This is a retrospective, longitudinal quantitative study, drawing on secondary programmatic data from a TB treatment programme implemented by a non-governmental organisation (NGO) in Southern Madagascar. The data collection spanned from January to November 2022.

Madagascar is an island nation in sub-Saharan Africa, with approximately 29 million inhabitants [37]. It is one of the least developed countries globally, with 81% of its population living below the international national poverty line of 2.15 USD (2017 purchasing power parity) [37]. Over 60% reside in rural areas, with over 74% engaged in the primary sector, predominantly agriculture. [38]

In Madagascar, TB prevalence is high, with 233 cases per 100,000 people (2022 data), notably higher than the global average of 134 cases per 100,000 people [39]. In 2020, only half of the people with TB in Madagascar were notified [1], with effective treatment coverage estimated at only 59% [1]. The reasons for these figures are manifold and include low health literacy [40], long distances to treatment sites, limited diagnostic equipment and drugs, a shortage of trained TB healthcare workers, and TB-associated stigma and social exclusion [41,42,43]. Lastly, food insecurity poses a significant barrier to accessing and completing TB treatment, as people with TB often experience a loss of livelihood after diagnosis, leading to reduced nutrition intake and creating a cycle of vulnerability. [43]

In Southern Madagascar, encompassing Atsimo-Andrefana, Anosy, and Androy regions, the aforementioned challenges are exacerbated by poverty rates exceeding 90% and specifically weak infrastructure [38, 42]. In 2021 and 2022, the region endured a severe famine, impacting more than 1 million people suffering from acute food insecurity [44]. As of 2023, more than 334,000 people in Southern Madagascar continue to face emergency food insecurity [45]. Concurrently, TB prevalence in the region has been found to significantly surpass the national average, alongside elevated rates of patients lost to follow-up during treatment [41, 46]. In response to these challenges and in accordance with its most recent country strategic plan, the WFP prioritises nutritional support for the most vulnerable, including women, children, and people with TB, specifically in Southern Madagascar [46].

In this setting, the German Malagasy NGO Doctors for Madagascar has been collaborating with Madagascar’s National TB programme since 2019 to promote community-based TB care, aiming to enhance access and improve the quality of TB care for rural populations [41]. The programme is built on four pillars: (1) training and capacity building of healthcare workers across all levels of TB care, (2) facilitation of mobile TB screening and treatment clinics to reach populations in hard-to-reach areas, (3) community outreach through mass sensitisation and collaboration with local leaders to strengthen health literacy and knowledge about service availability, and (4) training and motivation of community health workers (CHWs) for patient screening and follow-up. Since 2020, the programme has partnered with the WFP to provide nutritional support for undernourished people with TB. Eligible people with TB (those with a BMI of less than 18.5 kg/m2 at TB screening; designated as timepoint M0) receive monthly rations of 0.6 L of vegetable oil and 6 kg of enriched soy- and wheat-based flour. According to the WFP, these ratios are calculated to include provisions for sharing, for instance, with family members potentially also experiencing malnutrition [47]. The flour rations are calculated to provide up to 939 kcal/day, no precise information is provided by the WFP on the energy content of the vegetable oil. These rations are provided during regular follow-up visits at treatment initiation (M1), after the second (M2), third (M3), fourth (M4), fifth (M5) months, and at treatment completion (M6). Nutritional support is contingent on treatment continuity and attendance at follow-up visits. In 2022 the programme provided nutritional support in the catchment areas of three TB diagnostic and treatment centres in the Atsimo-Andrefana region: Ampanihy-Ouest, Androka, and Bezaha. Figure 1 shows a map of these sites.

Map of the study zone in the Ampanihy Ouest and Betioky Atsimo districts within the Atsimo-Andrefana region in Madagascar. Blue markings: intervention districts within the Atsimo-Andrefana region, grey markings: non-intervention districts within the Atsimo-Andrefana region, asterisks: tuberculosis care centers [French centre de diagnostic et de traitement (CDT)], dots: sites where nutritional support was provided, square: capital of the Atsimo-Andrefana region, triangle: capital of Madagascar

This study used routine programme data collected on structured data entry forms by the programme team for each patient’s follow-up visit. The data covered basic socio-demographic information (community of residence, travel distance to a screening point, age, gender), clinical details [smear-positive pulmonary TB (TBP +), smear-negative pulmonary TB (TBP−), extrapulmonary TB (TEP)], treatment specifics, laboratory test results, nutritional status measurements [height and weight at the beginning of treatment and at each follow-up visit for adults, weight and mid-upper arm circumference (MUAC) for children], and recorded the type and weight of nutritional support received at each visit, with beneficiaries confirming the reception by signature or fingerprint. Height measurements for children and adults were only taken at screening but not repeated during treatment follow-up or at treatment completion.

A standardised data entry mask (i.e., an interface that asked a standard set of questions to all users to capture patient data) including data checks for completeness and value limits was configured in EpiData [48]. All data were independently transcribed twice by two programme team members, supervised by the project lead. The data, originally in French, were translated to English for analysis. The data were stored in a password-protected database and anonymised before being passed on for analysis.

The research team cross-verified data for discrepancies between entries. Verification was sought from the programme team for any discrepancies and extreme values (e.g., BMI < 12.0 kg/m2, BMI variation > 0.5 kg/m2 between two consecutive visits). Twenty-seven participants were excluded due to unresolved inconsistencies. Participants with missing data points for weight were included in the analysis, patients with missing data points for socio-demographic data, type of TB, or height were excluded (n = 12).

Data were included from all people who were diagnosed with TB between January 1, 2022, and November 30, 2022 in the catchment area of three TB diagnostic and treatment centres (Ampanihy-Ouest, Androka, Bezaha). Patients were included on a rolling basis, meaning that people were at different points of treatment between M0 (the screening visit) and M6 (treatment completion) at the time of data extraction (December 1, 2022). People diagnosed with TB outside the period from January 1, 2022 to November 30, 2022 were excluded.

Participants were classified into different levels of undernutrition depending on their age.

Adults were classified into well-nourished, undernourished, moderately undernourished, or severely undernourished categories, according to the WFP guidelines [49]: (1) Well-nourished (BMI equal to or above 18.5); (2) Undernourished (BMI below 18.5 and equal to or above 17.0) (hereafter: mildly undernourished); (3) Moderately undernourished (BMI below 17.0 and equal to or above 16.0); (4) Severely undernourished (BMI < 16.0).

Children under 5 were classified based on the National Malagasy guidelines, using both z-score for height-for-weight and mid-upper-arm circumference (MUAC) [50]: (1) Well-nourished with MUAC > 125 mm and a z-score > − 2 standard deviations (SD); (2) Moderate acute malnutrition (MAM) with MUAC between 115 and 125 mm and a z-score between − 3 and − 2 SD; (3) Severe acute malnutrition (SAM) with MUAC < 115 mm and a z-score < − 3 SD.

Children aged 5 to 18 were classified into moderately undernourished, severely undernourished, or well-nourished, according to the National Malagasy reference tables for weight for height [50].

We performed descriptive statistics for participants’ nutritional status and socio-demographic characteristics, including continuous variables that were not normally distributed, medians, and interquartile ranges (IQR). We performed chi-square-tests for nutritional status throughout the course of treatment, for (i) people with TB who received nutritional support through the WFP and those who did not, (ii) male and female people with TB, and (iii) by clinical type of TB (TBP + , TBP−, TEP). All analyses were performed in R Studio (Version: 2023.06.1). [51]

Ethical approval was received from the London School of Hygiene and Tropical Medicine ethics committee under registration number 29394. Local approval for the use of secondary data was obtained from the Comité Malgache d’Éthique pour les Sciences et les Technologies (CMEST) in Madagascar on April 26, 2023.