What Health Looks Like in Matrilineal Societies of Northeast India?

ABSTRACT

Matrilineal kinship systems are often assumed to confer health advantages, yet their health implications remain poorly understood. We examined the associations between matriliny‐related arrangements and nutritional, cardiometabolic, and child health outcomes in Meghalaya, Northeast India. Data from the Meghalaya NFHS‐5 included 8802 women, 1409 men, and 5265 children from the Khasi and Garo populations. Matriliny was operationalized using matrilocal residence and female ownership of house and land. Multivariable logistic regression assessed associations with nutritional status, central obesity, hypertension, diabetes, anemia, and childhood morbidity. Meghalaya had a higher prevalence of matrilocal residence and female ownership of house and land than most Indian states. Matriliny‐related indicators were associated with higher levels of women's empowerment, lower tobacco use, and reduced unhealthy food consumption. Compared with female land ownership, women in households with male and joint land ownership had 2.5‐fold and 2.4‐fold higher odds of underweight, respectively, while joint land ownership was associated with 42% lower odds of high waist circumference. Joint house ownership was associated with 4.7‐fold higher odds of underweight among men and 43% lower odds of diabetes among women. Children living in non‐matrilocal and male‐owned house had 32% and 38% lower odds of reported sickness, respectively. Hypertension and anemia were independently associated with broader demographic and socioeconomic factors. Matriliny was associated with gendered, outcome‐specific patterns of health, with asset ownership central to nutritional security, whereas cardiometabolic and child health outcomes were more consistently associated with broader demographic, socioeconomic, and structural factors, highlighting kinship institutions as consequential but contingent influences on health.

1. Introduction

Kinship systems are foundational social institutions that organize descent, residence, inheritance, and obligations, thereby shaping access to resources, patterns of cooperation, and the distribution of power within households and communities (Radcliffe‐Brown ¹⁹⁵⁰; Lowes ²⁰²⁰). Far from being culturally neutral arrangements, kinship structures are increasingly recognized as central to gender relations, social inequality, and population health, influencing nutritional status, disease vulnerability, and healthcare access through pathways that extend beyond individual behaviors or biomedical factors (Reynolds et al. 2020; Lowes ²⁰²⁰). Understanding how kinship systems structure health is therefore critical for explaining heterogeneity in disease burden and health disparities in low‐ and middle‐income countries (LMICs), particularly amid ongoing social and epidemiological change.

Matrilineal kinship systems, where descent, inheritance, and often residence are traced through women, have received attention for their potential influence on gendered resource allocation, caregiving practices, and intra‐household power dynamics in ways that differ fundamentally from patrilineal systems (Mattison 2011; Surowiec et al. ²⁰¹⁹). Empirical studies have linked matriliny to improved outcomes for women and children, including greater autonomy, lower exposure to domestic violence, improved reproductive outcomes, and more favorable cardiometabolic profiles (Leonetti et al. 2007; Reynolds et al. ²⁰²⁰; Lowes ²⁰²⁰; Lamarque et al. ²⁰²⁵). However, matrilineal societies are neither uniformly egalitarian nor insulated from broader structural inequalities, challenging assumptions of automatic or universal health advantages (Singh 2020; Thongni and Subudhi ²⁰²⁵).

Classic anthropological scholarship anticipated this more nuanced understanding of matriliny. Ethnographic studies among the Khasi and Garo communities characterized matriliny as an integrated system of descent, residence, inheritance, and kinship obligations rather than a simple inversion of patriarchy, with lineage continuity maintained through the mother and ancestral property inherited by the youngest daughter (ka khadduh among the Khasi and Nokna among the Garo) (Gurdon 1907; Nakane ¹⁹⁶⁸). Crucially, authority within these systems remained distributed across complementary gendered domains rather than monopolized by either women or men. Gough further conceptualized matriliny as a configuration of residence patterns, property transmission, and authority relations that structure access to material resources and social security over the life course (Gough 1961, 1968). This perspective suggests that matriliny is likely to generate selective and outcome‐specific effects rather than uniform advantages across all domains of social life and health. This institutional perspective informs the present study's operationalization of matriliny through patterns of residence and ownership of immovable assets.

These questions are particularly urgent given the scale and gendered nature of contemporary health burdens in LMIC settings characterized by overlapping burdens of undernutrition and non‐communicable diseases (NCDs). India exemplifies this transition, with persistent nutritional deficiencies coexisting alongside rising obesity, hypertension, and diabetes (Rahman and Talukdar 2025; Ramesh and Kosalram ²⁰²³). While anemia remains highly prevalent, especially among women, it represents a broader, gendered health regime shaped by life‐course exposures, socioeconomic position, and access to care rather than a condition uniquely structured by kinship arrangements alone (Gardner et al. 2023; WHO ²⁰²⁵; International Institute for Population Sciences (IIPS) and ICF ^2021a). Understanding whether and how kinship institutions shape vulnerability across multiple health outcomes therefore requires an integrated, multi‐outcome perspective.

India also provides a distinctive setting in which predominantly patrilineal systems coexist with a small number of long‐standing matrilineal societies. Among the most prominent are the Khasi and Garo populations of Meghalaya in Northeast India, where matrilocal residence and female‐centered inheritance of house and land persist within a largely patrilineal national framework (Nakane 1968; Leonetti et al. ²⁰⁰⁷; Brulé and Gaikwad ²⁰²¹). Simultaneously, Meghalaya is undergoing a rapid epidemiological transition, marked by the persistence of anemia and undernutrition alongside rising burdens of obesity, hypertension, and diabetes (International Institute for Population Sciences (IIPS) and ICF ^2021b; Gupta et al. ²⁰²³; Rahman and Talukdar ²⁰²⁵), providing a crucial setting to examine whether, and how, matriliny operates as a contemporary institutional context shaping health.

Existing research has not resolved this question. Much of the literature on matriliny and health remains either descriptive or narrowly focused, often examining single outcomes such as reproductive health or political participation without examining multiple health domains or the mechanisms through which kinship arrangements operate (Leonetti et al. 2007; Brulé and Gaikwad ²⁰²¹). Studies that do address health frequently treat matriliny as a binary cultural attribute rather than as a set of institutional practices, such as residence patterns, property ownership, and household authority, that may exert differential effects across genders and generations (Lowes 2020; Reynolds et al. ²⁰²⁰). Public health research in India, meanwhile, has largely examined anemia and NCDs through biomedical, behavioral, or socioeconomic lenses, with limited attention to kinship systems, gendered property relations, or intra‐household power dynamics (Ramesh and Kosalram 2023; Seenappa et al. ²⁰²⁴).

An important pathway through which these institutional arrangements may shape health is women's empowerment and structural access to healthcare. While empowerment is increasingly recognized as a determinant of health, it is often operationalized in ways that overlook local kinship arrangements, shared asset ownership, and collective decision‐making processes (Zimmerman 1995; Cattaneo and Chapman ²⁰¹⁰). While standardized indices such as the Survey‐based Women's Empowerment (SWPER) Index have advanced comparative research, their applicability within matrilineal contexts remains insufficiently examined (Ewerling et al. 2017, 2020). Whether matriliny improves health primarily through enhanced material control, greater decision‐making autonomy, reduced structural barriers to healthcare, or whether these mechanisms operate selectively across health outcomes remains an open empirical question.

This study addresses these gaps by examining matriliny as a set of measurable, gendered institutional arrangements, focusing on matrilocal residence, female ownership of house and land. Using nationally representative data from the NFHS‐5, the study situated Meghalaya within the broader Indian context and examined how these arrangements relate to socio‐demographic characteristics, socioeconomic status, women's empowerment, structural barriers to healthcare access, and health outcomes among women, men, and children. By analyzing nutritional status, central obesity, hypertension, diabetes, anemia, and childhood morbidity together, the study assesses where matriliny confers protection or risk, where its influence is limited, and where broader structural determinants predominate.

In doing so, this paper advances three contributions: it demonstrates that matriliny in Meghalaya is internally heterogeneous and not uniformly protective across health domains; it shows that women's empowerment and structural access to healthcare are closely related to matriliny‐related arrangements, although their effects vary across health outcomes; and by analyzing women, men, and children simultaneously, the study moves beyond women‐centred narratives to reveal how matrilineal systems redistribute health risks and protections across genders and generations.

2. Materials and Methods

2.1. Study Design and Data Source

This study used data from the fifth round of the NFHS‐5 (2019–2021), a nationally representative, cross‐sectional survey conducted under the Demographic and Health Survey (DHS) programme by the Ministry of Health and Family Welfare (MoHFW), Government of India (GoI), in collaboration with the International Institute for Population Sciences (IIPS). NFHS‐5 provides demographic, socioeconomic, health, and nutrition data for 707 districts across 28 states and 8 union territories in India (International Institute for Population Sciences (IIPS) and ICF ^2021a).

NFHS‐5 employed a two‐stage stratified cluster sampling design, with villages in rural areas and Census Enumeration Blocks (CEBs) in urban areas selected as primary sampling units (PSUs), followed by systematic random sampling of households within each PSU. As the primary objective of this study was explanatory inference focused on associations rather than population‐level prevalence estimation, survey weights were not applied in either descriptive or regression analyses. However, variables related to the sampling design, including socio‐demographic and socioeconomic variables, behavioral factors, women's empowerment, and structural barriers, were included as covariates to account for differential sampling across strata, consistent with DHS analytical practice.

The cross‐sectional design allows assessment of associations between matriliny‐related indicators and health outcomes at a single point in time but precludes causal inference; findings should therefore be interpreted as evidence of association rather than causation.

2.2. Sample Selection and Inclusion Criteria

The analysis focused on Meghalaya, a state in Northeast India characterized by long‐standing matrilineal systems among Khasi and Garo populations. Three NFHS‐5 datasets were used: the women's (Individual Recode), the men's (Household Member Recode), and the children's (Children's Recode) datasets, all obtained from the DHS Program repository (https://dhsprogram.com/data/available‐datasets.cfm).

The analysis was restricted to Scheduled Tribe respondents in Meghalaya reporting Khasi or Garo as their native language in order to capture populations historically characterized by institutionalized matrilineal systems. Women aged 15–49 years were included. Women who were pregnant at the time of the survey, had missing biomarker data (blood glucose, hemoglobin, or blood pressure), or reported chronic illness (thyroid disorders, respiratory disease, kidney disease, heart disease, or cancer) were excluded, yielding a final sample of 8802 women. Men aged 15–54 years were selected from the household members dataset. Exclusion of individuals with missing biomarker data or reported chronic illness yielded a final analytic sample of 1409 men. Children under 5 years of age residing in Meghalaya were included. Children with missing health or anthropometric information or reported chronic illness were excluded, yielding a final sample of 5265 children. Information on child health and care‐seeking was reported by mothers. All analyses were conducted using complete‐case observations for each outcome.

2.3. Health Data and Definition of Health Endpoints

Multiple health domains were examined.

Nutritional status was assessed using body mass index (BMI; kg/m²) and categorized using Asia‐Pacific cut‐off values as underweight (< 18.5), normal (18.5–22.9), overweight (23.0–24.9), or obese (≥ 25.0) (World Health Organisation. Regional Office for the Western Pacific ²⁰⁰⁰). Central Obesity was assessed using waist circumference (WC) and waist‐to‐hip ratio (WHR). High WC was defined as > 80 cm for women and > 90 cm for men, while high WHR was defined as > 0.80 for women and > 0.90 for men, following WHO Asian‐specific guidelines (World Health Organisation ²⁰¹¹). These cut‐offs were used to account for the South Asian metabolic phenotype, characterized by increased visceral adiposity and cardiometabolic risk at relatively lower BMI and waist measures, reflecting population‐specific biological and sociocultural determinants of body composition (World Health Organisation Expert Consultation ²⁰⁰⁴; Misra et al. ²⁰²⁵). Receiver operating characteristics (ROCs) analyses further indicated lower optimal thresholds for cardiometabolic outcomes in this population (Table S1).

Blood pressure status was classified according to JNC‐7 criteria as normal (< 120/80 mmHg), pre‐hypertension (120–139/80–89 mmHg), or hypertension (≥ 140/90 mmHg), based on the average of second and third readings (Chobanian et al. 2003). Glycemic status was classified using glucose measurements. Fasting status was derived from self‐reported time since last food or caloric beverage intake (≥ 8 h; plain water permitted). Fasting glucose levels were classified as normal (< 100 mg/dL), pre‐diabetic (100–125 mg/dL), or diabetic (≥ 126 mg/dL), according to WHO guidelines (World Health Organization ²⁰⁰⁶). For non‐fasting participants, blood glucose was treated as random and categorized as normal (< 140 mg/dL), elevated (140–199 mg/dL), or diabetic (≥ 200 mg/dL). For analytical purposes, pre‐diabetic and elevated categories were combined into a single group. Anemia was defined using hemoglobin concentration following WHO guidelines, with standard adjustments for altitude and smoking (World Health Organization ²⁰²⁴). As pregnant women were excluded from the analysis, anemia among women was defined using the non‐pregnant cut‐off of < 120 g/L, compared with < 130 g/L among men. Severity classifications were identical for both sexes for moderate (80–109 g/L) and severe anemia (< 80 g/L), while mild anemia was defined as 100–119 g/L among women and 110–129 g/L among men.

Child health outcomes included recent sickness and care‐seeking behavior. Children were coded as sick if they experienced diarrhea, fever, and/or symptoms suggestive of acute respiratory infection in the 2 weeks preceding the survey. These conditions were selected as they represent the leading causes of preventable child morbidity in low‐ and middle‐income settings and are consistently captured across DHS surveys, enabling comparability. Care‐seeking behavior was assessed among children experiencing illness and defined as whether treatment was sought from any public or private healthcare provider and/or medication was administered.

2.4. Variables and Measures

Consistent with the study's biocultural framework, explanatory variables were selected to explore associations between institutional dimensions of matrilineal kinship systems, together with socio‐demographic, socioeconomic, behavioral, women's empowerment, and healthcare access factors, and multiple health outcomes.

Matriliny‐related indicators were conceptualized as an institutional arrangement governing residence, inheritance, and control over immovable assets (Stivens 2025). Given the absence of direct kinship measures in NFHS‐5, matriliny was operationalized using three proxy indicators. Matrilocal residence was approximated using the variable “years lived in current place of residence,” with women reporting that they had “always” lived in their current location classified as matrilocal. This variable was available only for women and children, and absent in the men's dataset. This proxy may misclassify some cases and does not fully capture the normative complexity of matrilocal residence. Additional indicators included land ownership and house ownership, coded as male‐owned, female‐owned, or jointly owned.

Socio‐demographic and socioeconomic variables common to both sexes included place of residence (urban or rural), sex of household head, age group, religion, marital status, household size, respondent's education, and household wealth index in quintiles.

Behavioral factors common to both sexes included substance use variables, including alcohol and tobacco consumption, coded as binary indicators. Dietary variables, including frequency of consumption of selected food groups and composite indices of healthy and unhealthy food consumption based on reported intake frequency, were available only for women and not captured in the men's dataset. Healthy food consumption comprised reported consumption of milk or curd, pulses or beans, dark green leafy vegetables, fruits, eggs, fish, and chicken or meat, whereas unhealthy food consumption comprised aerated drinks and fried foods.

2.4.1. Measurement of Women's Empowerment (SWPER‐M) and Structural Barriers to Healthcare Access

A modified Survey‐based Women's Empowerment Index (SWPER‐M) was constructed to capture empowerment dimensions relevant to matrilineal contexts. The index included three domains: decision‐making autonomy, attitudes toward violence, and social independence. Items were coded following established SWPER conventions, with adaptations to reflect local context.

The index demonstrated acceptable internal consistency (Cronbach's α = 0.761). Sampling adequacy was supported by the Kaiser‐Meyer‐Olkin statistic (KMO = 0.789), and Bartlett's test of sphericity was significant (χ ² = 4310.18, p < 0.001), indicating suitability for principal component analysis (PCA). PCA was used to derive domain scores, which were standardized and categorized into low, moderate, and high empowerment based on interquartile ranges. The SWPER‐M was included only in women's models, as no equivalent or conceptually comparable empowerment measure was available for men in the NFHS‐5 dataset.

Structural barriers to healthcare access were assessed only among women, as comparable variables were not available in the men's NFHS‐5 dataset. These included reported difficulties in obtaining permission to seek care, financial constraints, distance to health facilities, transport, need for accompaniment, and concerns regarding availability of female health providers, general healthcare providers, and medicines. Variables were treated as categorical indicators reflecting perceived severity of barriers to healthcare access.

2.5. Statistical Analysis

All statistical analyses were performed using IBM SPSS Statistics version 28. Descriptive statistics summarized matriliny‐related indicators (matrilocality, land, and house ownership), socio‐demographic and socioeconomic characteristics, behavioral factors, women's empowerment measures, structural barriers to healthcare, and health outcomes. Categorical variables are presented as frequencies and percentages. Pearson's chi‐square (χ ²) test examined bivariate associations.

Separate binary logistic regression models were fitted for each health outcome. Multivariable analyses were conducted as follows. Separate models were fitted for women and men using explanatory variables common to both sexes, including matriliny‐related indicators, socio‐demographic, socioeconomic, and behavioral characteristics. The models for women were expanded to include additional variables available only in the women's dataset, including matrilocality, women's empowerment (SWPER‐M), structural barriers to healthcare access, and dietary characteristics. Separate multivariable models were fitted for children's health outcomes using explanatory variables available in the children's dataset. For explanatory variables common to both sexes, sex‐by‐covariate interaction terms were evaluated in binary logistic regression models. Odds ratios (ORs), adjusted odds ratios (aORs), and 95% confidence intervals (CIs) are reported.

ROC curve analyses were conducted separately for women and men to evaluate the ability of BMI, WC, and WHR to discriminate individuals with hypertension and diabetes. The area under the curve (AUC) was calculated to assess the diagnostic performance of each anthropometric indicator. Optimal sex‐specific cut‐off values were determined using the Youden Index (sensitivity + specificity −1), which identifies the threshold that maximizes the combined sensitivity and specificity of each measure.

Statistical significance was defined as p < 0.05. Given the cross‐sectional design, the findings are interpreted as adjusted associations rather than causal effects.

3. Results

3.1. Mapping Matriliny in India

Figure 1 illustrates the spatial distribution of three matriliny‐related indicators measured across Indian states: matrilocal residence, female ownership of house and land. Compared with other Indian states, Meghalaya had the highest prevalence of matrilocal residence (75.4%) and among the highest levels of female ownership of house (44.4%) and land (31.9%) (Figure 1). Across Indian states, female ownership of house was consistently more common than female ownership of land.

FIGURE 1.

State‐wise distribution of matrilocal residence and female ownership of assets in India. (A) Percentage of households with matrilocality regions. (B) Percentage of households with female house ownership. (C) Percentage of households with female land ownership. Meghalaya is highlighted in each panel because it is the study area. Maps were generated using NFHS‐5 (2019–2021) state‐level estimates.

3.2. Ethnic Context, Matrilocality and Asset Ownership in Meghalaya (Figure 2)

FIGURE 2.

Distribution of matriliny‐related characteristics in Meghalaya. (A) Distribution of matrilocal and non‐matrilocal residence among children and women. (B) Distribution of house ownership (male, female, and joint ownership) among children, women, and men. (C) Distribution of land ownership (male, female, and joint ownership) among children, women, and men. (D) Distribution of matrilocal residence, female house ownership, and female land ownership among Khasi and Garo households for women, men, and children. Percentages are based on NFHS‐5 (2019–2021) data. Information on children was reported by their mothers. The sample study was evenly divided between Khasi and Garo respondents (Khasi women: 52.3%; Garo women: 47.7%; Khasi men: 51.8%; Garo men: 48.2%). Among children, 70.6% had Khasi‐speaking mothers and 29.4% had Garo‐speaking mothers.

Within Meghalaya, matrilocal residence and female‐centred asset ownership were common, although their distribution varied by sex and ethnic group. Among women, 77.9% reported matrilocal residence (Figure 2A), whereas ownership of houses and land was less universal, though still substantial, with 68.4% of women reporting ownership of a house and 50.6% reporting ownership of land (Figure 2B,C). Among men, exclusive ownership of house and land was relatively uncommon. Only 12.9% of men reported male‐owned house and 8.3% reported male‐owned land. The majority of men reported that houses (62.4%) and land (54.9%) were owned by women, with joint ownership comprising a substantial minority. Sex‐disaggregated estimates of matrilocal residence among men were unavailable in the NFHS dataset, precluding direct assessment of men's residential positioning within matrilineal households.

Among children, nearly two‐thirds (64.1%) resided in households where mothers reported matrilocal residence (Figure 2A). More than half (56%) lived in houses owned by women, with an additional 20.6% residing in jointly owned houses, whereas fewer than one‐quarter (23.4%) lived in male‐owned households (Figure 2B). Land ownership was similarly distributed, with 37.8% associated with female‐owned land, 17.1% with joint ownership, and 45.1% with land owned by men (Figure 2C).

Matrilocal residence was more common among Garo women than Khasi women (87.3% vs. 67.7%) and among Garo children than Khasi children (78.4% vs. 58.8%). In contrast, male respondents from Khasi households more frequently reported female ownership of house (70.1% vs. 50.1%) and land (67.2% vs. 38.1%) (Figure 2D).

3.3. Socio‐Demographic and Socioeconomic Correlates of Matriliny

Tables 1 and 2 summarize the associations between socio‐demographic and socioeconomic characteristics and matriliny‐related indicators among women and men. Among women, matrilocal residence and female ownership of house and land were significantly more common in rural than urban areas (Table 1, all p < 0.001). Female‐headed households had a significantly higher prevalence of female‐owned houses in both sexes, and also in female‐owned land among males (Tables 1 and 2, all p < 0.001).

TABLE 1.

General characteristics of women in Meghalaya according to matriliny‐related indicators.

General Characteristics	Matrilocality		House ownership			Land ownership
	Matrilocal	Non‐matrilocal	Male‐owned	Female‐owned	Jointly‐owned	Male‐owned	Female‐owned	Jointly‐owned
	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
Sociodemographic and socioeconomic factors
Place of residence
Urban	642 (9.4)	344 (17.7)	78 (18.4)	19 (6.7)	62 (9.8)	119 (18)	5 (3.3)	35 (6.7)
Rural	6218 (90.6)	1598 (82.3)	345 (81.6)	266 (93.3)	569 (90.2)	542 (82)	147 (96.7)	491 (93.3)
χ 2, p‐value	106.218, < 0.001*		27.341, < 0.001*			48.129, < 0.001*
Sex of household head
Male	4405 (64.2)	1057 (54.4)	231 (54.6)	142 (49.8)	436 (69.1)	349 (52.8)	82 (53.9)	378 (71.9)
Female	2455 (35.8)	885 (45.6)	192 (45.4)	143 (50.2)	195 (30.9)	312 (47.2)	70 (46.1)	148 (28.1)
χ 2, p‐value	61.535, < 0.001*		39.214, < 0.001*			47.158, < 0.001*
Age group (in years)
15–24	2644 (38.5)	585 (30.1)	257 (60.8)	46 (16.1)	173 (27.4)	312 (47.2)	19 (12.5)	145 (27.6)
25–34	2029 (29.6)	718 (37)	113 (26.7)	101 (35.4)	228 (36.1)	195 (29.5)	59 (38.8)	188 (35.7)
35–44	1540 (22.4)	486 (25)	42 (9.9)	97 (34)	174 (27.6)	118 (17.9)	52 (34.2)	143 (27.2)
45–49	647 (9.4)	153 (7.9)	11 (2.6)	41 (14.4)	56 (8.9)	36 (5.4)	22 (14.5)	50 (9.5)
χ 2, p‐value	64.144, < 0.001*		206.626, < 0.001*			96.840, < 0.001*
Religion
Christian	6529 (95.2)	1797 (92.5)	398 (94.1)	266 (93.3)	610 (96.7)	620 (93.8)	142 (93.4)	512 (97.3)
Other	331 (4.8)	145 (7.5)	25 (5.9)	19 (6.7)	21 (3.3)	41 (6.2)	10 (6.6)	14 (2.7)
χ 2, p‐value	20.644, < 0.001*		6.230, 0.044*			9.057, 0.011*
Current marital status
Married	4438 (64.7)	1531 (78.8)	189 (44.7)	209 (67.2)	523 (86.4)	385 (58.2)	395 (75.1)	141 (92.8)
Never married	2422 (35.3)	411 (21.2)	234 (55.3)	102 (32.8)	82 (13.6)	276 (41.8)	131 (24.9)	11 (7.2)
χ 2, p‐value	138.692, < 0.001*		202.709, < 0.001*			84.652, < 0.001*
Family size
< 5	2173 (31.7)	820 (42.2)	129 (30.5)	130 (45.6)	230 (36.5)	226 (34.2)	74 (48.7)	189 (35.9)
≥ 5	4687 (68.3)	1122 (57.8)	294 (69.5)	155 (54.4)	401 (63.5)	435 (65.8)	78 (51.3)	337 (64.1)
χ 2, p‐value	75.041, < 0.001*		16.789, < 0.001*			11.327, 0.003*
Educational level
No education	850 (12.4)	306 (15.8)	37 (8.7)	58 (20.4)	93 (14.7)	88 (13.3)	33 (21.7)	67 (12.7)
Primary	1268 (18.5)	460 (23.7)	68 (16.1)	73 (25.6)	129 (20.4)	120 (18.2)	40 (26.3)	110 (20.9)
Secondary	4204 (61.3)	1024 (52.7)	273 (64.5)	134 (47)	370 (58.6)	390 (59)	76 (50)	311 (59.1)
Higher	538 (7.8)	152 (7.8)	45 (10.6)	20 (7)	39 (6.2)	63 (9.5)	3 (2)	38 (7.2)
χ 2, p‐value	52.574, < 0.001*		40.298, < 0.001*			22.898, < 0.001*
Wealth index
Lowest	1248 (18.2)	479 (24.7)	71 (16.8)	62 (21.8)	118 (18.7)	137 (20.7)	28 (18.4)	86 (16.3)
Lower	1416 (20.6)	424 (21.8)	80 (18.9)	54 (18.9)	138 (21.9)	127 (19.2)	29 (19.1)	116 (22.1)
Middle	1274 (18.6)	377 (19.4)	78 (18.4)	54 (18.9)	114 (18.1)	117 (17.7)	35 (23)	94 (17.9)
Higher	1410 (20.6)	347 (17.9)	96 (22.7)	66 (23.2)	141 (22.3)	141 (21.3)	38 (25)	124 (23.6)
Highest	1512 (22)	315 (16.2)	98 (23.2)	49 (17.2)	120 (19)	139 (21)	22 (14.5)	106 (20.2)
χ 2, p‐value	64.092, < 0.001*		7.369, 0.497			10.080, 0.259
Behavioral factors
Alcohol consumption
No	6814 (99.3)	1922 (99)	420 (99.3)	283 (99.3)	629 (99.7)	657 (99.4)	150 (98.7)	525 (99.8)
Yes	46 (0.7)	20 (1)	3 (0.7)	2 (0.7)	2 (0.3)	4 (0.6)	2 (1.3)	1 (0.2)
χ 2, p‐value	2.626, 0.105		0.972, 0.615			3.044, 0.218
Tobacco consumption
No	5369 (78.3)	1330 (68.5)	329 (77.8)	239 (76.8)	433 (71.6)	482 (72.9)	417 (79.3)	102 (67.1)
Yes	1491 (21.7)	612 (31.5)	94 (22.2)	72 (23.2)	172 (28.4)	179 (27.1)	109 (20.7)	50 (32.9)
χ 2, p‐value	79.600, < 0.001*		6.022, 0.049*			11.594, 0.003*
Healthy food frequency score
High	2573 (37.5)	704 (36.3)	163 (38.5)	112 (39.3)	221 (35)	260 (39.3)	53 (34.9)	183 (34.8)
Medium	2367 (34.5)	630 (32.4)	146 (34.5)	89 (31.2)	226 (35.8)	225 (34)	53 (34.9)	183 (34.8)
Low	1920 (28)	608 (31.3)	114 (27)	84 (29.5)	184 (29.2)	176 (26.6)	46 (30.3)	160 (30.4)
χ 2, p‐value	8.341, 0.015*		3.088, 0.543			3.566, 0.468
Unhealthy food frequency score
High	2107 (30.7)	851 (43.8)	172 (40.7)	88 (30.9)	210 (33.3)	265 (40.1)	45 (29.6)	160 (30.4)
Medium	3914 (57.1)	949 (48.9)	214 (50.6)	165 (57.9)	331 (52.5)	329 (49.8)	85 (55.9)	296 (56.3)
Low	839 (12.2)	142 (7.3)	37 (8.7)	32 (11.2)	90 (14.3)	67 (10.1)	22 (14.5)	70 (13.3)
χ 2, p‐value	128.583, < 0.001*		14.173, 0.007*			15.345, 0.004*

Note: p < 0.05 is considered statistically significant (*). Values are presented as frequencies (n) with corresponding percentages (%). Percentages were calculated column‐wise.

TABLE 2.

General characteristics of men in Meghalaya according to matriliny‐related indicators.

Socio‐demographic, socioeconomic, and behavioral variables	House ownership			Land ownership
	Male‐owned	Female‐owned	Jointly‐owned	Male‐owned	Female‐owned	Jointly‐owned
	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
Sociodemographic and socioeconomic Factors
Place of residence
Urban	18 (13.6)	59 (9.2)	17 (6.7)	1 (2.2)	5 (1.7)	1 (0.5)
Rural	114 (86.4)	579 (90.8)	235 (93.3)	44 (97.8)	294 (98.3)	200 (99.5)
χ 2, p‐value	4.929, 0.085			1.648, 0.439
Sex of head of household
Male	119 (90.2)	299 (46.9)	188 (74.6)	40 (88.9)	160 (53.5)	162 (80.6)
Female	13 (9.8)	339 (53.1)	64 (25.4)	5 (11.1)	139 (46.5)	39 (19.4)
χ 2, p‐value	117.379, < 0.001*			50.637, < 0.001*
Age group (in years)
15–24	40 (30.3)	224 (35.1)	89 (35.3)	14 (31.1)	109 (36.5)	69 (34.3)
25–34	38 (28.8)	181 (28.4)	74 (29.4)	13 (28.9)	75 (25.1)	56 (27.9)
35–44	36 (27.3)	145 (22.7)	44 (17.5)	11 (24.4)	65 (21.7)	46 (22.9)
45–54	18 (13.6)	88 (13.8)	45 (17.9)	7 (15.6)	50 (16.7)	30 (14.9)
χ 2, p‐value	7.185, 0.304			1.266, 0.974
Religion
Christian	123 (93.2)	597 (93.6)	242 (96.0)	41 (91.1)	286 (95.7)	197 (98.0)
Others	9 (6.8)	41 (6.4)	10 (4.0)	4 (8.9)	13 (4.3)	4 (2.0)
χ 2, p‐value	2.221, 0.329			5.161, 0.076
Current marital status
Married	91 (68.9)	149 (59.1)	396 (62.1)	30 (66.7)	114 (56.7)	173 (57.9)
Never married	41 (31.1)	103 (40.9)	242 (37.9)	15 (33.3)	87 (43.3)	126 (42.1)
χ 2, p‐value	3.568, 0.168			1.521, 0.467
Family size
< 5	49 (37.1)	166 (26.0)	68 (27.0)	24 (53.3)	69 (23.1)	47 (23.4)
≥ 5	83 (62.9)	472 (74.0)	184 (73.0)	21 (46.7)	230 (76.9)	154 (76.6)
χ 2, p‐value	6.816, 0.033*			19.644, < 0.001*
Educational level
No education	21 (15.9)	145 (22.7)	33 (13.1)	6 (13.3)	54 (18.1)	28 (13.9)
Primary	27 (20.5)	130 (20.4)	61 (24.2)	14 (31.1)	63 (21.1)	51 (25.4)
Secondary	78 (59.1)	326 (51.1)	132 (52.4)	21 (46.7)	167 (55.9)	104 (51.7)
Higher	6 (4.5)	37 (5.8)	26 (10.3)	4 (8.9)	15 (5.0)	18 (9.0)
χ 2, p‐value	18.780, 0.005*			7.553, 0.273
Wealth index
Lowest	36 (27.3)	130 (20.4)	62 (24.6)	10 (22.2)	54 (18.1)	51 (25.4)
Lower	25 (18.9)	160 (25.1)	57 (22.6)	10 (22.2)	80 (26.8)	45 (22.4)
Middle	25 (18.9)	138 (21.6)	66 (26.2)	13 (28.9)	73 (24.4)	50 (24.9)
Higher	24 (18.2)	132 (20.7)	40 (15.9)	9 (20.0)	63 (21.1)	34 (16.9)
Highest	22 (16.7)	78 (12.2)	27 (10.7)	3 (6.7)	29 (9.7)	21 (10.4)
χ 2, p‐value	12.299, 0.138			6.060, 0.641
Behavioral Factors
Alcohol consumption
No	87 (65.9)	170 (67.5)	377 (59.1)	31 (68.9)	127 (63.2)	187 (62.5)
Yes	45 (34.1)	82 (32.5)	261 (40.9)	14 (31.1)	74 (36.8)	112 (37.5)
χ 2, p‐value	6.339, 0.042*			0.680, 0.712
Tobacco Consumption
No	51 (38.6)	92 (36.5)	198 (31.0)	19 (42.2)	72 (35.8)	104 (34.8)
Yes	81 (61.4)	160 (63.5)	440 (69.0)	26 (57.8)	129 (64.2)	195 (65.2)
χ 2, p‐value	4.328, 0.115			0.942, 0.624

Note: p < 0.05 is considered statistically significant (*). Values are presented as frequencies (n) with corresponding percentages (%). Percentages were calculated column‐wise.

Among women, matriliny‐related indicators varied significantly according to age, religion, marital status, household size, educational attainment, and household wealth (Table 1, all p < 0.05). Younger women were more likely to report matrilocal residence, and women aged 25–44 years were more frequently represented in ownership categories (Table 1). Christian women were more frequently represented within matrilocal residence and jointly owned house and land categories. Married women were more likely to be in non‐matrilocal households and joint ownership categories, whereas never‐married women constituted larger proportions of male‐owned house and land categories. Larger households were more likely to report matrilocal residence, whereas female ownership of house and land was more common in smaller households.

Socioeconomic gradients further distinguished matriliny‐related indicators. Educational attainment varied across all matriliny‐related indicators among women (Table 1, all p < 0.001), with matrilocal women more likely to have secondary education and less likely to have no formal education than non‐matrilocal women. Across house and land ownership categories, secondary education constituted the largest educational group. However, women in female‐owned houses and those in female‐owned land categories showed relatively higher proportions with no education or only primary education compared with women in male‐owned or jointly owned households. Household wealth was associated only with matrilocal residence (χ ² = 64.09, p < 0.001) among women, with matrilocal women more frequently represented in the higher and highest wealth quintiles, whereas non‐matrilocal women were more commonly represented in the lowest wealth category (Table 1).

In contrast, among men, fewer socio‐demographic and socioeconomic characteristics were associated with matriliny‐related indicators (Table 2). Household size was associated with house and land ownership, with larger households more common in female and jointly owned categories. Educational attainment varied across house ownership categories, with the highest proportion of men attaining higher education observed in jointly owned houses (χ ² = 18.780, p < 0.05). Age, religion, marital status, place of residence, and household wealth showed no significant associations with ownership patterns among men.

3.4. Behavioral Correlates of Matriliny

Behavioral characteristics varied according to matriliny‐related indicators, particularly among women (Table 1). Tobacco consumption was less common in women living in matrilocal households than those in non‐matrilocal households (21.7% vs. 31.5%, χ ² = 79.60, p < 0.001). Across ownership categories, tobacco use was highest in jointly owned houses (28.4%) and land (32.9%) and lowest among those reporting female‐owned land (20.7%). Alcohol consumption among women was low overall and was not associated with matriliny‐related indicators. Dietary frequency scores were available for women only. Matrilocal women were less likely to report high unhealthy food consumption than non‐matrilocal women (30.7% vs. 43.8%, χ ² = 128.58, p < 0.001). Although healthy food consumption also differed by matrilocal residence (χ ² = 8.34, p < 0.05), these differences were comparatively modest. Ownership of house and land varied across unhealthy food consumption categories (both p < 0.01), with female‐house ownership showing the lowest unhealthy food consumption.

Among men (Table 2), alcohol consumption was significantly associated with house ownership (χ ² = 6.34, p < 0.05), with the highest prevalence observed among those reporting joint house ownership and the lowest in female ownership (40.9% vs. 32.5%). Tobacco consumption was not significantly associated with any matriliny‐related indicators (all p > 0.05).

3.5. Matriliny, Women's Empowerment, and Healthcare Access

Matriliny‐related indicators were significantly associated with multiple dimensions of women's empowerment and reported barriers to healthcare access (Table 3). Matrilocal residence and female ownership of house and land were associated with higher decision‐making autonomy and stronger rejection of attitudes condoning violence (all p < 0.001). Social independence varied significantly by matriliny status (all p < 0.05), with matrilocal women and those in male‐owned house and maled‐owned land more likely to be represented in the low social independence category, whereas women in female‐owned house and female‐owned land were least frequently represented in this category.

TABLE 3.

Empowerment and structural barriers of health in Meghalaya according to matriliny‐related indicators.

Empowerment and structural barriers to healthcare access	Matrilocality		House ownership			Land ownership
	Matrilocal	Non‐matrilocal	Male‐owned	Female‐owned	Jointly‐owned	Male‐owned	Female‐owned	Jointly‐owned
	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
Empowerment
Attitude toward violence
High Empowerment	1761 (25.7)	452 (23.3)	280 (66.2)	195 (68.4)	383 (60.7)	430 (65.1)	106 (69.7)	322 (61.2)
Moderate Empowerment	3500 (51)	911 (46.9)	125 (29.6)	57 (20)	178 (28.2)	190 (28.7)	26 (17.1)	144 (27.4)
Low Empowerment	1599 (23.3)	579 (29.8)	18 (4.3)	33 (11.6)	70 (11.1)	41 (6.2)	20 (13.2)	60 (11.4)
χ 2, p‐value	34.444, < 0.001*		24.508, < 0.001*			19.836, < 0.001*
Decision making
High empowerment	1875 (27.3)	306 (15.8)	88 (20.8)	76 (26.7)	115 (18.2)	158 (23.9)	31 (20.4)	90 (17.1)
Moderate empowerment	3151 (45.9)	1005 (51.8)	335 (79.2)	209 (73.3)	516 (81.8)	503 (76.1)	121 (79.6)	436 (82.9)
Low empowerment	1834 (26.7)	631 (32.5)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
χ 2, p‐value	110.610, < 0.001*		8.482, 0.014*			8.214, 0.016*
Social independence
High Empowerment	1686 (24.6)	524 (27)	35 (8.3)	36 (12.6)	71 (11.3)	62 (9.4)	16 (10.5)	64 (12.2)
Moderate Empowerment	3269 (47.7)	1126 (58)	190 (44.9)	222 (77.9)	414 (65.6)	364 (55.1)	120 (78.9)	342 (65)
Low Empowerment	1905 (27.8)	292 (15)	198 (46.8)	27 (9.5)	146 (23.1)	235 (35.6)	16 (10.5)	120 (22.8)
χ 2, p‐value	134.144, < 0.001*		131.169, < 0.001*			50.551, < 0.001*
Structural barriers to healthcare access
Getting medical help for self: getting permission to go
No problem	3842 (56)	1361 (70.1)	279 (66)	188 (66)	329 (52.1)	447 (67.6)	89 (58.6)	260 (49.4)
Big problem	1080 (15.7)	116 (6)	39 (9.2)	43 (15.1)	94 (14.9)	53 (8)	32 (21.1)	91 (17.3)
Not a big problem	1938 (28.3)	465 (23.9)	105 (24.8)	54 (18.9)	208 (33)	161 (24.4)	31 (20.4)	175 (33.3)
χ 2, p‐value	167.345, < 0.001*		33.551, < 0.001*			55.262, < 0.001*
Getting medical help for self: getting money needed for treatment
No problem	1337 (19.5)	488 (25.1)	111 (26.2)	46 (16.1)	114 (18.1)	147 (22.2)	28 (18.4)	96 (18.3)
Big problem	2846 (41.5)	627 (32.3)	154 (36.4)	117 (41.1)	213 (33.8)	258 (39)	58 (38.2)	168 (31.9)
Not a big problem	2677 (39)	827 (42.6)	158 (37.4)	122 (42.8)	304 (48.2)	256 (38.7)	66 (43.4)	262 (49.8)
χ 2, p‐value	60.506, < 0.001*		21.128, < 0.001*			15.098, 0.005*
Getting medical help for self: distance to health facility
No problem	1332 (19.4)	539 (27.8)	117 (27.7)	43 (15.1)	132 (20.9)	161 (24.4)	27 (17.8)	104 (19.8)
Big problem	2721 (39.7)	615 (31.7)	143 (33.8)	128 (44.9)	184 (29.2)	236 (35.7)	66 (43.4)	153 (29.1)
Not a big problem	2807 (40.9)	788 (40.6)	163 (38.5)	114 (40)	315 (49.9)	264 (39.9)	59 (38.8)	269 (51.1)
χ 2, p‐value	75.080, < 0.001*		35.993, < 0.001*			21.822, < 0.001*
Getting medical help for self: having to take transport
No problem	1532 (22.3)	606 (31.2)	132 (31.2)	58 (20.4)	152 (24.1)	193 (29.2)	29 (19.1)	120 (22.8)
Big problem	2528 (36.9)	516 (26.6)	118 (27.9)	119 (41.8)	178 (28.2)	208 (31.5)	64 (42.1)	143 (27.2)
Not a big problem	2800 (40.8)	820 (42.2)	173 (40.9)	108 (37.9)	301 (47.7)	260 (39.3)	59 (38.8)	263 (50)
χ 2, p‐value	96.040, < 0.001*		27.659, < 0.001*			24.560, < 0.001*
Getting medical help for self: not wanting to go alone
No problem	2657 (38.7)	875 (45.1)	182 (43.1)	111 (38.9)	249 (39.5)	284 (43)	55 (36.2)	203 (38.6)
Big problem	1335 (19.5)	228 (11.7)	72 (17.1)	65 (22.8)	90 (14.3)	109 (16.5)	34 (22.4)	84 (16)
Not a big problem	2867 (41.8)	839 (43.2)	168 (39.8)	109 (38.2)	292 (46.3)	267 (40.5)	63 (41.4)	239 (45.4)
χ 2, p‐value	66.598, < 0.001*		13.554, 0.009*			6.972, 0.137
Getting medical help for self: concern no female health provider
No problem	2052 (29.9)	751 (38.7)	150 (35.5)	86 (30.2)	196 (31.1)	225 (34)	51 (33.6)	156 (29.7)
Big problem	1864 (27.2)	377 (19.4)	112 (26.5)	93 (32.6)	119 (18.9)	173 (26.2)	47 (30.9)	104 (19.8)
Not a big problem	2944 (42.9)	814 (41.9)	161 (38.1)	106 (37.2)	316 (50.1)	263 (39.8)	54 (35.5)	266 (50.6)
χ 2, p‐value	72.606, < 0.001*		30.477, < 0.001*			20.344, < 0.001*
Getting medical help for self: concern no provider
No problem	595 (12.9)	1064 (25.3)	103 (24.3)	56 (19.6)	126 (20)	150 (22.7)	33 (21.7)	102 (19.4)
Big problem	1825 (39.6)	1731 (41.2)	155 (36.6)	123 (43.2)	209 (33.1)	242 (36.6)	62 (40.8)	183 (34.8)
Not a big problem	2184 (47.4)	1403 (33.4)	165 (39)	106 (37.2)	296 (46.9)	269 (40.7)	57 (37.5)	241 (45.8)
χ 2, p‐value	287.003, < 0.001*		14.193, 0.007*			5.467, 0.243
Getting medical help for self: concern no drugs available
No problem	1108 (16.2)	351 (18.1)	94 (22.2)	46 (16.1)	109 (17.3)	137 (20.7)	28 (18.4)	84 (16)
Big problem	3242 (47.3)	864 (44.5)	178 (42.1)	136 (47.7)	249 (39.5)	287 (43.4)	67 (44.1)	209 (39.7)
Not a big problem	2510 (36.6)	727 (37.4)	151 (35.7)	103 (36.1)	273 (43.3)	237 (35.9)	57 (37.5)	233 (44.3)
χ 2, p‐value	6.158, 0.046*		12.329, 0.015*			10.135, 0.038*

Note: p < 0.05 is considered statistically significant (*). Values are presented as frequencies (n) with corresponding percentages (%). Percentages were calculated column‐wise.

Women's experiences of healthcare access varied significantly across matriliny‐related indicators (all p < 0.05). Compared with women in non‐matrilocal and male‐owned or jointly owned houses, women in matrilocal and female‐owned houses reported greater difficulties accessing healthcare, reflecting cost, distance, and limited availability of services. For land ownership, a similar pattern was observed: women in households with female‐owned land reported greater challenges accessing healthcare compared with those in male‐owned or jointly owned houses, with the exceptions of concern about going alone for getting medical help (χ ² = 6.97, p = 0.137) and lack of available providers (χ ² = 5.47, p = 0.243).

3.6. Nutritional Status and Cardiometabolic Indicators

Nutritional and cardiometabolic indicators differed markedly between women and men (Table 4). Overall, 26.3% of adults were classified as overweight and/or obese, with a higher prevalence among men than women (30.4% vs. 25.6%), while underweight prevalence was similar between sexes (overall: 10.6%; women: 10.9%; men: 9.1%). Central obesity showed marked sex differences. High WC was more common among women than men (26.0% vs. 3.1%, χ ² = 359.98, p < 0.001), and a similar pattern was observed for high WHR (86.9% vs. 27.0%, χ ² = 2590.86, p < 0.001).

TABLE 4.

Nutritional and cardiometabolic indicators in Meghalaya according to sex.

Nutritional and cardiometabolic indicators	Overall (N = 10 211)	Females (n = 8802)	Males (n = 1409)
	n (%)	n (%)	n (%)
BMI
Normal	6441 (63.1)	5589 (63.5)	852 (60.5)
Underweight	1085 (10.6)	957 (10.9)	128 (9.1)
Overweight and Obese	2685 (26.3)	2256 (25.6)	429 (30.4)
χ 2, p‐value		16.14, < 0.001*
WC
Normal	7876 (77.1)	6511 (74.0)	1365 (96.9)
High	2335 (22.9)	2291 (26.0)	44 (3.1)
χ 2, p‐value		359.98, < 0.001*
WHR
Normal	2183 (21.4)	1154 (13.1)	1029 (73)
High	8028 (78.6)	7648 (86.9)	380 (27)
χ 2, p‐value		2590.86, < 0.001*
Blood pressure
Normal	4744 (46.5)	4266 (48.5)	478 (33.9)
Pre‐hypertension	4099 (40.1)	3428 (38.9)	671 (47.6)
Hypertension	1368 (13.4)	1108 (12.6)	260 (18.5)
χ 2, p‐value		109.24, < 0.001*
Glycemic Status
Normal	9265 (90.7)	8048 (91.4)	1217 (86.4)
Pre‐diabetes	766 (7.5)	601 (6.8)	165 (11.7)
Diabetes	180 (1.8)	153 (1.7)	27 (1.9)
χ 2, p‐value		42.24, < 0.001*
Anemia
Normal	5007 (49.0)	3996 (45.4)	1011 (71.8)
Mild	2507 (24.5)	2178 (24.7)	329 (23.3)
Moderate and severe	2697 (26.5)	2628 (29.9)	69 (4.9)
χ 2, p‐value		459.48, < 0.001*

Note: p < 0.05 is considered statistically significant (*). Values are presented as frequencies (n) with corresponding percentages (%). Percentages were calculated column‐wise.

Abbreviations: BMI, body mass index; WC, waist circumference; WHR, waist‐to‐hip ratio.

Hypertension was present in 13.4% of the study population and was more common among men than women (18.5% vs. 12.6%). An additional 40.1% of participants were classified as pre‐hypertensive, with higher prevalence among men (47.6%) than women (38.9%). Diabetes prevalence was low (overall: 1.8%; women: 1.7%; men: 1.9%), while pre‐diabetes was more common (7.5%) and occurred more frequently among men than women (11.7% vs. 6.8%). Anemia showed a marked sex difference (χ ² = 459.48, p < 0.001). Overall, 51% of participants were anemic, with moderate‐to‐severe anemia reported by 29.9% of women compared to 4.9% of men.

3.7. Sex Differences and Interaction Effects

Women and men differed significantly in several matriliny‐related, socio‐demographic, socioeconomic, and behavioral characteristics (Table S2). Interaction analyses were conducted to assess associations between explanatory variables common to both sexes and health outcomes differed by sex. Evidence of effect modification varied across outcomes (Table S3). Accordingly, all subsequent multivariable analyses were stratified by sex and are presented separately for women and men.

3.8. Adjusted Associations With Nutritional Status and Cardiometabolic Indicators

Figures 3, 4, 5 summarize the adjusted multivariable logistic regression analyses. Figures 3 and 5 present Model 1 for women and men, respectively, using explanatory variables measured common to both sexes, whereas Figure 4 presents the expanded model for women (Model 2), which additionally included matrilocal residence, dietary characteristics, women's empowerment, and structural barriers to healthcare access. Complete ORs, aORs, 95% CIs, and p‐values are presented in Tables S11–S16.

FIGURE 3.

Forest plots showing factors associated with underweight (A), overweight/obesity (B), high waist circumference (C), high waist‐to‐hip ratio (D), hypertension (E), diabetes (F), and anemia (G) among women in Meghalaya (Model 1). Plots display adjusted odds ratios (aORs) and 95% confidence intervals for variables significantly associated with each outcome (see Tables S13 and S14). Model 1 was adjusted for matriliny dimensions, sociodemographic and socioeconomic characteristics, and behavioral factors available for both women and men. Odds ratios > 1 indicate higher odds of the outcome relative to the reference category, whereas odds ratios < 1 indicate lower odds. Filled circles indicate statistically significant associations (p < 0.05), and open circles indicate non‐significant associations (p ≥ 0.05). Horizontal lines represent 95% confidence intervals.

FIGURE 4.

Forest plots showing factors associated with underweight (A), overweight/obesity (B), high waist circumference (C), high waist‐to‐hip ratio (D), hypertension (E), diabetes (F), and anemia (G) among women in Meghalaya. Plots display adjusted odds ratios (aORs) and 95% confidence intervals for variables significantly associated with each outcome in Model 2 (see Tables S13 and S14). Model 2 included all variables in Model 1 and was additionally adjusted for women's empowerment, structural barriers to healthcare access, and dietary characteristics available only in the women's dataset. Odds ratios > 1 indicate higher odds of the outcome relative to the reference category, whereas odds ratios < 1 indicate lower odds. Horizontal lines represent 95% confidence intervals.

FIGURE 5.

Forest plots showing factors associated with underweight (A), overweight/obesity (B), high waist circumference (C), high waist‐to‐hip ratio (D), hypertension (E), diabetes (F), and anemia (G) among men in Meghalaya. Plots display adjusted odds ratios (aORs) and 95% confidence intervals for variables significantly associated with each outcome (see Table S15). Model was adjusted for matriliny dimensions, sociodemographic and socioeconomic characteristics, and behavioral factors available in both the women's and men's datasets. Odds ratios > 1 indicate higher odds of the outcome relative to the reference category, whereas odds ratios < 1 indicate lower odds. Horizontal lines represent 95% confidence intervals.

The factors associated with underweight differed between women and men and between the two models for women (Figures 3A, 4A, 5A). In Model 1, Figure 3A, underweight among women was associated with Khasi ethnicity and marital status. Following adjustment for matrilocal residence, dietary characteristics, women's empowerment, and healthcare access, these associations were no longer evident (Model 2, Figure 4A). Instead, land ownership, educational attainment, decision‐making empowerment, and healthcare access emerged as independent correlates. Compared with female land ownership, both male‐owned (aOR = 2.54) and jointly owned land (aOR = 2.43) were associated with higher odds of underweight. Moderate decision‐making empowerment (aOR = 1.76) and reporting that obtaining money for medical treatment was not a big problem (reference: no problem; aOR = 1.83) were also associated with higher odds of underweight. Women attaining secondary or higher education (aOR = 0.45) and those reporting big problem for drug availability (aOR = 0.37) were associated with significantly lower odds. Among men (Figure 5A), joint house ownership was the only independent correlate of underweight, with higher odds than female ownership (aOR = 4.67).

BMI‐defined overweight/obesity was primarily associated with demographic and socioeconomic characteristics, whereas matriliny‐related indicators showed comparatively limited independent associations (Figures 3B, 4B, 5B). Among women, increasing age, Khasi ethnicity, urban residence, and higher household wealth were independently associated with overweight/obesity in Model 1 (Figure 3B). The associations remained largely unchanged in Model 2 (Figure 4B), which additionally identified non‐matrilocal residence as protective (aOR = 0.65), while currently married women had higher odds than never‐married women (aOR = 1.84). Men (Figure 5B) showed a similar age‐ and household wealth‐related pattern. Higher odds were observed in male‐headed households (aOR = 1.87), those aged 35–44 years (aOR = 5.04) and 45–54 years (aOR = 3.52), and across all wealth categories. Compared with the lowest wealth quintile, men in the lower, middle, higher, and highest wealth quintiles had aORs of 3.68, 2.76, 3.42, and 7.36, respectively.

Among women, WC‐defined central obesity was independently associated with increasing age, marital status, household wealth, land ownership, family size, and sex of household head (Model 1, Figure 3C). Compared with women aged 15–24 years, the odds increased progressively among those aged 25–34 years (aOR = 3.27), 35–44 years (aOR = 4.28), and 45–49 years (aOR = 4.13). Married women (aOR = 1.77) and those in the middle (aOR = 1.56) and highest (aOR = 2.45) wealth quintiles had higher odds than their respective reference groups. In contrast, joint land ownership (aOR = 0.58) and smaller household size (aOR = 0.75) were associated with lower odds. These associations remained stable in Model 2 (Figure 4C), which further identified medium unhealthy food consumption (aOR = 0.64) and moderate social independence (aOR = 0.64) as being associated with lower odds of high WC. Among men (Figure 5C), Khasi ethnicity was associated with lower odds of high WC than Garo ethnicity (aOR = 0.17), whereas urban residence was associated with higher odds (aOR = 13.25) of high WC.

Relatively few factors were independently associated with WHR‐defined central obesity, and the pattern differed for women and men (Figures 3D, 4D, 5D). In Model 1 (Figure 3D), urban residence and increasing age were the only independent correlates of high WHR. Urban women (aOR = 0.55) had lower odds of high WHR than rural women, whereas women aged 25–34 years (aOR = 1.87) had higher odds than those aged 15–24 years. These associations were no longer significant in Model 2 (Figure 4D). Instead, religion, women's empowerment, and dietary characteristics emerged as independent correlates. Christian women had higher odds of WHR‐defined central obesity than women of other religions (aOR = 2.03), whereas women with low empowerment regarding attitudes toward violence had higher odds than those with high empowerment (aOR = 2.29). Medium and high unhealthy food consumption were both associated with lower odds of WHR‐defined central obesity (aOR = 0.29 and 0.31, respectively). Among men, Khasi ethnicity (aOR = 2.22) and current marriage (aOR = 2.58) were independently associated with higher odds of WHR‐defined central obesity (Figure 5D).

Hypertension was primarily associated with demographic characteristics, with few independent associations observed for matriliny‐related indicators (Figures 3E, 4E, 5E). Among women, Khasi ethnicity was associated with lower odds of hypertension (aOR = 0.45), whereas urban residence (aOR = 1.60) and age were associated with progressively higher odds (25–34 years: aOR = 2.70; 35–44 years: aOR = 3.08, and 45–49 years: aOR = 6.56) relative to those aged 15–24 years (Model 1, Figure 3E). Women in the lower wealth quintile had lower odds of hypertension than those in the lowest wealth quintile (aOR = 0.63). These associations persisted in Model 2 (Figure 4E) and additionally identified lower odds of hypertension among women living in male‐owned households (aOR = 0.68) and among those reporting that the lack of healthcare providers was a major barrier to accessing healthcare (aOR = 0.47). Among men (Figure 5E), Khasi ethnicity was similarly associated with lower odds of hypertension than Garo ethnicity (aOR = 0.27). In addition, Christian men had higher odds of hypertension than men belonging to other religions (aOR = 3.80), while those in the higher wealth quintile had higher odds than those in the lowest wealth quintile (aOR = 2.53).

Independent correlates of diabetes were limited (Figures 3F, 4F, 5F). Among women, current marriage was the only independent correlate, with married women having higher odds of diabetes than never‐married women (aOR = 3.29) in Model 1 (Figure 3F). This association was no longer evident after adjustment in Model 2 (Figure 4F), which identified joint house ownership (aOR = 0.56) and reporting distance to a health facility as a major barrier (aOR = 2.63) as independent correlates. Among men (Figure 5F), alcohol consumption was associated with lower odds of diabetes (aOR = 0.40).

Few independent associations were observed for anemia (Figures 3G, 4G, 5G). Among women, Khasi ethnicity was consistently associated with lower odds of anemia in both Model 1 (aOR = 0.66, Figure 3G) and Model 2 (aOR = 0.69, Figure 4G). Among men (Figure 5G), secondary or higher educational attainment was associated with lower odds of anemia than no formal education (aOR = 0.33).

3.9. Children's Sickness and Care Seeking Behavior

Children's sickness and care‐seeking behavior are presented in Figure 6 and Tables S10 and S17. Reported childhood sickness was associated with both matriliny‐related indicators and broader socio‐demographic characteristics. Compared with children from matrilocal and female‐owned houses, children from non‐matrilocal households (aOR = 0.68) and male‐owned house (aOR = 0.62) had lower odds of reported sickness. In contrast, Khasi ethnicity (aOR = 2.05), residence in a male‐headed household (aOR = 1.52), age below 1 year (aOR = 1.75), and household size of fewer than five members (aOR = 1.47) were associated with higher odds of reported sickness. Household wealth showed a graded inverse association with reported childhood sickness, with children from the lower (aOR = 0.53), middle (aOR = 0.53), higher (aOR = 0.51), and highest (aOR = 0.21) wealth quintiles having lower odds than those from the lowest wealth quintile.

FIGURE 6.

Forest plot showing factors associated with childhood sickness among children in Meghalaya. The figure presents adjusted odds ratios (aORs) for childhood sickness (significant in the Table S16). The model was adjusted for matriliny‐related indicators and the sociodemographic and socioeconomic variables available in the children's dataset (see Table S10). Odds ratios > 1 indicate higher odds of the outcome relative to the reference category, whereas odds ratios < 1 indicate lower odds. Horizontal lines represent 95% confidence intervals.

Care‐seeking behavior was not significantly associated with most matriliny‐related, socio‐demographic, or socioeconomic factors (Table S10). Although female children were more likely than male children to receive care in the bivariate analysis (57.6% vs. 42.4%; χ ² = 3.997, p = 0.046), this association was not retained after multivariable adjustment. No explanatory variable was independently associated with care‐seeking behavior in the adjusted logistic regression model (Table S17).

4. Discussion

This study examined associations between matriliny‐related indicators and health outcomes in a matrilineal society undergoing social and epidemiological transition in Meghalaya, Northeast India. The findings demonstrate that matriliny does not confer uniform health advantages. Instead, it operates through selective, outcome‐specific, and gendered mechanisms that redistribute health risks and protections across women, men, and children (Gough 1961; Mattison ²⁰¹¹; Lowes ²⁰²⁰; Reynolds et al. ²⁰²⁰).

4.1. Matriliny as a Living Institutional System

Within India's predominantly patrilineal kinship landscape, Meghalaya is distinctive for its high prevalence of matrilocal residence and female‐centred ownership of house and land (Nakane 1968; Leonetti et al. ²⁰⁰⁷; Brulé and Gaikwad ²⁰²¹). Our descriptive findings and comparisons with national and state‐level data (Figure 1) indicate that these institutional features are more pronounced in Meghalaya than in other states, enacted through everyday residential and property relations rather than confined to symbolic descent. At the same time, matriliny in Meghalaya is internally heterogeneous: large proportions of women do not own house or land, and ownership is frequently joint rather than exclusive. These findings are consistent with classic anthropological arguments that matriliny is not a simple inversion of patriarchy but a complex institutional system in which authority, responsibility, and resource control are distributed across gendered domains (Radcliffe‐Brown ¹⁹⁵⁰; Gough ¹⁹⁶⁸).

Ethnographic accounts of the Khasi and Garo emphasize the central role of the youngest daughter, ka khadduh among the Khasi and nokna among the Garo, in maintaining lineage continuity and custodianship of the ancestral home (Nakane 1968; Gurdon ¹⁹⁰⁷; Nongbri ²⁰⁰³; Mylliemngap ²⁰²⁴; Sangma ²⁰²⁴). The high prevalence of female‐owned houses observed in this study is consistent with this institutional logic. However, this custodial role does not translate into monopolization of authority, which remains shared with male kin, usually the maternal uncle, and lineage councils (Nakane 1968; Nongbri ¹⁹⁸⁸; Singh ²⁰²⁰; Mylliemngap ²⁰²⁴; Sangma ²⁰²⁴). Matriliny therefore appears to redistribute certain forms of material security and residential stability without fully displacing gendered hierarchies or insulating households from broader political‐economic transformations (Stivens 2025).

4.2. Nutritional and Cardiometabolic Health in a Transitional Context

The overall health profile observed in Meghalaya reflects an early‐to‐intermediate stage of nutritional and epidemiological transition, characterized by the coexistence of undernutrition, persistent anemia, and emerging cardiometabolic risk (Popkin 2006; Popkin et al. ²⁰²⁰). Underweight and anemia remain highly prevalent, particularly among women, while overweight, obesity, hypertension, and pre‐diabetes are increasing but remain below national averages (International Institute for Population Sciences (IIPS) and ICF ^2021a; Gupta et al. ²⁰²³).

The high prevalence of female central adiposity is particularly notable and highlights the growing dual burden of malnutrition, a pattern increasingly documented in low‐ and middle‐income populations undergoing uneven nutrition transition (Doak et al. 2005; Popkin and Ng ²⁰²²). These findings suggest that matriliny does not offset broader structural processes linked to urbanization, changing diets, market integration, and declining physical activity. Instead, kinship organization appears embedded within wider transformations shaping contemporary health risks. Persistent sex disparities in nutrition and cardiometabolic risk indicate that gendered inequalities remain pronounced despite the matrilineal context (Reynolds et al. 2020; Lamarque et al. ²⁰²⁵; Krieger ²⁰²⁴).

4.3. Ethnic Heterogeneity Within Matriliny

The contrasting health profiles of the Khasi and Garo populations underscore that matriliny is not uniformly associated with health outcomes. Khasi women had higher odds of both underweight and general obesity, and Khasi men demonstrated higher odds of elevated WHR but lower odds of WC‐defined central obesity. This pattern is characteristic of the double burden of malnutrition associated with nutritional transition, whereby undernutrition persists alongside increasing overweight and obesity (Popkin et al. 2020). At the same time, Khasi populations showed a comparatively reduced cardiometabolic risk profile, characterized by significantly lower odds of hypertension risk. Khasi women also appeared relatively protected against anemia, showing lower odds compared to their Garo counterparts. The comparatively higher odds of anemia among the Garo population may, at least in part, reflect the greater malaria endemicity of the Garo Hills, where ecological conditions, geographical remoteness, and constraints to healthcare accessibility have contributed to higher malaria burden relative to other districts of Meghalaya (Kessler et al. 2018). Such differences in disease ecology may provide one contextual explanation for the higher odds of anemia observed among the Garo population, although this hypothesis warrants further investigation.

These findings suggest that health variation within matrilineal populations may reflect differing stages of nutritional and epidemiological transition rather than shared kinship structure alone. Differences in subsistence strategies, urban exposure, and dietary practices likely mediate how matriliny intersects with nutritional and metabolic risk (Lowes 2020; Reynolds et al. ²⁰²⁰; Fortunato ²⁰¹⁹). Rather than constituting a uniform biological or social environment, the health implications of matrilineal organization appear contingent on broader ecological, socioeconomic, and sociocultural contexts operating within and across populations in Meghalaya.

4.4. Matrilineal Institutions and Health Outcomes

Among the matriliny‐related indicators examined, matrilocal residence showed little evidence of independent associations with health outcomes after adjustment. Non‐matrilocal residence was associated only with lower odds of overweight and obesity among women. This suggests that residential organization alone provides limited explanatory value beyond socioeconomic, behavioral, empowerment, and healthcare access factors. Although matrilocal residence is a defining characteristic of matriliny, it primarily reflects residential organization and does not necessarily determine control over household resources or everyday decision‐making authority (Agarwal 1997; Doss ²⁰¹³; Lowes ²⁰²⁰). Anthropological studies have shown that matriliny comprises multiple institutional components, including descent, inheritance, residence, and authority, that need not operate synchronously, particularly in societies experiencing rapid socioeconomic change (Schneider and Gough 1961; Mattison ²⁰¹¹; Fortunato ²⁰¹⁹).

In contrast, ownership of house and land showed more consistent associations with health, particularly undernutrition. In Meghalaya, female ownership may reflect forms of household organization in which women retain greater continuity of residence and stronger claims over material assets. Such arrangements may shape everyday practices surrounding food allocation, caregiving practices, and economic security, particularly in contexts characterized by persistent nutritional vulnerability and ongoing socioeconomic transition. Compared with female ownership, male and joint land ownership were associated with higher odds of underweight among women, while joint house ownership was associated with higher odds of underweight among men. The lower odds of underweight observed among women in households with female‐owned land and among men in households with female‐owned houses suggest that female‐centered ownership may reflect household environments that support nutritional well‐being beyond women themselves (Lamarque et al. 2025).

Associations with cardiometabolic outcomes were less consistent. Joint ownership was associated with lower odds of central obesity and diabetes among women but higher odds of underweight, while male‐owned households were associated with lower odds of hypertension among women. This pattern suggests that female‐centred ownership may be more effective in buffering against material deprivation and nutritional insecurity than against emerging cardiometabolic risks linked to aging, urbanization, and dietary transitions. Notably, ownership was associated with few health outcomes among men after adjustment, indicating that asset ownership may be more closely linked to women's social and economic positions within matrilineal households.

These findings suggest that asset ownership constitutes one pathway through which kinship organization intersects with nutritional inequality and cardiometabolic health. These patterns are consistent with evidence from Indian and cross‐cultural contexts linking women's control over assets with improved nutritional outcomes for both women and other household members (Reynolds et al. 2020; Dutta et al. ²⁰²⁵). However, the associations remained selective and outcome‐specific, suggesting that the biological consequences of matriliny are shaped by broader socioeconomic and epidemiological conditions rather than by kinship structure alone (Reynolds et al. 2020; Lamarque et al. ²⁰²⁵).

4.5. Nutritional and Cardiometabolic Health Beyond Matriliny

Beyond the associations observed for matriliny‐related indicators, nutritional and cardiometabolic outcomes were more consistently patterned by demographic and socioeconomic characteristics, indicating that kinship institutions operate within broader social and epidemiological processes accompanying nutritional transition (Popkin et al. 2020; Gupta et al. ²⁰²³; Nikolic Turnic et al. ²⁰²⁴).

Underweight and anemia were associated primarily with lower educational attainment and socioeconomic disadvantage, suggesting that nutritional deprivation remains closely linked to broader structural inequalities. Among women, higher educational attainment was associated with lower odds of underweight, whereas among men, secondary and higher secondary education was associated with lower odds of anemia, consistent with findings widely reported across low‐ and middle‐income settings, where education improves nutritional status through greater health literacy, economic opportunities, and access to resources (Amugsi et al. 2017; Muchomba ²⁰²²).

General and central obesity were more consistently associated with increasing age, urban residence, and household wealth, reflecting the nutritional transition in which excess adiposity increasingly accompanies socioeconomic development while undernutrition persists (Popkin et al. 2020; Jaacks et al. ²⁰¹⁹; Wells et al. ²⁰²⁰; Gupta et al. ²⁰²³). Similarly, hypertension was associated predominantly with demographic and socioeconomic characteristics than with matriliny‐related institutions, particularly among women, supporting the influence of aging, urbanization, and socioeconomic transition (Popkin et al. 2020; Mohammad and Bansod ²⁰²⁴), while diabetes showed comparatively fewer independent associations beyond marital status and selected healthcare access indicators.

Household characteristics contributed relatively little once other factors were considered. Marriage was associated with lower odds of underweight but higher odds of adiposity among women, consistent with shifts in food security and lifestyle accompanying marital life in many low‐ and middle‐income settings (Aung et al. 2021; Jaacks et al. ²⁰¹⁹). The attenuation of the association with diabetes after adjustment suggests that these relationships are partly explained by broader household and social contexts. By contrast, marriage among men was associated only with higher odds of WHR‐defined central obesity, indicating that marital transitions may influence adiposity differently by sex. Household size, headship, and religion showed few and selective associations, which probably reflect broader social and contextual influences rather than independent effects of these characteristics. Similarly, the associations observed for Christianity are unlikely to reflect the effects of religious affiliation per se. The higher odds of WHR‐defined central obesity among Christian women and hypertension among Christian men probably capture the broader sociocultural and socioeconomic contexts of Khasi and Garo communities, where Christianity is closely intertwined with urbanization, market integration, and nutritional transition (Gupta et al. 2023; Popkin et al. ²⁰²⁰).

These findings indicate that while matriliny contributes to specific aspects of nutritional inequality, the broader distribution of nutritional and cardiometabolic disease in Meghalaya is more consistently structured by demographic aging, socioeconomic conditions, urbanization, and nutritional transition.

4.6. Behavioral Regulation and Uneven Cardiometabolic Protection

Behavioral patterns varied across matriliny‐related contexts, with women residing in matrilocal households and households with female‐owned assets reporting lower tobacco use and less frequent consumption of unhealthy foods than women in non‐matrilocal, male‐owned, or jointly owned households. These findings are consistent with the possibility that female‐centred households promote stronger social regulation of health behaviors through kin‐based support and everyday social norms (Dyson and Moore 1983; Lowes ²⁰²⁰; Reynolds et al. ²⁰²⁰).

However, these behavioral differences did not translate consistently into more favorable cardiometabolic outcomes. The inverse associations between unhealthy food consumption and central obesity in women are unlikely to represent true protective effects and probably reflect limitations of the NFHS dietary indicators, which capture only the frequency of fried food and aerated drink consumption rather than overall dietary quality, energy intake, or portion size.

Among men, alcohol consumption was associated with lower odds of diabetes, although evidence for this relationship remains inconsistent and is strongly influenced by drinking patterns and beverage type (Knott et al. 2015). In Northeast India, alcohol consumption often includes traditional fermented rice beverages that differ from commercial alcoholic drinks in their fermentation and composition (Loying et al. 2024). As the NFHS does not distinguish beverage types or quantify alcohol intake, this association should be interpreted cautiously.

4.7. Empowerment and the Limits of Health Conversion

Matrilocal residence and female ownership of house and land were strongly associated with higher levels of women's empowerment and reduced reported barriers to healthcare access. These findings accord with evidence that women's control over residence and assets may reduce constraints related to mobility, permission‐seeking, and financial dependence that are commonly observed in patrilocal systems (Brulé and Gaikwad 2021; Lamarque et al. ²⁰²⁵). However, empowerment‐related advantages did not translate uniformly into improved health outcomes and, in some cases, were associated with counterintuitive findings (Lalnuneng et al. 2026).

Women reporting lower empowerment regarding attitudes toward violence had higher odds of adverse WHR‐defined central obesity, while women reporting moderate decision‐making empowerment demonstrated higher odds of underweight. Additionally, moderate social independence were protective for higher WC‐defined central obesity. These findings are consistent with the growing recognition that empowerment is multidimensional and that gains in one domain do not necessarily extend to others (Kabeer 1999; Quisumbing et al. ²⁰²¹). These findings suggest that increased social and structural agency does not necessarily result in proportional health gains, echoing critiques of linear empowerment‐health models in global health research (Zimmerman 1995; Cattaneo and Chapman ²⁰¹⁰). Consequently, empowerment within matrilineal settings appears enabling but not uniformly protective (Lowes 2020).

4.8. Structural Barriers to Healthcare Access

Perceived barriers to healthcare showed heterogeneous associations with nutritional and cardiometabolic health outcomes, indicating that different dimensions of healthcare access may have distinct implications for health. Greater perceived distance to health facilities was associated with higher odds of diabetes, consistent with evidence that geographical access influences opportunities for diagnosis and long‐term disease management (Peters et al. 2008; Levesque et al. ²⁰¹³; Biswas and Kabir ²⁰¹⁷). Similarly, the association between perceived drug unavailability and hypertension may reflect difficulties in maintaining treatment or, alternatively, greater awareness of medicine shortages among women already receiving care for hypertension (WHO ²⁰²¹; Mills et al. ²⁰¹⁶). In contrast, the associations with underweight were less straightforward, as undernutrition is influenced primarily by nutritional, infectious, and socioeconomic factors rather than healthcare access alone (Loechl et al. 2023; Peters et al. ²⁰⁰⁸). These findings suggest that self‐reported healthcare barriers capture subjective experiences of healthcare utilization and illness, as well as structural accessibility, and therefore should not be interpreted as uniform indicators of poorer health (Levesque et al. 2013).

4.9. Children and the Intergenerational Reach of Matriliny

Reported childhood sickness was associated with selected matriliny‐related characteristics, household organization, and broader sociodemographic factors, while care‐seeking showed limited variability due to high overall utilization. As childhood sickness was based on maternal reports, the observed associations may reflect differences in illness recognition and healthcare‐seeking behavior, as well as underlying morbidity.

The lower odds of reported children's sickness in non‐matrilocal and male‐owned houses suggest that the associations between matriliny‐related institutions and child health are likely context‐dependent rather than uniformly beneficial, potentially reflecting differences in caregiving practices, intra‐household resource allocation, or illness reporting (Lowes 2020). At the same time, children's sickness was more consistently associated with age and socioeconomic conditions. The concentration of reported sickness among infants is consistent with greater biological vulnerability of early childhood to infectious illness, while the graded inverse association with household wealth categories accords with extensive evidence that socioeconomic resources shape child health through their influence on nutrition, living conditions, sanitation, and access to healthcare (Black et al. 2013; Kollmann et al. ²⁰¹⁷). Likewise, the associations with smaller household size, male household headship, and Khasi ethnicity are more plausibly interpreted as reflecting broader differences in caregiving and support networks, household composition; although this interpretation cannot be confirmed in the present study (Treleaven 2023; Wendt et al. ²⁰²¹).

By examining women, men, and children simultaneously, this study moves beyond women‐centred narratives to show how matrilineal systems redistribute both health risks and protections across genders and generations. Integrating anthropological perspectives on kinship with epidemiological analysis, the findings demonstrate that kinship institutions remain consequential for health, not as static cultural artifacts but as dynamic social systems embedded within broader structures of inequality and change.

5. Conclusion

Matriliny in Meghalaya functions not as a uniformly protective kinship system for health, but as a dynamic institutional system whose health effects are selective, gendered, and context dependent. Matriliny‐related advantages are most evident in domains related to nutritional security, particularly through women's land and house ownership. This benefit did not hold uniformly across cardiometabolic outcomes, where joint and male ownership were sometimes more favorable than exclusive female ownership, indicating that matriliny's protective capacity is neither absolute nor linear. Across many major health outcomes, including general obesity, anemia, and hypertension, age, education, ethnicity, household wealth, and the broader ongoing nutritional transition remained more consistent correlates of health, underscoring the constraints on kinship institutions in offsetting broader structural influences. Matriliny is shown to be internally heterogeneous, with variation across households and between Khasi and Garo populations, cautioning against assumptions of uniform matrilineal advantage. Although matriliny was associated with greater levels of women's empowerment and fewer barriers to healthcare access, these advantages did not translate uniformly into improved health, highlighting that greater agency and access to resources do not necessarily yield proportional health gains. These findings advance a biosocial understanding of kinship as a consequential but contingent influence on health in contexts of rapid social and epidemiological change.

6. Limitations and Future Directions

The cross‐sectional nature of the data precludes causal inference, and several dimensions of matriliny were necessarily operationalized using proxy measures available in the survey. The use of secondary data limited behavioral detail and statistical power for outcomes with low prevalence. Future research employing longitudinal and mixed‐methods designs would be better positioned to examine life‐course processes and institutional mechanisms. Within these constraints, the study documents patterned associations between kinship‐related arrangements and health outcomes, indicating that kinship institutions continue to intersect with health through specific, context‐dependent pathways rather than uniform effects.

Author Contributions

Abigail Lalnuneng contributed to the study's conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, software, supervision, validation, visualization, original draft writing, writing – review and editing. Banrida Langstieh contributed to conceptualization, supervision, validation, manuscript review and editing. Thiyam Seityajit Singh contributed to data curation, formal analysis, methodology, writing of the original draft, and manuscript review and editing. Roshni Tripathy contributed to data curation, formal analysis, writing of the original draft, and manuscript review and editing. Madhurima Samanta contributed to data curation, formal analysis, writing of the original draft, and manuscript review and editing. V. Souzhi Pao contributed to data curation, formal analysis, and writing of the original draft, and manuscript review and editing. Jenny Jami and Naorem Kiranmala Devi contributed to the investigation, manuscript review and editing. All authors have read and approved the final manuscript.

Funding

The authors have nothing to report.

Ethics Statement

This study utilized publicly available, de‐identified data from the NFHS‐5. Access to the dataset was obtained from The DHS Program after approval of the data request. The original NFHS‐5 survey protocol was reviewed and approved by the Institutional Review Boards of the IIPS, Mumbai, and ICF.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Sex‐stratified optimal anthropometric cutoffs and discriminatory performance for hypertension and diabetes.

Table S2: Sex Differences in matriliny dimensions, socio‐demographic, socioeconomic and behavioral characteristics.

Table S3: Interaction analysis for associations between sex, matriliny dimensions, socio‐demographic, socioeconomic, and behavioral factors for health Outcomes in Meghalaya.

Table S4: Distribution of women's health outcomes (nutritional status, central obesity) in relation to matriliny dimensions, socio‐demographic, socioeconomic, and behavioral factors in Meghalaya.

Table S5: Distribution of women's health outcomes (hypertension, diabetes, and anemia) in relation to matriliny dimensions, socio‐demographic, socioeconomic, and behavioral factors in Meghalaya.

Table S6: Distribution of women's health outcomes (malnutrition, central obesity) in relation to empowerment and structural barriers to healthcare access in Meghalaya.

Table S7: Distribution of women's health outcomes (hypertension, diabetes, and anemia) in relation to empowerment and structural barriers to healthcare access in Meghalaya.

Table S8: Distribution of men's health outcomes (malnutrition, central obesity) in relation to matriliny dimensions, socio‐demographic, socioeconomic, and behavioral factors in Meghalaya.

Table S9: Distribution of men's health outcomes (hypertension, diabetes, and anemia) in relation to matriliny dimensions, socio‐demographic, socioeconomic, and behavioral factors in Meghalaya.

Table S10: Distribution of children's sickness and care seeking behavior in relation to matriliny dimensions, socio‐demographic and socioeconomic factors in Meghalaya.

Table S11: Odds ratio (ORs) of women's health outcomes (malnutrition, central obesity) in relation to matriliny dimensions, socio‐demographic, socioeconomic, behavioral factors, and empowerment and structural barriers to healthcare access in Meghalaya.

Table S12: Odds ratio (ORs) of women's health outcomes (hypertension, diabetes, and anemia) in relation to matriliny dimensions socio‐demographic, socioeconomic, behavioral factors, and empowerment and structural barriers to healthcare access in Meghalaya.

Table S13: Adjusted odds ratio (aORs) of women's health outcomes (malnutrition, central obesity) in relation to matriliny dimensions, socio‐demographic, socioeconomic, behavioral factors, and empowerment and structural barriers to healthcare access in Meghalaya.

Table S14: Adjusted odds ratio (aORs) of women's health outcomes (hypertension, diabetes, and anemia) in relation to matriliny dimensions socio‐demographic, socioeconomic, behavioral factors, and empowerment and structural barriers to healthcare access in Meghalaya.

Table S15: Odds ratio (ORs) and adjusted odds ratio (aORs) of men's health outcomes (malnutrition, central obesity) in relation to matriliny dimensions, socio‐demographic, socioeconomic, and behavioral factors in Meghalaya.

Table S16: Odds ratio (ORs) and adjusted odds ratio (aORs) of men's health outcomes (hypertension, diabetes, and anemia) in relation to matriliny dimensions socio‐demographic, socioeconomic, and behavioral factors in Meghalaya.

Table S17: Odds ratio (ORs) and adjusted odds ratio (aORs) of children's sickness and care seeking behavior in relation to matriliny dimensions, socio‐demographic, and socioeconomic factors in Meghalaya.

Data Availability Statement

This study is based on secondary data from the National Family Health Survey (NFHS‐5), India. The datasets are publicly available from the Demographic and Health Surveys (DHS) Program repository (https://dhsprogram.com) upon registration and approval.