Chronic patients as retirement-aged workers: the impact of employment-based health insurance and chronic conditions on health-related working capacity and late-life career participation

Abstract

Retirement-aged workers with chronic conditions are increasingly engaged in late-life careers in the policy context of delayed retirement initiative. However, it remains uncertain as to how chronic conditions and employment-based social health insurance interact to affect health-related working capacity and late career participation in this group of people. Using data from the China Health and Retirement Longitudinal Study (CHARLS) and the discrete choice model, this study finds that chronic conditions are negatively associated with health-related working capacity (– 0.400, p < 0.01) and late-life career participation (– 0.170, p < 0.01). Employment-based health insurance is positively associated with health-related working capacity of retirement-aged workers (0.432, p < 0.01), but is negatively associated with their late-life career participation (– 1.027, p < 0.01). Moreover, employment-based health insurance could weaken the negative associations between chronic conditions and health-related working capacity (interaction = 0.285, p < 0.05) and late-life career participation (interaction = 0.251, p < 0.05). More fine-grained policies for delayed retirement are needed to focus on the long-neglected health of retirement-aged workers with chronic conditions.

Introduction

Policymakers encourage retirement-aged workers to participate in late-life careers to relieve the pressure of population aging and address the shrinking workforce, declining dependency ratio, and the sustainability of pension systems (Le and Lhuissier 2017; Litwin et al. 2009). Population aging brings a decline in the working-age cohort, leading to the labor deficit (Cai and Wang 2006). Existing policies address labor shortages by raising the retirement age and encouraging retirement-aged workers to participate in the labor market (Zeng 2011). Many countries have introduced phased or progressive retirement system to encourage retirement-aged workers to continue working (Le and Lhuissier 2017). For example, Britain has adopted a progressive retirement scheme and planned to delay the retirement age up to 66 by 2020 and 69 by the end of 2040 (Department for Work and Pensions of the UK, 2014). The USA has raised the legitimate retirement age from 65 to 67, when full pensions will be paid (The United States Social Security Administration, 1983). South Korea’s parliament passed the Law to Extend the Retirement Age to extend the statutory retirement age to 60 years by 2013 and further to 65 by 2033. Japanese government, in 2013, passed the Elderly Employment Security Law to delay the statutory retirement age to 65. France whose retirement systems are considered the most generous raised the legal retirement age from 60 to 62 (Le and Lhuissier 2017). Similar challenges happen to China where there are 249 (and 167) million people aged > 60 (and > 65), accounting for 17.9% (and 11.9%) of the total population (National Bureau of Statistics of China 2020). Delayed retirement initiative is being discussed in China to cope with the challenges of population aging (Wang et al. 2019).

However, the disparity in health conditions among retirement-aged workers may be a key determinant of their working capacity and willingness to participate in late-life careers (e.g., Martin et al. 2017). Poor health can reduce individual willingness to work (Kessler et al. 2001). Retirement-aged workers with health problems may intend to take early retirement (Litwin et al. 2009). Of older people, over 23% (at least 33 million) are sufferers of chronic diseases, and this number is rising rapidly with the aging of the population (WHO 2019). The suffering of chronic conditions may become a non-negligible determinant of participation in late-life careers. As China enters a rapidly aging society, the health insurance system aims to address the growing needs for health care and inequality in healthcare access (Gong et al. 2016). There is still a gap in the level of health services between different groups of people, and the rapid health insurance reform in China has not yet achieved universal coverage (Zhang et al. 2017). Older adults who are beneficiaries of health insurance are more likely to have greater affordability of healthcare expenditure. Health insurance may thus play an important role in the working capacity and late-life career participation.

Prior research on delayed retirement has mainly focused on the impact of delayed retirement on fiscal policy and government pension pressure (e.g., Wang et al. 2019), and the impact on welfare receipt and expenditure (e.g., Oguzoglu et al. 2020). The impacts of health conditions, health insurance, and their interaction on career participation in late life have not been adequately discussed. Relevant research has found that the tendency to support late retirement is associated with health status, family characteristics, economic status, work status and ideology (e.g., Bidewell et al. 2006). Reduced work stress and the improvement of occupational environment and satisfaction can also facilitate the participation in late-life careers (e.g., Carr 2016).

It is of practical significance to investigate the interaction of chronic conditions and health insurance on the participation in late-life careers. The findings can help improve the applicability of late retirement to retirement-aged workers with chronic conditions and improve the policy design of delayed retirement initiative and social health insurance schemes. Research questions of this study are as follows.

RQ 1: How does the employment-based health insurance in China affect late-life career participation and health-related working capacity of retirement-aged workers?

RQ 2: How does the employment-based health insurance in China interact with chronic conditions to affect late-life career participation/health-related working capacity of retirement-aged workers?

Literature review

The relationships between chronic conditions and late-life career participation, health-related working capacity of retirement-aged workers

Faced with challenges of population aging, policymakers encourage older adults to participate in late-life careers. However, personal preferred retirement age could be subject to health conditions (Bidewell et al. 2006). The factors influencing retirement decision (working until retirement or early retirement) are shown to be discrepant between older workers with and without chronic conditions (Sewdas et al. 2018). Personal retirement plan is shown to be associated with changes in health status, and the deteriorating health conditions could discourage people from continuing to work in later life (Lee 2005). It is found that poor health may increase the disutility of one’s current job, reduce productivity and work returns (e.g., wages), make the non-wage incentives (e.g., disability benefits) more attractive, and inhibit the motivation to accumulate wealth for later life with a shorter life expectancy (Disney et al. 2006; French and Jones 2017). Older workers with poor health conditions are shown to be more likely to exit from the labor market before the retirement age (Thorsen et al. 2016; Van den Berg et al. 2010; Von Bonsdorff et al. 2010; Yuan et al. 2022b) and less likely to return to work after retirement even though they ever have a plan to postpone retirement (Flynn 2010). People who continue to work over age 65 have better self-reported health than those who retire at the age of 65 (Anxo et al. 2019). Older workers in good health may retreat from the labor market at an older age than their counterparts (Bidewell et al. 2006; Griffin and Hesketh 2008; Stenholm et al. 2014).

The incidence of chronic conditions is increasing among older adults at all levels of socioeconomic status (Pefoyo et al. 2015). Existing retirement policies aiming to encourage late-life career participation often lack the fine-grained design, without considering the specific needs of people with chronic conditions (Sewdas et al. 2018). Chronic conditions (e.g., cardiovascular diseases, hypertension, diabetes, hypercholesterolemia, asthma, chronic obstructive pulmonary diseases, and cancers; Wilper et al. 2008) can be accompanied by long-lasting symptoms which require ongoing medical care and limit daily activities, such as walking, sitting, or lifting heavy things (Warshaw 2006; French and Jones 2017). Treatment for and daily care of chronic conditions require a high level of cooperation between patients and medical staff. In addition to regular medical visits, health checks, and disease tracking, the self-management is also important for chronic disease patients, such as daily medication, physical exercise, and healthy diet (Cramm and Nieboer 2012). Chronic patients need to work fewer hours than healthy colleagues to adapt to their illness (Miah and Wilcox-Gök 2007) and are more likely to exit the labor market earlier than their counterparts (Fleischmann et al. 2018; Sewdas et al. 2018; Yuan et al. 2022a). Older adults with chronic conditions, like functional limitations and circulatory diseases, have been reported to plan to retire one to two years earlier on average (Dwyer and Mitchell, 1999). Chronic conditions lead to decreased productivity and low quality of life, which are shown to be associated with early withdrawal from the labor market (Ranzi et al. 2013). A meta-analysis shows that chronic condition is a risk factor for older adults to withdraw from paid employment (e.g., in a way of disability pension, unemployment, and early retirement; Rijn et al. 2014). Based on the findings above, older workers with chronic conditions are more likely to suffer from impaired health-related working capacity and are less willing to extend their working life.

The relationships between employment-based social health insurance and late-life career participation, health-related working capacity of retirement-aged workers

The establishment of social health insurance system in China aims to reduce health service spending and the risk of catastrophic health expenditure through a cost-sharing mechanism (Zhou et al. 2017). China’s current social health insurance system includes the Urban Employee Basic Medical Insurance (UEBMI) designed for urban workers, the Urban Resident Basic Medical Insurance for urban residents, and the New Rural Cooperative Medical System for rural residents. The UEBMI is the employment-based social health insurance in China. It is mandatory and funded by contributions from both the employers and employees. Its funding level is significantly higher than that of other health insurance schemes in China (Peng and Ling 2019). Beneficiaries of UEBMI can enjoy higher reimbursement rates of medical expenses and a broader drug reimbursement list than those of other schemes (Fan et al. 2019; Meng et al. 2015; Su et al. 2017). The grade of medical facilities where they are entitled to higher reimbursement levels is also much higher (Peng and Ling 2019). After reaching the minimum payment period (20 years for females, and 25 years for males), UEBMI beneficiaries will be entitled to lifetime health security (outpatient and inpatient services) with no further payment after retirement, whereas other health insurance schemes do not set a minimum payment period and require annual payments (Zhu et al. 2017).

Despite the higher coverage of the UEBMI, inequities in coverage, accessibility, and affordability of medical services pose a major challenge to this health insurance scheme (Zhao et al. 2018). The UEBMI relies heavily on the stable employment relationship and has weak restriction on informal employment, making it difficult to fully cover informal employees (Peng and Ling 2019) who account for a significant part of workforce in China (van Ginneken, 1999). Some enterprises have shirked their responsibilities to pay for health insurance for informal employees. As such, there are still a group of informal employees uncovered by the current employment-based health insurance system (UEBMI) in China. Such employees may turn to participation in other social health insurance schemes that are voluntary and designed for urban and rural residents. In some cases, a part of them may not even be insured by any of health insurance schemes. A typical example is a group of migrant workers without the stable employment. They are often faced with the separation of insured places (hometown) and actual places of medical treatments (place of work). They are required to pay in full for medical treatments incurred in work destinations and to get reimbursed for these expenses on their return to hometowns (Meng and Xu 2014), leading them to suffer from complicated reimbursement procedures and low reimbursement rates (Peng and Ling 2019). Besides, poor health resources of their hometowns and high costs of round-trip transportation could also be a major obstacle to returning to hometowns to seek medical treatments (Wang et al. 2018). As a result, they may drop the health insurance scheme they had before.

Given that UEBMI beneficiaries enjoy the higher reimbursement rate of medical expenses and the broader reimbursement list of drugs (Su et al. 2017), older adults with the UEBMI are reported to have higher utilization of healthcare services and therefore better health conditions, whereas those with other types of health insurance schemes have poor access to health services in China (Gong et al. 2016; Li and Yuan 2022). It is reported that older adults without social health insurance have fewer physical examinations or doctor visits which are found to be the prominent demands of older adults (Gong et al. 2016), and those who do not enjoy continuous beneficiary status are more likely than their counterparts to experience severe declines in physical and mental functioning (Baker et al. 2001). As such, the beneficiary status of the UEBMI could allow retirement-aged workers to have better access to health services, thereby mitigating their decline in health-related capacity.

The expectation of future out-of-pocket medical expenses is found to significantly influence retirement decisions (Johnson et al. 2003). The higher coverage of the UEBMI can help beneficiaries reduce the economic burden of high out-of-pocket expenditures to a greater extent (Li and Yuan 2019; Meng et al. 2015), whereas older people without the UEBMI have to pay for medical treatments out of their own pockets, and even might be trapped in poverty as a result (Li and Yuan 2019). Older people without the UEBMI would thus rely more on late-life career participation for financial income to pay for medical services. A prior study shows that retirement-aged workers who are eligible for the benefits of social health insurance after retirement are 44–68% more likely to consider retirement than their counterparts (Rogowski and Karoly 2000). A macro-level study shows that the enrollment rate of social health insurance can increase the retirement rate by 3–5% (Kapur and Rogowski 2011). The employer-sponsored health insurance is found to increase the average probability of retirement by 8% (Strumpf 2010). Such effect of employer-sponsored social security is greater for older workers aged 62 and 63, resulting in a 5.9% and 6.9% increase in retirement rates, respectively (Nyce et al. 2013). Similarly, the impact of employer-provided retiree health insurance on the annual rate of labor force exit increases with age (from 51 to 61), reaching 7.5% by age 61 (Blau et al. 2001). Besides, the enrolment of social health insurance can increase the retirement intention of public-school employees aged 60–64 by 5–10% (Fitzpatrick 2014) and of government employees aged 60–64 by 5.1% (Shoven and Slavov 2014), while those in the private sector are 26–38% more likely to retire than their counterparts (Nyce et al. 2013). It is thus conceivable that retirement-aged workers in the beneficiary status of UEBMI may have a higher likelihood of terminating their career participation in later life without serious concerns about the shortfall of out-of-pocket expenditure. Hence, we propose.

Hypothesis 1

The employment-based social health insurance (UEBMI) is negatively associated with the late-life career participation.

Hypothesis 2

The employment-based social health insurance (UEBMI) is positively associated with the health-related working capacity of retirement-aged workers.

The interaction of employment-based social health insurance (UEBMI) and chronic conditions on late-life career participation, health-related working capacity of retirement-aged workers

People living with chronic conditions struggle to balance the conflict between patient roles and work roles (Cheshire et al. 2021) and are unable to simultaneously cope with pressures of persistent ill-health and working demands. Health insurance may alleviate this conflict suffered by chronic patients, as it can meet the needs for health care and improve health status (Liu and Zhao 2006; Finkelstein et al. 2012), It is important for older workers to enjoy the benefits of health insurance which protect them from the risk of deteriorating health and associated financial burden (Yuan et al. 2022c). Health insurance can lessen the burden of medical expense on older workers with chronic conditions and prompt them to seek medical treatments in time and thus reduce the incidence of deterioration (Chiao et al. 2014). In this way, the beneficiary status of health insurance can alleviate the conflict between patient roles and work roles faced by chronically ill retirement-aged workers. The obstacle posed by chronic conditions to the willingness of older adults to participate in late-life careers can thus be overcome to some extent. In contrast, uninsured older workers with chronic conditions could not obtain the supportive physical and financial conditions from health insurance (Li and Yuan 2019). The concerns about health deterioration and incapability of work roles haunt themselves (Carr 2016), which may lead them to give up fulfilling work roles to end the competitive expectations between persistent medical treatments and job responsibility. As such, health insurance may buffer the negative effect of chronic conditions on the late-life career participation and health-related working capacity of older workers.

Accordingly, it can be conceivable that for retirement-aged workers in the beneficiary status of employment-based social health insurance (UEBMI), the troubles with chronic conditions that impair their competence to continue working in late-life careers could be alleviated, and they are thus more inclined to participate in late-life careers. However, for those without the UEBMI, such troubles are more likely to persist and could reduce the likelihood of extending working life and the ability to work in their later life accordingly. Thus, we proposed.

Hypothesis 3

There is an interaction of employment-based social health insurance (UEBMI) and chronic conditions on late-life career participation among retirement-aged workers, such that the negative relationship between chronic conditions and late-life career participation becomes weaker when they are in the beneficiary status of UEBMI.

Hypothesis 4

There is an interaction of employment-based social health insurance (UEBMI) and chronic conditions on health-related working capacity of retirement-aged workers, such that the negative relationship between chronic conditions and health-related working capacity becomes weaker when they are in the beneficiary status of UEBMI.

Materials and methods

Data description

This study used the data from the China Health and Retirement Longitudinal Study (CHARLS-2018) which tracked health status of older people and covered 150 counties (450 communities/villages) with stratified random sampling. Based on secondary data, no additional ethical approval was required for this study. The sample of this study includes older people who have reached the statutory retirement age in China (i.e., > 60 years for males and > 55 years for females).

Measures

Dependent variables

Late-life career participation. It was coded = 1 if respondents who were aged > 60 years for males or > 55 years for females reported that they were still in the status of working when surveyed. Otherwise, it was coded = 0.

Health-related working capacity. Respondents were asked about their health status classified into 3 categories including (1) “I can’t work at all, because of disability or relevant health problems”; (2) “I can’t work for long time, because of disability or relevant health problems”; (3) “I have no problem at work due to disability or relevant health problems.” In this study, this variable was coded as the 0–1 binary indicator (= 1, if respondents reported that “I have no problem at work due to disability or relevant health problems”; = 0, otherwise).

Independent variables

Employment-based social health insurance. It was coded = 1 if respondents were the beneficiaries of urban employee basic medical insurance (UEBMI) when surveyed. Otherwise, it was coded = 0.

Chronic conditions. It was coded = 1 if respondents had at least one of the following chronic conditions (including hypertension; dyslipidemia; diabetes or high blood sugar; cancer or malignant tumor; chronic lung diseases; liver disease; heart attack, coronary heart disease, angina, congestive heart failure, or other heart problem; stroke; kidney disease; stomach or other digestive disease; emotional, nervous, or psychiatric problems; memory-related disease; arthritis or rheumatism; asthma). Otherwise, it was coded = 0.

Covariates

A series of covariates were controlled in this study, including gender (1 = male; 2 = female), age, education (1 = less than lower secondary; 2 = upper secondary & vocational training; 3 = tertiary; 4 = graduate and above), marital status (coded as the 0-1 indicator for each of types listed), healthcare service (coded as the 0-1 indicator), doctor visit/outpatient (coded as the 0-1 indicator), lifestyle and health behavior (coded as the actual number of activities listed), out-of-pocket expenditure and economic burden (taken as the form of natural log).

More details about variables are shown in Table 1.

Table 1.

Description of variables

Variables	Description	Value	Freq	%
Dependent variable
Late-life career participation	= 1, if respondents (those aged > 60 years for males or > 55 years for females) reported that they were still in the status of working when surveyed. = 0, otherwise	= 1	3112	51.20
		= 0	2965	48.80
Health-related working capacity	In year 2018, which is the best fit of your health status, (1) “I can’t work at all, because of disability or relevant health problems”; (2) “I can’t work for long time, because of disability or relevant health problems”; (3) “I have no problem at work due to disability or relevant health problems.” It was coded as the 0-1 binary indicator (= 1, if respondents reported that “I have no problem at work due to disability or relevant health problems”; = 0, otherwise)	= 0	3084	50.76
		= 1	2993	49.24
Independent variable
Chronic conditions	= 1 if he/she has at least one of the following chronic conditions: hypertension; dyslipidemia; diabetes or high blood sugar; cancer or malignant tumor; chronic lung diseases; liver disease; heart attack, coronary heart disease, angina, congestive heart failure, or other heart problem; stroke; kidney disease; stomach or other digestive disease; emotional, nervous, or psychiatric problems; memory-related disease; arthritis or rheumatism; asthma. = 0 otherwise	= 1	2879	47.37
		= 0	3198	52.63
Employment-based social health insurance	= 1 if he/she has Urban Employee Basic Medical Insurance (UEBMI). = 0 otherwise	= 1	733	12.05
		= 0	5344	87.95
Covariates
Gender	= 1 male, = 2 female	= 1	2521	41.49
		= 2	3556	58.51
Age	The age of respondents	≤ 65	2432	40.01
		66-75	2574	42.36
		> 75	1071	17.63
Healthcare service	Respondents were asked about whether he/she has received healthcare services in the last two years (i.e., “physical examination, routine blood test, routine urine test, liver function test, kidney function test, lipids profile test, blood glucose test, surgical, internal medicine, electrocardiogram, B-type ultrasonic, chest fluoroscopy, ophthalmology and otorhinolaryngology, andrology or gynecology”). = 1, if yes; = 0, if no	= 0	2715	44.68
		= 1	3362	55.32
Education	= 1, less than lower secondary. = 2, upper secondary & vocational training. = 3, tertiary, = 4 graduate and above	= 1	4585	75.44
		= 2	1430	23.54
		= 3	61	1.00
		= 4	1	0.02
Doctor visit/outpatient	Respondents were asked about their doctor visit/outpatient last month. = 1, if yes; = 0, if no	= 0	4919	80.96
		= 1	1158	19.04
Out-of-pocket expenditure	In regressions, ln (1 + out-of-pocket expenditure) is used	< 100	2446	40.26
		100-200	1777	29.22
		201-500	1208	19.87
		501-1000	416	6.85
		> 1000	230	3.80
Economic burden	In regressions, ln (outstanding debt) is used	< 1000	5948	97.88
		1000-10,000	21	0.34
		> 10,000	108	1.78
Lifestyle and health behavior	The number of following activities taken in last month (1. Interacted with friends, 2. Played Ma-jong, played chess, played cards, or went to community club, 3. Provided help to family, friends, or neighbors who do not live with you, 4. Went to a sport, social, or other kind of club, 5. Took part in a community-related organization, 6. Done voluntary or charity work, 7. Cared for a sick or disabled adult who does not live with you, 8. Attended an educational or training course, 9. Stock investment, 10. Used the Internet, 11. Others)	= 1	4991	82.13
		= 2	710	11.69
		= 3	247	4.07
		= 4	85	1.40
		≥ 5	44	0.72
Marital status	Coded as 0-1 indicator for each of types, including married, partnered, separated, divorced, widowed, never married	Married	4844	79.72
		Partnered	264	4.34
		Separated	9	0.15
		Divorced	31	0.51
		Widowed	890	14.64
		Never married	39	0.64

The descriptive statistics are based on the sample where all independent variables and covariates have no missing values. Data of dependent and independent variables come from the data of 2018 wave

Analytic strategy

This study created a binary indicator from the 3 categories of the variable of health-related working capacity. Given that the dependent variables (late-life career participation and health-related working capacity) were binary indicators, the discrete choice model was applied to examine the relationships of variables. The logit model with n regressors (x₁, x₂,---, x_n) performs better than the probit model with n regressors in the case of larger sample size, because when the sample size increases, the probability of observes in tail increases too (Alsoruji et al. 2018). Thus, the logit model is used for the second regression (health-related working capacity) with the larger valid sample size. The formula of probit model with n regressors is provided as below.

$$\begin{array}{c}Probability\left[late-lifecareerparticipation=1\right]=\mathrm{\Phi }({\beta }_0+{\beta }_{1}employment-basedsocialhealthinsurance+{\beta }_{2}chronicconditions+{\beta }_{3}chronicconditions\times employment-basedsocialhealthinsurance+{\beta }_4\text{cov}ariates)\\ \end{array}$$

where Φ(•) indicates the cumulative distributive function of standard normal distribution.

The formula of logit model with n regressors is provided as below.

$$\begin{array}{c}Probability\left[health-relatedworkingcapacity=1\right]\\ =\frac{1}{1+ex{p}^{-\left(\begin{array}{c}{\beta }_0+{\beta }_{1}employment-basedsocialhealthinsurance+{\beta }_{2}chronicconditions\\ +{\beta }_{3}chronicconditions\times employment-basedsocialhealthinsurance+{\beta }_{4}covariates\end{array}\right)}}\\ \end{array}$$

Results

Results of Table 2 show that the employment-based social health insurance (UEBMI) in China is negatively associated with late-life career participation (– 1.027, p < 0.01). The positive interaction between UEBMI and chronic condition (0.251, p < 0.05) on the late-life career participation indicates that the beneficiary status of employment-based social health insurance can alleviate the negative relationship (– 0.170, p < 0.01) between chronic condition and late-life career participation. Hence, hypothesis 1 and 3 are supported. These results demonstrated that retirement-aged workers who are insured by the China’s employment-based social health insurance may be less willing to participate in late-life careers. In addition, the beneficiary status of employment-based social health insurance may help mitigate the decline in their willingness to participate in late-life careers due to chronic conditions.

Table 2.

The influence of chronic condition and employment-based social health insurance on late-life career participation

	Dependent variable: late-life career participation
	Testing the direct effect (Hypothesis 1)			Testing the interaction effect (Hypothesis 3)
	Coef	S.E	95% CI	Coef	S.E	95% CI
Direct effect
Chronic condition	– 0.170 **	0.035	[– 0.240, – 0.101]	– 0.193 **	0.037	[– 0.266, – 0.121]
Employment-based social health insurance	– 1.027 **	0.064	[– 1.153, – 0.901]	– 1.158 **	0.088	[– 1.331, – 0.985]
Interaction effect
Chronic condition × Employment-based social health insurance				0.251 *	0.119	[0.017, 0.485]
Covariates
Age	– 0.063 **	0.003	[– 0.068, – 0.057]	– 0.063 **	0.003	[– 0.068, – 0.057]
Gender
Male	REF			REF
Female	– 0.488 **	0.038	[– 0.562,– 0.414]	– 0.487 **	0.038	[– 0.561,– 0.413]
Education	– 0.241 **	0.044	[– 0.326,– 0.156]	– 0.240 **	0.044	[– 0.325,– 0.154]
Doctor visit/outpatient	– 0.021	0.044	[– 0.108, 0.066]	-0.021	0.044	[– 0.108, 0.066]
Health examination	0.084 *	0.036	[0.014, 0.154]	0.084 *	0.036	[0.014, 0.154]
Out-of-pocket expenditure	– 0.139 **	0.013	[– 0.165,– 0.113]	– 0.140 **	0.013	[– 0.166,– 0.114]
Economic burden	0.035 **	0.012	[0.011, 0.058]	0.035 **	0.012	[0.012, 0.059]
Lifestyle and health behavior	0.007	0.026	[– 0.044, 0.057]	0.008	0.026	[– 0.043, 0.059]
Marriage
Married	REF			REF
Partnered	0.072	0.087	[– 0.098, 0.242]	0.069	0.086	[– 0.101, 0.238]
Separated	– 0.377	0.420	[– 1.200, 0.446]	– 0.376	0.418	[– 1.194, 0.443]
Divorced	– 0.708 **	0.270	[– 1.237,– 0.180]	– 0.715 **	0.268	[– 1.240,– 0.189]
Widowed	– 0.214 **	0.053	[– 0.318,– 0.111]	– 0.215 **	0.053	[– 0.318,– 0.111]
Never married	– 0.764 **	0.212	[– 1.178,– 0.349]	– 0.764 **	0.212	[– 1.179,– 0.349]
Intercept	5.735 **	0.229	[5.286, 6.185]	5.756 **	0.229	[5.307, 6.206]
Num. of obs	6077			6077
F-statistics	1068.81			1094.73
P-values	[0.000]			[0.000]

The probit model is applied. Robust standard errors are reported. ^*p < 0.05, ^**p < 0.01

Besides, results of Table 3 show that the employment-based social health insurance (UEBMI) in China is positively associated with health-related working capacity (0.432, p < 0.01). The positive interaction between UEBMI and chronic condition (0.285, p < 0.05) on health-related working capacity indicates that the beneficiary status of employment-based social health insurance can alleviate the negative relationship (– 0.400, p < 0.01) between chronic condition and health-related working capacity. Thus, hypothesis 2 and 4 are supported. These results demonstrated that retirement-aged workers who are insured by the China’s employment-based social health insurance may be physically more capable of performing job functions in later life. Moreover, the beneficiary status of employment-based social health insurance may help alleviate the impairment of working capacity due to chronic conditions.

Table 3.

The influence of chronic condition and employment-based social health insurance on health-related working capacity

	Dependent variable: health-related working capacity [0-1 indicator]
	Testing the direct effect (Hypothesis 2)			Testing the interaction effect (Hypothesis 4)
	Coef	S.E	95% CI	Coef	S.E	95% CI
Direct effect
Chronic condition	– 0.400 **	0.046	[– 0.490,– 0.311]	– 0.429 **	0.048	[– 0.523,– 0.334]
Employment-based social health insurance	0.432 **	0.079	[0.276, 0.587]	0.274 *	0.109	[0.060, 0.488]
Interaction effect
Chronic condition × Employment-based social health insurance				0.285 *	0.145	[0.001, 0.570]
Covariates
Age	– 0.054 **	0.003	[– 0.059,– 0.049]	– 0.054 **	0.003	[– 0.059,– 0.049]
Gender
Male	REF			REF
Female	– 0.226 **	0.047	[– 0.318,– 0.134]	– 0.226 **	0.047	[– 0.318,– 0.134]
Education	0.310 **	0.051	[0.210, 0.410]	0.311 **	0.051	[0.211, 0.411]
Doctor visit/outpatient	– 0.374 **	0.057	[– 0.485,– 0.264]	– 0.374 **	0.057	[– 0.485,– 0.263]
Health examination	0.128 **	0.047	[0.036, 0.220]	0.128 **	0.047	[0.036, 0.220]
Out-of-pocket expenditure	– 0.284 **	0.018	[– 0.319,– 0.249]	– 0.285 **	0.018	[– 0.320,– 0.250]
Economic burden	0.020 *	0.010	[0.001, 0.039]	0.020 *	0.010	[0.001, 0.039]
Lifestyle and health behavior	0.226 **	0.033	[0.161, 0.290]	0.227 **	0.033	[0.163, 0.291]
Marriage
Married	REF			REF
Partnered	– 0.001	0.093	[– 0.183, 0.181]	– 0.003	0.093	[– 0.185, 0.180]
Separated	– 0.735	0.589	[– 1.890, 0.420]	– 0.734	0.589	[– 1.888, 0.420]
Divorced	– 0.357	0.260	[– 0.866, 0.152]	– 0.357	0.261	[– 0.868, 0.154]
Widowed	– 0.252 **	0.080	[– 0.409,– 0.094]	– 0.252 **	0.080	[– 0.410,– 0.094]
Never married	– 0.899 **	0.295	[– 1.477,– 0.320]	– 0.899 **	0.295	[– 1.478,– 0.320]
Intercept	4.624 **	0.217	[4.200, 5.049]	4.646 **	0.217	[4.220, 5.072]
Num. of obs	9666			9666
F-statistics	1266.09			1268.77
P-values	[0.000]			[0.000]

The logit model is applied. Robust standard errors are reported. ^*p < 0.05, ^**p < 0.01

Discussion

Findings of this study suggested that the employment-based social health insurance (UEBMI) in China may help maintain health-related working capacity of retirement-aged people and reduce their willingness to participate in late-life careers. It is worth noting that there still have a fair number of older workers uncovered by the current health insurance scheme (UEBMI) in China. Imperfections in the current health insurance scheme could make it difficult for those with informal jobs to get coverage. They may lack better access to healthcare services to maintain their health-related working capacity. Their willingness to extend working life may be driven by economic pressure of higher out-of-pocket medical expenditures.

Moreover, the employment-based social health insurance (UEBMI) in China could weaken the negative associations of chronic conditions and health-related working capacity and late-life career participation. By providing a supportive set of financial and physical conditions, the health insurance scheme may play an important role in reducing the conflict encountered by chronically ill retirement-aged workers between persistent treatment and rehabilitation and demands of work. Chronically ill older adults who are insured by the UEBMI may be physically more able to work and are more likely to enter the labor force market after the retirement age. Therefore, the government should consider the relationships between chronic conditions, health insurance and the willingness and ability to delay retirement in the formulation of policies for delayed retirement.

Targeted policy preparedness for improvement in health status of retirement-aged workers

As the poor physical health may prompt retirement-aged people to withdraw from the labor force market in later life, it could be helpful to emphasize the targeted policy preparedness for improvement in their health status. For example, the appropriate working hours for retirement-aged workers can be included in the labor protection statutes, along with tax levy or subsidy incentive provided for employers, to encourage them to participate in late-life careers. Besides, more opportunities of position transfer to brainwork can be provided for older adults who used to be engaged in physically demanding work. It could also be helpful to improve the level of health security for retirement-aged workers with chronic illness to provide them with regular and affordable health examination services. Integration of intra-organization management and the health examination project may be beneficial to the follow-up health checks for retirement-aged workers. It could be favorable to create a healthy and age-friendly working environment and help employees reduce the risk of chronic conditions through health lectures, regular health examinations and other ways in the workplace. Moreover, it can be necessary to increase government investments in health care, including optimization of medical resources and improvement of health-related supporting facilities. This could help promote patients with chronic diseases to seek medical treatments in time and improve health quality. As a result, it is likely that the willingness and the capability of retirement-aged people with chronic diseases to participate in late-life careers will increase, to some extent helping relieve the pressures of state pensions and working-age population shrinking.

The establishment of supportive policies for late retirement initiative

The decision about late-life career participation could be subject to the change in many aspects including chronic conditions as noted in studies. It could be important to establish supportive policies in favor of retirement-aged people to stimulate their labor participation in the age of population aging. These supportive policies can consider an increase in salary of late retirees or a reduction in part of tax on them. Relevant laws and regulations concerning late retirement can also be established to protect benefits of late retirees. Some potential practices could focus on setting appropriate work demands and providing greater security for late retirees to encourage them to choose late-life career participation.

The delayed retirement initiative in China focuses mainly on reliving the pressure of population aging and accompanied social issues, without fine-grained consideration of disparities in physical health among retirement-aged workers. In practice, there may be a wide spectrum of health status among retirement-aged workers, which may affect their plans for participation in late-life careers. The “one size fits all” ideology is probably not suitable in the design of incentives for retirement-aged workers (Flynn 2010). Due to the high proportion of people with chronic conditions in China, the practical working ability of this group could be taken into consideration in the formulation of policies for delayed retirement. When a firm policy is adopted to extend the retirement age, some people may be unable to continue working due to poor health conditions, which may increase a certain degree of resistance to implementation of delayed retirement policies. Therefore, this study suggests to follow the principle of incentive and voluntariness in the early implementation of delayed retirement policies. Incentives, such as higher wages and pensions, could be used to encourage the public to delay retirement. The old workers who have chronic diseases or other subjective needs could be allowed to retire normally or retire ahead of schedule according to actual physical conditions.

Limitations and future research

This study still has some limitations. First, there is concern about potential endogeneity resulting from the non-randomness of chronic condition in this study. The incidence of chronic diseases in older adults may not be random and those who are disadvantaged in healthcare services are more likely to suffer from chronic conditions. The associations between chronic condition and late-life career participation as well as health-related working capacity may be due to a fact that they are all driven by latent common factors (e.g., healthcare utilization or disadvantage in healthcare services). In this case, the results of this study may only prove the associations between chronic condition and late-life career participation as well as health-related working capacity rather than the causality.

Second, there may be disparities in late-life career participation and health-related working capacity among retirement-aged people, given the characteristics of different chronic conditions and working capacity required varying with occupations. However, since the dataset used in this study has not provided the exact types of chronic diseases suffered by respondents and not differentiated between the assessments of health-related working capacity in different occupations, it is not feasible to examine these cases. With more information available, the fine-grained research can be conducted and more vivid findings can be revealed in the future.

Third, health-related working capacity is measured through a self-report. Self-report data may cause response bias when participants either lack the knowledge to answer correctly or they do not want to answer a question correctly (Althubaiti 2016; Devaux and Sassi 2016), although various ways are used to avoid response bias while framing survey questionnaire (e.g., by keeping questions short and clear; avoiding leading questions; avoiding or breaking down difficult concepts; coding options into an interval scale of score 1–3, etc.). With relevant information measured with objective data available, future research can use multi-method assessment to provide a more global and thus more likely accurate picture of respondents’ health-related working capacity. Besides, considering the years of education is not available in the present dataset and the measures of education in this study followed the original dataset, we cannot perform a robustness check using years of education to accurately compare the results. Finally, the data of this study are from China, the conclusions of study may thus be subject to regional limitations. The universality of the conclusions can be verified with global data in the future.

Funding

We appreciate the support from the National Social Science Fund of China (20CTY017).

Declarations

Conflict of interest

Authors of this study has no competing interest to declare.

Ethical approval and informed consent

The ethical approval and informed consent are not required in this stage, as this study uses publicly available data source and authors have no contact to human related materials. Meanwhile, the original materials have obtained relevant ethical approval and informed consent. More specifically, the data applied in this study are publicly available and unrestricted re-use is permitted via an open license. Besides, the CHARLS research team obtained ethics approval (license numbers: IRB00001052–11015, IRB00001052–14030, and IRB00001052–17053) from the institutional review board of the Peking University National School of Development. All respondents provided written informed consent. If the respondent was illiterate, he/she would press the fingerprint after the interviewer dictated the content of the informed consent. The study was conducted in compliance with WMA declaration of Helsinki.

Consent for publication

Consent for publication is not required since there are no personally identifying materials included in this manuscript.