Indian Journal of Public Health

: 2021  |  Volume : 65  |  Issue : 3  |  Page : 231--236

Health insurance coverage and its determinants among middle-income households in Urban Puducherry: A mixed methods study

Poomathy Ponnusamy1, Vinayagamoorthy Venugopal2, Amol R Dongre3,  
1 Assistant Professor, Community Medicine, Government Tiruppur Medical College, Tamil Nadu, India
2 Associate Professor, Community Medicine, Sri Manakula Vinayagar Medical College and Hospital, Puducherry, India
3 Professor in Community Medicine, Head, Department of Extension Programme, Pramukhswami Medical College, Karamsad, Gujarat, India

Correspondence Address:
Vinayagamoorthy Venugopal
Department of Community Medicine, Sri Manakula Vinayagar Medical College and Hospital, Puducherry - 605 001


Background: In India, there exists public health insurance for government employees and poor people. However, the middle-income households (MIHs) remain neglected. Objectives: The study was conducted to find out the coverage of health insurance and its determinants among MIH. Methods: It was a community-based sequential exploratory mixed methods study. Group interview and key informant interview were conducted among various stakeholders related to insurance coverage. The quantitative survey was done among 400 randomly selected head of households (HoHs) in an urban area of Puducherry. Manual content analysis was done for qualitative data. Generalized linear model with Poisson distribution was used to calculate the adjusted prevalence ratio (aPR) using Stata software. Results: The coverage of health insurance among the MIH was 41% (95% confidence interval [CI]: 36.1–50). The major reasons reported for not having insurance were poor financial status (63.7%) and lack of felt needs (59.4%). The significant determinants for the lack of health insurance among MIH were unskilled occupation (aPR: 1.62, 95% CI: 1.13–2.34) and lower education status (aPR: 1.79, 95% CI: 1.22–2.64) of HoH and less monthly family income (aPR: 2.19, 95% CI: 1.18–4.08). Conclusion: The health insurance coverage of 41% among MIH is better despite the fact that there was no public insurance scheme available for them in Puducherry. The MIH with the identified determinants might be considered in future for including them under publicly sponsored health insurance scheme.

How to cite this article:
Ponnusamy P, Venugopal V, Dongre AR. Health insurance coverage and its determinants among middle-income households in Urban Puducherry: A mixed methods study.Indian J Public Health 2021;65:231-236

How to cite this URL:
Ponnusamy P, Venugopal V, Dongre AR. Health insurance coverage and its determinants among middle-income households in Urban Puducherry: A mixed methods study. Indian J Public Health [serial online] 2021 [cited 2021 Dec 2 ];65:231-236
Available from:

Full Text


The Insurance Regulatory and Development Authority in India defines the term “health insurance” as “A type of insurance that essentially covers the medical expenses and it is a contract between an insurer and an individual/group in which the insurer agrees to provide specified health insurance cover at a particular “premium” subject to terms and conditions specified in the policy” (p. 2).[1] The health insurance usually covers the expenditure during inpatient hospitalization such as medicine, diagnostic investigation, hospital charges, major surgery, and critical care illness and reduces the out-of-pocket payment at the time of hospitalization.[2] Those people working under formal sector/government employees are covered either by government or employer-provided insurance.[3] The below poverty line people are covered under publicly sponsored health insurance schemes such as Ayushman Bharat Pradhan Mantri Jan Arogya Yojana (AB-PMJAY) scheme.[4] Rich sectors of the population are affluent enough to afford for their health expenses irrespective of the availability of public health insurance. However, middle-income group which constitutes nearly 54% of present population and projected expansion to 78% by 2030[5] was not covered under any public health insurance scheme. This group will drive 47% of total Indian consumption and forms the backbone of the Indian economy.[5],[6] Therefore, their health and health insurance is crucial to ensure the productivity of the country. To achieve universal health coverage, which is one of the sustainable development goals, it is mandatory to cover the middle-income group's middle-income households (MIH) under the government health insurance scheme. To the best of our knowledge, there is very limited research done on health insurance and the factors determining its coverage among this group. Hence, the present study was done among the MIH in urban Puducherry with the following objectives: to find out the coverage of health insurance at the household and at the individual level, to identify the determinants of lack of health insurance, and to explore the reasons for the lack of health insurance.

 Materials and Methods

Study setting and design

The present study was undertaken in an urban area of Puducherry, namely, Villianur, a commune panchayat. The people avail health-care services from Urban Primary Health Centre of Puducherry Government, private hospitals, and clinics situated in the town. It was a community-based sequential exploratory type of mixed methods design.[7] The purpose of qualitative component was to generate items for questionnaire development and the quantitative part for doing survey using that questionnaire.

Phase I – Qualitative part (qual)

Two group interviews and four key informant interviews (KII) were conducted among various stakeholders such as head of households (HoHs), health-care providers, and health insurance agents. Those who were vocal and willing to share the information were purposively selected. Theses interviews were carried out by the postgraduate student using interview guide and supervised by the authors of this article. These interviews and discussions generated enough information for developing the questionnaire. Each interview lasted for 20–30 min till the point of attainment of saturation. These interviews were audio recorded and transcribed within a week time to avoid the loss of information. The result of manual content analysis was used to generate items for survey tool.

Phase II – Quantitative part (QUAN)

The study participants were HoHs of middle-income group of family. The data were collected during the 6 months period (September 2018–February 2019).

Sample size and sampling

The sample size was calculated using software OpenEpi (Version 3) taking into consideration, the prevalence of health insurance coverage in the urban Puducherry as 38% based on the National Family Health Survey-4 finding,[8] absolute precision of 5%, design effect of 1, with 95% confidence interval (CI). Considering nonresponse rate as 10%, the sample size was inflated to 394 and finally rounded off to nearest high figure of 400 households. Among the four Anganwadi centers (AWCs) situated in Villianur town, two AWCs were selected by lottery method. The study participants were selected using systematic random sampling.[9] If any of the houses was locked/unavailable, the next house was selected.

Data collection procedure

After obtaining the Institutional Ethics Committee (Human Studies) (IEC approval number: 51/2017 dated November 11, 2017), the study was carried out. Written informed consent was obtained from all the study participants before the initiation of interview at selected households. National Council of Applied Economic Research defines middle-income group based on monthly income.[10] Since, asking questions related to monthly income was a sensitive issue, we used alternate way of assessing the socioeconomic status (SES). We used alternate way of assessing the socioeconomic status (SES). Asset-based scale was developed using principal component analysis (Eigenvalue 6.7 accounting for 33.6% of variance in the data). Finally, reduced 10-item based SES scale was created and scoring system was formulated based on regression coefficient. If the household was found eligible, written informed consent was obtained. Data were collected by face-to-face interview using a predesigned and pretested questionnaire. To ensure the availability of HoH, the survey was carried out at evening hours and during weekends.

Operational definitions

Currently insured family was defined as “any family or individual in the family who is currently enrolled in any health insurance schemes.” Past insured family was termed as “any family where no one in the given family was currently insured but any individual in the family was enrolled in any of the health insurance in the past.” Never insured family was defined as “any family where no one in the given family was insured in their lifetime.” The past and never insured family constituted the uninsured group of households.

Data analysis

The data were entered using EpiData (version software package and analyzed using SPSS 24 software (SPSS Inc., Chicago, Illinois, USA) package. Categorical variables were presented as frequencies and percentages. The prevalence of insurance coverage was reported with 95% CI. Adjusted Prevalence ratio and 95% CI was calculated to identify the determinants of lack of health insurance at household level. Log binomial regression was used for adjusted analysis using Stata software (StataCorp, College Station, TX, USA).

Manual thematic analysis of qualitative data was done. It was done by two independent coders and was then compared. In case of any discrepancy, it was settled after discussion with the third person. The good reporting of a mixed methods study guidelines was used for reporting the study findings.[11]


Qualitative component

Manual thematic analysis of KII (n = 4) and group interview (n = 2; Group-I with six and Group-II with eight members) resulted in two themes, namely, client's perspective and provider's perspective. Each of these themes had three categories. Lack of awareness, monetary aspects, and issues related to availing the health-care services were the categories under the theme of client's perspective. Lack of reliability of insurance agents, constraints in terms and conditions, and difficulty in the process of claiming health insurance were the categories under the other theme [Table 1]. The codes having maximum frequency under these six categories were used to construct survey questionnaire of the next phase of this mixed methods study. Statements in italics indicate direct quotations or verbatim from the respondents. Only salient text codes that help explain what respondents shared are reported below, and the rest are given in [Table 1].{Table 1}

Theme-1: Client's perspective

Under category-1: Lack of awareness, two of the participants mentioned, “I have heard of it but not sure for what disease condition we can use this insurance scheme” and “Many people are not aware the usefulness of health insurance.” Under category-2: Monetary aspects, the statements made by the respondents were “We keep on paying the premium but if not used that money will not be returned back to us” and “As elders and people with chronic disease require insurance, but the amount to be paid for them is higher.” Under category-3: Issues with services, “In our area we have fewer hospitals that are included in insurance” and “Money paid for outpatient department services are not returned back to us” were the statements uttered by the interviewees.

Theme-2: Provider's (insurance agents) perspective

Under the category-1: Lack of reliability of insurance agents, the statements conveyed by the participants was, “These agents never provide clear information. They always confuse us and speak nicely till we take the insurance” and “They pressurize us for taking insurance in order to achieve their targets.” Under the category-2: Constraints in terms and conditions, “The amount to be paid increases as the age increases, which is really worrying” and “Dependents are excluded once they achieve 18 years of age” were the statements made by them. Under category-3: Disappointing process, the sentences given by the participants were, “We face so much difficulty in obtaining no objection certificate from the treating doctor” and “They ask for more documents when we go for reimbursement and that really irritates us.

Quantitative component

The prevalence of health insurance coverage among the middle-income group of family was 41% (95% CI: 36.1–50) at household level (N = 400) and 38.2% (95% CI: 35.7–40.7) at individual level (N = 1517). Among the 236 uninsured households, 212 (53%; 95% CI: 47.9–57.9) were never insured and the rest 24 (6%; 95% CI: 4–9) were past insured. Among the currently insured 164 households, majority (92; 56.1%) had owned private insurance and the remaining 72 (43.9%) had government insurance. About 58 (35.4%) households had owned health insurance to meet their medical expenses, 104 (63.4%) owned because of compulsion of employers in their working place, and 2 (12%) for purpose of deduction in tax amount [Table 2].{Table 2}

The major reasons for not having health insurance among the 212 never users were poor financial status (63.7%), lack of felt need to have health insurance (59.4%), difficulty in application process (54.7%), lack of awareness (25.5%), high premium amount (21.7%), perceived delay in reimbursement (16%), and provision of low-quality medical service for insured (11.3%). The reported reasons for discontinuing the health insurance among 24 past users were high premium amount and poor financial status (50% each), not availed any service (37.5%), low-quality service provided (33.3%), and delay in reimbursement (20.8%) [Table 3].{Table 3}

The major determinants found in unadjusted analysis were HoH having jobs related to unskilled and semiskilled occupation, had education up to primary and secondary level, and income <40,000 Indian Rupee (INR). The sociodemographic details of HoH and the results of unadjusted analysis are shown in [Table 3]. Multivariate analysis showed that the HoH belonging to unskilled (prevalence ratio [PR]: 1.62; 95% CI: 1.13–2.34) and semiskilled occupation (PR: 1.82; 95% CI: 1.29–2.58) had significantly higher prevalence of not having health insurance coverage compared to retired. HoH with <10 years of schooling had 1.79 times (PR: 1.79; 95% CI: 1.22–2.64) higher risk for not having health insurance. Those households with family monthly income <20,000 INR had 2.19 times (PR: 2.19; 95% CI: 1.18–4.08) more likely to remain uninsured as compared with those having family income more than 40,000 INR [Table 4].{Table 4}


The prevalence of health insurance coverage among MIH was 41% in the present study. Around two-third of households stated that the poor financial status and high premium rate as the major reasons for not having health insurance. The significant determinants of not having health insurance were <16 years of schooling of HoH, semi and unskilled occupation of HoH, and <20,000 INR monthly family income of MIH.

The coverage of health insurance in the current study (41%) was higher than the prevalence at national level for urban area (28%).[12] However, it was lower than insurance coverage at Tamil Nadu (59.2%), the neighboring state.[13] There is a state-owned health insurance scheme called Chief Minister's Comprehensive Health Insurance Scheme in operation at Tamil Nadu and that could be the reason for increase in coverage. The other reason for the disparity could be the inclusion of urban MIH only in the present study. In Puducherry, there was no such scheme available; however, the present coverage was encouragingly better.

The common reasons reported for not having health insurance in our study were poor financial status, absence of perception of health threat to them, and higher premium amount. Similar reasons were reported in a previous study done at Lucknow.[14] Such perception of denouncing health insurance for various reasons in MIH would lead to risk of out-of-pocket health expenditure. A previous study showed that MIHs were unable to save the income to afford for their annual health insurance premium, as they need to spend most of income on family and household matters.[15] To make it affordable, certain modifications in health insurance policy such as annual family income-based premium and choice of payment either on a monthly or biannual basis have been suggested.[15],[16] India's recent flagship program AB-PMJAY has announced cashless benefits through publicly funded health insurance scheme to cover the below poverty people.[4] Unfortunately, the MIH is deprived of this scheme due to policy constraints. Hence, in future, the policy implications and design of publicly sponsored health insurance scheme need to make certain changes for the inclusion of them, to protect them from being pushed into poverty due to catastrophic health-care expenditure.

The major determinants of lack of health insurance identified in our study were the unskilled and semiskilled occupation of HoH, lower level of education, and less monthly income. These findings were similar to the previous study done by Desai et al. and Bawa and Ruchita.[17],[18] Although the level of education determines the knowledge about health insurance schemes that can be improved through frequent mass media campaign, the attitudinal change that decides the felt need to have insurance is complex to modify.[19] The other determinants, namely, income and occupation, could not be modified much. Inclusion of all MIHs in the publicly funded insurance at national level might not be affordable for a developing nation like India with huge and diverse population. Hence, policy makers and designers of public health insurance schemes need to consider certain changes in the enrollment criteria such as including the middle-income group of people with semi and unskilled occupation and those with lesser monthly income. Besides this, government can make certain modification by making MIH to pay a part of premium based on their capacity to pay to avail the publicly sponsored health insurance scheme.[20]

To the best of our knowledge, there was no community-based prevalence study done in the middle-income group of people at national level and this is one among the rare kind. For the identification of middle-income group of people, we used objectively verifiable asset-based SES assessment scale that was prepared to the context of study participants that ensures validity. Information on some of the study variables was self-reported; however, we ensured privacy and confidentiality to minimize the possibility of reporting bias.


Middle-income group of family had better health insurance coverage; nevertheless, there was no publicly funded schemes operating in Puducherry. The high premium amount, lack of perceived threat, and poor financial status were the common reason for not having health insurance. MIH involved in unskilled and semiskilled occupation, lower level of education, and less family income to be given special consideration for publicly sponsored health insurance schemes as these were the major determinants identified.


We sincerely acknowledge the Institutional Ethics Committee and Research Committee for providing the permission and support to carry out the study.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Handbook on Health Insurance; 2011. Available from: [Last accessed on 2017 Aug 25].
2Devadasan N, Seshadri T, Trivedi M, Criel B. Promoting universal financial protection: Evidence from the Rashtriya Swasthya Bima Yojana (RSBY) in Gujarat, India. Health Res Policy Syst 2013;11:29.
3Ahlin T, Nichter M, Pillai G. Health insurance in India: What do we know and why is ethnographic research needed. Anthropol Med 2016;23:102-24.
4Angell BJ, Prinja S, Gupt A, Jha V, Jan S. The Ayushman Bharat Pradhan Mantri Jan Arogya Yojana and the path to universal health coverage in India: Overcoming the challenges of stewardship and governance. PLoS Med 2019;16:e1002759.
5Future of Consumption in Fast-Growth Consumer Markets: India. World Economic Forum; 2019. Available from: [Last accessed on 2020 May 10].
6Ahalya R, Paul SB. Identification and Characterization of Middle Class in India and its Comparison with Other Economic Classes; 2017. Available from: [Last accessed on 2019 Aug 24].
7Creswell JW, Clark VL. Designing and Conducting Mixed Methods in Research. 3rd ed. Los Angeles: SAGE Publication Ltd; 2009.
8Ministry of Health and Family Welfare. National Family Health Survey-4. Union Territory Fact Sheet Puducherry; 2015-2016. Available from: [Last accessed on 2017 Oct 12].
9Abramson JH, Abramson ZH. Survey Methods in Community Medicine: Epidemiological Research, Programme Evaluation, Clinical Trials. 5th ed. Philadelphia: Churchill Livingstone; 2004. p. 89-104.
10Center for Global Development. New Estimates of India's Middle Class; 2012. Available from: [Last accessed 2017 Aug 24].
11Brown KM, Elliott SJ, Leatherdale ST, Robertson-Wilson J. Searching for rigour in the reporting of mixed methods population health research: A methodological review. Health Educ Res 2015;30:811-39.
12Ministry of Health and Family Welfare. National Family Health Survey – 4 (NFHS-4) - India Fact Sheet. Mumbai: International Institute for Population Sciences; 2015-16. Available from: [Last accessed on 2019 Sep 22].
13Ministry of Health and Family Welfare. National Family Health Survey – 4 (NFHS-4) - Tamil Nadu Fact Sheet. Mumbai: International Institute for Population Sciences; 2015-16. Available from: [Last accessed on 2019 Sep 22].
14Mathur T, Paul UK, Prasad HN, Das SC. Understanding perception and factors influencing private voluntary health insurance policy subscription in the Lucknow region. Int J Health Policy Manag 2015;4:75-83.
15Madhukumar S, Sudeepa D, Gaikwad V. Awareness and perception regarding health insurance in Bangalore rural population. Int J Med Public Health 2012;2:18-22.
16Hadaye RS, Thampi JG. Catastrophic health-care expenditure and willingness to Pay for health insurance in a metropolitan city: A cross-sectional study. Indian J Community Med 2018;43:307-11.
17Desai S, Sinha T, Mahal A, Cousens S. Understanding CBHI hospitalisation patterns: A comparison of insured and uninsured women in Gujarat, India. BMC Health Serv Res 2014;14:320.
18Bawa SK, Ruchita V. Awareness and Willingness to pay for health insurance: An empirical study with reference to Punjab India. Int J Humanit Soc Sci 2011;1:9-17.
19Bredenkamp C, Evans T, Lagrada L, Langenbrunner J, Nachuk S, Palu T. Emerging challenges in implementing universal health coverage in Asia. Soc Sci Med 2015;145:243-8.
20Wallace LS. A view of health care around the world. Ann Fam Med 2013;11:84.