More public funding? A choice experiment on the healthcare funding system for older populations

doi:10.21203/rs.3.rs-4858586/v1

Download PDF

Research Article

More public funding? A choice experiment on the healthcare funding system for older populations

https://doi.org/10.21203/rs.3.rs-4858586/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

We conducted unique discrete choice experiments on the healthcare funding system for older populations and examined several relevant attributes. The experiments, which comprised a sample of 1,112 individuals in Japan, focused on the need for public funding. We studied the valuation of several insurance attributes, namely, the share of public funding, income equity, intergenerational equity, and the local or national burden. This study elicited respondents’ preferences using a random parameter probit model and assessed the heterogeneity of choice responses using demographic, socioeconomic, and attitudes towards government bonds based on latent class responses. The results show that the public overall has a negative preference for an increase in public funding and positive preferences for an income-proportional burden, universal burden regardless of age, and an increase in the national government burden relative to the local burden. However, the latent class responded heterogeneously to insurance choices. For example, the latent segment comprising 23% of the total sample—the second-largest group consisting mainly of young males—preferred a greater increase in public funding.

The demand for healthcare for the older population is growing because of the aging population. Additional funding is required to satisfy this growing demand. Most developed countries have multiple financial sources: public (e.g., tax and social insurance) and private (e.g., private insurance and out-of-pocket payments). The funding system and public–private mixture of healthcare for older populations have been shaped by various factors. The social norms and preferences of the general public towards income equity, healthcare finance, healthcare utilization, and health in itself are important factors.

Health economics research has used discrete choice experiments (DCEs) to elicit preferences regarding various topics. A strength of DCEs lie in their ability to elicit individual preferences when faced with multiple discrete alternatives. By presenting respondents with a set of hypothetical choice scenarios, each characterized by varying attributes and levels, DCEs allow us to quantify the relative importance of different attributes influencing decision-making. Akkazieva et al. [1] elicited patient preferences regarding healthcare system reforms. DCE is also used to extract public preferences on the issue of the financial burden on the healthcare system. Gyrd-Hansen and Slothuus [2] elicited public preferences in Denmark for financing healthcare. Brau and Bruni [3] used DCEs to determine the demand for long-term care coverage.

Furthermore, DCEs explicitly consider the heterogeneity of preferences by classifying individuals into groups with similar preferences. Akaichi et al. [4] found by using these types of DCEs (Latent Class Model: LCM) that major segments in the US negatively evaluate the introduction of compulsory private insurance. This approach enables us to identify variations in decision making and specific groups of individuals with similar preferences. The LCM provides more nuanced insights into individuals’ preferences and better information-targeted interventions and policy recommendations.

Previous studies investigated the preferences of the general public in the US and Western European countries. However, no related studies have been conducted in Asian countries facing rapid aging.

Japan has the fastest aging population among developed countries. Currently, over 28% of the population is aged 65 years or older, and this number is projected to reach nearly 40% by 2060. Baby boomers in Japan were born between 1947 and 1949. Children of this generation are called second-generation baby boomers (baby boomer juniors). By 2035, all baby boomers will be at least 85 years old, and the same will be true for junior baby boomers by 2060. Thus, the healthcare demand of older generations will continue to grow for the next 40 years.

During the longstanding economic turndown since the 1990s, income inequality among the working generations has expanded in Japan. The Gini coefficient increased from 0.43 (0.36 after redistribution) in 1990 to 0.56 (0.37 after redistribution) in 2017 [5], indicating that Japanese society has become more inequitable. Therefore, more public support for healthcare expenses for working-age people (junior baby boomers are the main component) is indispensable to prevent more inequitable health in the future.

Japan has maintained a universal public healthcare system since 1961. The finances of healthcare for older people have been paid by taxes before a new medical care system for those 75 years and older started in 2008. Since then, social employment-based insurers have subsidized healthcare for those over 75 years, meaning that the working-age population supports the healthcare of older populations [6]. Moreover, rapid aging has increased the tax devoted to healthcare for older adults. Healthcare expenses for older adults increased from 1.32% (1990) to 3.0% (2019) of the total GDP.

Currently, public healthcare for those over 75 years is financed in part by the government and in part by health insurance schemes. Among government schemes, the ratio of central to local governments is two to one. Out-of-pocket payments for those over 75 years are income proportional, ranging from 10–30% of the total healthcare costs. The combination of these financial resources is one of the most important social issues, comprising taxes (central and local governments), social insurance, and out-of-pocket payments for older adults.

The current study aims to elicit the preference of the Japanese public towards finance allocation to the medical care of older populations and considers how respondent preferences for healthcare funding attributes vary across respondent groups. This approach will be helpful in designing funding policies tailored to the needs of each group and in addressing them more efficiently.

Our results using DCEs suggest that the public overall has a negative preference for an increase in public funding but has positive preferences for income-proportional burdens, universal burdens regardless of age, and the increase in central government burdens relative to local government burdens. We also relied on the latent class type of DCE to reveal the heterogeneity of public preferences.

2.1. Experimental design

The survey consisted of questionnaires and a DCE section. Before the DCE questions, respondents were asked to answer personal demographic questions regarding age, gender, household income, and their attitudes towards public bonds, and so on. In the DCE section, participants were successively provided with 12 different choice sets and repeatedly asked to choose between two different policy plans. The attributes, levels, and their descriptions are listed in Table 1.

Table 1

(1): Attributes and levels used in DCEs
Attributes	0	1
A. More Public Funds	Status quo	More Public Funds
B. Flat Income Burden	Status quo (income proportional)	Flat Burden
C. The Elderly Pay More	Status quo	The Elderly Pay More
G. More National Burden	Status quo	More National Burden

Table 1

(2): Descriptions of attributes
Attributes	Descriptions
A. Public Funds	Currently, 50% is publicly funded and 50% is self-funded (base). Do you think we need more public funding?
B. Flat Income Burden	People currently pay in proportion to their income (base). Do you think that what you pay should not depend on your income?
C. The Elderly Pay More	Currently, the intergenerational burden between those under 75 years and over 75 years is 4:1. Should young people pay less and older people (75 or over) pay more?
G. More National Burden	Currently, the ratio of the burden of the national government to the local burden is 2:1. Is a greater national burden needed?

The first attribute of the current design is More Public Funds. Currently, healthcare expenses for those over 75 years (HE75) are covered by both public funds (50%) and insurance premiums (50%). Some people may prefer more public spending, which means less subsidization from working generations.

The second attribute is the Flat Income Burden. This variable is related to income equity. Currently, HE75 is an income-proportional burden. Most Japanese people believe that people with higher incomes should pay more; in this case, this variable will be negative and significant.

The third attribute is The Elderly Pay More. This variable is related to intergenerational equity. Currently, the ratio of intergenerational burdens for those under 75 and over 75 years is four to one. Some respondents believed that older people should pay more.

The fourth attribute is More National Burden. This variable is related to the central government burden compared to the local government burden. Currently, the central government pays twice as much as local governments. Some respondents may have preferred smaller local government burdens, which would lead to a smaller budget share.

2.2. DCE modelling

Based on the utility model reported by Louviere et al. [7], individual i’s value function (U_ijt), which is specific for each jth alternative at the tth choice occasion among the 12 pairwise choice sets, takes the following form:

$$\:{U}_{ijt}={V}_{ijt}+{\epsilon\:}_{ijt}$$

V_ijt captures the deterministic term of respondents’ utility and ɛ_ijt the random term. ɛ_ijt is assumed to be independent and identically distributed. Assuming that the deterministic term is linear-in-parameter, Eq. (1) can be written as

$$\:{U}_{ijt}={\beta\:X}_{ijt}+{\epsilon\:}_{ijt}$$

where β means the vector of unknown utility parameters that are associated with the product attributes X_ijt.

The variables corresponding to the attributes More Public Funds, Flat Income Burden, The Elderly Pay More, and More National Burden were coded as zero or one, where zero means status quo and one means non-status quo. The positive signs of the beta coefficient show that respondents preferred more public funding, a flat income burden, for the elderly to pay more, and an increased national burden.

Respondents were asked to answer the questionnaires after reading the following explanation of Fig. 1.

Figure 1. The explanation of the DCE survey

Which plan do you think we should prioritize when our society must choose one? Respondents were requested to choose the preferred one while looking at a discrete-choice diagram as Table 1:

Table 2

Example DCE question
	Plan A ○	Plan B ○
A. More Public Funds	Status quo	More Public Funds
B. Flat Income Burden	Status quo (income proportional)	Flat Burden
C. The Elderly Pay More	The Elderly Pay More	Status quo
D. More National Burden	Status quo	Increased National Burden
Note: This Table is one of the examples in our survey.

In this analysis, we estimated the above parameters using two models. Based on the Random Parameter Probit (RPP) model proposed by Train [8], individuals’ preferences are assumed to be heterogeneous, and the Independence of Irrelevant Alternatives is assumed to be relaxed. In the RPP, the conditional choice probability that individual i selects alternative j at choice occasion t is given by

$$\:\text{Pr}\left({y}_{ijt}=1|{X}_{ijt}\right)=\varPhi\:\left(\beta\:{\prime\:}{X}_{ijt}\right)$$

The solution is approximated using simulation methods. The simulated log-likelihood (SLL) for all respondents is estimated via maximum likelihood procedures and is calculated as

$$\:L\left(\beta\:\right)=\sum\:_{i-1}^{n}\left\{{y}_{ijt}\text{log}\varPhi\:\left({X}_{i}^{{\prime\:}}\beta\:\right)+\left(1-{y}_{ijt}\right)\text{log}\left(1-\varPhi\:\left({X}_{i}^{{\prime\:}}\beta\:\right)\right)\right\}$$

Although the RPP model considers the heterogeneity of preferences, the type of heterogeneity that occurs is unclear. However, understanding the heterogeneity of preferences among individuals limits their practical use. The LCM can help us understand the sources of heterogeneity across individuals. The LCM sets several groups, where individual preferences within a group are homogeneous but those across groups are heterogeneous. Individuals’ membership in a specific group is assumed to be based on their preferences for funding attributes and characteristics.

The log likelihood for all respondents in the LCM for discrete choice analysis is

$$\:InL=\sum\:_{i=1}^{N}ln\left[\sum\:_{q=1}^{Q}{H}_{iq}\left(\prod\:_{t=1}^{Ti}{P}_{it|q}(j\right)\right]$$

where H_iq denotes the prior probability for class q for individual i.

Participants were recruited from among Japanese people aged 20 years or older who were registered with the survey panel managed by NTT Com. Online Marketing Solutions, Inc., a consumer research company. Informed consent was obtained from all individual participants included in the study. The participants were asked to provide their consent to participate in the survey after reading the objectives and a document explaining confidentiality protection on the survey website. This study was approved by the Institutional Review Board of Konan University on June 10, 2022.An invitation to participate in the survey was sent to 4,000 monitors. A total of 1,112 respondents were provided with a consent form and completed the survey. The response rate was 27.8%. Table 3 shows some differences between the distribution of the collected respondents and the actual populations. Our sample included fewer women than the national average. There was a slight under-representation of people in their 20s and 50s and a slight over-representation of people in their 60s.

Table 3

Demographic characteristics of respondents
		Samples	Population (%) (2022)
		n = 1127
Female (%)		49.8	51.4
Age	20 ~ 29	12.0	13.7
	30 ~ 39	14.4	14.7
	40 ~ 49	18.8	18.8
	50 ~ 59	16.4	18.9
	60 ~ 69	20.7	16.2
	70~	17.8	17.7
Level of education (%)	Junior high school	2.0
	High school	37.4
	2- or 4-year university	55.3
	Graduate school	4.9
Household income (%)	< 2 500 000 yen	18.0
	2 500 000 yen to 4 000 000 yen	23.7
	4 000 000 yen to 6 000 000 yen	22.1
	6 000 000 yen to 8 000 000 yen	15.0
	8 000 000 yen to 10 000 000 yen	10.3
	10 000 000 yen and more	10.7

Table 4 presents the results estimated by the RPP model. The results showed that the coefficients corresponding to the main effects were significant. Unexpectedly, we found a negative sign for the More Public Funds coefficient.

Table 4

Estimation results using RPP
	RPP (Total)
	Coef.	Std. Err	p value
More Public Funds	-0.0626	0.0172	< 0.001
Flat Income Burden	-0.5126	0.0152	< 0.001
The Elderly Pay More	-0.1381	0.0161	< 0.001
More National Burden	0.0427	0.0142	0.003
Constant	-0.0234	0.0138	0.088
n	1,112
Log likelihood	-8497.84
Wald chi square	1389.53
Prob > chi square	0.0000

3.1. More public funds

The results from the RPP model displayed in Table 4 show that respondents were more likely to choose the status quo, which means no more public funds. The subsample results displayed in Table 5 show that female and less-educated respondents were more likely to choose the status quo. That is, female and high school-educated respondents were not likely to require more public funds.

Table 5

(1): Estimation results for subsamples
	Male			Female
	Coef.	Std. Err	p value	Coef.	Std. Err	p value
More Public Funds	-0.0155	0.0240	0.519	-0.1123	0.0246	< 0.001
Flat Income Burden	-0.4765	0.0212	< 0.001	-0.5509	0.0217	< 0.001
The Elderly Pay More	-0.1033	0.0227	< 0.001	-0.1738	0.0230	< 0.001
More National Burden	0.0665	0.0199	< 0.001	0.0182	0.0202	0.368
Constant	-0.0555	0.0187	0.003	0.0100	0.0202	0.621
n	558			554
Log likelihood	-4316.30			-4165.82
Wald chi square	599.14			809.78
Prob > chi square	0.0000			0.0000

Table 5

(2): Estimation results for subsamples
	College			High School
	Coef.	Std. Err	p value	Coef.	Std. Err	p value
More Public Funds	-0.0291	0.0220	0.187	-0.1147	0.0276	< 0.001
Flat Income Burden	-0.5135	0.0195	< 0.001	-0.5137	0.0242	< 0.001
The Elderly Pay More	-0.0460	0.0207	0.027	-0.2800	0.0258	< 0.001
More National Burden	0.0420	0.0183	0.022	0.0436	0.0225	0.053
Constant	-0.0214	0.0176	0.226	-0.0271	0.0222	0.222
n	670			442
Log likelihood	-5133.48			-3337.82
Wald chi square	813.67			622.93
Prob > chi square	0.0000			0.0000

Table 5(3): Estimation results for subsamples

	Younger			Older (50 years and over)
	Coef.	Std. Err	p value	Coef.	Std. Err	p value
More Public Funds	-0.0704	0.0254	0.006	-0.0531	0.0234	0.024
Flat Income Burden	-0.4041	0.0222	<0.001	-0.6088	0.0210	<0.001
The Elderly Pay More	-0.0148	0.0239	0.536	-0.2419	0.0220	<0.001
More National Burden	-0.0172	0.0211	0.415	0.0927	0.0193	<0.001
Constant	-0.0047	0.0220	0.83	-0.0381	0.0175	0.03
n	502			610
Log likelihood	-3937.69			-4496.98
Wald chi square	407.45			1063.09
Prob > chi square	0.0000			0.0000

3.2. Income equity

Table 4 shows that the respondents tended to prefer an income-proportional burden. Table 5 shows this tendency in all subgroups regardless of gender, educational background, and age. For example, both male and female respondents were more likely to choose an income-proportional burden.

3.3. Intergenerational equity

Respondents in general tended to have negative preferences for ‘the older the more’. That is, they preferred a universal burden regardless of age. We found this tendency in all subgroups except for younger respondents. Table 5 shows that only the younger population had no significant tendency toward the universal burden.

3.4. More national burdens (more central government burdens)

Respondents tended to have a significantly positive preference for increased central government burdens. However, Table 5 shows that female, less-educated, and younger respondents had no significant preference for a greater national burden.

3.5. Heterogeneity

We have described how individual preferences for funding policy attributes varied across respondents. They were classified into several classes based on their preferences. A Consistent Akaike Information Criterion (CAIC) was used to determine the appropriate number of classes. All estimations were performed using Stata ver. 16. The values corresponding to the LCM with two to seven classes are listed in Table 6. We report the results of the estimation of the LCM with six classes, which has the lowest CAIC value (Table 6).

Table 6

Consistent Akaike Information Criterion (CAIC)
Class number and CAIC
2	16182.5
3	15507.2
4	15517.2
5	14993.3
6	14857.8 ***
7	14871.6

Below, we highlight the features that characterize the preferences of the members of each of the six classes (which are named according to their characteristics). This can help design funding policies that match the needs of the members of each class.

3.5.1 More public funds

Responses displayed significant heterogeneity. Table 7(1) shows that Class 2 respondents (23% of respondents) were willing to choose more public funds, while members of Classes 3 and 4 (22% and 19% of respondents, respectively) were willing to remain in the status quo.

Table 7

(1): Estimation results using Latent type of DCE
	RP (Total)			Class 1		Class 2		Class 3		Class 4		Class 5	Class 6
	Coef.		Std. Err	Coef.		Coef.		Coef.		Coef.		Coef.	Coef.
More Public Funds	-0.0626	***	0.0172	0.0332		0.3899	***	-0.2776	***	-0.4737	***	-4.3405	-0.2231
Flat Income Burden	-0.5126	***	0.0152	-2.5351	***	0.1175		-2.6935	***	0.1355		-3.1168	0.1745
The Elderly Pay More	-0.1381	***	0.0161	-1.9134	***	-0.6690	***	1.6491	***	0.4387	**	2.4805	0.7912
More National Burden	0.0427	**	0.0142	0.2553	***	0.4991	***	0.0223		-0.5940	***	1.4016	-0.4933
Constant	-0.0234	*	0.0138	0.0965		-0.1939	***	-0.0200		0.1764	***	-10.2530	3.7215	***
Class Share	1.00			0.300		0.233		0.223		0.189		0.031	0.023

Table 7(2): The latent class results by characteristics

	Class 1			Class 2			Class 3
	Coef.	Std. Err	p value	Coef.	Std. Err	p value	Coef.	Std. Err	p value
Constant	-	-	-	0.8928	0.4828	0.064	0.7680	0.4302	0.074
Female	-	-	-	-0.6500	0.2180	0.003	-0.2124	0.2033	0.296
Age	-	-	-	-0.0293	0.0079	0.000	-0.0272	0.0072	0.000
College	-	-	-	0.1833	0.2174	0.399	0.7345	0.2110	0.001
Married	-	-	-	0.5806	0.2433	0.017	0.0179	0.2179	0.935
Public Bonds	-	-	-	0.7085	0.2515	0.005	0.0354	0.2605	0.892
High Income	-	-	-	0.7260	0.2792	0.009	0.3623	0.2848	0.203
%	30.0%			23.3%			22.3%

	Class 4			Class 5			Class 6
	Coef.	Std. Err	p value	Coef.	Std. Err	p value	Coef.	Std. Err	p value
Constant	2.2265	0.4465	0.000	-1.0361	0.7879	0.189	-0.6856	1.0208	0.502
Female	-0.2383	0.2374	0.315	-0.2807	0.3899	0.472	0.7360	0.5221	0.159
Age	-0.0560	0.0089	0.000	-0.0317	0.0138	0.022	-0.0458	0.0155	0.003
College	0.1193	0.2297	0.604	0.3778	0.3959	0.340	0.2022	0.4923	0.681
Married	0.2897	0.2508	0.248	0.2950	0.4133	0.475	-0.1390	0.4971	0.780
Public Bonds	-1.0208	0.3854	0.008	1.0694	0.4064	0.009	0.4527	0.6354	0.476
High Income	1.0631	0.2946	0.000	-0.2922	0.5879	0.619	-1.5614	1.3108	0.234
%	18.9%			3.1%			2.3%

Note: 'Public Bonds' means those who have the positive attitudes towards public bonds for supporting the future burden.

3.5.2 Income equity

We found some heterogeneity across classes. The largest group (Class 1, 30% of respondents) and Class 3 (22% of all respondents) tended to choose an income-proportional burden, whereas the other classes had no significant preference.

3.5.3 Intergenerational equity

Table 7(1) shows some heterogeneity across classes. Classes 1 and 2 (30% and 23% of respondents, respectively) tended to prefer a universal burden regardless of age, while Classes 3 and 4 (22% and 19% of respondents, respectively) were willing to ask the older population to pay more.

3.5.4 More national burdens (more central government burdens)

There was some heterogeneity across classes. The two largest groups (Classes 1 and 2 comprising 30% and 23% of respondents, respectively) tended to prefer more central government burdens, while Class 4 (19% of respondents) preferred the status quo.

Senior-friendly group (Class 1):

This group comprised the largest proportion with 30% of the sampled respondents. Members of this class placed more emphasis on the universal burden, regardless of age. That is, they had a negative preference for ‘the older the more’. Additionally, they preferred an income-proportional burden and an increase in the national government burden relative to the local burden.

Public spending advocates (Class 2):

Members of this segment (23% of respondents) were characterized by respondents with a positive preference for public funds, a universal burden regardless of age, and an increase in the national government burden. Class 2 is a unique class whose members preferred the More Public Funds policy.

Class 2 members were more likely to be male, younger, married, and to have a high-income household. They were also more likely to have positive attitudes toward public bonds for elderly care.

Harsh group to the older population (Class 3):

This class (22% of respondents) stood out as harsh towards the older population. They had a positive preference for ‘the older the more’. Furthermore, members of Class 3 were more likely to remain with the status quo for public funds and prefer income-proportional burdens. Members of Class 3 tended to be younger and highly educated.

Austere group (Class 4):

These class members had austere attitudes towards elderly care. Members of Class 4 were more likely to have negative attitudes towards public funding and central government spending. Furthermore, they were harsh towards the older population. They had a positive preference for ‘the older the more’. Members of Class 4 tended to be younger and live in high-income households. They also had negative attitudes towards public bonds.

Minority groups (Classes 5 and 6):

Classes 5 and 6 were the smallest groups, accounting for only 3% and 2% of respondents, respectively. Four attributes, More Public Funds, Flat Income Burden, The Elderly Pay More, and More National Burden, were all statistically insignificant, which means Classes 5 and 6 are status quo groups.

Considering the above results, the Japanese population can be divided into four groups, except for the two 2–3% minority groups that prefer the status quo.

We employed a DCE on the funding system of healthcare for older populations and examined several of its relevant attributes, namely, the share of public funds, income equity, intergenerational equity, and the local or national burden.

The results show that the general public has a negative preference for an increase in public funds and a positive preference for income-proportional burdens, a universal burden regardless of age, and an increase in the national government burden relative to the local burden.

Several studies have analyzed public preferences on healthcare financing in developed countries. In this study, the general public preferred equal burdens regardless of age. This finding was observed in similar research in other countries [9]. This preference was also observed among doctors in Switzerland [10]. In other studies, the older population was allocated more healthcare financing than the younger population (Wagstaff [11] ; Van Doorslaer et al. [12] ). Thus, this topic may not raise a severe policy debate, because the public and experts have similar preferences.

In this study, the general public also preferred a more progressive burden on the high-income group. Based on the Comprehensive Survey of Living Conditions by the Ministry of Health, Labor and Welfare, the median household income of all households in Japan has drastically fallen over the past 25 years (from 5.45 million yen in 1995 to 4.4 million yen in 2020; one-dollar values were around 120 yen in 2020). This means that the average income of the middle class has declined significantly. A study [13] found that the burden of “social insurance premiums” and “patient co-payments” are regressive. These findings show that, while the middle class is shrinking and the average household income is sluggish, the burden of social insurance premiums is increasing [14]. Lowering the burden on the middle class and raising the burden on the high-income class is an effective strategy for reviving the middle class. Lowering the burden of social insurance premiums is a useful means of raising the disposable income of the middle class [15]. The study revealed that the burden ratio (social insurance premium 6.8% and self-payment 3.0%) was higher in the lower-income group (lowest in the 1st decile) than the burden ratio on the higher-income group (4.3% and 0.7%, respectively) (highest in the 10th decile).

Prior studies examined public preferences by focusing on income and intergenerational equity. Regarding income equity, those who have lower income should be given more healthcare financing than those with higher incomes [9]. In terms of intergenerational equity, older populations should be given more healthcare financing than younger populations [11, 12].

It was also found that people prefer for the central government to have more financial commitments than local governments. In Japan, community-based public health insurance is managed by first-tier local (municipal) governments, which leads to severe budget crises. There have been discussions about the integration of community-based public health insurance into a national entity in developed East Asian countries such as Korea and Taiwan. The general public may agree with this unification policy.

Another important finding of this study is that individual preferences are heterogeneous. In particular, preferences towards the volume of tax subsidies were diverse. The results showed that the second largest segment, mainly young males, comprising 23% of the total respondents, preferred an increase in public funding, while the general public had a negative preference for more public funding. The reason for this is because the national burden rate has risen to around 50%. While the public overall is against increasing public funding, roughly one-quarter of people favor expanding them.

This study has several limitations. First, our results were based on public responses to a hypothetical survey, meaning that some respondents may not have fully understood the consequences of the scenario. However, DCEs are useful methods that allow us to study various scenarios and attributes related to funding models for public preferences. The second limitation is related to the survey sample, which may differ in certain characteristics from the Japanese population. However, the distribution of our survey sample was almost the same as the actual distribution for some characteristics such as gender and age.

Ethics approval:

This study was approved by the Institutional Review Board of Konan University on June 10, 2022.

Author Contribution

Contributors Concept and design: TM and RG. Acquisition, analysis or interpretation of data: TM and RG. Drafting of the manuscript: TM. Critical revision of the manuscript for important intellectual content: TM and RG. Statistical analysis: TM. Obtained funding: TM and RG. Administrative, technical or material support: TM and RG. Supervision: TM and RG.

Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Akkazieva B, Gulacsi L, Brandtmuller A, Péntek M, Bridges JF. Patients’ preferences for healthcare system reforms in Hungary: a conjoint analysis. Appl Health Econ Health Policy. 2006;5:189–98. https://doi.org/10.2165/00148365-200605030-00005
Gyrd-Hansen D, Slothuus U. The citizen’s preferences for financing public health care: a Danish survey. Int J Health Care Fin Econ. 2002;2:25–36. https://doi.org/10.1023/a:1015345429726
Brau R, Lippi Bruni M. Eliciting the demand for long‐term care coverage: a discrete choice modelling analysis. Health Econ. 2008;17:411–33. https://doi.org/10.1002/hec.1271
Akaichi, F., Costa-Font, J., & Frank, R. Uninsured by choice? A choice experiment on long term care insurance. Journal of Economic Behavior & Organization. 2020;173:422-434. https://www.sciencedirect.com/science/article/abs/pii/S016726811930232X
Ministry of Health, Labour and Welfare. Annual Report on Health, Labor and Welfare. 2020.
Ikegami N, Yoo BK, Hashimoto H, Matsumoto M, Ogata H, Babazono A, et al. Japanese universal health coverage: evolution, achievements, and challenges. Lancet. 2011;378:1106–15. https://doi.org/10.1016/S0140-6736(11)60828-3
Louviere JJ, Pihlens D, Carson R. Design of discrete choice experiments; 2011 a discussion of issues that matter in future applied research. J Choice Model. 2011;4:1–8.
Train, K. E. Discrete choice methods with simulation. Cambridge university press. 2009.
Lu, H., Burge, P., & Sussex, J. Measuring public preferences between health and social care funding options. Journal of choice modelling. 2021;38: 100266.
Jannot AS, Perneger TV. What should be the basis for compulsory and optional health insurance premiums? Opinions of Swiss doctors. Swiss Med Wkly. 2014;144:w13918. https://doi.org/10.4414/smw.2014.13918
Wagstaff A, van Doorslaer E, Calonge S, Christiansen T, Gerfin M, Gottschalk P, et al. Equity in the finance of health care: some international comparisons. J Health Econ. 1992;11:361–87. https://doi.org/10.1016/0167-6296(92)90012-p
Wagstaff A, van Doorslaer E, van der Burg H, Calonge S, Christiansen T, Citoni G, et al. Equity in the finance of health care: some further international comparisons. J Health Econ. 1999;18:263–90. https://doi.org/10.1016/s0167-6296(98)00044-7
Tanaka S, Shikata M, Komamura K. Analysis of tax and social security burdens on senior citizens: using individual questionnaire data of the national survey of family income and expenditure. Financ Rev. 2013;115:117–33 (in Japanese)
Tanaka S, Shikata M. The middle class in Japan, 1994–2009: trends and characteristics (No. 2019-001). Institute for Economic Studies, Keio University; 2019
Ando M, Furuichi M, The middle class and the welfare state: examination of various studies on the redistributive effects of taxes, social insurance premiums, and social security. Rengo Research Institute Report No. 32(6); 2019, 14–8 (in Japanese)

No competing interests reported.

Download PDF

Reviews received at journal
22 Sep, 2024
Reviewers agreed at journal
15 Sep, 2024
Reviewers agreed at journal
11 Sep, 2024
Reviewers agreed at journal
28 Aug, 2024
Reviewers invited by journal
20 Aug, 2024
Editor assigned by journal
16 Aug, 2024
Submission checks completed at journal
14 Aug, 2024
First submitted to journal
04 Aug, 2024

You are reading this latest preprint version

More public funding? A choice experiment on the healthcare funding system for older populations

Status:

Version 1

Abstract

Figures

1. Introduction

2. Method

2.1. Experimental design

2.2. DCE modelling

3. Results

3.1. More public funds

3.2. Income equity

3.3. Intergenerational equity

3.4. More national burdens (more central government burdens)

3.5. Heterogeneity

3.5.1 More public funds

3.5.2 Income equity

3.5.3 Intergenerational equity

3.5.4 More national burdens (more central government burdens)

4. Discussion and conclusion

Declarations

References

Additional Declarations

Status:

Version 1