The syndemic of structural vulnerabilities & HIV risk among men who inject drugs in India: Reflections from a population-based study

Sir,

The article by Chakrapani et al¹, titled ‘The syndemic of incarceration, violence victimisation, needle/syringe sharing & HIV infection: A population-based study of men who inject drugs in India’, published in the March 2025 issue of the Indian Journal of Medical Research addresses a much-needed application of syndemic theory to a nationally representative dataset of people who inject drugs (PWID) in India.

However, some methodological concerns limit the generalisability and contemporary relevance of the findings. The reliance on data collected in 2014-15 raises questions about the study’s applicability in 2025, especially given the substantial developments in India’s HIV prevention landscape over the past decade. Expanded opioid substitution therapy, broader needle–syringe programmes, and evolving legal and social responses to drug use likely influenced behaviours and risk patterns, potentially altering the relationships explored in the study^2,3. Additionally, the cross-sectional design provides valuable prevalence data but cannot establish causal relationships between incarceration, violence, needle sharing, and HIV infection. This limitation risks conflating correlation with causation, particularly in a complex syndemic framework, where temporality is critical. For example, it remains unclear whether incarceration precedes or follows needle-sharing behaviours, a distinction vital for designing targeted interventions. Additionally, reliance on self-reported data for sensitive behaviours such as needle sharing and violence victimisation introduces recall bias and social desirability bias. Given the stigma surrounding drug use and HIV in India, underreporting of these behaviours is likely, which may underestimate the true prevalence of risk factors and weaken the observed associations⁴. Also, the operationalisation of ‘severe violence victimisation’ as six or more incidents in a year seems arbitrary and may have misclassified individuals who experienced fewer but equally traumatic events.

While the study offers valuable insights, the exclusion of female PWID, who often face compounded risks including sexual violence, limits the generalisability of findings. Moreover, selection bias cannot be completely ruled out, the time-location cluster sampling approach may have excluded PWID in more private or hidden settings, thereby underrepresenting more marginalised subgroups. The potential influence of unmeasured confounders such as mental health conditions, recent harm reduction access, or socioeconomic deprivation further complicates the interpretation of findings⁵.

Although the authors tested additive and multiplicative interactions, the interpretation of semi-elasticities as percentage point changes may confuse readers unfamiliar with such statistical methods. The subgroup driving the significant three-way interaction represented only 0.3 per cent of the sample, raising concerns about statistical power and the stability of estimates. In addition, while HIV programme exposure was included as a covariate, the study did not examine its potential moderating effects on the syndemic relationships, an omission that reduces programmatic relevance given India’s longstanding targeted interventions for PWID.

To advance this important line of inquiry, future research should prioritise the use of recent, longitudinal data to capture evolving patterns and establish causal pathways. Inclusion of female and rural PWID would broaden applicability, and triangulating self-reported behaviours with objective measures such as HIV test records would strengthen validity. Advanced statistical models, including structural equation modelling, could better capture complex syndemic interactions. Finally, ethical considerations, especially in research involving criminalised populations must be revisited in light of changes in India’s legal and policy environment.