Tuberculosis (TB), caused by Mycobacterium tuberculosis, remains a major global health concern, particularly in regions burdened by poverty and under nutrition (Mendes et al., 2025). Despite being declared a global emergency by the World Health Organization in 1993, TB continues to claim millions of lives annually, with the highest incidence reported in Asia, followed by Africa and the Eastern Mediterranean (FAO, 2023; Jain et al., 2024). While only about 10% of those infected develop active TB, the risk is significantly elevated among individuals suffering from nutritional deficiencies and compromised immunity. India accounts for nearly 28% of the world's TB cases, making it the highest-burden country globally (Ministry of Health & Family Welfare 2023). Within India, Madhya Pradesh stands out as a state with persistently high TB prevalence, particularly in districts where malnutrition indicators such as anemia, stunting, and underweight status are also alarmingly elevated (NITI Aayog, 2022). This intersection of infectious disease and nutritional deprivation calls for a deeper, data-driven investigation into the underlying determinants of TB (Doke et al., 2023).

Madhya Pradesh continues to grapple with a high burden of tuberculosis (TB), posing a significant challenge to the state's public health infrastructure. Despite intensified efforts under the National TB Elimination Programme, including mobile screening units and community outreach, recent data from a 100-day detection drive revealed over 19,000 new TB cases, raising concerns about the feasibility of achieving the 2025 TB-free target. Tribal-dominated districts such as Jabalpur, Sheopur, and Gwalior remain hotspots, with the Saharia tribe exhibiting some of the highest pulmonary TB rates in the country. While over 5,000 gram panchayats have been declared TB-free, experts warn that underreporting, limited access to healthcare, and persistent malnutrition continue to fuel transmission, especially in remote and vulnerable populations (Ministry of Health & Family Welfare 2023).Recent research has emphasized the utility of Bayesian regression models in analyzing complex health data. Mohammed and Asfaw (2018) demonstrated the effectiveness of Bayesian Gaussian regression in identifying nutritional risk factors among children under five in Ethiopia. Similarly, Lasisi et al. (2015) applied Bayesian techniques to assess the nutritional status of Nigerian children, highlighting the method's robustness in capturing subtle associations in public health datasets.

Building on this foundation, the present study aims to quantify the nutritional determinants of TB in high-burden districts of Madhya Pradesh using Bayesian regression analysis. By integrating district-level data on anemia, stunting, underweight prevalence, and TB incidence, we seek to identify statistically significant predictors of TB burden. This approach not only enhances the precision of epidemiological insights but also informs targeted public health interventions that address both nutritional deficiencies and TB control. The findings are expected to contribute to evidence-based policymaking and resource allocation in one of India's most vulnerable regions.

Methodological Framework

This study aims to quantify the nutritional determinants of tuberculosis (TB) prevalence in high-burden districts of Madhya Pradesh using Bayesian regression analysis. By integrating district-level data on TB incidence and key malnutrition indicators, the study seeks to identify statistically significant predictors that contribute to elevated TB rates.

A. Data Sources and Collection

Secondary data were compiled from two primary sources:

Tuberculosis Data: TB case data from 2015 to 2022 were retrieved from the NIKSHAY portal maintained by the Central TB Division, Ministry of Health and Family Welfare, Government of India.

Nutritional Indicators: Malnutrition-related data were sourced from the National Family Health Survey (NFHS-5) and the State Nutrition Profile for Madhya Pradesh (NITI Aayog, 2022). These datasets include district-level figures for: Anemic women (ages 15–49), Anemic children (under 5 years), Stunted children, Underweight women and children

The analysis focuses on ten districts identified as high-burden: Bhopal, Shivpuri, Jabalpur, Rewa, Sagar, Satna, Betul, Mandla, Chhindwara, and Ujjain.

B. Data Preparation and Tabulation

Collected data were systematically organized and tabulated to facilitate statistical modeling. Published statistics were reclassified into uniform categories to ensure consistency across districts and variables. This step enabled the construction of a clean dataset suitable for regression analysis and visual interpretation.

C. Statistical Analysis

To explore the relationship between malnutrition and TB prevalence, the following statistical techniques were employed:

Descriptive Statistics: Mean plots and scatter diagrams were used to visualize distributions and preliminary associations between variables.

Correlation Analysis: Pearson and Spearman’s rank correlation coefficients assessed the strength and direction of relationships between TB incidence and nutritional indicators.

Bayesian Regression Modeling: Bayesian linear regression was applied to quantify the influence of each nutritional variable on TB prevalence. This method was selected for its ability to incorporate prior knowledge and manage uncertainty in parameter estimation. The model comparison metrics included posterior probabilities, Bayes Factors, and R² values to evaluate predictive strength.

All analyses were conducted using statistical software such as JASP (Version 0.19.3)  and validated through online platforms like StatsKingdom to ensure robustness and reproducibility. By combining descriptive, correlational, and Bayesian modeling approaches, this study aims to uncover the nutritional determinants most strongly associated with TB prevalence. The insights derived will support evidence-based public health strategies targeting both nutritional deficiencies and TB control in Madhya Pradesh.

The data from Table 1 reveals a significant burden of malnutrition across districts in Madhya Pradesh. Rewa, Satna, and Sagar stand out with the highest numbers of anemic women and children, stunted children, and underweight individuals. For instance, Rewa reported 441,000 anemic women and 203,000 anemic children, alongside 107,000 stunted children and 91,000 underweight children. These figures align with findings from the National Family Health Survey (NFHS-5), which identified Rewa and Satna among the top districts with public health concerns related to anemia and stunting (IIPS, 2021).

Table 1: Nutritional Status and Prevalence of Anaemia in Women and Children Across Selected Districts during 2019 to 2021 along with Data of TB  from Madhya Pradesh.

TB prevalence appears to correlate with nutritional deficits. Bhopal and Jabalpur, despite being urban centers, show high TB averages (9,154 and 6,055 respectively), possibly due to population density and undernutrition. Goswami and Jain (2023) found a strong positive correlation between malnutrition and TB incidence in Madhya Pradesh, with a Pearson coefficient of 0.817 and Spearman’s rank of 0.867. This supports the hypothesis that nutritional deficiencies especially anemia and underweight status are significant predictors of TB burden.  Bayesian regression analysis further confirms the strength of this relationship as given in Table 2. The best-fitting model includes all five nutritional predictors anemic women, anemic children, stunted children, underweight women, and underweight children and explains 95% of the variance in TB prevalence (R² = 0.950). This model had the highest posterior probability and Bayes Factor, indicating strong evidence for its predictive power. The use of Bayesian methods in this context reflects a growing trend in public health research. For example, Mohammed and Asfaw (2018) applied Bayesian Gaussian regression to assess malnutrition among Ethiopian children under five, highlighting its robustness in handling complex health data. Similarly, Lasisi et al. (2015) demonstrated the effectiveness of Bayesian regression with Gaussian and Binomial links in analyzing nutritional status among Nigerian children, reinforcing its relevance in low-resource settings.

Table  2: Comparative Bayesian Linear Regression of TB Prevalence Using Anemia, Stunting, Underweight, and Malnutrition Indicators.

Note.  Table displays only a subset of models; to see all models, select "No" under "Limit No. Models Shown".

To understand the individual contributions of each variable, Table 3 presents the posterior summaries of coefficients from the Bayesian regression: These results show that anemia in children and underweight status in women have the strongest positive associations with TB prevalence. Interestingly, stunting and underweight in children show negative coefficients, which may reflect complex interactions or confounding factors in the dataset. The inclusion probabilities and Bayes Factors suggest that all five indicators are relevant, with anemic children emerging as the most influential predictor.

India continues to face a severe nutrition crisis. According to the UN’s State of Food Security and Nutrition in the World report, 37.4 million children under five are stunted, and 53.7% of women aged 15–49 suffer from anemia (FAO, 2023). These figures have worsened over the past decade, despite interventions like POSHAN Abhiyaan. The persistent high rates of anemia and stunting in Madhya Pradesh mirror national trends, especially among marginalized communities where access to nutritious food and healthcare remains limited.

Table 3: Posterior Summaries of Coefficients from Bayesian Linear Regression Predicting TB Prevalence Using Nutritional Indicators.

The intersection of TB and malnutrition demands integrated public health strategies. Districts like Rewa, Satna, and Sagar should be prioritized for nutrition-based TB interventions. Evidence from Madhya Pradesh’s State Nutrition Profile and the Central TB Division suggests that targeted programs addressing anemia and child undernutrition could significantly reduce TB incidence (NITI Aayog, 2022; Ministry of Health & Family Welfare 2023). Strengthening maternal and child health services, improving food security, and ensuring timely TB diagnosis and treatment are critical steps forward.

This study highlights the strong link between malnutrition and TB in Madhya Pradesh, suggesting future research should expand to other high-burden regions, include more socioeconomic factors, and use longitudinal data for deeper insights. Bayesian models can guide targeted interventions, policy planning, and mobile health tools to improve TB control through nutritional support.