TB Reduction Through ART and TB Screening Project

1. STUDY OBJECTIVES AND DESIGN The overall Aim of this project is to measure the impact of a combined TB/HIV intervention of population level screening for TB, combined with universal testing and treatment (UTT) for HIV, delivered over 4 years, on TB incidence, prevalence and incidence of TB infection.

1.1 The Study Objectives

1. To compare the population impact of a combined TB/HIV intervention of population level screening for tuberculosis, combined with universal testing and treatment (UTT) for HIV on:
2. The prevalence of TB disease in a randomly selected sample of individuals ≥15 years b. The incidence of infection with TB in a randomly selected cohort of adolescents and young adults (15-24 years) c. The effect on notified bacteriologically confirmed pulmonary TB incidence in the parent HPTN071 trial Population Cohort participants aged 18-44 years over the last 24 months of follow-up (2017-2018) d. The effect on bacteriologically confirmed pulmonary TB case notification rates among adults (≥18 years) residing in the study communities, during the period 2017-2018 e. The clinical characteristics and treatment outcomes of bacteriologically confirmed pulmonary TB cases in Population Cohort participants aged 18-44 years over 36 months f. The clinical characteristics and treatment outcomes of bacteriologically confirmed pulmonary TB cases among adults (≥18 years) residing in the study communities over 60 months (2014-2018)
3. To understand the pathway to impact of the PopART combined TB/HIV intervention on TB by measuring the yield of case finding by HIV status, change in TB notifications and reported TB incidence.
4. To use qualitative methods to assess the delivery of the PopART interventions:
5. describing TB stigma and popular understanding of TB among community members drawing on longitudinal data (2004-18) from the selected communities
6. describing the experience of diagnosed TB patients and their households across intervention and control sites, and household intervention and standard of care health facility services
7. To use mathematical modelling to assess:
8. The effect of the different PopART intervention components (universal ART and screening for TB) on the impact on TB b. The projected impact of this intervention on TB in other settings and what factors influence this
9. Use economic analysis methods combined with mathematical modelling to measure the cost-effectiveness of the PopART intervention in terms of its impact on TB
10. To evaluate newer methods for measuring the impact of public health interventions on TB a. To evaluate the role of QuantiFERON® Gold Plus in assessing epidemiological impact of TB interventions on infection with TB b. To evaluate new biomarkers for assessing the predictors of progression from infection with M. tuberculosis to active disease ("incipient TB") c. To compare a novel method of measuring the prevalence of active TB disease, including Xpert® MTB/RIF Ultra combined with computer aided digital chest x-ray (CAD4TB), with traditional methods in terms of costs and efficiency
11. To increase the capacity of African field sites to conduct large cluster randomised trials through south-south as well as north-south collaboration
12. To inform global policy:
13. On the costs and effectiveness of the PopART intervention to control TB b. On strategies to measure the burden of TB at population level
14. To raise awareness amongst researchers, funders, programme managers and civil society of the impact of a combined TB/HIV intervention of population level screening for TB and universal HIV testing and treatment to inform future approaches to addressing TB effectively

1.2 The Study Design

The HPTN 071(PopART) Trial is a three-arm matched cluster-randomised trial of combination HIV prevention including universal testing and treatment for HIV (UTT) being conducted in 21 communities in Zambia and South Africa. The primary outcome is HIV incidence at community level measured in a randomly selected cohort of approximately 40,000 adults (HPTN071 population cohort). The trial is sponsored by DAIDS and is managed via the HIV Prevention Trials Network (HPTN). Trial registration ClinicalTrials.gov NCT01900977.

The TREATS project is nested within the HPTN 071(PopART) trial. The HPTN071 had two intervention arms (arm A and B) that differed in that both had the same community intervention (see below) but in arm A, HIV positive individuals received universal ART regardless of CD4 count whereas in arm B national guidelines on CD4 count were used. During the trial the thresholds for commencement of ART have changed such that universal ART regardless of CD4 count is provided in all arms of the trial.

Randomisation of the 21 communities (clusters) for the HPTN071 trial was done using matching of triplets. Within each triplet, one community was randomly allocated to each of the three study arms, using a process of restricted randomisation to ensure balance on key covariates. Communities were matched into triplets by geography and estimated baseline HIV prevalence. The randomisation was done in a public ceremony in 2013. For practical reasons there is no blinding of which cluster is in which intervention arm, however laboratory staff analysing biological samples for the primary outcomes are blinded to the arm of the study, and investigators are blinded to outcome data beyond baseline.

1. STUDY INTERVENTION

2.1 Intervention

The HPTN071(PopART) intervention (arm A) consists of a package of combination TB/HIV prevention activities, including active case finding (ACF) for TB and universal test and treat (UTT) for HIV. The intervention was delivered over a period of four years by community health workers called community HIV care providers (CHiPs) who were responsible for a zone of their community. They went door-to door within their zone ensuring that every member of the population had access to HIV and TB prevention messages, annual TB symptom screening, HIV testing and sexually transmitted infection screening, condoms, referral to voluntary male medical circumcision and, for HIV positive individuals, referral to care and ART regardless of CD4 count. During each annual visit, each member of the community was screened for symptoms of TB and sputum samples taken from any who were symptomatic for TB diagnosis, and tested using Xpert® MTB/RIF (Cepheid, Sunnyvale, CA) in individuals who were HIV-positive and smear microscopy for those who were HIV-negative. All symptomatic individuals were followed up to ensure that results were available and appropriate referrals were made for further investigation if required. Diagnosed cases of TB were referred to care through routine government health care facilities (HCF) and were treated with standard 6-month TB regimens for drug-susceptible TB and referred for treatment according to national guidelines if they had drug-resistant disease. All TB treatment was provided by government HCF. All HIV positive individuals who presented to the clinic for HIV care were screened for TB using a symptom screen, with symptomatic individuals being tested using Xpert® MTB/RIF according to standard national policy, and were offered isoniazid preventive therapy (IPT) if they were non-symptomatic. The CHiPs collected data on the uptake of the intervention (CHiP data).

The HPTN071(PopART) intervention in arm B was the same as in arm A, consisting in a package of combination of TB/HIV prevention activities, including active case finding (ACF) for TB and UTT for HIV delivered in the community. Differently to arm A, in arm B national guidelines on CD4 count were used for ART initiation. During the trial the thresholds for commencement of ART changed from 500 CD count to 350 CD count and then, in 2016, to universal ART regardless of CD4 count.

2.2 Standard of Care

The standard of care arm C communities had access to HIV testing services (usually at the health facility), HIV care and ART provision according to national guidelines, which since mid-2016 includes ART initiation regardless of CD4 count. TB case finding is "passive" i.e. relies on individuals to present with symptoms to the health facility. TB diagnostics were as per national guidelines including Xpert®TB/RIF for those who were HIV positive. All diagnosed cases of TB were treated as in the intervention arm. TB screening and prevention at HIV care clinics were as is in the intervention arm.

1. STUDY PROCEDURES

3.1 Infection Cohort

The study will be conducted in the 7 arm A (intervention) communities and the 7 arm C (standard of care) communities of the HPTN071 (PopART) trial. A cohort of 300 adolescents and young adults (15-24 years of age) will be recruited in each community (total 4200).

3.1.1 Sampling/Recruitment of Infection Cohort

In 2013 the location of every household in the community was recorded. This provides a sampling frame for the selection of households. The study community area will be subdivided into blocks whereby every block consists of approximately 40 households in Zambia and approximately 55 households in South Africa, on the basis that this is expected to result in an average of ~15 individuals aged 15-24 years enrolled in each block in both countries. The blocks will be visited in a random order. An expected number of 20 blocks will be visited to reach the required sample size in each community with a minimum of 10 blocks and stratified sampling (having divided each community into "zones") to ensure good geographical representation.

All households within a block are eligible and will be approached using a door-to-door approach. After explanation of the study, permission will be sought from the responsible adult to enumerate all household members. Household members of 15-24 years of age that have lived in the community for 2 years or more and are expecting/planning to be resident for at least 9 months per calendar year for the next 2 years, are given an invitation card and invited for participation in the study.

Recruitment

Eligible individuals responding to our invitation will be seen either in their home, at a central site for each community, this may be the local health facility or some other clinical research site or at a mutually agreed convenient place. At this site further information about the study will be provided, eligibility will be confirmed, and informed consent and assent can be provided.

Participants aged > 18 years will provide written consent. Individuals <18 years will provide written assent and will need written consent from their parent or guardian before participating in the study.

3.1.2 Procedures and Activities

Baseline assessment

1. A structured questionnaire will be administered to the consenting young people by trained research assistants (RAs) which will include information on socio-demographic data, previous TB and HIV history, risk factors for TB and HIV and social networks. Locator information including contact numbers of the participant and their family will be taken to facilitate mid-year contact calls.
2. Physical measurement of height and weight will be taken.
3. A TB symptom screen will be conducted, and Xpert MTB/RIF will be done on sputum samples provided for all symptomatic participants.
4. Up to 15mls of blood will be drawn tests including QFT Plus test, HIV testing and for storage for additional testing for TB biomarkers or other respiratory pathogens.
5. Any participant who tests QFT Plus positive at baseline will be assessed by study staff for active TB disease, counselled regarding their need for isoniazid preventative therapy (IPT), currently offered to HIV positive individuals.
6. All participants will be encouraged to accept HIV testing using rapid tests as per national guidelines at baseline and throughout the follow-up period. Any newly diagnosed HIV positive will be linked to ART care.
7. IPT will be offered according to national guidelines

Follow Up

All participants will be asked to come back for follow-up visits at month 12 and 24.

Participants will be educated on the signs and symptoms of TB and asked to return to the study site if any symptoms suggestive of TB are noted. Follow up contacts will be made for any participants identified as HIV-positive or diagnosed with TB to ensure linkage to care has occurred

Procedures at follow up visits at 12 and 24 months

1. Participants will be interviewed using a standardized questionnaire including activities and social networks as well as contacts with TB.
2. All participants will be asked additional questions about TB transmission/tracing and social contact patterns
3. All participants will have blood drawn to be retested using QFT-plus and for storage for other tests as at baseline.
4. Participants will also be screened for TB, using a standard symptom screen and asked to provide one sputum sample, so that we can test for TB disease using Xpert MTB/RIF.
5. All participants will be encouraged to accept HIV testing using rapid HIV testing

Any participant found to have TB or HIV at any follow up visit will be referred to the clinic for further assessment and care. All HIV+ participants will have a CD4 count taken at the HCF and the participant will be asked if the research team can have access to the clinical records including the results of CD4 counts and HIV viral loads. Any participant identified as HIV+ attending the HCF will also be referred for IPT for 6 months as per national guidelines. This will be given at the ART clinic. During the study visit any other medical issue encountered by the research team will be referred for appropriate HCF services.

Contact phone calls will be made between visits (e.g. at months 6 and 18) to check on the whereabouts of the participants and to check that they are well. It is anticipated that this will reduce losses to follow-up.

Laboratory procedures

Up to 15 mls of blood will be taken at each visit for tests including Quantiferon Plus testing, HIV testing and for storage for additional tests to explore host and Mycobacterium tuberculosis (MTB) biomarkers that can predict progression to active TB disease, improve TB disease/infection diagnosis and monitor response to TB treatment, and to explore the role of different viral infections in the aetiology of TB.

3.2 Prevalence Survey

Data collection will be implemented in the two countries in all TREATS 21 communities in arms A, B and C commencing in 2019. In each community approximately in arms A and B approximately 2000 people will be enrolled and 4000 people in arm C.

3.2.1 Sampling/Recruitment of Prevalence Survey Participants.

The 21 communities consist of census zones each with an estimated 300-500 households and an anticipated ~1300 adults (15 years and above) per zone. Census zones will be further subdivided into "blocks" that each include ~40-60 households, having 100-150 adults (15 years and above). The blocks will be grouped geographically to enable stratified sampling, which will ensure there is good geographical coverage of the whole community. The blocks will be randomly sampled, with stratified sampling, and then all households within a sampling block will be eligible for inclusion in the study until the targeted sample size of either 2000 for arms A and B and 4000 participants for arm C is reached in the different communities.

Recruitment

Centrally in each census zone a mobile field site (MFS) will be set up where the OneStopTB Platform (a truck, containing a digital X-ray and Xpert® instrument) will be stationed and tents will be erected to conduct the different survey procedures. Community mobilization activities will be conducted to sensitize community members and then each selected zone will be enumerated block by block. The census listing will be done by systematically moving house-to-house to cover all dwellings in the selected blocks. In each household all those eligible for the TB prevalence survey (resident, and age ≥15 years) will be listed. Those eligible will be invited to the MFS. At this site further information about the study will be provided and eligibility will be confirmed.

Participants aged > 18 years will provide written consent. Individuals <18 years will provide written assent and will need written consent from their parent or guardian before participating in the study.

3.2.2 Procedures and Activities

All enrolled participants, following consent/assent will be asked a set of questions to determine socio-demographic and other key characteristics of the survey population for analysis. Individual factors may include smoking, alcohol, education, marital status, history of working in the mines, being a health care worker. Questions around TB stigma as well as questions about HIV testing, HIV status, ART use and exposure to the PopART intervention will also be included. Socio-economic status will also be determined at a household level. Using an invitation card with barcode all enrolled participants will be invited for TB screening at the MFS.

Upon arrival at the MFS participants will be logged in at the reception desk using the barcode where their identity and eligibility will be checked after which TB screening will be conducted. TB screening will consist of symptom screening and chest X-ray for all enrolled participants. Detailed questions around health seeking behaviour will be asked to all those reporting symptoms. Digital chest X-ray (CXR) images will be read automatically using CAD4TB, a software product that takes a single frontal chest radiograph as input, in the form of a DICOM image, and produces several outputs.

All individuals who either have a positive symptom screen and/or an abnormal CXR (based on their CAD4TB score) are eligible for sputum examination. Participants refusing to undergo CXR or symptom screen will automatically be sputum eligible. All sputum eligible will be asked to provide sputum samples for TB diagnosis using Xpert® MTB/RIF. Xpert testing will be conducted in the mobile truck by a trained technicians and results should be available within 2 hours. All those diagnosed with TB after confirmatory testing, will be referred to the nearest health facility for treatment initiation/ linking to care, following the country's national TB treatment guidelines. Confirmatory testing is needed in a TB prevalence survey setting to avoid false positive results given the lower level of TB prevalence then the routine clinic setting. In case of error results, the test will be repeated.

All participants will be offered HIV counselling testing using rapid tests as per standard protocol at the mobile field site. Those with known HIV positive status will not be retested. Any newly diagnosed HIV positive will be linked to ART care.

Participants living with HIV and any newly diagnosed HIV positive will be asked permission for finger-stick blood sampling for viral load testing and HIV sequencing for phylogenetic and HIV drug resistance analysis. Other viruses such as hepatitis C virus will be tested for as well as HIV

The research team will ensure the results of the CXR are given back to participants. This can be in real time or potentially as requested by the participant through a different route for those who don't wish to wait for their CXR results. These results can be given back for the participant while attending for a second visit, through a household visit by the research team or by text message.

To ensure other findings in the lungs besides TB are properly addressed, all images with a high CAD4TB score (to be confirmed after the pilot) will be read by an experienced radiologist at a central location via tele reading to investigate whether there is need for referral for other lung conditions than TB. Referrals will be made to the nearest appropriate HCF depending on the findings. Any participant not attending for clinical follow up will be checked by the research team within 2-4 weeks from the test.

3.3 Implementation Science Studies

3.3.1 Secondary data Analysis

Secondary data analysis will use two data sources:

• Routine TB case notification data (held in TB registers and containing information on all patients starting TB treatment, including microbiology results, HIV-status and treatment outcomes), for all adults residing in the Pop-ART communities (N=330,000) from 2014-2018.
• Population cohort data (N=28,000) linked to TB case notification data. Detailed individual-level data are collected annually on the cohort, over approximately 36 months of follow-up for each participant (between 2014-2018). At each visit, cohort members are asked if they have started TB treatment in the preceding year and, if so, unique TB case identifier numbers from patient cards are obtained. Population Cohort data will be linked to TB case notification data allowing all TB diagnoses in the cohort to be verified and microbiology results and TB treatment outcomes to be determined.

Data collection and management: TB case notification data between 01/2011-12/2018, with treatment outcomes to 06/2019, will be collected. This allows historical trends and data during and immediately following the PopART intervention to be investigated. Population cohort data will be linked to TB case notification data. Anonymised datasets (Population cohort data linked to TB case notification data and overall TB case notification data) will be generated and used to address the study objectives.

As part of HPTN071, data on the intervention uptake in the intervention sites only are captured electronically and summarised annually. These intervention process measures will be used to document how uptake of the intervention, especially TB screening and diagnosis, varies across communities and to explore the association between these measures and TB case notification rates at community-level. In addition, the intervention process measures will be analysed by age, gender and other socio-demographic characteristics to explore inequities in intervention uptake/distribution by different population groups.

3.3.2 Qualitative Studies

Qualitative studies will be conducted in both Zambia and South Africa

The following activities will be performed:

• Semi-structured interviews with individuals aged ≥18 and their households who are part of an existing qualitative cohort to explore the relationship between TB and HIV, popular understanding of TB, options for managing TB, TB stigma and the implications of missing TB cases. A cohort representing different HIV decisions and outcomes (to test/not test, HIV-negative/HIV-positive, People living with (PLHIV) on treatment/PLHIV not on treatment) and different gender and age groups was established by the PopART social science team from 2016-2018. Approximately 8 individuals in one arm A and C community per country (n=32) and their households will be revisited and asked TB-specific questions.
• Semi-structured interviews with stakeholders providing TB services, to determine TB services available (arm A and C), including a historical perspective of the history of TB services in each particular site. Investigators will liaise closely with the community engagement team for this activity.
• Focus Group Discussions with 10-12 Community HIV Care Providers (CHiPs) in each Arm A site (n=84 approximately), to document restrospectively their experiences of delivering TB services within the wider PopART intervention and with particular communities.
• In-depth interviews with key TB staff to describe their experience with TB service delivery and the role of other stakeholders (n=14; arm A and C).
• In-depth interviews with TB patients and their households in intervention sites (n=49 - 7 per Arm A Site across both countries). In-depth interviews will be conducted using participatory techniques including body mapping and network charts (for mobility and TB pathways) to determine the role of CHiPs, the patient's experience of falling ill with TB and the enablers and barriers to diagnosis and treatment from a household and patient perspective.
• In-depth interviews with TB patients and their households diagnosed through passive case finding (standard of care at health facility level) in health facilities in intervention sites (n=49 - 7 per Arm A Site across both countries). The research tool will be similar to the comparative interview of those diagnosed through CHiPs but explore, rather, the role of the health facility and other services. TB patients will be identified through the TB services at the health facility and in each site it would be aimed to recruit a range of age and gender groups (aged 18 and over). Investigators will ask TB patients if they can accompany them home and interview them and their household members at home.

3.4 Mathematical modelling

The model software will be developed through the project as an open source R package, with the core deterministic compartmental transmission model written in Fortran 90.

There are three main components to this model development work:

1. Adaptation of an existing underlying demographic and HIV/ART model that has been developed for the PopART trial in discussion with the PopART HIV-modelling team
2. Adapting an existing TB transmission model in this framework to represent the TB interventions in TREATS, and contact the TB data collected by PopART and TREATS
3. Adapting this product to generate output data and perform analyses required for health economic evaluation of the TB intervention

Component 1. HIV transmission will be modelled phenomenologically, with HIV incidence and changes in ART coverage derived from the estimates from the PopART HIV-modelling to ensure a consistent account of HIV epidemiology. To facilitate this, the underlying demographic/HIV/ART structure (currently 81 age categories, 2 sexes, HIV-uninfected + 7 CD4 categories for PLHIV, and 4 ART durations) and the HIV/ART natural history parameters will be modified to ensure the demographic/HIV/ART model substrate aligns with the approach taken in the PopART HIV modelling. The focus will be in reproducing in the TB model substrate the patterns of HIV infection by age and sex, and the estimated distributions of CD4 cell count and time-since-ART initiation, which strongly influence TB risk, to capture the influence of the PopART HIV interventions on TB incidence.

Component 2. The TB transmission model will follow established model structures (including latent TB infection, fast and slow routes of progression to disease, re-infection and progression, increased risks of relapse after treatment), and sit on top of the HIV/demographic layer, accounting for increased risks of TB progression due to HIV-related immune-compromisation (i.e. linking TB risk to CD4 count and time on ART). HIV/ART status will also affect the natural history of TB disease. The model will also be modified to include the TB active case-finding aspects of PopART, and to include a likelihood that contacts the routine data and actively collected data (disease prevalence, infection incidence).

Component 3. To be used for projections and health economic evaluation, additional outputs will be generated to, first, estimate incremental health benefits of future TB infections averted in terms of disability-adjusted life years (DALYs), second, estimate healthcare cost savings due to TB infections averted, and third, estimate incremental costs of active case-finding for TB in the context of UTT projected into the future. In addition, the model will need to be modified to project forward beyond the end of the intervention, and be made flexible enough to consider other settings.

There will be continuous effort to collate and analyse process and outcome data from PopART and TREATS for use in modelling.

Literature review work will be undertaken to generate data for modelling, and to inform the economic evaluation. This will focus on two areas: recent epidemiological literature informing on the natural history of TB for individuals on ART; and then later the epidemiological characteristics of other settings identified as relevant to generalizing findings beyond the PopART communities.

3.5 Cost-effectiveness Analysis (CEA).

The basic premise of the CEA of TREATS is to provide important information to decision-makers on the following two aspects:

• Is a combination of UTT for HIV along with active case finding for TB worth implementing as compared to standard care, in South Africa and Zambia, but also in other high endemic countries?
• Are there economies of scope from jointly providing UTT and active case finding for TB compared to only UTT for HIV?

The epidemiological and economic analyses will be closely integrated. The CEA will take a "Health systems" perspective and include all costs and benefits associated with provision of HIV and TB care. Outcomes will be valued in terms of TB infections and DALYs averted, and will be projected into the future.

3.5.1 CEA data.

Calculation of the cost and benefits will use primary data collected from health service facilities, data and results from other costing studies that have already been implemented and published literature sources.

Economic costs will be collected via a bottom-up micro-costing approach from intervention expenditure data of Zambart and implementing partners, and compared against facility level costing of HIV and TB prevention under Standard of Care. Costs components considered are salary, test kits (Xpert), other equipment, and travel. Results of a time-and-motion study conducted under HPTN071(PopART) will estimate the time spent by CHiPs on the TB component.

The TREATS CEA will collect health facility costs of providing TB services in some of the health facilities each in Zambia and South Africa, to ensure good representation. The objective of the facility level costing survey is to estimate per-patient-per-year unit costs of typical TB related tests and treatments. In both countries, two facilities from the control communities of PopART and 1 facility from the intervention communities will be chosen according to predefined criteria. Health facility costs of HIV related care have already been collected under PopART from all trial facilities.

From the surveyed facilities, costs for prevention under standard of care will be collected (such as household contact tracing and IPT provision), and components considered are salary, test kits, and IPT. The costs of caring for TB patients at all stages of the disease will also be collected from healthcare facilities, to calculate the healthcare costs saved of TB infections averted. Shared expenditures on equipment, building, and general overheads will be apportioned to TB prevention and care according to activity data. Unit costs estimates of the cost per person tested and cost per TB case identified will be estimated. Costs of care for drug-resistant TB, outside the surveyed health facilities, such as at tertiary health care units will be gathered from secondary data sources.