SCAF-TB - Saving Children and Their Families From Tuberculosis: Implementation Research to Improve TB Preventive Treatment Initiation and Completion Rates Among Contact Persons in Tajikistan and Tanzania

Despite renewed global commitments to end TB, as reinforced by the UNHLM on TB, progress remains insufficient to meet TB elimination targets. A critical barrier is the limited reach and uptake of TPT, which continues to rely heavily on facility-based models that are often inaccessible or burdensome for those at risk. A fundamental shift toward broader, community-delivered TPT is urgently needed to close this implementation gap.

One key limitation in current TBI diagnostics, such as the traditional tuberculin skin test (TST), is the requirement for a follow-up visit 48-72 hours after administration to measure the result. This return visit is usually at the clinic, typically performed by a healthcare worker (HCW), which introduces significant challenges including loss to follow-up, logistical burden, and variability in result interpretation. These issues hinder timely initiation of TPT. The SCAF-TB study aims to evaluate a novel, home-based TB infection testing model using the WHO-endorsed SIILTIBCY skin test combined with a skin patch diagnostic paired with an AI-driven smartphone application to enable community-based result interpretation. This approach eliminates the need for clinic visits and may reduce inter-reader variability by leveraging technology for standardized, home-based result reading.

Apart from that, TB screening currently depends largely on chest X-rays, which are not always easily accessible and often require visits to healthcare facilities. New, more portable screening technologies-such as the AudibleHealth AI platform, which uses machine learning to analyze cough sounds and flag potential respiratory infections such as TB-could offer an innovative alternative. This study aims to evaluate the feasibility of this tool in detecting TB patients, providing evidence on whether such digital solutions can support or enhance future TB screening strategies at community level.

Moreover, prior studies have demonstrated that shifting TPT services from healthcare facilities to community-based delivery-particularly through community health workers (CHWs)-is not only feasible but also preferred, highly acceptable to patients, and cost-effective in various settings (32). Building on this evidence, SCAF-TB will assess the effectiveness, acceptability, and scalability of this integrated, decentralized TB prevention algorithm.

Objectives The overall goal of SCAF-TB is to improve access to innovative TB preventive services among children and their families in Tanzania and Tajikistan, both settings with a high burden of TB.

The investigators will utilize a type 2 hybrid effectiveness-implementation study design, which has a dual focus on effectiveness and implementation outcomes (33, 34). This multi-country, mixed-methods study includes, an inter-reader agreement study of reading SIILTIBCY results with and without AI-assisted reading, 2) in-depth interviews of key interest-holders, 3) a comparison of TPT uptake and completion rates in close contacts between clinics with and without home-based skin testing and AI-assisted reading, and 4) budget impact assessments.

The investigators will utilize an adaptation of the RE-AIM framework. The investigators propose an integrated work stream to determine the impact of the SCAF-TB model compared to the standard of care.

This study will be conducted in high-volume primary health care facilities located in Tajikistan and Tanzania. The selection of clinics within each country will be carried out in close consultation with the respective National TB Programs (NTPs) to ensure alignment with national priorities and operational feasibility. Tajikistan and Tanzania were chosen based on a set of strategic criteria: Tajikistan represents a high-burden country for multidrug-resistant tuberculosis (MDR-TB), while Tanzania faces a significant burden of both TB and TB/HIV co-infection. Both countries have demonstrated a clear willingness through their NTPs to introduce TBI testing using SIILTIBCY. Additionally, they have programmatic access to short-course TPT regimens, which are actively prescribed within their health systems. Their national guidelines also support the initiation of TPT for contact persons regardless of age, reinforcing the feasibility of broad implementation. Importantly, both countries host KNCV-affiliated offices, which will serve as key implementing partners throughout the study.

Tajikistan, a landlocked country in Central Asia, continues to face a high burden of TB, with one of the highest incidence rates in the WHO European Region. In 2023, an estimated 8,200 people were living with TB disease in 2023 out of which 620 were estimated to be children, out of whom 273 (44%) were notified. Contributing factors include poverty, malnutrition, limited access to quality healthcare, and significant labor migration, which increases transmission risk. MDR-TB poses a growing challenge, accounting for a considerable proportion of new and retreatment patients. In response, the government, in collaboration with international partners such as the WHO and the Global Fund, is focusing on early case detection, improved diagnostics (including GeneXpert), and expanded access to MDR-TB treatment. Community-based interventions and cross-border collaboration are also being strengthened to control TB transmission and improve treatment outcomes. 52% of household contacts (or all close contacts) of all ages started on TPT out of those eligible in 2023. In Tajikistan, the study is planned to be implemented in Dushanbe, which will cover 16 medical institutions (in collaboration with Center for Protection of Population from TB in Dushanbe) and 15 city polyclinics in Dushanbe.

Tanzania, located in East Africa, remains among the high TB burden countries globally, with an estimated 122,000 people developing TB disease in 2023, including approximately 20,000 children. Key contributing factors to TB transmission in Tanzania include high HIV prevalence, poverty, undernutrition, and limited healthcare access in rural and hard-to-reach areas. In response, the Tanzanian government, supported by partners is intensifying efforts on early diagnosis through decentralized services and advanced diagnostics like GeneXpert and digital chest X-rays. Treatment for both drug-sensitive and MDR-TB has been expanded, and the country is also scaling up preventive strategies such as TPT for eligible contacts-nevertheless, just 16% of household contacts of all ages initiated TPT in 2023. Community-based TB care models and integration with HIV services are further strengthening the national response.

In Tanzania, the study will be conducted in two districts of Ilala and Kinondoni located in Dar es Salaam region. Tanzania started implementation of TPT back in 2005 whereby it was given to under-five years of age household contacts of bacteriologically confirmed TB patients and provided at TB clinics. Upon 2011, the TPT was extended to include all people living with HIV (PLHIV) and Isoniazid Preventive Therapy (IPT) was solely given as the TPT. Following the policy and guidelines reviewed in 2025, TPT is recommended for all household contacts of bacteriologically confirmed PTB patients, adults and children living with HIV, prisoners, miners, people who use drugs (PWUDs), and elders after ruling out active TB disease.

At the community level, CHWs are engaged to conduct community sensitization as well as to create demand creation for TPT to help improve health seeking behavior. It is recommended that TPT sensitization be integrated with other routine services done at the community level, including active case finding, TB contact investigation, Reproductive, Maternal, Newborn, Child, and Adolescent Health outreach services, immunization campaigns, HIV services campaigns. CHWs have been trained in how to evaluate TPT eligibility for those at risk using standardized TB screening questionnaires. Following TB screening, all presumptive TB patients are referred for TB testing at health facility level. Those who test TB negative will be eligible for TBI testing and recorded in TPT register for close monitoring and evaluation. Those who will test positive for TB will be enrolled in the TB register and initiated anti TB medication. CHWs are asked to ensure close follow-up of TPT clients (and TB patients) for drugs adherence and completion at their catchment areas of implementation.

Contact persons of bacteriologically confirmed TB patients (so called index patients) who are documented in contact registers at participating Primary Health facilities will be considered for the study. CHWs will proactively visit TB-affected households to screen eligible contacts, particularly children and vulnerable family members. CHWs will also follow up with contact persons after TPT initiation to provide necessary support and manage potential adverse reactions, to improve TPT adherence and completion rate.

In the study clinics, index patients will first agree to home-visits by a CHW. Qualifying TB index patients are defined by the Tajikistan and Tanzanian National TB guidelines and are reflected in the inclusion/exclusion criteria by country. The TB index patient's diagnosis will be kept confidential and only disclosed to household members with permission from the index patient. The consent process will document the index patient's agreement to a home visit and their consent for abstracting their clinical data. Community health teams will make up to three attempts to visit the household. All adults and children under 15 years whose parent or legal guardian consent to study participation (and the child provides assent when appropriate) will be screened for both TB and HIV at home (as per national guidelines). HIV testing will be offered and conducted in the household for all adults and children 18 months and older with unknown HIV status using a rapid HIV test (per local practice and standard of care). Children under 18 months old will be assessed for HIV exposure by both report and evaluation of personal health card. Their care relating to prevention of mother-to-child transmission (PMTCT) will be reviewed including maternal antiretroviral status, infant prophylaxis, and infant testing. If they are not up to date with testing or have been lost to follow up, the community health team will refer that infant/mother pair to the clinic.

Recruitment The visit will begin with contact tracing where the index patient is asked to list all household members. Clinic staff will then ask index patients if the household can be approached by research staff for study participation. The home visit will include (1) TB symptom screening for all children under 5 years, (2) TB disease screening using a combination of symptom screening, AudibleHealth, and portable digital CXR (if available) for all contacts ≥5 years, (3) TBI screening using SIILTIBCY in eligible contacts, (4) TPT initiation for asymptomatic child contacts under 5 years, PLHIV (regardless of age) and contacts diagnosed with TBI, and (5) clinic referral for all symptomatic child contacts under 5 years, contacts ≥5 years of age with presumptive TB or who are ineligible for SIILTIBCY testing (Figure 2). An individual is defined as having presumptive TB if they are symptomatic, screened positive by AudibleHealth, or present with any chest X-ray abnormalities suggestive of TB.

Rule-out TB For children, TB symptom screening will be conducted using the WHO endorsed pediatric symptom screen that will focus on major pediatric symptoms including fever, cough of two weeks or more, wheeze, reduced playfulness or unusual fatigue, lethargy, visible mass in the neck, weight loss or failure to thrive based on clinical report and/or the personal health card. For adults symptom screening will assess whether participants are experiencing cough of any duration, sputum, haemoptysis, fever, night sweats and weight loss (36). In addition to symptom screening to rule-out TB disease the RAIsonance technology will be used to analyze forced coughs recorded via smartphone to detect disease-specific respiratory signatures. CHWs will use a smartphone for which they receive a secure link via SMS or email for each participant and submit an 8-second cough sample through a web-based interface; results are processed in the cloud and returned within one minute. The solution requires no hardware or consumables, operates on any smartphone with internet or offline capability. Providers manage the process through the TestHub portal, which also supports laboratory information management system and public health system integrations.

Presumptive TB patients will be referred to the local clinic or district referral hospital for further investigation of TB disease. As a result of those investigations, all children under 5 should be started on TPT (if TB disease is ruled out) or TB treatment, whereas for contacts ≥5 years, the decision to initiate TPT will be guided by the TBI testing result (per local practice and standard of care).

TBI testing Contacts are eligible for SIILTIBCY testing once active TB disease has been excluded. They may choose to undergo testing either at home or at the health facility where the index-patient is under treatment.

Home-based testing At-home testing will be carried out by a trained CHW, and when required by national policy, under the supervision of a healthcare provider capable of delivering basic life-saving care. The procedure will follow the SIILTIBCY Product Information (SIILTIBCY Product Information). SIILTIBCY will be prepared and administered via intradermal injection using the Mantoux technique (SIILTIBCY Product Information). The injection will be given under aseptic conditions with proper hand hygiene. Participants will be observed for at least 15 minutes after administration to monitor any anaphylactic reactions. A follow-up household visit will take place 48-72 hours later for measurement and recording of the test result.

Facility-based testing At the facility, testing will be performed by a trained healthcare provider following the SIILTIBCY Product Information using similar scrutiny as home-based testing. Participants will be requested to return to the facility 48-72 hours after injection for measurement and recording of the test result.

SIILTIBCY interpretation If there is an induration 48-72 hours after conducting the test, it can be seen as a raised area with clearly defined margin at and around the injection site. Although erythema can accompany the induration, only the induration will be measured by measuring the diameter of the induration transversely to the long axis of the forearm with a ruler. To allow ease with measurement, using a flexible (or easily bendable) ruler is recommended (can be provided from KNCV Global). Normally the induration and erythema will decrease after 4 days and disappear within 28 days after the injection. An induration of ≥ 5 mm will be considered as a positive test result, which indicates infection with Mycobacterium tuberculosis. Performing a test before 6 to 8 weeks from Mycobacterium tuberculosis exposure might result in a false-negative result. The risk of false-positive test results may increase if SIILTIBCY is repeated within 6 weeks. Therefore, an interval of at least 6 weeks should be observed between repeated tuberculosis skin tests. Within SCAF-TB no repeat testing is foreseen, but the added value of retesting might be explored in a small study sample if allowed by national policies.

Smart patch and TB Connect App The project will start with a 6-month run-in phase of SIILTIBCY testing for CHWs to get familiar with administering the test and reading the test results. After this phase, the investigators will introduce the MyLab Smart patch and AI-driven smartphone TB Connect app to assist CHWs in determining the test result (Figure 1). The TB Smart Patch is an adhesive patch that is to be placed on the forearm of the test subject by CHW. A QR code which is unique to each patch can be seen printed on the patch. Also, a circular flap which is demarcated by dotted line can be seen on the patch. This circular portion is removable and acts as the site for administration of test. The TB Connect app is used to register and monitor the subjects. The app has an integrated system for recording data and tracking each individual. It is to be used by the CHW providing SIILTIBCY. It is also used to check their eligibility for performing SIILTIBCY.

TPT initiation Participants who test TBI positive will be counselled and managed in the community by the community health teams. TPT dosing will be determined using a light-weight scale the community health teams bring to the household. Because TB drugs are regulated in both Tajikistan and Tanzania, including medications for TB prevention, the community health teams will not be allowed to maintain a drug supply in the community setting. Drug dispensing will need to occur in a two-step process where the community health team first evaluates/screens the contact in the home and then returns with the appropriate regimen and dose. If country regulations do not allow CHWs to prescribe TPT, the CHW will refer the participant to the clinic for initiation of TPT. People initiated on TPT will be followed at home monthly for the duration of treatment, as is standard of care. Communication of all community-based TPT initiations and referrals will be provided to the clinic TB nurse within one week and real-time when using the TB Connect app. Clinical outcome data will be extracted from the client records maintained by the community health teams and the clinic.

Descriptive statistics will be used to summarize the study cohort and detail the degree of engagement with the intervention. The investigators will calculate the proportion of eligible contacts who: (i) accepted to participate in the study (ii) have a registered SIILTIBCY test result (iii)had a positive result on either the AudibleHealth app or in the SIILTIBCY test (with and without the MyLab patch and TB Connect app), (iv) initiated on TPT after a positive SIILTIBCY test, (v) were lost to follow-up or prematurely stopped TPT treatment. Outcomes will be stratified by age, sex, socioeconomic status, and geographic region when data permits. Within the Adoption domain, the proportion of trained CHWs who delivered the intervention model will be computed. In the Maintenance domain the percentage of CHWs continuing to use the model between 6- and 12-months post-implementation will be reported using descriptive intervals.

Agreement SIILTIBCY reading CHW, HCW, and TB Connect app During the first 6 months of the study, the investigators will compare the readings of CHWs (without AI-assisted reading) and HCWs (e.g. nurses or TB focal persons). For these analysis the index test is SIILTIBCY reading by CHWs and the reference test is SIILTIBCY with HCW reading result. An induration size of ≥5 mm, whether measured by the health worker (reference standard) or by the CHW (index) is regarded as a positive test result. A positive test result in principle will be regarded as TBI present. True positive test results are defined as positive SIILTIBCY results (i.e., induration size ≥5 mm), both according to the CHW and to the HCW. False negative test results have an CHW induration size of <=5 mm (i.e. negative) while the HCW´s SIILTIBCY result is test-positive. False positive test results have an CHW´s induration size of ≥5 mm (i.e. positive) while the HCW´s SIILTIBCY result is test-negative. True negative results are scored negative both by the CHW and the HCW. Blinding: CHW´s who will read the SIILTIBCY result will be blinded to the HCW´s reading result.

After 6 months, if adequate agreement is observed between CHWs' and HCWs' readings of SIILTIBCY, the Smart Patch and TB Connect App will be introduced to facilitate the reading and assess the agreement between HCW reading (reference) and CHW with (index 1) and without (index 2) reading with the Smart Patch and TB Connect App using the similar approach as listed above. To ensure effective implementation, CHWs will be trained on the use of the Smart Patch and TB Connect App by the vendor, with ongoing mentorship provided by KNCV in-country staff. These applications will be customized, such as through translation and contextual adaptation, for use in their respective settings, enabling CHWs to confidently navigate the tools and contribute to accurate assessments.

CHWs will independently measure and document induration size, then perform a second measurement with the TB Connect application. Both readings will be performed on each participant diagnosed through the app. Agreement between the two methods will be computed at study conclusion. A positive test will mean that TBI is present. Classification rules will proceed as follows:

• True positives: induration size ≥5 mm according to both AI-assisted and CHW measurements.
• False negatives: induration size ≤5 mm by AI-assisted measurement but ≥5 mm by CHW measurement.
• False positives: induration size ≥5 mm by AI-assisted measurement but ≤5 mm by CHW measurement.
• True negatives: induration size ≤5 mm by both AI-assisted and CHW measurement.

Sample size Reliability will be quantified using Cohen's kappa; values of agreement equal to or greater than 0.80 will be interpreted as an indicator of adequate reliability, establishing AI-assisted SIILTIBCY readings as an acceptable screening tool for TBI. For assessing agreement in SIILTIBCY reading between CHWs and HCWs via Cohen's Kappa, an 0.8 expected kappa was assumed, with an accepted minimum of 0.6, a 25% prevalence of SIILTIBCY positivity in the population, a significance level of 0.05, and 80% statistical power, and adjusting for 10% drop-out, resulting in 110 participants per age category (5-15 years and ≥15 years) and a total sample size of 220 per country. For assessing agreement in SIILTIBCY reading between HCWs and CHWs with and without AI-assisted reading another 220 participants per country will be required using similar assumptions.

Evaluation of the AudibleHealth AI Cough Screening Application All participants who consent and are capable of providing a cough sample will be screened with AudibleHealth cough application to rule out TB disease. This cough screening will be in parallel to standard of care screening for TB disease (e.g. symptom screening with or without CXR). Community health workers (CHWs) will receive training on the use of the AudibleHealth application from the vendor, with ongoing mentorship provided by KNCV in-country staff. The application will be customized, such as through translation and contextual adaptation, for use in their respective settings, ensuring CHWs are equipped to carry out the screening effectively. The investigators will compare agreement between the cough application and standard of care TB disease screening. Results will be stratified by country, sex, age group, and presence of comorbid conditions.

Comparison TPT care cascade This study will compare TPT care cascade indicators across different SILTIBCY testing delivery modalities: (1) home-based testing with results read by CHWs, (2) home-based testing with results read by CHWs supported by the Smart Patch and TB Connect app, and (3) facility-based testing by HCWs. The indicators to be assessed include: proportion of eligible contacts identified (out of those estimated), proportion tested for TBI, proportion with a registered TBI result, proportion testing TBI positive, proportion of TBI-positive individuals offered TPT, proportion initiating TPT among those offered, and proportion completing TPT among those initiated. If TPT for household contacts above a certain age is introduced in non-participating clinics during the project, those clinics will serve as control sites. A before-and-after design is not possible, as TPT had not been provided to this group of contacts before this study.

Feasibility and Acceptability

To assess the feasibility and acceptability of the SCAF-TB work package activities, a mixed-methods sub-study will be conducted. The quantitative component will rely on indicators derived from routinely collected data throughout the study. The qualitative component will involve in-depth interviews with policymakers, healthcare workers, community health workers, and contact persons participating in the study.

An adapted RE-AIM framework combined with key constructs from the Implementation Outcomes Framework (IOF), adding feasibility and acceptability to better understand intervention delivery and reception, will be used. According to the IOF framework, feasibility is defined as "The extent to which a new treatment, or an innovation, can be successfully used or carried out within a given agency or setting" while acceptability is understood as "The perception among implementation stakeholders that a given treatment, service, practice, or innovation is agreeable, palatable, or satisfactory". In this extended version of the RE-AIM framework, feasibility and acceptability are understood as antecedents to implementation outcomes and are transversal components of adoption, implementation, and maintenance of an intervention.