Longitudinal NCD surveillance aims to fill a major evidence gap in Ghana, India and Nepal
Non-communicable diseases are already the dominant cause of death worldwide, and the burden is rising fastest in low- and middle-income countries where health systems are often built around acute care rather than chronic follow-up. In Ghana, India, and Nepal, the pressure is being driven by urbanisation, demographic change, and the steady spread of lifestyle-related risk factors. Cross-sectional datasets have documented the problem, but they cannot show how disease develops, clusters, or evolves across households and generations. This protocol for the GIN NCDs Study is designed to answer precisely those questions through a prospective, population-based cohort.
The study will enroll all residents aged 18 years and older in the selected study areas at baseline, with younger age groups planned for later waves. By anchoring recruitment in households rather than clinics, the cohort is positioned to capture both diagnosed and undiagnosed disease, along with the social and environmental exposures that shape risk before patients ever reach a facility. That matters in settings where people often present late, after hypertension, diabetes, or chronic respiratory disease has already caused irreversible harm.
Household recruitment and repeated measurement are built to show disease progression, not just prevalence
Data collection will occur at baseline, Year 2, and Year 4 using harmonised tools adapted from the World Health Organization’s STEPwise approach to Surveillance, or WHO STEPS. The protocol spans socio-demographic factors, climate-related exposures and impacts, lifestyle behaviors, health-seeking patterns, mental and physical health, and physical and anthropometric measurements. This breadth gives the cohort a chance to link structural determinants with biological outcomes rather than treating chronic disease as a narrow biomarker problem.
A longitudinal design is especially valuable for conditions such as hypertension, diabetes, chronic respiratory disease, mental health disorders, dementia, and multimorbidity. These are not static diagnoses. Blood pressure rises or falls with treatment access and adherence, glycemic control changes with diet and medication, respiratory symptoms can worsen with pollution or occupational exposure, and mental health often shapes whether patients can engage in care at all. Repeated measurement is what turns a descriptive survey into a platform for etiologic research.
One sentence captures the study’s central advantage: it can observe change where earlier datasets could only record a snapshot.
Sample size calculations target precision in rural, peri-urban, and urban strata
The protocol provides detailed calculations for the baseline cohort. The study is powered to estimate the prevalence of raised blood pressure with ±3% precision, assuming a design effect of 1.5, 10% to 15% non-response, and 10% annual attrition. Required baseline sample sizes are listed as 2,211 in India, 2,126 in Nepal, and 2,262 in Ghana. To preserve precision within rural, peri-urban, and urban strata, those totals will be tripled to 6,628 in India, 6,378 in Nepal, and 6,786 in Ghana.
These numbers are not just statistical formalities. They reflect the practical reality that chronic disease risk does not distribute evenly across geography. Urban households may face greater processed food intake, sedentary behavior, and air pollution exposure, while rural communities may encounter different barriers to diagnosis, treatment continuity, and specialist access. By stratifying across settlement types, the cohort can compare risk and care pathways across environments that are often grouped together in national estimates.
Sampling will use a two-stage cluster design with household-level enrollment. That approach supports population representativeness while keeping fieldwork feasible across three countries with different health system structures and population densities. It also creates the kind of linked household data that can reveal shared exposures inside families, including diet, tobacco exposure, activity patterns, and possible intergenerational health effects.
Multimorbidity, mental health, and dementia sit alongside cardiometabolic outcomes
The endpoint structure is broad and clinically relevant. Primary outcomes include hypertension, diabetes, chronic respiratory diseases, mental health disorders, dementia, and multimorbidity. This is an important decision because people living with chronic disease rarely fit into a single diagnostic silo. A patient with uncontrolled blood pressure may also live with depression, obesity, or impaired lung function; someone with diabetes may carry unrecognized cognitive decline that complicates medication adherence and self-management.
By placing mental health and dementia alongside cardiometabolic disease, the study recognizes that chronic illness is shaped by both biology and behavior. Depressive symptoms can reduce care-seeking and medication adherence. Cognitive impairment can blunt the capacity to follow treatment plans. Chronic respiratory disease can limit physical activity, which in turn worsens metabolic risk. A cohort that tracks these conditions together can show where comorbidity begins and how it clusters over time.
The protocol also includes climate-related exposures and impacts, an addition that has immediate relevance in South Asia and sub-Saharan Africa. Heat, pollution, food insecurity, and displacement can alter activity patterns, worsen respiratory symptoms, and interrupt continuity of care. Integrating those variables into a chronic disease cohort gives investigators a way to study how environmental stressors translate into long-term non-communicable disease burden.
Directed acyclic graph methods aim to strengthen causal interpretation
Analyses will use multivariable mixed-effects models guided by a directed acyclic graph, or DAG, framework. In practical terms, that means investigators are not relying on associations alone; they are specifying the pathways and confounders that matter before modeling begins. For a cohort of this size and complexity, that is a sensible guardrail against overadjustment and misleading estimates.
Mixed-effects modeling is particularly well suited to repeated measurements across households and communities. Individuals nested within clusters share exposures, services, and social conditions, so the statistical approach must account for correlation within the data. The same structure will allow investigators to examine within-person change over time, a key feature when studying disease progression, treatment access, and evolving risk profiles.
Policy value lies in planning for care before complications accumulate
The study is explicitly aligned with the WHO Global Monitoring Framework and Sustainable Development Goal Target 3.4, which calls for reducing premature mortality from non-communicable diseases through prevention and treatment. The practical value of the cohort will come from its ability to identify where risk concentrates, which groups remain underdiagnosed, and how multimorbidity emerges in real households rather than in clinic-based samples.
For health systems in Ghana, India, and Nepal, the findings should help with workforce planning, screening strategies, and the design of chronic care services that can follow patients over years rather than visits. For policymakers, the cohort may provide one of the few longitudinal datasets able to connect social conditions, climate exposure, mental health, and cardiometabolic disease in the same framework. For communities, the household design offers a direct view of how shared environments shape risk across life stages.
The unanswered question is not whether NCDs are rising; it is how quickly they are accumulating within families, which exposures are accelerating that trajectory, and where prevention can still interrupt it.