Background

Family physician (also known as the general practitioner) system is considered as the core of primary care and has gained rapid ground worldwide [1,2,3]. Primary care services provided by family physician enhance the continuity and coordination of care, reduce the inappropriate use of specialty services, and improve a population’s health [4,5,6]. The Chinese government has explored and established a hierarchical medical system to suit the national conditions in recent years [7]. The hierarchical medical system emphasizes the importance of a clear division of health care roles and responsibilities for different levels of hospitals, and highlights the significance of taking strategies to encourage the utilization of primary care [8]. The implementation of the family physician-contracted service (FPCs) is one of the key actions, which improve people’s access to primary health care, minimize the unfair distribution of medical resources and improve the health of community residents [9,10,11].

Family physician-contracted services were officially launched throughout China in June 2016 after pilot projects in 200 selected areas [12]. Family physician-contracted services target the entire population, focusing on the elderly, pregnant women, children, people with disabilities, and patients with hypertension, diabetes, tuberculosis and other chronic diseases. The public is being encouraged and guided to contract with nearby family physician teams who are responsible for the provision of proactive, continuous, and comprehensive health care [12]. Family physician-contracted services are being explored and implemented in various regions of China. For example, in Shaanxi province (part of Northwest China), the FPCs—fully launched in 2016—provides ‘free service package + paid service package’ and its operational cost is partly subsidized from the regional funds of medical insurance and basic public health program [13]. The free service package was designed to provide basic public health services: such as establishing and managing personal health records, health education, health literacy promotion, and primary care management of chronic diseases. The paid service package includes basic medical services and personalized health management services. The residents who are contracted with a family physician can enjoy free or low-cost healthcare services from the family physician team, which include family physicians, nurses, public health practitioners.

The effects of FPCs have been widely reported in previous studies. Family physician-contracted services can improve the continuity, comprehensiveness, and coordination of primary care services [14,15,16]. This service also helps individuals with chronic diseases to improve their health-related awareness, self-management behaviors, and treatment compliance [17, 18]. For example, a population-based retrospective cohort study in Canada suggested that family physician care can reduce hospitalizations in elderly people with diabetes [19]. Another randomized controlled trial study in Norway identified that in addition to geriatrician, clinical assessments and medication reviews from collaborative team including family physicians can improve health-related quality of life (HRQoL) in older patients exposed to polypharmacy [20]. However, one Iranian study indicated that family physician program had a positive effect on the proximal health indicators in maternal and child health (e.g., birth weight), but no substantial effect on mortality [21]. One study from South Africa did not find a positive association between the supply of family physician and health indicators (i.e., maternal mortality, perinatal mortality and under-five mortality) [22]. The inconsistencies among existing studies could be partly attributed to different health indicators that were used.

Existing studies concerning the effects of FPCs on health in China are mainly limited on improving health outcomes in patients with chronic diseases. For example, previous studies have shown that FPCs have positive effects on hypertension control rates among patients with hypertension [18, 23]. Unfortunately, in China, whether family physician-contracting services could improve HRQoL and health equity in the general population has not been extensively studied.

Therefore, we conducted this study using the data from a household health survey in Shaanxi Province, China, with the aim to (1) explore whether being contracted with family physician services can improve HRQoL, (2) investigate whether the inequity in HRQoL would be lessened under the implementation of FPCs, which might provide implications for policymakers in terms of advancing family practice contract services in China. Based on previous research [15, 24,25,26], we hypothesize that FPCs would help improve community residents’ HRQoL and lessen the inequity in HRQoL.

Methods

Data sources

Data were derived from the 2018 cross-sectional household health survey that was conducted in three districts and two counties in Shaanxi Province, an area about 205,800 km2 in northwest China with a population of 37.9 million.

Participants were recruited by multistage, stratified cluster random sampling. In short, out of 5 districts or counties, 25 sub-districts or townships were randomly selected. Out of those sub-districts, we randomly selected 50 communities or villages (two for each sub-district or township). All members from 3000 households including 7819 individuals were collected. Only participants aged 15 years and above were included in this study, since children who were below 15 years could not answer several questions regarding socioeconomic characteristics and health status. Finally, 6503 individuals were included in the analyses.

Measures

Pre-trained interviewers conducted face-to-face interviews by structured questionnaire, and collected the following information: sociodemographic characteristics, economic characteristics, health status, health services coverage, and health services utilization.

Sociodemographic and economic characteristics include the following information: sex, age group (15–44 years/45–59 years/60 years and above), educational level (primary and below/junior middle school/senior middle school/ college and above), marital status (unmarried/married/others), employment status (unemployed/employed), commercial medical insurance (with/without), minimum travel time to the nearest health-care facility (within 15 min/more than 15 min), basic medical insurance (Urban Employee Basic Medical Insurance [UEBMI]/Urban-Rural Resident Basic Medical Insurance [URRBMI]), chronic conditions (yes/no), and residential areas (urban or rural areas). The household consumption expenditure per equivalent adult was used as a proxy measure of economic status. We divided the household consumption expenditure per equivalent adult into five groups, the first quintile represents the poorest economic group (i.e., the lowest 20%) while the fifth quintile represents the wealthiest economic group (i.e., the highest 20%).

We inquired whether responders had contracted with a family physician by asking “Are you contracted with a family physician?”, which they could respond with “yes” (scored as 1) and “no” or “I’ve heard nothing of this services” (scored as 0).

Health-related quality of life (HRQoL) combines physical and mental health into a summary score, which ranges from 0 (death) to 1 (perfect health). Notably, certain scores can also measure states worse than death [27], such as an index score of EQ-5D ranges − 0.59 (worst possible health state) to 1 (best possible health states) according UK tariffs, and by construction, the value of 0 is equal to death and negative values represent HRQoL worse than being dead [28]. In our study, HRQoL was measured by the health utility values for the validated Chinese version of the EuroQol five-dimensional questionnaire-three-level version (EQ-5D-3L) [29, 30]. The questionnaire includes five questions and consists of five dimensions, i.e., responder’s mobility, self-care, usual activities, pain or discomfort, and anxiety or depression; each dimension has three response levels (1 = no problems, 2 = moderate problems and 3 = extreme problems). We employed the Chinese time trade-off values for EQ-5D-3L to measure the utility values of the EQ-5D-3L [31], which has also been widely used in other studies [32, 33]. The overall utility values of EQ-5D-3L ranges from − 0.149 (having extreme problems) to 1 (no problems).

Statistical analyses

Matching method

Coarsened exact matching (CEM)

The pretreatment covariates differ between the groups with and without FPCs for the observational data, the Coarsened Exact Matching (CEM) technique is designed to improve the estimation of causal effects via a powerfully matching method to keep a better balance of distributions of the covariates between groups, and thereby reducing the bias [34]. The idea of CEM is to temporarily coarsen each variable into substantively meaningful groups, match on these coarsened data exactly, and then only retain the original (uncoarsened) values of the matched data [35]. We applied the CEM method to control confounding variables between the two groups to investigate whether a contracted family physician could improve HRQoL and promote health equity. The detail of technique is well described in detail in previous literatures [34, 35]. The STATA cem command was used to perform the CEM algorithm [35].

There is a comprehensive imbalance measure L1 statistic, which was used to check the overall imbalance. The calculation for L1 is as follows: first, we coarsened the covariates into bins, then cross-tabulated the discretized variables as X1 × … × Xk for the groups with and without FPCs separately, and recorded the k-dimensional relative frequencies for the group with FPCs \( {f}_{e_1\dots {e}_k} \) and for the control \( {g}_{e_1\dots {e}_k} \) units. Finally, the absolute difference over all the cell values is defined as the measure of imbalance L1 :[35].

$$ {L}_1\left(f,g\right)=\frac{1}{2}{\sum}_{e_1\dots {e}_k}\left|{f}_{e_1\dots {e}_k}-{g}_{e_1\dots {e}_k}\right| $$
(1)

L1 ranges from 0 (perfect global balance) to 1 (maximal imbalance). A substantial reduction in L1 means a good matching performance. In this study, we focused on matching the following covariates: sex, age, chronic conditions, economic status, educational level, employment status, marital status, medical insurance, spatial accessibility of health-care facility, and residential areas.

Tobit regression models

(a) Concentration index (C)

We measured the degree of socioeconomic-related inequality in HRQoL between the two groups in our study using the concentration index (C), which is a widely used parameter [36, 37]. The C takes values in the range [− 1, 1]; a value of 0 means perfect equality, a positive value signifies pro-rich bias while a negative value signifies pro-poor bias. The C formula is

$$ \mathrm{C}=\frac{2}{\mu}\mathit{\operatorname{cov}}\left({y}_i,{r}_i\right) $$
(2)

where y is the health variable (i.e., the utility values of the EQ-5D in this study), μ is mean of the EQ-5D utility value, ri (rang [0, 1]) is the fractional rank of the i th individual in the economic distribution.

(b) Decomposition analysis for C.

The decomposition analysis was based on a regression model that decomposes C into its contributing factors. Tobit regression was applied to calculate the partial effects of regressors, owing to the outcome variables (i.e., the utility values of EQ-5D) which were limited variables [38]. Health (y) is modelled as follows: [36].

$$ {y}_i={\sum}_k{\beta}_k{x}_{ki}+{\varepsilon}_i $$
(3)

βk are the partial effects, ε is the error term, while xk are the means of explanatory variables. C(y) can be decomposed as: [36].

$$ \mathrm{C}={\sum}_k\left(\frac{\beta_k{\overline{x}}_k}{\mu}\right){C}_k+\frac{GC_{\varepsilon }}{\mu } $$
(4)

Ck is the concentration index of explanatory variables and GCε is the generalized concentration index of ε. The deterministic component is \( {\sum}_k\left(\frac{\beta_k{\overline{x}}_k}{\mu}\right){C}_k \), while the residual component is \( \frac{GC_{\varepsilon }}{\mu } \). The contribution of each factor depends on its impact on health and the degree of unequal distribution across the economic gradient. The percentage contribution of each regressor (100Qk/C) was also calculated, and it can take both positive and negative values. We further calculated the age-sex adjusted C (also known as the horizontal inequity index)—which represents the potentially avoidable inequality—by subtracting the contributions of age and gender from the total C [39].

Standard errors were adjusted for clustering at the family level for all models. Statistical analyses were conducted using the STATA version 14.0 (Stata Corporation, College Station, Texas, USA) with the significance level as a P <  0.05 (two-tailed).

Results

Demographic characteristics of responders and matching performances

Table 1 represents basic sociodemographic and economic characteristics of the respondents before and after data matching. A total of 6503 respondents were included, and data from 4612 individuals were further analyzed after data matching. Prior to data matching, the proportion of respondents aged 60 and above, having chronic conditions, with middle and below economic status, with junior middle school and below education, having Urban-Rural Resident Basic Medical Insurance, and living in rural region in group with FPCs was greater than group without FPCs. After matching using CEM method, covariates imbalances were eliminated between the two groups. Table 2 shows L1 statistics of each variable, and multivariate L1 statistics before and after data matching. Consistently, in the two groups, L1 values of each variable decreased and were close to zero after data matching.

Table 1 Characteristics of respondents among those with and without a contracted family physician before and after data matching
Full size table
Table 2 The L1 measure of imbalance between two groups before and after Coarsened Exact Matching
Full size table

Description of EQ-5D and its concentration index

The utility values of EQ-5D and its each dimension are presented in Table 3. After data matching, the mean EQ-5D utility values between respondents with FPCs were 0.9355 (95% CI: 0.9302–0.9409), and without FPCs were 0.8995 (95% CI: 0.8926–0.9063). The Cs of EQ-5D score for the respondents with a contracted family physician was significantly lower than those without a contracted family physician (0.0084, 95% CI: 0.0047–0.0122 vs. 0.0263, 95% CI: 0.0187–0.0340; P <  0.001), suggesting that the pro-rich bias was slightly reduced by contracting family physician services.

Table 3 Description of EQ-5D health state and economic-related inequality in EQ-5D scores between respondents with and without contracted family physician
Full size table

Decomposition of inequality of HRQoL

Table 4 presents the decomposition results of Cs of the overall EQ-5D scores. The partial effects, absolute contribution and percentage contribution of each determinant to the inequality of the overall EQ-5D score are presented.

Table 4 Decomposition of Concentration Index of the EQ-5D Scores between respondents with and without contracted family physician
Full size table

Partial effect estimates indicated that among respondents with a contracted family physician, those aged 45 years and above, having chronic conditions, and married, divorced and widowed status were more likely to have lower overall HRQoL. Among those who did not contract with a family physician, the following factors were negatively associated with the overall HRQoL: aged 60 years and above, presence of chronic conditions, low economic status and educational level, divorced and widowed status, living in rural areas and far from health-care facilities.

A positive contribution to inequality means that the relevant determinant increases pro-rich inequality and vice versa. For those with a contracted family physician, the top three categories of residents with the greatest inequality in HRQoL were: aged 60 years and above (90.49%), having at least college-level education (79.44%) and richest economic status (68.10%). For those without a contracted family physician, the following factors had the largest positive contributions in explaining the inequality of HRQoL: richest economic status (58.62%), richer economic status (40.61%), and having at least college-level education (27.01%). The age-sex adjusted Cs in HRQoL among those with a contracted family physician was lower (0.003) than those without (0.022).

Discussion

Benefiting from the CEM technique and a representative dataset (the 2018 household health survey in Shaanxi Province, northwest China), our study found that individuals with a contracted family physician had significantly better HRQoL that was measured with EQ-5D-3L than those without. Moreover, it also suggested that the inequities in HRQoL were lower among those who contracted with a family physician than those who did not.

Family physician services are favorably helpful in promoting HRQoL [20, 40]—which was reconfirmed in this study showing that the EQ-5D-3L utility value among individuals who contracted with family physician services was greater than those who did not (P <  0.001). HRQoL is a generic outcome measure to assess the effectiveness of interventions and a way of reflecting one’s subjective perceptions and experiences [41]. Impaired HRQoL denotes perceived difficulties and functional limitations in daily life caused by illness [27]. The results also suggested that individuals contracted with a family physician were less likely to have functional limitations in moderate mobility, activity, pain and moderate anxiety. On the one hand, the health care services delivered by the family physicians include medicine and physical therapies, rehabilitation services, health counseling and the referral services. These proactive, continuous and comprehensive health care services could help enhancing the physical ability and mental health of contracted patients, such as relieving pain, improving mobility and activity, relieving depression and anxiety. On the other hand, the long-term cooperative relationships between contracted family physicians and residents promotes trust in primary care physicians among residents, which were identified to be associated with lower depression and anxiety in patients [42]. FPCs is a comprehensive strategy that strives to ensure equitable access, continuity of care, coordination and comprehensiveness of care in primary healthcare. Previous literature has identified the remarkable effects of FPCs in improving the quality and utilization of primary healthcare services [16, 43], which are highly correlated with better health outcomes, including total and specific mortality, life expectancy, and other health outcome indices, e.g., self-rated health, HRQoL [24].

Our study found that the pro-rich bias was slightly reduced by family physician-contracted services; age-sex adjusted Cs in HRQoL were 0.003 among the group with FPCs, and 0.022 among the group without FPCs—indicating the role of FPCs in reducing health disparities. Striking differences in health outcome still exist within and between populations, so equity is an important goal in health sectors. The primary healthcare system is important to promote equity in access and health outcome of a population, yet primary healthcare services are not always readily accessible for some populations in China [44]. The FPCs in China established a long-term contracted relationship between residents and family physicians. They provide comprehensive clinical care and health management services to the contracted populations, and promote information exchange; thus helping to improve the accessibility of primary health services [45]. The cost of FPCs is partially subsidized by public finances, which may reduce barriers to healthcare utilization for disadvantaged groups and thereby improving their health status [46].

From a policy perspective, improvement in health equity needs a broader focus that addresses socioeconomic determinants of health and health equity. The current FPCs gives the high-need populations, such as the elderly, pregnant women, children, people with disability, and people with chronic diseases, the top priority to primary healthcare services in our sample areas. Our findings suggested that lower socioeconomic status, i.e., lower educational level, unemployment, poor spatial accessibility, and living in rural areas were associated with lower HRQoL. In addition, economic status and educational level were the main sources of inequities in HRQoL among both individuals with and without FPCs. Therefore, authorities need to pay more attention to the economically and educationally disadvantaged groups to further enhance health equity.

Improving health status and the longstanding inequity of health and healthcare still remain as challenges nevertheless there was progress in implementing FPCs and primary healthcare services [10]. Firstly, further efforts to improve people’s awareness and willingness to contract with a family physician are required. In Shaanxi province, the signing contract rate among general residents was 54.2% (according to our full sample data) in 2018. Previous studies showed that the signing contract rates varied greatly in different Chinese regions (30–60%) [47,48,49]. Secondly, the family physician service team should be expanded further to include other allied health professionals such as counselling psychologists, health managers, pharmacists and social workers; this would enhance the capability of the FPCs team to provide better health services [50]. General practitioners, community nurses and public health practitioners are the main components of the current FPCs team, which may not be capable of meeting residents’ multifaceted healthcare needs, such as psychological counseling and care services which as we found no significant effects in HRQoL in the dimension of self-care and extreme anxiety. Well-implemented primary healthcare services tend to bring benefits to the most vulnerable in the communities as well as to those with complex healthcare needs [44].

One of our study merits is that we examined the effects of FPCs on health and performed the CEM technique to guarantee better balance of covariate distribution between individuals with and without FPCs. The second is that we used the EQ-5D-3L instrument that was validated among the Chinese general population [30] as a health outcome measure to assess the effect of FPCs on our sample. Some limitations should be noted. Firstly, some unobservable or unmatched factors, such as health literacy and beliefs, may have potential effects on the results. Secondly, the results should be explained as associations rather than causal effects as we used cross-sectional data, thus further longitudinal studies should be conducted. In addition, the severe ceiling effect of EQ-5D-3L cannot be fully eliminated when measuring HRQoL among the general population. Specifically, although 71.7% of the residents in our study rated themselves as in full health, it was slightly smaller than results in another Chinese study [51]. Lastly, the samples were obtained from one province, and there were heterogeneities regarding FPCs policies and socioeconomic environments between regions, which might limit the study generalizability.

Conclusion

This study sheds light on the positive effects of FPCs on the HRQoL and socioeconomic-related equity by comparing the magnitude of inequalities in HRQoL between the individuals with and without FPCs. To further enhance health equity, authorities should pay more attention to the economically and educationally disadvantaged individuals. Our study provided evidence for policy-makers and healthcare institutions to promote and implement FPCs, extend the service coverage of family physicians and enhance the professional competency of family physicians service team.