANALA TOTAL 286120 46525 17163 268 6807 (40%) 191 (1%) 2212 (13%) UTHUNGULU I: King Dinizulu Clinic 24456 6058 2528 28 1152 (46%) 22 (1%) 420 (17%) C: Nkwalini Clinic 10434 2573 1089 11 609 (56%) 11 (1%) 232 (21%) I: Thokozani Clinic 42678 10657 3875 65 220 (6%) 15 (0%) 722 (19%) C: Nseleni CHC 71060 18273 6218 83 2118 (34%) 6 (0%) 1021 (16%) I: Buchanana Clinic 21944 3574 1323 12 722 (55%) 10 (1%) 239 (18%) C: Ntambanana Clinic 19103 3323 1416 17 769 (54%) 8 (1%) 99 (7%) UTHUNGULU TOTAL 189675 44458 16449 216 5590 (34%) 72 (0%) 2733 (17%) TOTAL 830 886 170 844 58586 789 20686 (35%) 820 (1%) 7569 (13%) Note: *Percent of total active on ART, 30 June 2016. As can be seen from the table above, the sites are very different in terms of geographic location, size and number of eligible subjects within each of the intervention cohorts. The number of subjects eligible for each intervention between June 1-30th of 2016 shows that for most cohorts and sites, a sufficient number of subjects become eligible to allow site specific enrollment to complete within a four month period though variation exists. For cohorts like ACs and DMD, the numbers eligible in June 2016 are quite large, allowing full enrollment to be completed quite quickly if required, while 5 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA the EAC cohort shows few new eligible subjects, particularly in uThungulu and parts of Mopani and Ekurhuleni. ENROLLMENT 6.1 Methods for identifying eligible individuals In both intervention and control sites, the study team needed to identify patients eligible for enrollment into the study. In control sites, this necessitated finding subjects who would have been eligible for the interventions if they had been at an intervention site. This could be achieved using TIER.net data. At intervention sites, this necessitated finding patients who actually got the interventions. In some case, registers were able to show us who got the interventions and who did not within an intervention site. In other cases we had to develop alternative methods to identify the cohorts. An exception to this was the enrollment in control sites for the DMD cohorts and the adherence clubs. As control sites needed to identify a group of patients who were eligible for these interventions and because the eligibility criteria for these two cohorts were the same, we needed to ensure we did not enrol the same patients into two cohorts at a control site and we needed to ensure we had roughly balanced patient characteristics within each cohort. To accomplish this, we identified eligible patients at the sites and randomly allocated them to one of the two cohorts to attempt to create the balance on predictors of the outcomes. 6.2 Intervention Sites At intervention sites identifying individuals who were eligible for specific interventions and received the interventions was achieved using one of three approaches described below. The aim was, as far as possible, to spread enrollment into each cohort over the enrollment period (June to October). Recent TIER.Net dispatches from each facility were used to produce lists of patients who met eligibility criteria for each specific HIV cohort. These lists were produced for each cohort prior to each facility visit during the enrollment period. If the list of eligible patients was long, patients were randomly selected from each list to produce a shortened list of eligible individuals. If there were only a few eligible individuals on the initial list, all patients remained on the list. The final password protected ‘Cohort creation list’ was sent to the Provincial study coordinators by the Data Manager along with targets for visit enrollment for each cohort. Then one of three approaches was used: 1. List to Register approach. At intervention sites, cohort creation lists were compared against intervention specific registers (e.g., FTIC register, AC register, etc.) to identify which 6 FIRST ENROLLMENT REPORT patients had received the intervention. Provincial teams checked each patient on the cohort creation list until they identified the target number of cohort patients to be enrolled. 2. List to File approach. In some cases where intervention registers were absent or incomplete, cohort creation lists were used to find patient files to find evidence that the patient had received the intervention e.g. record of specific AGL counselling sessions, presence of a patient adherence plan, etc. This was the approach which was used in later rounds of recruitment and was mainly used to complete enrollment of FTIC and EAC cohorts, predominantly in Ekurhuleni. 3. Direct from Register approach. If patients on the cohort creation list could not be found using the List to Register approach then (rather than using the cohort creation lists to identify eligible patients), patients were instead identified directly from the intervention specific registers and their file numbers noted and found. Regardless of the method by which eligible patients receiving the intervention were identified, patient files were reviewed and information extracted using the electronic CRF to confirm that these patients did in fact meet all eligibility criteria for that intervention cohort (including receipt of the intervention). 6.3 Control sites At controls sites, cohort creation lists of eligible individuals were created in the same way as described above from TIER.net data and sent to the provincial teams. Each cohort list was then reviewed and files for each patient were found until the cohort enrollment target was achieved (i.e., all patients enrolled in the cohorts at control sites were found using the List to File approach). Files were then reviewed and information extracted using the electronic case report form (CRF) to confirm eligibility. 6.4 Cohort Specific Enrollment Below we describe the approach to identifying eligible patients for each of the cohorts. For all cohorts at both intervention and control sites we first produced cohort lists from the most recent TIER dispatch for each facility identifying patients who were: ► Active in care (visit within last 124 days and no record that patient had “died”, “lost to follow up” or “transferred/moved out” ► Over 18 years of age ► Pregnancy status at last visit did not equal “Pregnant” We then followed specific procedures for each cohort as described below. 7 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Cohort 1: Fast track treatment initiation (Patients newly eligible for ART) Cohort lists were refined using TIER.net to select patients who met the general inclusion criteria. We then proceeded differently for intervention and control subjects. Control subjects were identified and enrolled using the ‘List to file’ approach. Patient files for eligible patients were found and screened using the electronic CRF to ensure FTIC eligibility criteria were met. As with intervention subjects, once eligibility had been confirmed an FTIC barcode was assigned and the patient enrolled. For intervention subjects, we first used a ‘List to register’ approach then a ‘Direct from register’ then the ‘List to file’ approach. In some facilities (predominantly in Ekurhuleni) FTIC registers were absent or inadequately completed, hence a ‘List to file’ approach was used. Eligible files were reviewed for evidence of FTIC counselling sessions recorded on clinical stationery and/or presence of a completed or partially completed AGL Patient Adherence Plan on the patient file. All selected files were then screened using the electronic CRF to confirm patients’ FTIC eligibility criteria that could not be verified through TIER.net (i.e., resident in the facility catchment area, no intention to transfer to another facility, no cryptococcal meningitis diagnosis at ART initiation). All patients who screened as eligible for the FTIC cohort and as receiving the intervention were assigned an FTIC barcode and followed. In total 79% of eligible patients enrolled in the FTIC cohort were identified through either the ‘List to register’ or ‘List to file’ approach, while 21% were identified direct from register. Cohorts 2: Adherence clubs (Patients stable on ART) Cohort lists were refined using TIER.net data to select patients who met the general inclusion criteria. Control subjects were identified and enrolled using the ‘List to file’ approach and patient files screened to ensure all AC eligibility criteria were met. Given the implementation of DMD at some control sites, further checks were made to ensure that selected patients were not currently receiving their medication through DMD. If these criteria were met the patient was enrolled and an AC barcode assigned. For intervention subjects, first a ‘List to register’ approach was used to identify patients recorded on the facility AC register(s). If the AC target number could not be enrolled then the ‘Direct from register’ approach was used to identify and find AC patient files. A ‘List to file’ approach was not used for this cohort. Eligible patient files were screened using the electronic CRF to confirm patients’ eligibility including confirmation of enrollment in an AC, and ensuring that patients had no record of picking up their medications separately from a DMD pick-up-point. All patients who eligible for the AC cohort and as receiving the intervention were assigned an AC barcode. In total 76% of eligible patients enrolled in the AC cohort were identified through the ‘List to register’ approach and 24% were identified direct from register. 8 FIRST ENROLLMENT REPORT Cohorts 3: Decentralized medicine delivery (Patients stable on ART) For DMD, cohort lists were further refined as per AC, although it was not possible to confirm enrollment into DMD using TIER.net data. Control DMD subjects were enrolled from any site that was not implementing DMD during the enrollment period (June-October) irrespective of randomization. Control subjects were identified and enrolled using the ‘List to file’ approach. Patient files were found and screened to ensure DMD eligibility. Given the enrollment of control DMD subjects at some of the intervention sites further checks were made to ensure that selected cohort patients were not enrolled in an AC. Assuming all criteria were met the patient was enrolled and a DMD barcode assigned. For intervention subjects (enrolled at any site currently implementing DMD irrespective of randomization), the ‘List to register’ approach was used to identify patients recorded on DMD register(s) as having been decanted to a DMD (CCMDD/CDU) medicine pick up point. The relevant patient file numbers were recorded and those patient files found. If the DMD target number could not be enrolled then the ‘Direct from register’ approach was used and patient files found. As with AC, a ‘List to file’ approach was not used for this cohort at DMD intervention sites. Patient files were then screened using the electronic CRF to confirm patients’ eligibility including confirmation that the patient was not also enrolled in an AC. Eligible patients were enrolled and assigned a DMD barcode. In total 49% of eligible patients enrolled in the DMD cohort at DMD intervention facilities were identified through the ‘List to register’ approach; 51% were identified direct from register. As noted, DMD was unexpectedly implemented during the enrollment period at many sites. At those intervention sites where DMD was unexpectedly implemented during the enrollment period and patients were transferred by the facility from AC to DMD, any patient in the AC cohort who was transferred to DMD was then replaced in the AC cohort. At control sites where DMD was unexpectedly implemented during the enrollment period, we checked the IDs of all patients enrolled in the AC cohort and replaced any patient in that cohort who was now receiving DMD. Cohort 4: Enhanced adherence counseling (patients on ART with poor adherence) For EAC cohort lists were first refined using TIER.net. Control subjects were then identified and enrolled using only the ‘List to file’ approach. Once eligibility had been confirmed the patient was enrolled and an EAC barcode assigned. For intervention subjects, first a ‘List to register’ approach was used to identify patients on the facility EAC register(s) and to find those patient files. If the EAC target could not be enrolled then the ‘Direct from register’ approach was used. In some intervention facilities where EAC registers were absent or inadequately completed (predominantly Ekurhuleni) the ‘List to file’ approach was used. Patient files were found and reviewed to see if there was any evidence of EAC counselling sessions recorded and/or presence of a completed or partially completed or updated AGL Patient 9 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Adherence Plan on the patient file. All patients who screened as eligible for the EAC cohort were enrolled and assigned an EAC barcode. In total 51% of eligible patients enrolled in the EAC cohort were identified through either the ‘List to register’ or ‘List to file’ approach, while 49% were identified direct from register. Cohort 5: Early patient tracing (Patients lost from ART programs) For TRIC, cohort lists were again refined for this cohort using TIER. Control subjects were identified and enrolled using just the ‘List to file’ approach. Eligible files were screened using the electronic CRF to ensure that patients had in fact missed an appointment by 5-90 days as indicated on TIER (sometimes a backlog in TIER data entry could have resulted in patients being identified as having missed a visit when in fact they had not). Once eligibility was confirmed a TRIC barcode was assigned. For intervention subjects, initially a ‘List to register’ approach was used to identify patients from the TRIC registers who needed tracing or who had been traced. The relevant patient files were then found. If the TRIC target number could not be enrolled then the ‘Direct from register’ approach was used to identify TRIC eligible patients and their patient files. In some facilities (particularly Mopani and uThungulu), TRIC registers were not held by the facility but by implementing partners and/or WBOT teams. Information on these registers was not sent back to the facility or shared between stakeholders. This was discussed with both the facilities, district management and support partners and led to delays in enrolling this cohort at some intervention sites. In order to complete enrollment at these sites, visits are now being made to partner offices and meetings with WBOT teams organized in order to gain access to the registers and tracing information. This is similar to the ‘Direct from register’ approach although can require review of several different registers in order to identify patients. The names and file numbers (if recorded) of patients recorded on these lists are then noted and taken back to the facility by the team and these patient files found. The ‘List to file’ approach was not used at intervention sites for this cohort. Patient files were screened to confirm eligibility (including confirmation that they had missed a scheduled appointment by 5-90 days). Assuming patients screened as eligible they were enrolled and assigned a TRIC barcode. In total 46% of eligible patients enrolled in the AC cohort were identified through the ‘List to register’ approach while 54% were identified direct from register. Cohort 6: TBHD cohort (Tuberculosis, hypertension, and diabetes patients) TBHD screening cohort (Cohort 6A) At each facility visit the study team reviewed patient files from the previous or current day to exclude patients <18 years and pregnant women. All of those who remain are eligible for the TBHD screening cohort (Group A) and baseline data is extracted from their files using the electronic CRF. This is repeated until 100 patients eligible for Group A have been identified. 10 FIRST ENROLLMENT REPORT Once patient information is captured it is reviewed by the research team to determine the status of each patient: ► Not eligible (i.e., not screened for TB, hypertension or diabetes; no evidence of TB, hypertension or diabetes/no known disease/screened negative; prevalent case of TB or hypertension (diagnosed more than 1 month ago), prevalent case of diabetes (diagnosed more than 6 months ago), MDR or XDR TB) ► Screened positive for TB (at last visit) ► Screened positive for hypertension (at last visit) ► Screened positive for diabetes (at last visit) ► Newly diagnosed with TB (in last month) ► Newly diagnosed with hypertension (in last month) ► Newly diagnosed with diabetes (in last 6 months) The Data Manager then returns a list of file numbers of all patients who have screened positive for TBHD to the provincial teams and these files are found and marked with a barcode label to signify their enrollment in the TBHD screening cohort. TBHD diagnosed cohort (Cohort 6B) There are three ways in which patients are identified as eligible for the TBHD diagnosed cohort and enrolled: 1. Patients from the TBHD screening cohort who are diagnosed with TB, hypertension or diabetes during the TBHD cohort enrollment period (June-December 2016) are then eligible for inclusion in Cohort 6B. These file numbers are confirmed by the Data Manager and then be marked with a Cohort 6B (TBHD Diagnosed) barcode label. 2. Cases of TB and hypertension diagnosed within the last month of the site visit, or diabetes cases within the last 6 months of the site visit that were identified during the creation of Cohort 6A are also eligible. These are identified by the Data Manager from the baseline CRF data previously collected. A list of file numbers is then returned to the provincial teams so that patient files can be marked with a Cohort 6B barcode label. 3. TB patients (diagnosed after May 2016) are identified from the TB suspect register or the TB register, and hypertension (diagnosed after May 2016) and diabetes patients (diagnosed after Jan 2016) are identified from the PHC tick register. File numbers are noted and patient files pulled to verify these patients have been diagnosed within the required time period. If eligibility is confirmed files are marked with the Cohort 6B barcode label. 11 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA The process of identification of eligible patients for Cohort 6B will be repeated until Cohort 6B includes approximately: 35 newly diagnosed TB patients, 35 newly diagnosed hypertension patients and 35 newly diagnosed diabetic patients. 6.5 Enrollment process/Identification of cohorts While enrollment generally went smoothly after initiation of enrollment procedures began on June 20th, 2016 the team did identify some barriers to completing the process, and in some cases, starting the enrollment process. This was often cohort specific, with sites having difficulty enrolling into one or two of the cohorts, while the others were able to enrol without problem. Barriers for each cohort, therefore required different solutions. The main barriers to enrollment fall into four general categories: 1) delays in implementation at the sites and delays in implementation per specifications of the adherence guidelines; 2) changes in rollout of the intervention due to the pressure to “decant” patients; 3) incomplete or not used registers for the interventions; and 4) delay in enrollment at control sites due to delays in implementation of the interventions at intervention sites. Below we discuss each issue as it arose and what our strategies were for alleviating the problem. Delays in Implementation or not implemented as per AGL The main barrier to enrollment was delays in the implementation of the rollout of some of the interventions as per the AGL. The most commonly delayed interventions were decentralized medicine delivery and ACs. This largely occurred as it took some sites more time to determine which patients would be appropriate for each of the interventions and for the sites to decide how to organize these approaches. These mainly affected sites in Gauteng province, though this also was seen in uThungulu. Some of the delay in implementing ACs in uThungulu was also associated with existing implementation of pre-AGL models in the district and the need to change the approach to align with the adherence guidelines as new clubs were created. Due to delays at King Dinizulu we made the decision to enrol patients into the AC cohort from any club established after June 2016. In Mopani, delays implementing TRIC as per the guidelines has also been a barrier to cohort enrollment as TRIC has tended to focus on patients who have defaulted rather than those who have missed appointments. As the BU/HE2RO team are not service providers, we were not able to exert much influence on the timing of the rollout or supporting implementation of the interventions as per the adherence guidelines. Instead we kept in close contact with the treatment sites, the support partners and our contacts in government to discuss progress on the rollout. Ultimately all interventions were rolled out and enrollment was able to begin at all sites with a few exceptions. By mid-October, these include ACs in Motsamai and Buchanana and TRIC in Giyani and King Dinizulu. Changes in Implementation Another challenge the team faced around enrollment stemmed from changes made to the interventions within the sites. First, in many sites, pressure to decant as many patients as possible meant that patients who were initially designated to be in ACs were then moved to DMD and spaced 12 FIRST ENROLLMENT REPORT fast lane appointments (in Mopani). This was a challenge for the study team as we had to ensure that subjects enrolled in the intervention sites actually got the interventions for the cohort they were enrolled in. Second, at the same time as our enrollment was beginning there was a strong push to implement DMD in as many sites as possible, again in order to meet the targets for the government decanting strategy. This led to control sites being slated to rollout this intervention as well as the other decanting interventions (ACs and Spaced Fast Lane appointments) despite previous commitments to delay the rollout of other AGL interventions. To attempt solve this problem and minimize the delay in enrollment of the full cohort, the team worked with the World Bank and NDoH PIs but were not successful in delaying the implementation of the DMD intervention at control sites. Instead we agreed to investigate alternative analytic options for how to deal with this issue. These are currently under consideration. So far the evaluation team does appear to have managed to encourage delaying any further implementation of the other decanting strategies at control sites. Incomplete Registers for Interventions While in most cases, our data enhancement plans allowed us to work with sites to ensure improved record keeping and data recording, not all sites had the capacity to use the registers for documenting who got the interventions. This issue occurred most often with early patient tracing particularly in Mopani, as this intervention was largely managed by the implementing partners who kept their own records, often not at the treatment site. In Ekurhuleni, the shortage of Lay Counsellor staff resulted in inadequate completion of FTIC and EAC registers as counsellors were short staffed and felt they had insufficient time to complete the registers as well. Where patients who got the intervention could not be identified directly through registers at the sites, the study team worked closely with the treatment support partners to ensure access to the registers in order to identify a suitable intervention cohort. For example, we have worked with Anova, Casipo and WBOTs in Mopani to solve these issues. In Ekurhuleni we have tried to encourage register completion at each visit to sites where there is a problem with EAC or FTIC registers. Delay in control sites due to delays in intervention sites Our initial desire was to ensure that a matched pair of intervention and control sites enrolled roughly concurrently to limit the influence of any secular trends. These should be minimal in general but if possible it was our preferred approach and at most of the sites we were able to achieve this. At some sites however, delay in implementation of the interventions meant we delayed the enrollment at both the intervention and control site within the pair. Eventually we took the decision that we needed to continue with enrollment and did not wait until the interventions were completely rolled out (e.g. AC and EAC at Motsamai, Tamaho, Buchanana and Ntambanana). Thus at these sites we enrolled the control site first and then the intervention site once the intervention were rolled out. In addition while we had always intended not to enrol the initial group of patients who got the interventions (as we wanted to give the sites some “learning time”) this was not possible if the interventions were delayed and therefore, in discussions with the World 13 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Bank PI agreed to enrol the first subjects given the interventions. This includes AC at Motsamai in Ekurhuleni and Buchanana in KZN. REVIEW OF SAMPLE SIZE Total sample size for the study was estimated to be 2,880 subjects in the five intervention cohorts and 4,800 in the TBHD cohort, for a total sample size of 6,680. The table below shows the sample size that was estimated to be required for each cohort. We determined each sample size to be sufficient to measure the short term outcome for that cohort. For all the cohorts except the TBHD cohort, calculations assumed a site-clustered design with the clinic as the cluster and 24 clusters evenly randomized between intervention and comparison groups. They also assumed a coefficient of variation of 0.1, 80% power; and an alpha of 0.05. In table 2 below we describe the remaining assumptions behind the sample size for each cohort. Table 2 Assumptions and sample sizes for each cohort Cohort Sample Size % with the outcome in control arm Detection limit Fast Track ART Initiation Counseling 720 patients (30 per clinic) 60% of patients with initiate ART 15% change Adherence Clubs 576 patients (24 per clinic) 80% of patients will make all medication pickups 15% change Decentralized Medicine delivery 576 patients (24 per clinic) 80% of patients will make all medication pickups 15% change Enhanced Adherence Counseling 1008 patients 52% of patients with a detectable viral load will re- (42 per clinic) suppress after one session 15% change Early tracing of patients lost to follow 576 patients 30% of patients will be loss to follow up without up (24 per clinic) intervention 15% change TB, hypertension, and diabetes 4800 patients Descriptive in nature, no specific sample size calculation was done We note that we are currently in discussion to determine whether we should increase the overall sample size to account for ineligible subjects, changes at the sites and any other differences from anticipated data collection (i.e. subjects who got two interventions when they should have gotten only one). This will be particularly important for the DMD cohort where Central Chronic Medicine Dispensing and Distribution (CCMDD) has been rolled out at many facilities. It may also affect the ACs where enrollment is more challenging because patients who started on an AC may be pushed into CCMDD. The general consensus is that we should increase the maximum allowable sample size by 20% in each cohort. While we would not need to enroll the additional subjects, this would allow us the flexibility to do so if we find our number of ineligible subjects is higher than we anticipated or if too many subjects started in an intervention like ACs but then were moved to DMD. 14 FIRST ENROLLMENT REPORT ENROLLMENT 8.1 Timing of cohort initiation As noted, cohort enrollment began on June 20th, 2016. Table 3 below describes the timing of initiation of site assessments, and data collection for each cohort. Table 3 Timing of cohort initiation by site and cohort Cohort 6 Cohort 1 Cohort 2 Cohort 3 Cohort 4 Cohort 5 (TBHD): Site (FTIC): Data (AC): Data (DMD): Data (EAC): Data (TRIC): Data Data Assessment collection collection collection collection collection collection Facility date start date start date start date start date start date start date gp Motsamai Clinic 28-Oct-15 23-Jun-16 Pending 23-Jun-16 16-Aug-16 23-Jun-16 23-Jun-16 gp Tamaho Clinic 25-Nov-15 20-Jun-16 31-Aug-16 21-Jun-16 20-Jun-16 20-Jun-16 20-Jun-16 gp Phola Park CHC 10-Nov-15 25-Aug-16 20-Jul-16 20-Jul-16 26-Aug-16 15-Aug-16 20-Jul-16 gp Ramokonopi CHC 23-Nov-15 29-Aug-16 21-Jul-16 30-Aug-16 29-Aug-16 29-Aug-16 20-Jul-16 gp Khumalo Clinic 17-Nov-15 18-Aug-16 29-Jun-16 29-Jun-16 19-Aug-16 19-Aug-16 30-Jun-16 gp Zonkizizwe 1 Clinic 16-Nov-15 05-Aug-16 04-Jul-16 04-Jul-16 05-Aug-16 08-Aug-16 01-Jul-16 lp Grace Mugodeni CHC 16-Nov-15 24-Jun-16 27-Jun-16 27-Jun-16 13-Jul-16 Pending 27-Jun-16 lp Motupa Clinic 20-Nov-15 28-Jun-16 14-Jul-16 15-Jul-16 28-Jun-16 28-Sep-16 28-Jun-16 lp Giyani CHC 17-Nov-15 30-Jun-16 11-Aug-16 21-Jul-16 30-Jun-16 Pending 30-Jun-16 lp Dzumeri Clinic 14-Dec-15 04-Jul-16 11-Aug-16 05-Jul-16 05-Jul-16 04-Oct-16 04-Jul-16 lp Tzaneen Clinic 19-Nov-15 21-Jun-16 06-Jul-16 20-Jun-16 20-Jul-16 06-Oct-16 20-Jun-16 lp Nkowankowa CHC 18-Nov-15 22-Jun-16 22-Jun-16 22-Jun-16 08-Jul-16 23-Sep-16 22-Jun-16 nw Letlhabile CHC 24-Nov-15 28-Jun-16 28-Jun-16 17-Aug-16 29-Jun-16 30-Jun-16 28-Jun-16 nw Wonderkop Clinic 25-Nov-15 27-Jun-16 27-Jun-16 02-Sep-16 28-Jun-16 14-Jul-16 27-Jun-16 nw Hebron Clinic 25-Nov-15 20-Jun-16 21-Jun-16 26-Aug-16 20-Jun-16 10-Aug-16 20-Jun-16 nw Majakaneng Clinic 24-Nov-15 22-Jun-16 22-Jun-16 16-Aug-16 22-Jun-16 15-Jul-16 22-Jun-16 nw Tlhabane CHC 26-Nov-15 25-Jul-16 22-Aug-16 06-Sep-16 08-Aug-16 25-Jul-16 24-Jun-16 nw Bafokeng CHC 26-Nov-15 12-Jul-16 12-Jul-16 29-Aug-16 12-Jul-16 12-Jul-16 23-Jun-16 kz King Dinizulu Clinic 09-Dec-15 28-Jun-16 07-Sep-16 06-Sep-16 28-Jun-16 Pending 28-Jun-16 kz Nkwalini Clinic 24-Dec-15 30-Jun-16 22-Aug-16 26-Sep-16 21-Jul-16 30-Jun-16 30-Jun-16 kz Thokozani Clinic 27-Nov-15 04-Aug-16 17-Aug-16 17-Aug-16 21-Sep-16 14-Sep-16 04-Aug-16 kz Nseleni CHC 21-Dec-15 25-Aug-16 26-Aug-16 26-Aug-16 25-Aug-16 13-Oct-16 04-Aug-16 kz Buchanana Clinic 11-Dec-15 21-Jun-16 Pending 21-Jun-16 28-Jul-16 22-Jun-16 20-Jun-16 kz Ntambanana Clinic 10-Dec-15 23-Jun-16 31-Aug-16 23-Jun-16 24-Jun-16 27-Jun-16 23-Jun-16 Note: More than one month between enrollment at control and enrollment at intervention site 2-4 weeks between enrollment at matched pair 8.2 Enrollment by cohort The team has tracked enrollment into each cohort over time to ensure progress towards our sample size targets and to ensure we did not exceed our total sample size allowed. Table 4 below 15 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA shows the results of that tracking, demonstrating accrual into each cohort at each site through October 6th. The table is also stratified by treatment facility and district. Table 4 Enrollment by cohort as of October 6th Enrolled Enrolled Enrolled Enrolled Enrolled Enrolled Enrolled Cohort 1 Cohort 2 Cohort 3 Cohort 4 Cohort 5 Cohort 6a Cohort 6b Facility (FTIC) (AC) (DMD) (EAC) (TRIC) (TBHD) (TBHD) Total Target per facility 30 24 24 42 24 100 100 344 GAUTENG Motsamai Clinic 25 0 7 12 20 90 19 173 Tamaho Clinic 25 19 19 35 17 32 18 165 Phola Park CHC 25 20 20 10 19 45 37 176 Ramokonopi CHC 23 20 20 35 20 53 26 197 Khumalo Clinic 23 5 20 3 11 42 12 116 Zonkizizwe 1 Clinic 13 20 20 34 19 73 16 195 LIMPOPO Grace Mugodeni CHC 25 20 20 10 0 77 2 154 Motupa Clinic 25 11 12 27 10 100 3 188 Giyani CHC 25 20 20 33 0 100 2 200 Dzumeri 24 20 20 34 20 100 3 221 Tzaneen Clinic 21 20 20 15 10 100 2 188 Nkowankowa CHC 17 11 12 35 5 100 4 184 NORTH WEST Letlhabile CHC 25 20 20 35 20 10 15 145 Wonderkop Clinic 25 20 20 35 20 17 2 139 Hebron Clinic 25 20 20 33 20 74 17 209 Majakaneng Clinic 25 20 20 35 20 100 26 246 Tlhabane CHC 25 20 20 35 20 41 17 178 Bafokeng CHC 24 20 20 35 20 25 6 150 KWAZULU NATAL King Dinizulu Clinic 25 9 20 30 0 29 10 123 Nkwalini Clinic 25 20 20 20 20 28 17 150 Thokozani Clinic 25 14 20 10 16 42 15 142 Nseleni CHC 25 18 20 20 0 40 16 139 Buchanana Clinic 13 0 15 9 3 16 8 64 Ntambanana Clinic 24 20 20 35 20 41 14 174 Cohort 1 Cohort 2 Cohort 3 Cohort 4 Cohort 5 Cohort 6a Cohort 6b District (FTIC) (AC) (DMD) (EAC) (TRIC) (TBHD) (TBHD) Total Ekurhuleni 134 84 106 129 106 335 128 1022 Mopani 137 102 104 154 45 577 16 1135 Bojanala 149 120 120 208 120 267 83 1067 uThungulu 137 81 115 124 59 196 80 792 Study Total 557 387 445 615 330 1375 307 4016 Study Target 720 576 576 1008 576 2400 2400 8256 Percent of target achieved 77% 67% 77% 61% 57% 57% 13% 49% Intervention enrollment 282 168 186 235 139 666 156 1868 Control enrollment 275 219 259 380 191 709 151 2148 16 FIRST ENROLLMENT REPORT 8.3 Enrollment over time While each site was required to recruit the same number of subjects per cohort, due to logistical constraints at some sites and, more commonly, due to delays in implementation of the interventions at the sites, enrollment was not evenly distributed across time with some cohorts enrolling faster than others. Figure 1a-g below demonstrate this for the six cohorts. These figures show that the FTIC, AC and DMD cohorts have been the fastest to enrol, reaching targets in about 16 weeks. While enrollment has been largely steady week on week in these cohorts, enrollment accelerated between week 8 and 9 for the AC and DMD cohorts. This acceleration was largely due to enrollment at control sites in pairs where enrollment of the cohort at the intervention site was delayed and resolution of some register issues that had previously delayed enrollment. Each of these cohorts required screening more subjects than we ultimately enrolled. Subjects who were not eligible are described below. The EAC and TRIC cohorts took longer, although they should be complete by 20 weeks. The most prominent reasons that enrollment was slower for these cohorts was related to: 1) these interventions not being implemented to AGL specifications (e.g. focussing EAC on defaulters rather than unsuppressed patients, or only tracing pre-ART or patients who had defaulted rather than those who had early or late missed appointments); and 2) incomplete or unavailable registers at intervention sites. We also noted that for the TRIC cohort, a larger number of patients who were screened were found to be ineligible compared to other cohorts. This was due in large part to the intervention not being delivered as per the AGL specifications. As anticipated, the TB, Hypertension and Diabetes cohorts are taking the longest. We have currently identified 1375 (57%) subjects as eligible for enrollment into the TBHD screening cohort and 307 (13%) subjects as eligible for enrollment into the TBHD diagnosed cohort. However as described in the previous ‘Enrollment’ section we have not yet initiated cohort enrollment via the third method for the TBHD diagnosed cohort but will be able to do this retrospectively once it is initiated. We anticipate that enrollment of these two cohorts will be completed by end February 2016. 17 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Figure 1 (Cohort 1‒6a, 6b) Screening and enrollment by cohort over time compared to the target total through October 6th (lines for HIV cohorts have been updated to reflect 20% increased enrollment target) 18 FIRST ENROLLMENT REPORT 8.4 Enrollment by facility/district In addition to variation in enrollment over time by cohort, we also encountered variation in enrollment by site and province. While overall all five intervention cohorts have completed or are nearing completion of enrollment within each cohort, Figure 2 shows that enrollment was uneven with facilities in North West province enrolling fastest and Limpopo and KZN taking somewhat longer. This was largely due to delays in being able to enrol for one or more cohorts because of implementation or register issues. Figure 3 shows enrollment by cohort and district. Figure 2a‒d Enrollment over time by province and clinic a b c d Figure 3 below shows enrollment into each cohort within each district. Note that the large difference between 6A and 6B cohort numbers in Mopani is due to only having screening information for TB patients (for hypertension and diabetes cases, no records are created until patients are diagnosed). Because the third route of detecting the diagnosed patients has not been implemented yet, the number of diagnosed patients in Figure 3 only contains those Mopani patients diagnosed TB positive who were identified from those screened. This will change as we implement the 3rd method of identifying diagnosed patients. 19 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Figure 3 Enrollment by District and Study Cohort 8.5 Ineligible Patients Through the course of enrollment, some subjects have been identified who have been ineligible for enrollment following file review. Below in Table 5 we describe the number of ineligible subjects per cohort. The most common reason for not being eligible was a patient not actually receiving the intervention they were eligible for. For all cohorts, missing files was between 8 and 13% of reasons for exclusions. Table 5 Ineligible subjects by cohort and reason for exclusion Cohort 1: FTIC Cohort 2: AC Cohort 3: DMD Cohort 4: EAC Cohort 5: TRIC Control Intervention Control Intervention Control Intervention Control Intervention Control Intervention Reason not eligible N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) Total screened (eligible + ineligible) 347 465 255 310 343 306 435 543 334 332 STUDY CRITERIA File not found 6 (8%) 27 (15%) 3 (8%) 11 (8%) 9 (11%) 15 (13%) 2 (4%) 36 (12%) 12 (8%) 27 (14%) Not 18yrs 0 (0%) 0 (0%) 0 (0%) 0 (0%) 9 (11%) 0 (0%) 9 (16%) 10 (3%) 4 (3%) 7 (4%) Not resident 4 (6%) 1 (1%) 0 (0%) 0 (0%) 0 (0%) 1 (1%) 1 (2%) 1 (0%) 2 (1%) 0 (0%) Intention to transfer 4 (6%) 2 (1%) 0 (0%) 2 (1%) 1 (1%) 1 (1%) 4 (7%) 1 (0%) 12 (8%) 4 (2%) Pregnant 8 (11%) 10 (5%) 2 (6%) 2 (1%) 5 (6%) 0 (0%) 5 (9%) 7 (2%) 9 (6%) 11 (6%) Unknown 0 (0%) 0 (0%) 2 (6%) 0 (0%) 6 (7%) 0 (0%) 3 (5%) 1 (0%) 10 (7%) 2 (1%) COHORT-SPECIFIC CRITERIA COHORT 1: FTIC Not HIV positive 3 (4%) 2 (1%) TB Diagnosis 6 (8%) 5 (3%) Not eligible in last 30d 27 (38%) 17 (9%) Not ART Naïve 14 (19%) 31 (17%) Not receiving intervention 0 (0%) 88 (48%) 20 FIRST ENROLLMENT REPORT Table 5 Ineligible subjects by cohort and reason for exclusion (continued) Cohort 1: AC Cohort 2: AC Cohort 3: DMD Cohort 4: EAC Cohort 5: TRIC Control Control Control Intervention Control Intervention Control Intervention Control Intervention N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) COHORT 2: AC & COHORT 3: DMD Not 12-36 mos on ART 5 (14%) 25 (18%) 3 (4%) 23 (19%) ART change in last 12mos 6 (17%) 6 (4%) 2 (2%) 7 (6%) Last 2 VL not suppressed 15 (42%) 27 (19%) 25 (30%) 35 (29%) No VL in last 3 mos 3 (8%) 19 (13%) 5 (6%) 8 (7%) Not in AC / DMD 0 (0%) 50 (35%) 0 (0%) 30 (25%) Receiving other RPS than randomized 0 (0%) 0 (0%) 19 (23%) 0 (0%) COHORT 4: EAC No ART 3 mos 7 (13%) 10 (3%) No VL > 400 18 (33%) 44 (14%) Not on EAC register 0 (0%) 192 (62%) No VL in last 3 mos 6 (11%) 6 (2%) COHORT 5: TRIC Not ART Initiated 5 (3%) 7 (4%) Didn't miss last visit* 89 (62%) 55 (28%) Did not receive tracing 0 (0%) 80 (41%) Total ineligible 72 183 36 142 84 120 55 308 143 193 Total enrolled 275 282 219 168 259 186 380 235 191 139 Note: * For this group, often these patients appear on TIER as eligible but by the time we visit the site, clinical data has been updated and/or the patient has returned and so they are no longer eligible. COHORT DESCRIPTIVES Baseline data has been collected on each of the enrolled cohorts. Data come from TIER.net and patient files and have been entered into an electronic database for descriptive analysis in STATA. Below in Table 6 we describe the cohorts in terms of their baseline characteristics, both overall and by intervention arm. The FTIC cohort participants are mostly under age 40 (64%). As with most ART clinics, the cohort is more likely to be female (59%) though slightly less than the 66% we typically see. The cohort is also barely above the initiation threshold when the ART program started (<200) with an average of 230 cells/ml3 at ART initiation. Despite being sick, there was no TB in this cohort by design. For FTIC, the intervention and controls cohorts were largely balanced on baseline characteristics with only minor differences between groups, much like what would be expected in a moderate sized individually randomized trial. 21 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Table 6 Baseline characteristics of the cohorts COHORT 1: FAST TRACK INITIATION COUNSELLING FTIC Control FTIC Intervention FTIC Total N=275 N=282 N=557 Characteristic n (%) n (%) n (%) Age (n=556)* 18-29 68 (25%) 69 (25%) 137 (25%) 30-39 103 (37%) 116 (41%) 219 (39%) 40-49 67 (24%) 64 (23%) 131 (24%) 50+ 37 (13%) 32 (11%) 69 (12%) Gender (n=557) Male 120 (44%) 117 (41%) 237 (43%) Female 155 (56%) 165 (59%) 320 (57%) CD4 Count (at ART initiation) (n=514) 246 (2-516) 214 (2-509) 230 (2-516) TB status (n=551) Current TB diagnosis 0 (0%) 0 (0%) 0 (0%) No current TB diagnosis 273 (100%) 278 (100%) 551 (100%) Time eligible for intervention prior to enrollment (days) (Median, Range) (n=516) 29 (2-144) 29 (2-147) 29 (2-147) Note: *Totals for individual variables may differ because some observations have been dropped due to missing or out-of-range values. These are undergoing review to determine the correct value. Table 7 Adherence clubs AC Control AC Intervention AC Total N=219 N=168 N=387 Characteristic n (%) n (%) n (%) Age (n=387) 18-29 48 (22%) 26 (15%) 74 (19%) 30-39 80 (37%) 59 (35%) 139 (36%) 40-49 44 (20%) 50 (30%) 94 (25%) 50+ 47 (21%) 33 (20%) 80 (21%) Gender (n=387) Male 65 (30%) 40 (24%) 105 (27%) Female 154 (70%) 128 (76%) 282 (73%) CD4 Count (at ART initiation) (n=340) 287 (2-1141) 249 (5-1404) 270 (2-1404) Viral Load (copies/ml) (Median, Range) (n=381) 52 (0-400) 54 (0-384) 52 (0-400) log10 Viral Load (copies/ml) (Median, Range) (n=381) 1.71 (0-2.60) 1.73 (0-2.58) 1.72 (0-2.60) TB status (n=386) Current TB diagnosis 1 (0%) 0 (0%) 1 (0%) No current TB diagnosis 217 (100%) 168 (100%) 385 (100%) Time on ART at enrollment (days) (Median, Range) (n=314) 544 (381-1086) 559 (367-1095) 548 (367-1095) For the adherence clubs, about half of patients were under age 40 (55%) and nearly 75% were female, slightly more than what we typically see enrolling on ART (66%). These patients were somewhat healthier at ART initiation than those in the FTIC cohort at an average of 270 cells, but with a wide range. The intervention and control cohorts were similar with respect to most variables (age, sex and log viral load at time of eligibility) though the control cohort was somewhat healthier at ART initiation (287 vs. 249). 22 FIRST ENROLLMENT REPORT Table 8 DMD DMD Control DMD Intervention DMD Total N=259 N=186 N=445 Characteristic n (%) n (%) n (%) Age (n=445) 18-29 48 (19%) 28 (15%) 76 (17%) 30-39 86 (33%) 65 (35%) 151 (34%) 40-49 76 (29%) 59 (32%) 135 (30%) 50+ 49 (19%) 34 (18%) 83 (19%) Gender (n=445) Male 79 (31%) 51 (27%) 130 (29%) Female 180 (69%) 135 (73%) 315 (71%) CD4 Count (at ART initiation) (n=397) 284 (5-882) 247 (5-1428) 268 (5-1428) Viral Load (copies/ml) (Median, Range) (n=440) 35 (0-373) 124 (0-390) 50 (0-390) log10 Viral Load (copies/ml) (Median, Range) (n=440) 1.54 (0-2.57) 2.09 (0-2.59) 1.70 (0-2.59) TB status (n=438) Current TB diagnosis 0 (0%) 1 (1%) 1 (1%) No current TB diagnosis 254 (100%) 183 (99%) 437 (99%) Time on ART at enrollment (days) (Median, Range) (n=363) 548 (371-1044) 782 (387-111) 603 (371-1111) The DMD cohort also had roughly half of the patients below age 40 (51%) and had many more females (71%) than males. The cohort was also similar to the Adherence Clubs cohort with respect to baseline ART initiation CD4 count (268 cells). The intervention and control cohorts were well balanced with respect to sex and age but there were some imbalances in CD4 count at ART initiation and log viral load at eligibility. Table 9 EAC EAC Control EAC Intervention EAC Total N= 380 N=235 N=615 Characteristic n (%) n (%) n (%) Age (n=615) 18-29 43 (11%) 34 (14%) 77 (13%) 30-39 136 (36%) 90 (38%) 226 (37%) 40-49 115 (30%) 68 (29%) 183 (30%) 50+ 86 (23%) 43 (18%) 129 (21%) Gender (n=614) Male 147 (39%) 99 (42%) 246 (40%) Female 232 (61%) 136 (58%) 368 (60%) CD4 Count (at ART initiation) (n=526) 168 (2-1264) 137 (4-663) 152 (2-1264) Viral Load (copies/ml) (Median, Range) (n=600) 3490 (402-6,290,000) 20300 (403-2,149,970) 6248 (402-6,290,000) log10 Viral Load (copies/ml) (Median, Range) (n=600) 3.54 (2.60-6.80) 4.30 (2.60-6.33) 3.79 (2.60-6.80) TB status (n=614) Current TB diagnosis 2 (1%) 2 (1%) 4 (1%) No current TB diagnosis 378 (99%) 232 (99%) 610 (99%) Time between last viral load and enrollment (days) (Median, Range)* (n=601) 51 (1-168) 46 (4-173) 49 (1-173) Note: * Time between last VL and enrollment. It is thought that patients are only enrolled at their next visit unless there is a specific intervention at the site to follow up patients with high VL (if the patient is only returning every other month to pick up medication then these times seem reasonable, as the EAC SOP says “an EAC identified file should trigger referral for EAC as soon as the patient comes back to the facility”). It is possible that an EAC patient may be contacted in between telling them they need EAC but this is not recorded. 23 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA The EAC cohort was very evenly divided between those below and above 40 (50%) but slightly fewer females than in the DMD cohort (60%). The cohort was quite sick at ART initiation with an average CD4 count below 200 cells/ml3 (152). The cohort was eligible for the intervention for an average of 49 days at enrollment. The intervention and control cohorts were well balanced with respect to age and sex and baseline TB. There were small differences in CD4 count (168 vs 137 cells/ml3). Log viral load was also somewhat different and lower in the control cohort. Table 10 TRIC TRIC TRIC Control Intervention TRIC Total N=191 N=139 N=330 Characteristic n (%) n (%) n (%) Age (n=330) 18-29 31 (16%) 38 (27%) 69 (21%) 30-39 70 (37%) 56 (40%) 126 (38%) 40-49 49 (26%) 34 (24%) 83 (25%) 50+ 41 (21%) 11 (8%) 52 (16%) Gender (n=330) Male 69 (36%) 53 (38%) 122 (37%) Female 122 (64%) 86 (62%) 208 (63%) CD4 Count (at ART initiation) (n=297) 203 (8-1039) 226 (3-680) 205 (3-1039) Viral Load (copies/ml) (Median, Range) (n=225) 124 (0-6,377,293) 124 (20-318,792) 124 (0-6,377,293) log10 Viral Load (copies/ml) (Median, Range) (n=225) 2.09 (0-6.80) 2.09 (1.30-5.50) 2.09 (0-6.80) TB status (n=323) Current TB diagnosis 1 (1%) 2 (1%) 3 (1%) No current TB diagnosis 187 (99%) 133 (99%) 320 (99%) Time between last missed visit and enrollment (days)* (Median, Range) (n=292) 29 (1-221) 79 (2-243) 35 (1-243) Note: * Likely that patients are only enrolled at their next time between last missed visit and enrollment (visit can be missed by up to 90 days). The patients need to be eligible for tracing from May through October. Therefore if someone missed a scheduled visit in February, they would be eligible for tracing through May, and only enrolled in October, resulting in 243 days between missed visit and enrollment. The TRIC cohort was also very evenly divided between those below and above 40 (51%) and nearly exactly what we expect for female at 63%. As with previous cohorts they were again quite sick at ART initiation with an average CD4 count of 205 cells/ml3, though higher than the EAC cohort. The cohort was eligible for the intervention for an average of 35 days at enrollment (=is days between last missed visit and enrollment to the cohort, not time between missed visit and tracing attempt). The intervention and control cohorts were well balanced with respect to CD4 at ART initiation and log viral load. However they did show some differences, with younger patients in the intervention cohort and the intervention cohort being eligible for the intervention substantially longer than the control cohort (29 vs. 79). 24 FIRST ENROLLMENT REPORT Table 11 TBHD TBHD Total N=1584 Characteristic n (%) Age (n=1581) 18-29 234 (15%) 30-39 385 (24%) 40-49 396 (25%) 50+ 566 (36%) Gender (n=1578) Male 555 (35%) Female 1023 (65%) Screening at last visit (enrolled in cohort 6a) (n=1584) No screening (Diagnosis only) 205 (13%) TB screening 1067 (67%) DM screening 34 (2%) HTN screening 141 (9%) TB and DM screening 4 (0%) TB and HTN screening 129 (8%) HTN and DM screening 3 (0%) TB, HTN, and DM screening 1 (0%) Recent diagnosis (enrolled in cohort 6a and 6b) (n=1584) No diagnosis (Screened only) 1277 (81%) TB diagnosis 41 (3%) DM diagnosis 37 (2%) HTN diagnosis 179 (11%) HTN and DM diagnosis 50 (3%) Overall the TBHD cohort was older than the previous cohorts (only 39% under age 39) but their sex distribution looked exactly like an ART cohort, suggesting that ART cohorts are much like the general population of patients seeking care at a PHC. TB screening was common in this cohort (with 67% receiving TB screening at last visit) with only 9% getting hypertension screening and only 2% receiving diabetes screening. Among all patients, 81% had no diagnosed condition at last visit, while 11% had hypertension diagnosed, 2% had diabetes diagnosed and 3% had TB diagnosed. FOLLOW UP PROCESSES With the HIV cohorts now nearing completion of enrollment, we will continue to follow up patients through passive record review while continuing to work with the sites to maintain quality data and to monitor data systems for study outcomes. The short term and long term outcomes are described in Table 6 below. 25 EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA Table 12 Study evaluation outcomes for protocol 1 Cohort Short-term Outcome Long-term Outcomes Fast track ART initiation % who initiate ART within 30 days of becoming ART % of patients virally suppressed (< 400 copies/ml3) within counseling eligible 9 months of ART eligibility % who receive all medications within the first four % virally suppressed (< 400 copies/ml3) at twelve months Adherence clubs months after eligibility after club eligibility Decentralized medicine % virally suppressed suppression (< 400 copies/ml3) 12 delivery % who receive all medications within 3 months months Enhanced adherence % who resuppress their viral load (< 400 copies/ml3) % who resuppress their viral load (< 400 copies/ml3) counseling within 3 months of eligibility within 12 months of eligibility Early tracing of patients lost to follow up % who return to care within 3 months of eligibility % who return to care within 12 months of eligibility TB, hypertension, and For each condition, % of patients who have 80% visit For each condition, % of patients who achieve disease diabetes compliance in first 3 months after diagnosis control at the six-month visit after diagnosis We will continue to monitor TIER.net and work with the sites to collect follow up data for patients enrolled in the study. The Gantt chart below in Figure 4 show the timeline for follow up and completion of data collection for protocol 1. Data collection for short term endpoint for FTIC The short term outcome for FTIC requires a 30 day follow up for each patient to assess the percentage who initiate ART. FTIC follow up occurs through TIER.net through identification of the initiation visit with verification through patient registers if needed. FTIC follow up has been completed for 74% of patients who have currently been enrolled. All subjects are expected to reach a short term outcome by December 2016. At that point we will need to verify data against patient files for any subjects with missing data in TIER (e.g. should a backlog occur) with an expected date of having all short term outcomes by February 2017. Below are outcomes for the patients who have reached a primary short term outcome for FTIC. Currently 86% of patients in intervention sites and 84% of patients in control sites initiated ART within 30 days of eligibility. 26 FIRST ENROLLMENT REPORT Table 13 Short-term outcomes (ART initiation within 30 days) for those eligible for FTIC cohort INTERVENTION CONTROL Initiated Initiated Eligible for within % Eligible for within % Facility Total outcome 30 days initiated Facility Total outcome 30 days initiated Motsamai Clinic 25 20 15 75% Tamaho Clinic 25 7 5 71% Phola Park CHC 25 6 4 67% Ramokonopi CHC 23 8 8 100% Khumalo Clinic 23 16 16 100% Zonkizizwe 1 Clinic 13 11 10 91% Grace Mugodeni CHC 25 22 19 86% Motupa Clinic 25 23 17 74% Giyani CHC 25 19 17 89% Dzumeri Clinic 24 16 16 100% Tzaneen Clinic 21 18 13 72% Nkowankowa CHC 17 14 10 71% Letlhabile CHC 25 25 23 92% Wonderkop Clinic 25 24 20 83% Hebron Clinic 25 25 21 84% Majakaneng Clinic 25 25 22 88% Tlhabane CHC 25 23 21 91% Bafokeng CHC 24 19 18 95% King Dinizulu Clinic 25 14 12 86% Nkwalini Clinic 25 16 11 69% Thokozani Clinic 25 13 12 92% Nseleni CHC 25 24 19 79% Buchanana Clinic 13 10 9 90% Ntambanana Clinic 24 18 17 94% Total 282 211 182 86% Total 275 205 173 84% 27 Figure 4 Gantt Chart of Timeline for Project EVALUATION OF THE NDOH’S NATIONAL ADHERENCE GUIDELINES FOR CHRONIC DISEASES IN SOUTH AFRICA 28 2015 2016 2017 Project Management (Protocol 1&2) September October November December January February March April May June July August September October November December January Activities, Deliverables and Milestones % Complete 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 Key Milestones Secure all permissions and approvals 100% Recruit and train provincial teams 100% Procure tablets for data collection 100% Site assessments 100% Data enhancement 70% Finalize Protocol 1&2 data collection tools 100% Protocol 1: Data collection Baseline 84% Protocol 1: Data collection Follow-up 0% Protocol 2: Data collection 0% ENHANCE Project Plan Gauteng Province- Ekurhuleni SD S2 Office space 100% Stakeholder engagement 100% Site assessments 100% Data enhancement meeting 100% Data enhancement 100% Data collection - Protocol 1 82% Data collection - Protocol 2 0% Limpopo Province- Mopani District Office space 100% Stakeholder engagement 100% Site assessments 100% Data enhancement meeting 100% Data enhancement 100% Data collection - Protocol 1 75% Data collection - Protocol 2 0% North West Province- Bojanala District Office space 100% Stakeholder engagement 100% Site assessments 100% Data enhancement meeting 100% Data enhancement 100% Data collection - Protocol 1 95% Data collection - Protocol 2 0% KwaZulu Natal Province- uThungulu District Office space 100% Stakeholder engagement 100% Site assessments 100% Data enhancement meeting 100% Data enhancement 100% Data collection - Protocol 1 76% Data collection - Protocol 2 0% CONCLUSIONS Enrollment into the HIV cohorts is nearing completion (estimated by end of October, 2016) and enrollment into the TBHD cohorts is ongoing and on track to complete by end of December 2016. While there have been challenges to getting the cohort enrolled, including delays in implementation of the interventions and changes to the way the interventions have been implemented, the team has been able to enrol subjects into the cohorts. These challenges have been overcome through a combination of support from the World Bank and NDoH PIs to working with the sites to ensure accurate record keeping and access to registers. The cohorts are now entering the follow up phase for the HIV cohorts and work will shift to a mix of maintaining control over the quality of the cohort data, completing enrollment into the TBHD cohorts and to beginning protocol 2 where we will collect largely qualitative data from patients and providers to better understand how the rollout is occurring and the impacts that it is having. 29