Examining the Association Between Occupational Strain and Risk of Angina Pectoris Among Older Working Adults in India Using Karasek’s Job Demand Control Model

Abstract

Background

Occupational strain is a well-known predictor of cardiovascular diseases (CVDs), yet limited evidence exists for older workers in India. Using the Job Demand–Control (JDC) model, this study examines the association between job strain and the risk of Angina Pectoris (AP) among older workers, focusing on psychosocial workplace conditions linked to later life AP risk.

Method

We used Wave 1 (2017–18) of the Longitudinal Aging Study in India (LASI), a nationwide survey, focusing on working older adults (n = 66,331). Occupational strain was conceptualised using the JDC model (high strain, active, passive, low strain), while AP was assessed using the Rose Angina Questionnaire. We employed a multivariable logistic regression model to examine the association between job strain and AP while adjusting for socioeconomic and health-related variables.

Results

After adjusting for socioeconomic and health factors, individuals in both active (OR=0.77, 95% CI: 0.64–0.92) and passive jobs (OR=0.84, 95% CI: 0.74–0.96) exhibited a significantly lower likelihood of AP, whereas low-strain jobs showed a marginally protective but non-significant association compared to high-strain jobs. Early labour force participation (before age 14) and poorer self-rated health were associated with a higher risk of AP. Regional variation was significant, while socioeconomic variables were not significant after adjustment.

Conclusion

This study highlights the role of psychosocial conditions, including occupational strain, in the development of angina in later life. The findings point to the need for better work environments that allow employees more control, decision-making autonomy, and greater skill discretion in their roles. At the same time, early identification and management of job stress and burnout may reduce the burden of angina and other cardiac events among India’s aging workforce.

Introduction

Cardiovascular diseases (CVDs) remain a major cause of death worldwide. In 2022, an estimated 19.8 million people lost their lives due to CVDs, which was responsible for roughly one in three deaths. Nearly 85% of these deaths were due to heart attacks or strokes alone. What stands out is that more than 75% of CVD-related deaths occurred in low- and middle-income countries (LMICs). Of the nearly 18 million early deaths (before age 70) from noncommunicable diseases (NCDs) in 2021, close to 38% were linked to CVDs. The countries with the highest number of CVD deaths include China, India, Russia, the United States, and Indonesia [2].

In 2016, NCDs accounted for 63% of all deaths in India, with CVDs alone resulting in 27% deaths [3]. The age-standardised CVD mortality rate in the same year stood at 282 deaths per 100,000 population, which was higher than the global average of 233 per 100,000, highlighting a serious public health issue [4].

Given the serious socio-economic consequences of these trends, India set a target of reducing premature mortality due to CVDs and other major NCDs by 25% by 2025 under the National Health Policy, 2017 [5]. However, achieving this target remains a major challenge given the estimated economic loss of US$4.58 trillion due to NCDs by 2030, of which nearly half (US$2.17 trillion) is attributed to CVDs alone [6].

One primary clinical manifestation of this burden is Angina Pectoris (AP), a condition characterised by chest pain or discomfort caused by coronary heart disease. AP occurs when the blood supply to the heart muscle does not meet its demand due to obstruction or spasm of the coronary arteries [7]. As angina precedes serious cardiac events, identifying and addressing its risk factors, including psychosocial stressors such as job strain, is critical for reducing the burden of CVDs and associated deaths in India.

Studies across the globe have shown mixed results about the association between an increase in the risk of CVDs and stress at work, which has been measured by the Demand-Control Model (DCM) [8–10]. Despite this variability, the association between this model and CHD has been found to be strong and consistent [12–14].

Several studies have demonstrated that high job strain, characterised by high psychological demands combined with low decision control, is associated with increased cardiovascular risk. For instance, Schnall et al. (1994) observed that job strain (high psychological demands and low decision control) was associated with CVDs, especially AP [14]. Similarly, a Canadian community-based study established that job strain was associated with an increased risk of heart disease in older adults [15]. Likewise, a prospective cohort study further demonstrated that employees in high-strain jobs had twice the CVD mortality risk compared to their counterparts in low-strain jobs [16]. Karasek et al. (1981) also reported that higher job demands and low intellectual discretion are significant predictors of the development of CHD [17]. In addition, one study highlighted exogenous factors like unfavourable occupational conditions, excessive workload, lack of teamwork and endogenous pressures such as individual personality characteristics, are potential risk factors for acute coronary syndromes [18].

Several large scale studies have found an association between high job strain and risk of developing cardiovascular conditions, including the Copenhagen Psychosocial Questionnaire (COPSOQ-1) study, INTERHEART case-control Study, NHANES1 study, Danish Work Environment Cohort Study (DWECS), Finnish Public Sector (FPS) study, WOLF Stockholm study, Intervention Project on Absence and Well-being (IPAW) study, Northern Italian Study, Women’s Health Study, and Mid-life in the United States (MIDUS) cohort study [19–28].

On the contrary, other major studies have reported that job strain was not correlated with cardiovascular conditions among workers. These include the Framingham Offspring Study, Nurses’ Health Study, and WHI-OS (Women’s Health Initiative Observational Study), European cohort studies, Swedish Primary Prevention Study (PPS), Swedish Västerbotten Intervention Programme, and Belgian Job Stress Project (BELSTRESS) [29–35]. A similar association was documented in previous studies [36–38].

In the Indian context, several studies have identified low-density lipoprotein-cholesterol (LDL-C) [39], type-2 diabetes mellitus (T2DM) [40,41], obesity [42,43], smoking [44], insomnia or sleep disorder [45,46], hypertension [47,48], family history of CVD [49,50] and physical inactivity or sedentary lifestyle [51] as major modifiable risk factors for CVDs.

However, the cause of AP due to occupational risk factors is not fully understood in India, particularly among the older working population. To address this gap, the present study examines the association between job strain and AP using nationally representative data. To our knowledge, this is the first study in India to use two globally validated instruments: the Rose Angina Questionnaire (RAQ) for assessing AP and Karasek’s Job Demand-Control (JDC) model for measuring job strain. In addition, our study includes several understudied occupational variables, such as age at entry into the workforce and intention to leave current occupation, among others that have rarely been explored in previous studies. This combination of validated tools, representative data, and unique occupational variables adds novel evidence on the relationship between job strain and AP in India, with important implications for both research and policy.

Karasek’s Job Demand-Control (JDC) Model

The JDC model, introduced by Robert Karasek in 1979, uses 18-item scale of the Job Content Questionnaire (JCQ) to combine psychological demands at work with job decision latitude or control in four categories, or quadrants: 1) low strain (low psychological demands and high decision latitude); 2) passive (low psychological demands and low decision latitude); 3) active (high psychological demands and high decision latitude); and 4) high strain (high psychological demands and low decision latitude) [52,53].

Psychological demands refer to the mentally demanding aspects of a main job, such as working under time pressure, heavy workload, strenuous tasks, and job complexity. In contrast, job decision latitude, also known as job control, refers to the level of autonomy and skill discretion an employee has in performing their work or assigned activities [54–56]. The high strain quadrant is related to the highest risk of adverse health, and the low strain quadrant is proposed to have the lowest risk, whereas the active and passive quadrants have intermediate risks [54].

Based on the above, this study examines the association between job strain and AP using nationally representative data from India. We hypothesize that the risk of AP is higher among individuals employed in high-strain jobs compared to those in other jobs.

Method

Study Design and Setting

We used Wave 1 of the Longitudinal Aging Study in India (LASI), a nationally representative study in India. LASI covers 73,396 individuals aged 45 and above, including their spouses, regardless of age. The sample included 31,902 older adults (aged 60 and above) and 6,880 oldest-old individuals (aged 75 and above) drawn from all states and union territories of India. Data collection for LASI Wave 1 took place between April 2017 and December 2018.

Households in LASI were considered eligible if they included at least one member aged 45 years or older. LASI offers detailed information on the socio-economic and health dimensions of the aging population in India and is harmonized with the HRS in the United States and other aging surveys in the world, for instance, CHARLS, SHARE, MHAS, KLoSA and JSTAR. The study employed a multistage, stratified, area-probability cluster sampling design to arrive at the final sample [57,58].

All demographic, occupational, and health-related variables used in this study were derived and recoded from items included in the survey instrument, which is available for research purposes and provides a detailed description of items and response categories relevant to each construct (lasiindia.org/public/documentation/LASI_Questionnaire.pdf#page = 182.05). Moreover, this study is a secondary analysis and does not involve any intervention; therefore, clinical trial registration is not applicable.

Outcome variable

We assessed the presence of AP using the Rose Angina Questionnaire (RAQ) from the World Health Organisation, which is a validated instrument for detecting angina in population-based studies [59]. Respondents were confirmed to have angina if they ever experienced exertional chest pain or discomfort that met the following criteria: 1) pain occurring during physical activity such as “walking uphill, hurrying, or walking at a normal pace on level ground”; 2) pain located in the “sternal region or the left side of the chest, with or without radiation to the left arm”; 3) pain severe enough to “cause the individual to stop or slow down”; and 4) symptoms that “resolved within 10 minutes after stopping or reducing activity.” We summed these four cardiovascular conditions and coded them into a binary variable (0 = No, 1 = Yes) [60].

Explanatory variables

Further, to account for the strain in the main job/occupation we used an adapted Job Content Questionnaire (JCQ), which is based on the Job Demand-Control (JDC) model by combining nine occupational demand items available in the original survey, wherein respondents were asked how often their job required the following tasks: 1) “a lot of physical effort”, 2) “lifting heavy loads”, 3) “stooping, kneeling, or crouching”, 4) “good eyesight”, 5) “intense concentration or attention”, 6) “skill in dealing with other people”, 7) “exposure to burning material, exhaust, or smoke (excluding car exhaust)”, 8) “exposure to chemicals, pesticides, or herbicides”, and 9) “exposure to noxious odors” [58]. These demand items were averaged, reverse-coded (1 = low, 4 = high) and then categorised into four quadrants: 1) “High Strain (high psychological demand, low control)”, 2) “Active (high demand, high control)”, 3) “Passive (low demand, low control)”, and 4) “Low Strain (low demand, high control) [61]. These items had acceptable internal reliability, with a Cronbach’s α of 0.71.

Covariates

Demographic and social background

Age was categorised into < 45 years, 45–59 years, and ≥ 60 years. Sex was coded as male or female. Marital status was classified as currently married, widowed, or not in marital union. Educational attainment was grouped into no education, primary, secondary, and higher education. Caste was categorised as Scheduled Caste, Scheduled Tribe, Other Backwards Classes, or Others. Religion was classified as Hindu, Muslim, Christian, or Other. Place of residence was coded as rural or urban, while geographic region was classified into six major regions based on state of residence.

Socioeconomic position

Household economic status was measured using an abridged version of the NSS consumption expenditure schedule, which collected information on 11 food and 29 non-food expenditure items on the basis of reference periods of 7 days and 30 days preceding the survey. This information was used to compute the “monthly per capita consumption expenditure (MPCE)”, which in turn represents overall household consumption levels. In the original survey, MPCE was categorised into five quintiles ranging from the poorest to the richest, which we subsequently regrouped into three categories in our analysis: “Poor (lowest two quintiles)”, “Middle (third quintile)”, and “Rich (top two quintiles)” [62].

Lifestyle behaviors

Smoking history was defined as a binary variable based on a history of smoking (cigarette, bidi, cigar, hookah, cheroot) or smokeless tobacco (chewing tobacco, gutka, pan masala, etc) products [63]. Alcohol use was similarly coded into a dichotomous variable based on lifetime consumption of both commercially and traditional or locally produced alcoholic beverages.

Occupational and work-history characteristics

In relation to the legal working age of 14 in India, age at first work was categorised as early (< 14 years) and late (≥ 14 years)

Further, the work status (0 = employed, 1 = self-employed) of individuals was assessed based on responses to two survey items. Respondents were classified as “self-employed” if they reported owning a business or farm or identified their occupation as private or entrepreneur. They were classified as “employed” if they reported working in government, public sector, private sector (non-entrepreneur), as a casual labourer, or in other employment types without owning a business or farm. Non-workers or those with incomplete data were coded as missing in our study. Further, respondents in seasonal or temporary jobs/occupations were also dropped from our study.

In addition, job turnover intention was generated as a binary variable (0 = Intention to Stay, 1 = Intention to Leave) based on responses to four survey items asked of respondents who reported ever working. Respondents were asked whether: 1) they were currently looking for another job, 2) they registered with an employment exchange, 3) they registered with the MGNREGA, which was asked only in rural areas, and 4) they had contacted prospective employers in the past month. The variable was categorised into “Intention to Leave (ITL)” if respondents answered yes to any of these questions, which indicates active job-seeking behaviour and “Intention to Stay (ITS)” if respondents gave no positive responses to these three questions. Non-workers and respondents with incomplete work histories were excluded from the present study.

Functional health and disability

We assessed difficulties in performing activities of daily living (ADL) and instrumental activities of daily living (IADL) because of physical, mental or cognitive issues using the Barthel Index (BI) [64]. For ADL, respondents were asked about difficulties faced by them in performing six daily activities: 1) “dressing, including putting on footwear”, 2) “walking across a room”, 3) “bathing”, 4) “eating”, 5) “getting in or out of bed”, and 6) “using toilet including sitting down and standing up”. For IADL, respondents were asked about difficulties in seven important activities: 1) “preparing a hot meal including cooking and serving”, 2) “grocery shopping”, 3) “making telephone calls”, 4) “taking prescribed medicines”, 5) “doing household or garden work”, 6) “managing finances such as paying bills or tracking expenses”, and 7) “moving around or navigating unfamiliar places”. Each item was coded as binary (0 = No difficulty, 1 = Difficulty) based on responses of “Yes” or “No”, excluding difficulties expected to last less than three months. An ADL score (0–6) was calculated by summing the six ADL items, and an IADL score (0–7) was calculated by summing the seven IADL items. We created three categories for the ADL score: “No” (score = 0), “Moderate” (score = 1–2), and “Severe” (score ≥ 3) difficulty. Similarly, the IADL score was categorised into three levels: “No” (score = 0), “Moderate” (score = 1–2), and “Severe” (score ≥ 3) difficulty.

Subjective health and well-being measures

We classified self-rated health (SRH) into three levels: Poor (coded as 0), Moderate (coded as 1), and Good (coded as 2) by merging “good” and “fair” into a single category. Further, we generated satisfaction with life using the “Satisfaction with Life Scale (SWLS)”, which measures overall quality of life. Respondents in the survey rated their life based on five evaluative statements: 1) “my life is close to ideal”, 2) “the conditions of my life are excellent”, (3) “I am satisfied with my life”, (4) “I have achieved the important things I want in life”, and (5) “I would change almost nothing if I could live my life over”. Each item was recorded on a seven-point Likert scale ranging from Strongly Disagree (1) to Strongly Agree (7). A total life satisfaction score was calculated by summing the responses to these five items, ranging from 5 to 35. The score was then categorised as “Low” (5–20), “Medium” (21–25), or “High” (26–35) [65]. These items demonstrated excellent internal reliability, with a Cronbach’s α of 0.89.

Mental health and sleep indicators

We evaluated the presence of depressive symptoms using a 10-item version of the “Centre for Epidemiologic Studies Depression (CES-D) Scale”. We generated a binary variable (0 = Not depressed, 1 = Depressed) based on a set of questions from LASI survey about symptoms experienced by respondents over a period of two weeks or more in a year. Respondents were asked whether they: 1) “felt depressed”, 2) “found everything to be an effort”, 3) “had sleep restlessness”, 4) “lost interest in most things”, 5) “felt unusually tired or lacking energy”, 6) “experienced loss of appetite”, 7) “had increased appetite” (if no loss of appetite was reported), 8) “faced greater difficulty concentrating”, 9) “felt worthless or down on themselves”, and 10) “thought a lot about death (their own, others’, or in general)”. Responses to feeling depressed, sleep restlessness and everything being an effort were based on frequency, ranging from every day, almost every day, less often, with respondents reporting these symptoms every day or almost every day being said to have depressive symptoms. Other symptoms were binary (Yes/No). We calculated a CES-D score (0–10) by summing the presence of these symptoms, with a score of 4 or higher indicating the presence of depression [66,67].

The presence of insomnia symptoms (coded as 0 for ‘No’ and 1 for ‘Yes’) was assessed using the Jenkins Sleep Scale (JSS-4), which remains a widely used scale for evaluating sleep disorders in population-based studies. In the LASI survey, respondents were asked four questions about sleep difficulties experienced over the past month, which included how often they had 1) “difficulty falling asleep”, 2) “waking up at night and struggling to return to sleep”, 3) “waking up too early and being unable to fall asleep again”, and 4) “feeling unrested during the day despite adequate sleep”. Each item was rated as “Never or rarely” (1–2 nights per week), “Occasionally” (3–4 nights per week), or “Frequently” (5 or more nights per week). Following Jenkins et al. (1988), respondents were classified as having insomnia if they reported experiencing any of these symptoms occasionally or frequently [68].

Clinical morbidity and subjective health complaints

Morbidity status was assessed using self-reported diagnoses of nine specific conditions: 1) hypertension or high blood pressure, 2) diabetes or high blood sugar, 3) cancer or a malignant tumor, 4) chronic lung disease such as asthma, chronic obstructive pulmonary disease, chronic bronchitis, or other chronic lung problems, 5) chronic heart disease such as coronary heart disease, congestive heart failure, or other chronic heart problems, 6) stroke, 7) musculoskeletal conditions such as arthritis, rheumatism, osteoporosis, or other disorders of the bones and joints, 8) neurological and mental health conditions including depression, Alzheimer’s disease, dementia, bipolar/unipolar disorders, seizures, or Parkinson’s disease, and 9) elevated cholesterol levels [69]. We recoded each condition as “Yes” or “No” based on self-reported conditions. We then defined the morbidity status as “no morbidity” (coded as 0) if respondents reported no conditions, “single morbidity” (coded as 1) if one condition was reported, and “multimorbidity” (coded as 2) if two or more conditions were reported.

Finally, the presence of subjective health complaints (SHC) captured on the basis of self-reported symptoms experienced in the past two years: 1) “pain/stiffness in joints”, 2) “persistent swelling of lower limbs”, 3) “breathlessness while awake”, 4) “persistent dizziness or light headedness”, 5) “back pain”, 6) “persistent headaches”, 7) “severe fatigue or exhaustion”, 8) “wheezing/whistling sound from the chest”, and 9) “cough with or without sputum”. If respondents reported any of these symptoms, they were classified into “Yes” (coded as 1) and “No” (coded as 0) if they reported otherwise [70].

Statistical analysis

We analysed the data in Stata version 17 (StataCorp, 2017. Stata Statistical Software: Release 17. College Station, TX: StataCorp LP).

Before we began statistical analysis, we checked for potential multicollinearity for all explanatory variables using the Variance Inflation Factor (VIF). The mean VIF was 1.56, and all individual VIFs were well below the conventional threshold of 10 (and even the conservative threshold of 5), which indicates no serious multicollinearity problem in our model. We then executed the svyset command to account for the complexity of the survey design of LASI and to adjust for sample weight in our analysis. We used the chi-square (χ²) test to examine the association between AP and each predictor variable under study and obtained p-values from this analysis. We employed a binary logistic regression model in our study to understand the association between the outcome and predictor variables. The equation for the logistic regression model can be defined as follows:

$$\text{log}\left(\frac{pi}{1-pi}\right)=\mathit{\text{logit}}\left(pi\right)={\beta }_0+{\beta }_{x1}+{\beta }_{x2}+\dots {\beta }_n{x}_n$$

In the equation given above, pi denotes the probability of an individual i experiencing AP, while 1 − pi is the probability of an individual i not experiencing AP. x1, x2..Xn are the predictors, β₀ is the intercept and β₁, β₂….β_n are the coefficients. The coefficients have been exponentiated to obtain odds ratios (OR), which represent the change in the odds of AP associated with a one-unit change in the predictor, while holding other factors constant.

We used two models in our study. In the first model, we used a univariate logistic regression model to derive crude odds ratios (ORs) to assess the independent effect of occupational strain on AP without adjusting for confounders. In the second model, we fitted a multivariable logistic regression model by including all covariates. This allowed us to estimate the adjusted odds ratios (AORs) while accounting for potential confounders. We obtained p-values based on logistic regression analysis. p < 0.05 is said to be statistically significant, while p < 0.001 is highly significant.

We also estimated the predicted probabilities of AP in relation to occupational strain by giving the margins command, and visualised these probabilities by applying the marginsplot command.

Ethical Considerations

In accordance with ethical guidelines, written informed consent was obtained from all age-eligible participants by the LASI survey team. The study protocol was reviewed and approved by the Institutional Review Boards of the Indian Council of Medical Research (ICMR, F.No.T.21012/07/2012-NCD), the International Institute for Population Sciences (Sr. No. 12/1054), the Harvard T.H. Chan School of Public Health (CR-16715–10), and the University of Southern California (UP-CG-14_00005) [70,71].

Results

Sample Characteristics and Regional Variation in the Prevalence of AP

Table 1 shows the prevalence of AP among older adults in India. Overall, about 11% (approximately 12 out of every 100) reported angina symptoms. Prevalence varied modestly by region and job quadrant. Regionally, AP was highest in the central region (17.65%), followed by the western region (12.86% and lowest in the south (9.68%) and northeast (9.75%). By job quadrants, the highest prevalence was seen among older adults in active jobs (12.02%), followed by high-strain (11.33%), low-strain (11.27%), and passive jobs (10.85%).

Table 1.

Background Characteristics of Older Adults with Angina Pectoris (AP) (n = 66,331), LASI Wave-1 (2017–18)

Characteristics	Absence of AP		Presence of AP
	(N = 58,616)		(N = 7,715)
	n	%	n	%	p-value
Age (years)
<45	6,129	90.44	648	9.56	< 0.001
45–59	30,834	89.03	3,801	10.97
60+	21,653	86.89	3,266	13.11
Sex
Male	24,993	90.09	2,749	9.91	< 0.001
Female	33,623	87.13	4,966	12.87
Marital Status
Currently in marital union	47,223	88.74	5,990	11.26	< 0.001
Widowed	9,419	85.91	1,545	14.09
Divorced/Separated/Deserted/Others	1,971	91.63	180	8.37
Level of Education
No education	25,297	86	4,118	14.00	< 0.001
Primary	14,377	87.62	2,031	12.38
Secondary	12,280	91.38	1,159	8.62
Higher	6,662	94.24	407	5.76
Caste
SC	9,681	87.32	1,406	12.68	< 0.001
ST	10,546	89.95	1,178	10.05
OBC	22,154	88.04	3,010	11.96
Others	14,127	88.52	1,832	11.48
Religion
Hindu	42,779	88.08	5,789	11.92	< 0.001
Muslim	6,890	87.16	1,015	12.84
Christian	5,931	91.02	585	8.98
Others	3,011	90.23	326	9.77
MPCE
Rich	23,407	87.58	3,320	12.42	< 0.001
Middle	11,800	88.66	1,510	11.34
Poor	23,409	89.03	2,885	10.97
Place of Residence
Rural	36,969	86.63	5,705	13.37	< 0.001
Urban	21,647	91.5	2,010	8.5
Region
North	12,834	88.01	1,749	11.99	< 0.001
South	14,232	90.32	1,526	9.68
East	10,320	89.3	1,236	10.70
West	7,837	87.14	1,157	12.86
Central	3,701	82.35	793	17.65
Northeast	8,776	90.25	948	9.75
Smoking
No	38,669	88.93	4,812	11.07	< 0.001
Yes	19,393	87.12	2,867	12.88
Drinking
No	48,328	88.29	6,411	11.71	0.614
Yes	9,748	88.46	1,272	11.54
Age at first job
Before 14	6,039	83.75	1,172	16.25	< 0.001
At or after 14	52,577	88.93	6,543	11.07
Employment Status
Employed	4,970	24.70	466	19.89	< 0.001
Self-employed	15,152	75.30	1,877	80.11
Current Main Job Quadrants
High Strain	11,791	88.67	1,506	11.33	0.007
Passive	6,957	89.15	847	10.85
Active	30,982	87.98	4,233	12.02
Low Strain	8,886	88.73	1,129	11.27
Job Turnover Intention
Intention to Stay	28,635	88.99	3,541	11.01	< 0.001
Intention to leave	1,024	85.05	180	14.95
ADL Limitation
No difficulty	52,294	89.46	6,162	10.54	< 0.001
Moderate	4,460	80.32	1,093	19.68
Severe	1,862	80.19	460	19.81
IADL Limitation
No difficulty	42,556	90.45	4,492	9.55	< 0.001
Moderate	9,157	84.77	1,645	15.23
Severe	6,903	81.39	1,578	18.61
Self-Rated Health (SRH)
Poor	7,386	78.25	2,053	21.75	< 0.001
Moderate	23,067	86.44	3,619	13.56
Good	27,618	93.23	2,004	6.77
Life Satisfaction
Low	16,981	87.52	2,421	12.48	< 0.001
Medium	14,300	87.86	1,975	12.14
High	26,005	89.08	3,187	10.92
Depressive Symptoms
No	54,527	89.11	6,665	10.89	< 0.001
Yes	3,191	76.49	981	23.51
Insomnia
No	49,279	90.61	5,108	9.39	< 0.001
Yes	9,121	77.78	2,605	22.22
Morbidity Status
No morbidity	34,388	91.67	3,124	8.33	< 0.001
Single morbidity	15,129	86.16	2,431	13.84
Multimorbidity	9,099	80.82	2,160	19.18
Subjective Health Complaints (SHC)
No	862	94.41	51	5.59	< 0.001
Yes	57,306	88.22	7,653	11.78

Notes: Reported p-values are from Pearson’s chi-square tests, where smaller p-values indicatestronger evidence of an association between the variables

Sociodemographic Predictors of AP among Older Adults in India

Table 2 summarises findings from the logistic regression model, which shows the association between AP and various sociodemographic, health, and work-related characteristics among older adults in India.

Table 2.

Univariate and multivariable logistic regression of Angina Pectoris (AP)

Characteristics	Univariate		Multivariable
	Unadjusted OR [95% CI]	P-value	Adjusted OR [95% CI]	P-value
Age (years)
<45	ref
45–59	1.17 [1.07–1.27]	0.001	0.81 [0.68–0.97]	0.02
60+	1.43 [1.31–1.56]	< 0.001	0.73 [0.60–0.90]	0.002
Sex
Male	ref
Female	1.34 [1.28–1.41]	< 0.001	1.26 [1.11–1.44]	< 0.001
Marital Status
Currently married	ref
Widowed	1.29 [1.22–1.37]	< 0.001	0.98 [0.85–1.13]	0.819
Divorced/Separated/Deserted/Others	0.72 [0.62–0.84]	< 0.001	0.71 [0.53–0.93]	0.015
Level of Education
No education	ref
Primary	0.87 [0.82–0.92]	< 0.001	0.91 [0.80–1.03]	0.137
Secondary	0.58 [0.54–0.62]	< 0.001	0.83 [0.71–0.96]	0.014
Higher	0.38 [0.34–0.42]	< 0.001	0.63 [0.51–0.78]	< 0.001
Caste
SC	ref
ST	0.77 [0.71–0.83]	< 0.001	0.88 [0.74–1.04]	0.138
OBC	0.94 [0.87–1.00]	0.053	0.96 [0.85–1.09]	0.553
Others	0.89 [0.83–0.96]	0.003	0.95 [0.82–1.12]	0.564
Religion
Hindu	ref
Muslim	1.09 [1.01–1.17]	0.02	1.05 [0.89–1.25]	0.55
Christian	0.73 [0.67–0.80]	< 0.001	0.97 [0.79–1.19]	0.768
Others	0.80 [0.71–0.90]	< 0.001	0.93 [0.74–1.17]	0.54
MPCE
Rich	ref
Middle	0.90 [0.85–0.96]	0.002	0.98 [0.86–1.12]	0.773
Poor	0.87 [0.82–0.92]	< 0.001	0.95 [0.85–1.06]	0.385
Place of Residence
Rural	ref
Urban	0.60 [0.57–0.63]	< 0.001	0.68 [0.61–0.76]	< 0.001
Region
North	ref
South	0.79 [0.73–0.85]	< 0.001	0.87 [0.75–1.02]	0.086
East	0.88 [0.81–0.95]	0.001	0.78 [0.66–0.92]	0.004
West	1.08 [1.00–1.17]	0.048	1.43 [1.21–1.69]	< 0.001
Central	1.57 [1.43–1.72]	< 0.001	2.14 [1.77–2.58]	< 0.001
Northeast	0.79 [0.73–0.86]	< 0.001	1.13 [0.93–1.38]	0.208
Smoking
No	ref
Yes	1.19 [1.13–1.25]	< 0.001	1.23 [1.11–1.37]	< 0.001
Drinking
No	ref
Yes	0.98 [0.92–1.05]	0.614	0.99 [0.87–1.13]	0.904
Age at first job
Before 14	ref
At or after 14	0.64 [0.60–0.69]	< 0.001	0.82 [0.73–0.93]	0.002
Employment Status
Employed	ref
Self-employed	1.32 [1.19–1.47]	< 0.001	1.07 [0.95–1.21]	0.252
Current Main Job Quadrants
High Strain	ref
Passive	0.95 [0.87–1.04]	0.293	0.84 [0.74–0.96]	0.007
Active	1.07 [1.00–1.14]	0.035	0.77 [0.64–0.92]	0.004
Low Strain	0.99 [0.92–1.08]	0.9	0.89 [0.80–1.01]	0.063
Turnover Intention
Intention to Stay	ref
Intention to leave	1.42 [1.21–1.67]	< 0.001	1.45 [1.18–1.79]	< 0.001
ADL Limitation
No difficulty	ref
Moderate	2.08 [1.94–2.23]	< 0.001	1.35 [1.15–1.58]	< 0.001
Severe	2.10 [1.89–2.33]	< 0.001	1.18 [0.87–1.59]	0.284
IADL Limitation
No difficulty	ref
Moderate	1.70 [1.60–1.81]	< 0.001	1.20 [1.06–1.36]	0.004
Severe	2.17 [2.03–2.31]	< 0.001	1.39 [1.18–1.63]	< 0.001
Self-Rated Health (SRH)
Poor	ref
Moderate	0.56 [0.53–0.60]	< 0.001	0.75 [0.66–0.85]	< 0.001
Good	0.26 [0.24–0.28]	< 0.001	0.42 [0.37–0.49]	< 0.001
Life Satisfaction
Low	ref
Medium	0.97 [0.91–1.03]	0.326	1.17 [1.03–1.32]	0.014
High	0.86 [0.81–0.91]	< 0.001	0.96 [0.85–1.07]	0.468
Depressive Symptoms
No	ref
Yes	2.52 [2.33–2.71]	< 0.001	1.28 [1.09–1.51]	0.002
Insomnia
No	ref
Yes	2.76 [2.62–2.90]	< 0.001	1.93 [1.73–2.15]	< 0.001
Morbidity Status
No morbidity	ref
Single morbidity	1.77 [1.67–1.87]	< 0.001	1.59 [1.42–1.78]	< 0.001
Multimorbidity	2.61 [2.46–2.77]	< 0.001	2.31 [2.02–2.65]	< 0.001
Subjective Health Complaints (SHC)
No	ref
Yes	2.26 [1.70–3.00]	< 0.001	2.04 [1.13–3.68]	0.018

Notes: Ref: Reference Category; OR: Odds Ratio; CI: Confidence Interval

Compared with those under 45 years, individuals aged 45–59 years were 19% less likely to have AP (AOR = 0.81; 95% CI: 0.68–0.97), and those aged 60 years or older were 27% less likely (AOR = 0.73; 95% CI: 0.60–0.90). Women were 26% more likely to have angina than men (AOR = 1.26; 95% CI: 1.11–1.44).

Relative to currently married respondents, those who were divorced, separated, deserted, or in other marital statuses were 29% less likely to have angina (AOR = 0.71; 95% CI: 0.53–0.93), while we found no association for widowed individuals.

Education showed a protective trend: secondary education was 17% less likely to have angina (AOR = 0.83; 95% CI: 0.71–0.96), and higher education was 37% less likely (AOR = 0.63; 95% CI: 0.51–0.78) compared with no education. We found no association for primary education. We also found no association between caste, religion, or socioeconomic status (MPCE) and AP after adjustment. Those living in urban areas had a 32% lower likelihood of angina compared with those residing in rural areas (AOR = 0.68; 95% CI: 0.61–0.76).

Regional differences were observed relative to the North. The East was 22% less likely to have AP (AOR = 0.78; 95% CI: 0.66–0.92), while the West was 43% more likely (AOR = 1.43; 95% CI: 1.21–1.69), and the Central region was 114% more likely (AOR = 2.14; 95% CI: 1.77–2.58). The South or Northeast regions were not significant.

Entering the workforce at or after age 14 was 18% less likely to have AP than entering before age 14 (AOR = 0.82; 95% CI: 0.73–0.93). The current employment status (self-employed vs. employed) was not associated with AP.

Compared with high-strain jobs, passive jobs (AOR = 0.84; 95% CI: 0.74–0.96) were 16% less likely to have AP, and active jobs were 23% less likely (AOR = 0.77; 95% CI: 0.64–0.92). On the other hand, low-strain jobs (AOR = 0.89; 95% CI: 0.80–1.01, p = 0.063) were not statistically significant but showed a marginally protective effect.

In accordance with the JDC model, predicted probabilities were highest in high-strain jobs and lowest in active jobs (Fig. 1), supporting the role of high job demands combined with low control in negative health outcomes.

Figure 1.

Predicted Probability of AP across Job Quadrants

Those with a higher intention to leave (ITL) their current occupation had 45% higher adjusted odds of angina than those without ITL (AOR = 1.45; 95% CI: 1.18–1.79).

For functional limitations, moderate ADL were 35% more likely to report AP than no limitations (AOR = 1.35; 95% CI: 1.15–1.58), while severe limitations were not significant. In contrast, moderate IADL limitations were 20% more likely (AOR = 1.20; 95% CI: 1.06–1.36), and severe IADL limitations were 39% more likely (AOR = 1.39; 95% CI: 1.18–1.63) compared to no IADL limitations.

A smoking history was 23% more likely to have AP (AOR = 1.23; 95% CI: 1.11–1.37), whereas alcohol consumption was not associated.

Moderate SRH was 25% less likely to experience AP (AOR = 0.75; 95% CI: 0.66–0.85), while good SRH was 58% less likely (AOR = 0.42; 95% CI: 0.37–0.49) compared to poor SRH. Medium life satisfaction was 17% more likely to report AP (AOR = 1.17; 95% CI: 1.03–1.32). We found no association with high satisfaction. Depressive symptoms had 28% more likelihood of AP (AOR = 1.28; 95% CI: 1.09–1.51), and insomnia symptoms were 93% more likely to have AP (AOR = 1.93; 95% CI: 1.73–2.15). Subjective health complaints in the past two years were also 104% more likely to have AP (AOR = 2.04; 95% CI: 1.13–3.68).

Finally, while individuals with single morbidity reported 59% higher likelihood of having AP (AOR = 1.59; 95% CI: 1.42–1.78), and those with multiple morbidities were 131% more likely (AOR = 2.31; 95% CI: 2.02–2.65) compared to those with no morbidity.

Discussion

Our study examined the relationship between occupational strain, psychosocial factors, behavioural risks, multimorbidity, and AP among working older adults in India using Karasek’s JDC model. We noted several key findings in our study.

First, those who entered the workforce early (before the age of 14) reported a higher likelihood of AP. Second, high job strain was linked to a higher risk of AP, whereas those in active jobs showed the lowest risk, as confirmed by predicted probabilities. Third, psychosocial factors such as poor self-rated health, depression, insomnia and subjective health complaints were associated with elevated risks of angina in our study. Fourth, limitations in IADL and ADL and single or multiple morbidities substantially raised the risk of AP. Fifth, rural residents reported a higher risk in comparison to urban residents. Sixth, subjective health problems stood out as a strong predictor of angina in our analysis. Finally, a clear regional variation emerged as a strong predictor of angina after adjustment.

We observed that the risk of AP was less common among older age groups (45–59 years and ≥ 60 years) compared with individuals younger than 45 years. Although the risk of CVDs typically increases with age, this pattern may be due to the characteristics of our working sample, where older adults who remain in the labour force tend to be healthier than those who have already withdrawn due to illness [72,73]. Second, older adults may experience underdiagnosis or under-reporting of angina [74], lower health-seeking behaviour [75], or normalisation or underestimation of symptoms as part of aging, particularly in resource-constrained settings [76].

Further, elderly women in our study were more likely to report angina than elderly men, a pattern widely reported in studies from both high- and middle-income countries [77–79]. Older adults who were not in a marital union were less likely to have angina than those who were currently married, which is in line with an earlier study that the prevalence of AP was higher among those in a marital union [80]. We found an inverse relationship between educational level and AP, where individuals with lower educational attainment exhibited a higher risk of CHD [81–84]. Interestingly, no significant association was observed between caste, religion, or MPCE and risk of AP after adjusting for key covariates. We assume that in large population surveys, these factors may not always be strong predictors of self-reported medical conditions, including angina, due to underreporting and underdiagnosis of CVDs in a country like India [85,86]. Older adults living in urban areas were less likely to experience angina than their rural counterparts, which is in agreement with national and international estimates [57,87–89]. Likewise, we observed a large state-level variation in angina burden in our study, which has been documented in various studies in the past [90–92]. These variations may be attributed to factors such as lower awareness, poor treatment-seeking behaviour and the presence of higher undiagnosed cases in rural settings.

Our results reaffirm that smoking is a major risk factor for AP, consistent with a large body of literature from across the globe [93–95]. We noted no association between drinking alcohol and elevated risk of AP. Studies have shown mixed results. For instance, Merry et al. (2011) found no clinically significant evidence to support this relationship [96], while others reported a meaningful association [97,98].

It was further noted that early initiation of the work significantly raises the risk of AP, which conceptually aligns with findings from previous studies [99]. In contrast, our analysis found no association between employment status and the risk of angina. We believe that the nature of employment alone may not capture the conditions that can increase the risk of AP [100]. We, therefore, recommend that future studies consider physical work environment, such as a noisy workplace, long working hours, variation in shift and physical demands, rather than employment status per se, as the more proximate determinants of CVD, including AP.

Furthermore, adjusted odds ratios revealed an unexpected pattern for passive and active jobs, which should be interpreted with caution in light of mixed evidence across studies. [27,55,101–104]. However, predicted probabilities derived from the same model revealed a pattern consistent with the JDC model, where the highest probability of angina was observed in high-strain and passive jobs. These roles are known to activate the stress response systems, e.g. HPA axis and sympathetic nervous system in the body [105,106], which trigger stress hormones and result in increased cardiovascular activity, including promotion of arterial inflammation [107,108]. Over time, repeated activation of these stress systems creates an accumulated physiological burden, referred to as ‘allostatic load’, which speeds up the buildup of plaque in the arteries and reduces cardiac blood flow. This in turn raises the risk of angina and other cardiovascular events [109]. In contrast, these probabilities were lowest in active and low-strain jobs due to lower allostatic load in the body and higher skill discretion and autonomy in jobs.

Further, in line with Sara et al. (2018) and Kachi et al. (2019), we noted that those with a higher turnover intention, possibly due to higher job dissatisfaction or burnout, were more likely to experience adverse health outcomes, including AP[110,111].

Likewise, older adults with functional limitations, especially limitations in performing IADL, were at higher risk of AP. This pattern is consistent with earlier studies10[112–114]. We also found an inverse association between self-rated health and angina, consistent with existing studies [115–117]. Accordingly, psychological distress was significantly associated with AP in our study, supported by studies across countries [11,73,100].

Moreover, older adults who exhibited insomnia symptoms were at higher risk of AP than those without these symptoms, which is in tune with studies that poor sleep and/or self-reported sleep disturbances are positively linked to CVD, including stable angina pectoris (SAP) and unstable angina pectoris (UAP) [118–121]. In accordance with previous literature, we observed that those with single or multiple morbidities were more likely to develop AP in our study [94,122,123]. Our findings further suggest that older adults who reported subjective health complaints over the last two years were more likely to experience AP, a finding consistent with prior studies [116–118].

In summary, our study contributes to a growing body of literature that behavioural and non-behavioural factors determine the risk of angina. In our study, geography emerged as a strong predictor of risk. These regional disparities are likely due to differences in access to healthcare, health-seeking behaviour, lifestyle factors, dietary patterns, socioeconomic and cultural background and awareness related to screening and diagnoses of AP, which highlights the need to introduce locally responsive public health interventions and target key determinants of CVDs. The absence of associations between caste, religion, and SES suggests that the risk of AP may be distributed across social groups in ways that may not be fully captured by traditional measures. By and large, our results resonate with the propositions outlined in the JDC that individuals exposed to high-strain jobs experience elevated risks of AP. In addition, subjective health problems emerged as an important predictor of angina in our study [124–126].

In conclusion, our study calls for improvements to existing laws in India, such as the NPSHEW, 2009, Factories Act, 1948, and OSH&WC Code, 2020. We also recommend early management of job-related stress and burnout, providing workplace counselling and improving the working conditions of female workers.

Strengths and limitations of the study

A major strength of this study lies in the use of a large and nationally representative sample, supporting the generalizability of the findings. To our knowledge, this is among the first studies in India to utilised the globally validated JDC model to examine AP risk among older adults. However, this study has some inherent limitations due to its cross-sectional design and the pending release of wave 2 data, which restricts our ability to establish a causal association. Furthermore, we cannot exclude the possibility of recall bias, particularly in self-reported variables such as MPCE, self-rated health and chronic morbidities. Finally, unmeasured confounders in this study, such as family history of heart disease, hypertension, BMI, physical inactivity and other work-related variables, can be considered in future analysis.

The LASI project was co-funded by the NIA/NIH (R01 AG042778), the Ministry of Health & Family Welfare, the Government of India, and the United Nations Population Fund India. It was developed and collected by the International Institute for Population Sciences (IIPS), Harvard T. H. Chan School of Public Health (HSPH), & University of Southern California (USC). The authors declare that they were not involved in the design of the study or the collection of data, and that they received no funding to do this research.

Abbreviations

ADL

Activities of Daily Living

AP

Angina Pectoris

CHARLS

China Health and Retirement Longitudinal Study

HPA

Hypothalamic-Pituitary-Adrenal Axis

HRS

Health and Retirement Study

IADL

Instrumental Activities of Daily Living

ITL

Intention to Leave

ITS

Intention to Stay

JDC

Job-Demand Control

JSTAR

Japanese Study of Aging and Retirement

Korean

Longitudinal Study of Aging

LASI

Longitudinal Aging Study in India

MGNREGA

Mahatma Gandhi National Rural Employment Guarantee Act, 2005

MHAS

Mexican Health and Aging Study

MPCE

Monthly Per Capita Consumption Expenditure

NPSHEW

National Policy on Safety, Health and Environment at Workplace

NSS

National Sample Survey

OSH&WC

Occupational Safety, Health and Working Conditions Code

SHARE

Survey of Health, Aging and Retirement in Europe

Biographies

Pravesh Kumar is a research scholar at the Jindal School of Public Health and Human Development (JSPH).

Dr. Yoshiko Ishioka Miyata is a faculty at the Jindal School of Liberal Arts and Humanities (JSLH), O.P. Jindal Global University (Institution of Eminence, Government of India), Delhi NCR, India.

Funding Statement

The LASI project was co-funded by the NIA/NIH (R01 AG042778), the Ministry of Health & Family Welfare, the Government of India, and the United Nations Population Fund India. It was developed and collected by the International Institute for Population Sciences (IIPS), Harvard T. H. Chan School of Public Health (HSPH), & University of Southern California (USC). The authors declare that they were not involved in the design of the study or the collection of data, and that they received no funding to do this research.

Availability of data and materials

Data Availability

The LASI data used in this study are freely available in the public domain. The LASI was released through the websites of the Gateway to Global Ageing Data (https://g2aging.org) and the International Institute for Population Sciences (IIPS) (www.iipsindia.ac.in/lasi). A data request form is available on the IIPS website (https://iipsindia.ac.in/content/data-request). To access the data, users can register and submit a statement of purpose for the use of the data.