Accepted on 14 Sep 2022
Submitted on 11 Jun 2021
Continuity of care is a crucial but often neglected component when we target strengthening primary care in low- and middle-income countries (LMICs) [1]. Continuity of care reflects the extent to which a series of discrete healthcare events is experienced by patients as coherent and interconnected over time and consistent with their health needs [2]. Continuity of care is a comprehensive concept, and its most important aspect is interpersonal continuity which results in trusting relationships between patients and their physicians [3]. As included in the Starfield 4C principle of primary care (first-contact access, care coordination, comprehensiveness, and continuity), continuity of care is particularly critical to the management of non-communicable diseases (NCDs) [4]. It has been widely reported that increased continuity improved the experience of care for both patients and providers, the quality of care, and health outcomes [5, 6]. However, continuity of care has received relatively little attention within LMICs. Historically, health systems in LMICs primarily provided acute, episodic care, with little capacity for longitudinal follow-up and care coordination [7]. Furthermore, much of the current evidence is from managed care settings in high-income countries [8, 9]. Little has been studied about continuity of care and its impact on policy-relevant outcomes (e.g., healthcare costs) in LMICs including China.
In the last decade, healthcare reform in China has focused on strengthening primary care to reshape its hospital-centric, fragmented, and inefficient delivery system [10]. Healthcare in China was primarily provided by hospitals and primary care facilities (i.e., township health centers and village clinics in rural areas and community health centers and stations in urban areas). While hospitals primarily provided acute and specialty care, primary care facilities were responsible for providing medical services as well as Basic Public Health Services including mostly preventive care [11]. In most circumstances, there were no formal gate-keeping and referral mechanisms between primary care facilities and hospitals. In 2016, a People-Centered Integrated Care (PCIC) model was introduced in a proposal regarding China’s future delivery system, a system providing the continuum of prevention, treatment, and rehabilitation services at different levels and sites of care throughout people’s life course [7, 12]. To achieve these goals, a series of policies have been issued to strengthen the workforce for primary care, realign incentives in provider payment [13], and reshape the healthcare delivery system. For instance, in 2016, China’s central government issued a policy to promote “family doctor contracting services” nationally [14]. “Family doctor contracting services” – including both medical and preventive care – were primarily provided by generalist physicians along with other healthcare providers (e.g., nurses, pharmacists) for target groups such as the elderly population and patients with chronic diseases in primary care settings. Each family doctor team can contract with up to 2000 residents [15]. By the end of 2020, there were more than 430 thousand family doctor teams serving millions of people in China [16].
However, although continuity of care is a core characteristic of PCIC, continuity of care and its association with policy-relevant outcomes (e.g., healthcare costs) remained unknown in China. Few studies have measured continuity of care in China. To our knowledge, only one study examined the association between continuity of care and health-related quality of life based on self-reported data [17]. Furthermore, findings from previous literature in other countries might not be generalizable to China for the following reasons. First, even for patients with contracted family doctors, referrals are usually not required for referring patients from primary care to hospital specialists or vice versa. This is the fundamental difference in the healthcare system between China and U.S. or U.K [18]. It is unclear whether the relationship between continuity of care and healthcare costs is still held with no formal gate-keeping and referral mechanisms in the healthcare system. Second, although about 95% of people in China were covered by social health insurance programs, out-of-pocket payments still consisted of about 28.36% of national health expenditure [19]. It is also unknown to what extent continuity of care was associated with total healthcare costs, insurance-reimbursed costs, and out-of-pocket costs, respectively.
This study aimed to examine the association of continuity of care with healthcare costs among patients with chronic disease, using social health insurance claims data in Yuhuan City, Zhejiang Province, China. Yuhuan was located on the east coast of China, with 631 thousand residents and a GDP per capita of 98094 RMB in 2019 [20]. Although Yuhuan’s GDP per capita was above the national average, its healthcare resources (e.g., hospital beds, physicians, nurses) per capita were below the national average [20]. Yuhuan established a team-based care model to enhance the management of chronic diseases. A patient with hypertension or diabetes could contract with a family doctor team to receive both preventive and curative care tailored to his or her needs. A family doctor team was multidisciplinary and usually comprised of physicians, nurses, pharmacists, and other healthcare workers who worked in a nearby primary care facility. By the end of 2020, family doctor teams had contracted with 43.57% of residents in Yuhuan. For each contracted resident, the team would be reimbursed with 120 RMB/year (1 RMB approximates 0.15 USD) on a per capita basis, and the payment was jointly financed through the patient (24 RMB), the government (48 RMB), and the health insurance program (48 RMB). Also, Yuhuan has consolidated public hospitals with primary care facilities to form an integrated care delivery system, as required by a national policy aiming at integrating healthcare services locally. Efforts have been made to facilitate information sharing between facilities and to provide technical support for providers at primary care facilities. But residents are still free to visit any hospitals and primary care facilities without a referral. In this study, we hypothesized that higher continuity of care was associated with lower healthcare costs.
This study was based on a social health insurance claims dataset collected from September 2017 to August 2019 in Yuhuan city in Zhejiang Province, China. This dataset included all 1421 patients with hypertension and/or diabetes who were managed by family doctors through contracts in 8 villages/communities. This claims dataset was initially created to evaluate the impact of a payment reform pilot program implemented in these 8 villages/communities from September 2018 to August 2019. The pilot program was designed for a cost-saving purpose, with no consideration of continuity of care in selecting the sample.
This dataset consolidated claims data of patients covered by the Urban Employee Basic Medical Insurance (UEBMI) or the Resident Basic Medical Insurance (RBMI). Like many other cities in China, these two types of social health insurance programs were available in Yuhuan, in combination covering nearly the entire population in the city. The UEBMI covered all public and private sector’s employees, including those who are self-employed, while the RBMI covered the remaining population. In general, UEBMI provided more generous coverage for its beneficiaries than RBMI. This dataset included patients’ healthcare encounters at all levels and sites of care, primarily in primary care facilities and public hospitals.
A cross-sectional study design was adopted to examine the association between continuity of care and healthcare utilization in the same year, using the claims data described above.
The study population included 1406 patients who had at least 2 outpatient encounters from September 2017 to August 2019.
We calculated four commonly used measures for interpersonal continuity of outpatient care which in this study included outpatient visits, outpatient emergency department visits, and visits to pharmacists [8, 21, 22]. Detailed formulas for the four continuous variables were presented in Table 1. All of these continuous measures range from 0 to 1, with the value close to 1 representing greater continuity of care.
Table 1
The formulas of four continuity of care measures.
| CONTINUITY MEASURES | FORMULAS |
|---|---|
| COC |
|
| HI |
|
| UPC |
|
| SECON |
|
p = total number of providers; n = total number of visits during the episode; ni = number of visits to provider i; cj = indicator of sequential visits to same providers, equal to 1 if visits j and j+1 are to the same provider, 0 otherwise.
COC indicates Bice-Boxerman Continuity of Care Index; HI, Herfindahl Index; SECON, Sequential Continuity Index; UPC, Usual Provider of Care.
As the identification for each physician was not available in this dataset, continuity was measured at the institutional level.
In addition, to approximate the continuity of primary care, one binary variable was created to indicate that a primary care physician was a patient’s usual provider of care (PCP-UPC). Here, we defined a primary care physician as a physician working in a primary care facility, including township health centers and village clinics in rural areas as well as community health centers and stations in urban areas. This variable was coded as 1 if a patient sought care from a primary care facility most often.
Two sets of outcome measures were used: (1) costs of all-cause outpatient care, including total outpatient costs, outpatient costs reimbursed by social health insurance and other public funding, and out-of-pocket outpatient costs; and (2) costs of all-cause inpatient care, including total inpatient costs, inpatient costs reimbursed by social health insurance and other public funding, and out-of-pocket inpatient costs. The original dataset was at the encounter level. To generate healthcare costs over the study period, costs of encounters were aggregated at the patient level.
Covariates in this study were age, sex, socioeconomic status (SES), the village/community, the number of outpatient encounters, and the presence of hypertension and/or diabetes. Whether a patient was covered by UEBMI or RBMI was a proxy for the patient’s SES, as patients with UEBMI were employed or self-employed and had relatively higher SES than those without employment. Whether or not the patient had hypertension and/or diabetes was ascertained by using patients’ medical records in their family doctors’ offices.
We used descriptive statistics and simple bivariate analyses to examine the association between the five continuity measures and patients’ characteristics. We estimated the association between continuity of care and outpatient costs using ordinary least squares (OLS) regression. In addition, we estimated the association between continuity of care and inpatient costs using two-part models: (1) a logistic regression with the presence of any inpatient costs as the dependent variable; and (2) an OLS regression among patients with inpatient costs. All regression models controlled for the aforementioned patient characteristics. A squared term for the number of outpatient encounters was also included in regression models.
These regression analyses were first conducted for the two-year period (from September 2017 to August 2019) with the average yearly costs as the outcomes. The regression analyses were further conducted to examine the association between continuity of care from September 2017 to August 2018 and healthcare costs from September 2018 to August 2019. These analyses aimed at eliminating the potential reverse causality which could not be addressed in a cross-sectional study design [8]. We also conducted stratified analyses by age (those aged 70 years and older vs. those younger than 70), sex (female vs. male), insurance program (UEBMI vs. RBMI), chronic disease (hypertension vs. diabetes vs. both conditions), and the number of outpatient visits (those with 20 outpatient encounters or more vs. those with fewer than 20 outpatient encounters), in order to better account for potential confounding and the heterogeneous impacts of continuity of care across subpopulations.
We reported post-estimation results for regression analyses except for the stratified analyses. For each continuity measure, we predicted the outcomes if the measure was set to 1 from the status quo. Confidence intervals were generated by the bootstrap method using 1000 replications. Regarding outpatient costs, we predicted reduced outpatient costs associated with improved continuity (i.e., if the continuity measure was set to 1 from the status quo) based on OLS models. Regarding inpatient costs, we predicted reduced hospitalization risks, unconditional costs (the net overall costs in all patients), and conditional costs (the costs conditioned on having inpatient costs) associated with improved continuity (as above) based on two-part models.
Among the 1406 patients, the yearly outpatient cost was on average 2567 RMB. During the two-year study period, 445 patients (31.65%) had been hospitalized, generating a yearly inpatient cost of 3495 RMB among all 1406 patients (unconditional) and 11042 RMB among 445 patients (conditional). The mean value of COC, HI, UPC and SECON were 0.58, 0.61, 0.71, and 0.68, while 43.03% of patients had a PCP-UPC (Table 2). Patients aged 70 years and older tended to have a higher level of continuity than those younger than 70 years of age. The continuity of patients who were covered by RBMI was also higher than those covered by UEBMI. Further, patients with ≥20 outpatient encounters had a lower continuity than other patients (Additional file 1).
Table 2
The detailed characteristics of 1406 patients in Yuhuan City between September 2017 and August 2019.
| VARIABLES (n = 1406) | N (%) |
|---|---|
| Age, mean (SD) | 65.57 (11.93) |
| Sex | |
| Female | 752 (53.49) |
| Male | 654 (46.51) |
| Medical insurance program | |
| Urban Employee Basic Medical Insurance | 140 (9.96) |
| Resident Basic Medical Insurance | 1266 (90.04) |
| Chronic diseases | |
| Hypertension | 972 (69.13) |
| Diabetes | 116 (8.25) |
| Both hypertension and diabetes | 318 (22.62) |
| Number of outpatient encounters, mean (SD) | 36.19 (32.88) |
| Village/community | |
| Village/community 1 | 140 (9.96) |
| Village/community 2 | 162 (11.52) |
| Village/community 3 | 265 (18.85) |
| Village/community 4 | 217 (15.43) |
| Village/community 5 | 85 (6.05) |
| Village/community 6 | 214 (15.22) |
| Village/community 7 | 137 (9.74) |
| Village/community 8 | 186 (13.23) |
| Total outpatient costs, mean (SD) | 2567 (5351) |
| Reimbursed | 1340 (4238) |
| Out-of-pocket | 1226 (1448) |
| Any inpatient costs, n (%) | 445 (31.65) |
| Total unconditional inpatient costs, mean (SD) | 3495 (10461) |
| Reimbursed | 1771 (6457) |
| Out-of-pocket | 1723 (5065) |
| Total conditional inpatient costs, mean (SD) | 11042 (16211) |
| Reimbursed | 5635 (10537) |
| Out-of-pocket | 5445 (7802) |
| Predictors of interest, mean (SD) | |
| COC | 0.58 (0.24) |
| HI | 0.61 (0.22) |
| UPC | 0.71 (0.19) |
| SECON | 0.68 (0.21) |
| Having a primary care provider as UPC, n (%) | 605 (43.03) |
COC indicates Bice-Boxerman Continuity of Care Index; HI, Herfindahl Index; SD, standard deviation; SECON, Sequential Continuity Index; UPC, Usual Provider of Care.
After controlling for covariates, we observed significant negative correlations between continuity of care measures and the yearly outpatient costs. For the four continuous measures of continuity of care, every 0.1-unit increase was associated with 5.87–8.88% lower total costs (151–228 RMB), 3.13–5.95% lower reimbursed costs (42–80 RMB), and 8.88%–12.54% lower out-of-pocket costs (109–154 RMB). Among the 4 indices, UPC and SECON were the most influential measures on outpatient costs. In addition, having a PCP-UPC was associated with 12.50% lower total costs (321 RMB), 1.61% lower reimbursed costs (22 RMB), and 24.42% lower out-of-pocket costs (299 RMB) compared to having no PCP-UPC (Table 3). The full regression results are presented in Additional file 2. Similar patterns were found in the direction and magnitude of associations between continuity of care and outpatient costs across subgroups by age, sex, insurance program, and chronic disease. Notably, the magnitude of associations between continuity of care and outpatient costs was much larger among patients with 20 outpatient encounters or more than those with fewer than 20 outpatient encounters. The results of stratified analyses are presented in Additional files 3–7.
Table 3
The association between continuity of care and outpatient costs between September 2017 and August 2019.
| PRIMARY PREDICTORS, COEF (95% CI) | COC | HI | UPC | SECON | PCP-UPC |
|---|---|---|---|---|---|
| Total costs | –151*** (–208, –94) |
–178*** (–240, –116) |
–228*** (–298, –158) |
–194*** (–259, –130) |
–321* (–627, –15) |
| Reimbursed costs | –42* (–82, –2) |
–52 * (–95, –9) |
–74** (–123, –25) |
–80** (–125, –35) |
–22 (–233, 190) |
| Out-of-pocket costs | –109*** (–132, –86) |
–126*** (–151, –101) |
–154*** (–182, –126) |
–115*** (–141, –88) |
–299*** (–424, –174) |
*p < 0.05, **p < 0.01, ***p < 0.001.
Ordinary least squares models adjusted for age, sex, village, medical insurance program, chronic diseases, number of total outpatient visits, and number of total outpatient visits squared.
CI indicates confidence interval; COC, Bice-Boxerman Continuity of Care Index; coef, coefficient; HI, Herfindahl Index; PCP-UPC, Having a primary care provider as the usual provider of care; SECON, Sequential Continuity Index; UPC, Usual Provider of Care.
Compared with the status quo, if optimum continuity of care measures were to be achieved (each index equals 1), 7.12–27.29% of total outpatient costs (183–700 RMB), 0.92–19.04% of reimbursed costs (12–255 RMB), and 13.91%–40.42% of out-of-pocket costs (171–496 RMB) per patient per year could be saved (Figure 1).
Saved outpatient costs when setting continuity of care to 1 compared to status quo.
COC indicates Bice-Boxerman Continuity of Care Index; HI, Herfindahl Index; PCP-UPC, Having a primary care provider as the usual provider of care; SECON, Sequential Continuity Index; UPC, Usual Provider of Care.
The first-year continuity of care measures were significantly associated with the second-year outpatient costs after controlling for covariates. For the four continuous measures of continuity of care, every 0.1-unit increase was associated with 4.49–7.23% lower total costs (134–215 RMB), 3.16–5.70% lower reimbursed costs (50–91 RMB), and 6.03–8.97% lower out-of-pocket costs (84–124 RMB). UPC had the largest impact on second-year outpatient costs. In comparison with no PCP-UPC, having a PCP-UPC was associated with a reduction of 15.04% in total costs (448 RMB), 8.23% in reimbursed costs (131 RMB), and 22.87% in out-of-pocket costs (317 RMB) (Additional file 8).
Assuming that optimum first-year continuity of care measures were to be reached, 7.47%–21.78% of total second-year outpatient costs (192–559 RMB), 4.19%–18.07% of reimbursed costs (56–242 RMB), and 11.07%–27.59% of out-of-pocket costs (136–338 RMB) per patient could be saved in comparison with the current costs (Additional file 9).
After controlling for covariates, we observed significant negative correlations between continuity of care measures and hospitalization risk, and out-of-pocket conditional inpatient costs. For the four continuous measures of continuity of care, every 0.1-unit increase was associated with a lower likelihood of hospitalization (odds ratio (OR) = 0.74–0.79) and 10.81%–11.91% (589–648 RMB) lower out-of-pocket costs among those who had been hospitalized. Moreover, having a PCP-UPC was associated with a lower likelihood of hospitalization (OR = 0.32) and 46.20% (2515 RMB) lower out-of-pocket costs, compared to having no PCP-UPC (Table 4). The full regression results are presented in Additional files 10 and 11. The magnitude of associations between continuity of care and inpatient costs was larger among women, patients aged 70 and older, patients with RBMI, patients with hypertension, and patients with fewer than 20 outpatient encounters. The results of stratified analyses are presented in Additional files 3–7.
Table 4
The association between continuity of care and inpatient costs between September 2017 and August 2019.
| PRIMARY PREDICTORS, COEF (95% CI) | COC | HI | UPC | SECON | PCP-UPC |
|---|---|---|---|---|---|
| Any cost, OR (95% CI) | 0.77*** (0.72, 0.82) |
0.74*** (0.69, 0.79) |
0.74*** (0.69, 0.80) |
0.79*** (0.74, 0.85) |
0.32*** (0.24, 0.44) |
| Total conditional costs (n = 445) | –773 (–1563, 18) |
–824 (–1657, 9) |
–830 (–1714, 53) |
–889* (–1758, –21) |
–4887* (–9146, –628) |
| Reimbursed conditional costs (n = 442) | –168 (–676, 341) |
–187 (–722, 349) |
–168 (–735, 399) |
–277 (–834, 280) |
–2286 (–5026, 454) |
| Out-of-pocket conditional costs (n = 445) | –589** (–971, –206) |
–620** (–1023, –217) |
–648** (–1076, –221) |
–598** (–1019, –177) |
–2515* (–4587, –443) |
*p < 0.05, **p < 0.01, ***p < 0.001.
Ordinary least squares models adjusted for age, sex, village, medical insurance program, chronic diseases, number of total outpatient visits, and number of total outpatient visits squared.
CI indicates confidence interval; COC, Bice-Boxerman Continuity of Care Index; coef, coefficient; HI, Herfindahl Index; OR, odds ratio; PCP-UPC, Having a primary care provider as the usual provider of care; SECON, Sequential Continuity Index; UPC, Usual Provider of Care.
When setting each continuity of care index to 1, hospitalization risk would be reduced by 11.55–18.41% (Additional file 12), while the total, reimbursed, and out-of-pocket unconditional inpatient costs per patient per year can be saved by 55.38–73.35%, 52.43–66.11%, and 55.63–80.61% (1936–2564 RMB, 929–1171 RMB, 959–1389 RMB) (Figure 2) compared to the status quo, respectively.
Saved unconditional inpatient costs when setting continuity of care to 1 compared to status quo.
COC indicates Bice-Boxerman Continuity of Care Index; HI, Herfindahl Index; PCP-UPC, Having a primary care provider as the usual provider of care; SECON, Sequential Continuity Index; UPC, Usual Provider of Care.
The first-year continuity of care measures had significant impacts on the second-year hospitalization. For the four continuous measures of continuity of care, every 0.1-unit increase was associated with a lower likelihood of hospitalization (OR = 0.91–0.93). Furthermore, having a PCP-UPC was associated with a lower likelihood of hospitalization (OR = 0.87) compared to no PCP-UPC. Nonetheless, no significant correlation was found between the first-year continuity of care measures and the second-year conditional inpatient costs per patient (Additional file 13).
Post-estimation results showed that the second-year hospitalization risk could be reduced by 0.90–4.38% and total, reimbursed, and out-of-pocket unconditional inpatient costs per patient could be saved by 8.84–40.22% (309–1406 RMB), 15.50–41.08% (274–727 RMB), and 1.74–40.91% (30–705 RMB) respectively, if the first-year continuity of care indices were to increase from the current level to the optimum level (Additional file 14 and 15).
This study is the first to report on continuity of care using claims data in mainland China. We used claims data from Yuhuan city to assess continuity of care and its association with healthcare costs among patients with hypertension and/or diabetes. Continuity measures in this study indicated relatively high levels of continuity of care among chronic disease patients in Yuhuan. Higher continuity was significantly associated with lower healthcare costs in the same year and the second year. When optimum continuity were to be achieved, 7.12–27.29% of total outpatient costs per patient and 55.38–73.35% of total inpatient costs per patient could be saved compared to the status quo during the two-year study period. If optimum continuity were to be achieved in the first year, 7.47%–21.78% of total outpatient costs and 8.84–40.22% of total inpatient costs could be saved in the second-year.
The levels of continuity of care were in the range of what was reported in managed care settings in high-income countries [27, 28]. and were higher among certain patient subgroups. Patients aged 70 years and older tended to have a higher level of continuity than their younger counterparts. As the elderly were one of the target populations of family doctor contracting services [15], the elderly might have closer relationships with their family doctors, which in turn boosted continuity of care. This might also explain why patients with RBMI had higher continuity of care than patients with UEBMI: the majority of elders were covered by RBMI as they were no longer employed. In addition, patients with ≥20 outpatient encounters had a lower continuity than other patients. These patients might have more complicated conditions which required healthcare from a diverse set of providers. However, some previous studies showed that patients with more healthcare needs valued continuity of care more and experienced greater continuity of care [29, 30]. This contrast suggested that more efforts are warranted to enhance the continuity of care among patients with complex healthcare needs.
The associations we observed in our study between continuity of care and reduced healthcare costs and decreased risks of hospitalization were consistent with the previous studies, though the contexts differed [31, 32]. Like other parts of China (except for Shenzhen and Dongguan city) [33], patients in Yuhuan have direct access to specialty care. However, gatekeeping is common in tax-funded health systems, social health insurance systems, and managed care settings in high-income countries. As gate-keeping could enhance continuity of care and reduce healthcare costs [34], it could possibly confound the relationship between continuity of care and healthcare costs. In this study, we ascertained the association between continuity of care and healthcare costs even without formal gate-keeping and referral mechanisms. Notably, our findings revealed that out-of-pocket costs were particularly sensitive to continuity of care. In China, the incidence of catastrophic health expenditure was as high as 15–20.3% [35, 36, 37]. Previous literature on catastrophic health expenditure has focused on the financial protection effects of social health insurance programs in China [38]. Our findings suggested that efforts to improve continuity of care through the delivery system reform might be another way to alleviate the risk of catastrophic health expenditure. Furthermore, stratified analyses suggested that the impacts of continuity of care on healthcare costs were even larger among vulnerable populations, such as women, older patients, and patients with lower socioeconomic status. These findings suggested that interventions to promote continuity of care might also be meaningful to improve health equity.
This study highlighted the role of primary care in providing continuous care in China. China’s primary care system faced a number of challenges to combat the growing disease burden of non-communicable chronic diseases, including the lack of well-trained human resources, fragmented health information systems, and distorting financial incentives [39]. The efficiency of primary health facilities in care provision has been questioned [40, 41]. We found that having a primary care provider as the usual source of care was associated with lower outpatient costs and hospitalization risks which in turn reduced inpatient costs. These findings suggested that for patients with direct access to specialty care, those who concentrated their care more in primary care settings might have lower costs and better outcomes. These findings underscored the potential role of primary care providers (e.g., family doctor teams) in enhancing the continuity of care as experienced by patients, particularly for patients with chronic diseases.
This study has several limitations. First, as an observational study using claims data, the reported associations might not imply causality. We cannot rule out the possibility of uncontrolled confoundings, such as patients’ socioeconomic status for which we used attendance of different types of health insurance programs as a proxy. We also conducted stratified analyses by age, sex, insurance program, chronic disease, and the number of outpatient visits to better account for potential confounding. However, due to data limitations, we were not able to adjust for comorbidities other than hypertension and diabetes. We acknowledged that patients with more healthcare needs might have lower continuity of care and higher healthcare costs. In this scenario (the confounder-exposure association and the confounder-outcome association are in opposite directions), the bias should be towards the null. Second, we were not able to compare provider-level continuity of care to institutional-level continuity of care, as the provider’s identification was unavailable. As facilities usually had more than one provider, continuity of care at the provider level should be lower than but highly correlated with continuity of care at the institutional level [42]. Third, also due to data limitations, we were not able to differentiate the purposes of outpatient encounters. By including all the outpatient encounters, instead of the outpatient encounters only related to chronic disease management, we tended to underestimate the continuity of care. Furthermore, the findings might not be generable to other parts of China and/or to the population without hypertension or diabetes. However, Yuhuan is a city with below-average healthcare resources per capita in China. Family doctor contracting services and other local efforts to enhance care integration were all required by national policies. Thus, to some extent, our findings may still be relevant to other parts of China.
In conclusion, we found relatively high continuity of care among the study population, and care continuity indicators were consistently associated with reduced healthcare costs and decreased risks of hospitalization. Future health reform should focus on improving healthcare continuity, in order to achieve a people-centered integrated care system.
The datasets generated and/or analyzed during the current study are not publicly available due to individual privacy protection but are available from the corresponding author on reasonable request.
The additional files for this article can be found as follows:
Additional file 1Table which demonstrates bivariate analyses of continuity of care and patients’ characteristics.docx. DOI: https://doi.org/10.5334/ijic.5994.s1
Additional file 2Table which presents the full regression results of the outpatient costs.docx. DOI: https://doi.org/10.5334/ijic.5994.s2
Additional file 3Table which presents the subgroup analyses of association between continuity of care measures and outpatient/inpatient costs based on age.docx. DOI: https://doi.org/10.5334/ijic.5994.s3
Additional file 4Table which presents the subgroup analyses of association between continuity of care measures and outpatient/inpatient costs based on gender.docx. DOI: https://doi.org/10.5334/ijic.5994.s4
Additional file 5Table which presents the subgroup analyses of association between continuity of care measures and outpatient/inpatient costs based on medical insurance program.docx. DOI: https://doi.org/10.5334/ijic.5994.s5
Additional file 6Table which presents the subgroup analyses of association between continuity of care measures and outpatient/inpatient costs based on disease type.docx. DOI: https://doi.org/10.5334/ijic.5994.s6
Additional file 7Table which presents the subgroup analyses of association between continuity of care measures and outpatient/inpatient costs based on number of visits.docx. DOI: https://doi.org/10.5334/ijic.5994.s7
Additional file 8Table which presents the association between the first-year continuity of care and the second-year outpatient costs.docx. DOI: https://doi.org/10.5334/ijic.5994.s8
Additional file 9Figure which demonstrates saved second-year outpatient costs when setting the first-year continuity of care to 1 compared to status quo.docx. DOI: https://doi.org/10.5334/ijic.5994.s9
Additional file 10Table which presents the full regression results of the hospitalization risk.docx. DOI: https://doi.org/10.5334/ijic.5994.s10
Additional file 11Table which presents the full regression results of the conditional inpatient costs.docx. DOI: https://doi.org/10.5334/ijic.5994.s11
Additional file 12Figure which demonstrates reduced hospitalization risk when setting continuity of care measures to 1 compared to status quo.docx. DOI: https://doi.org/10.5334/ijic.5994.s12
Additional file 13Table which presents the association between the first-year continuity of care and the second-year inpatient costs.docx. DOI: https://doi.org/10.5334/ijic.5994.s13
Additional file 14Figure which presents reduced second-year hospitalization risk when setting the first-year continuity of care to 1 compared to status quo.docx. DOI: https://doi.org/10.5334/ijic.5994.s14
Additional file 15Figure which presents saved second-year unconditional inpatient costs when setting the first-year continuity of care to 1 compared to status quo.docx. DOI: https://doi.org/10.5334/ijic.5994.s15
COC: Bice-Boxerman Continuity of Care Index; HI: Herfindahl Index; SECON: Sequential Continuity of Care Index; UPC: Usual Provider of Care; LMICs: low- and middle-income countries; NCDs: non-communicable diseases; PCIC: People-Centered Integrated Care; UEBMI: Urban Employee Basic Medical Insurance; RBMI: Resident Basic Medical Insurance; PCP-UPC: a primary care physician was a patient’s usual provider of care; SES: socioeconomic status; OLS: ordinary least squares; SD: standard deviation; CI: confidence interval; coef: coefficient; OR: odds ratio.
The institutional review board at Fudan University School of Public Health (IRB00002408 & FWA00002399) approved this study.
Xu Yi, Senior Nurse Clinician (Community Nurse), Population Health and Integrated Care Office, Singapore General Hospital, Singapore.
One anonymous reviewer.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
The authors have no competing interests to declare.
DL designed the study, performed the statistical analysis, and wrote the majority of the manuscript. WZ draw the tables and figures and wrote a part of the manuscript. YD conducted the payment reform pilot program in Yuhuan City, collected the data, and provided advice on the manuscript. YQ cleaned the data. DZ, JP, and WZ reviewed the draft and provided advice. JH administered the project, reviewed the draft, and provided advice.
Di Liang and Wenjun Zhu contributed equally to this work.
Schwarz D, Hirschhorn LR, Kim JH, Ratcliffe HL, Bitton A. Continuity in primary care: a critical but neglected component for achieving high-quality universal health coverage. BMJ global health. 2019; 4(3): e001435. DOI: https://doi.org/10.1136/bmjgh-2019-001435
Haggerty JL, Reid RJ, Freeman GK, Starfield BH, Adair CE, McKendry R. Continuity of care: a multidisciplinary review. British Medical Journal. 2003; 327(7425): 1219–21. DOI: https://doi.org/10.1136/bmj.327.7425.1219
Saultz JW, Albedaiwi W. Interpersonal continuity of care and patient satisfaction: a critical review. Annals of Family Medicine. 2004; 2(5): 445–51. DOI: https://doi.org/10.1370/afm.91
Starfield B. Is primary care essential? Lancet. 1994; 344(8930): 1129–33. DOI: https://doi.org/10.1016/S0140-6736(94)90634-3
Saultz JW, Lochner J. Interpersonal continuity of care and care outcomes: A critical review. Annals of Family Medicine. 2005; 3(2): 159–66. DOI: https://doi.org/10.1370/afm.285
van Walraven C, Oake N, Jennings A, Forster AJ. The association between continuity of care and outcomes: a systematic and critical review. Journal of Evaluation in Clinical Practice. 2010; 16(5): 947–56. DOI: https://doi.org/10.1111/j.1365-2753.2009.01235.x
World Health Organization. WHO global strategy on people-centred and integrated health services: interim report; 2015. Available from: https://apps.who.int/iris/handle/10665/155002. Accessed 3 March 2021.
Kao YH, Lin WT, Chen WH, Wu SC, Tseng TS. Continuity of outpatient care and avoidable hospitalization: a systematic review. The American journal of managed care. 2019; 25(4): e126–e34.
Espinel-Flores V, Vargas I, Samico I, Eguiguren P, Mogollón-Pérez A, López J, et al. Effectiveness of interventions related to continuity of health care in five Latin America countries. European Journal of Public Health. 2020; 30(Supplement_5). DOI: https://doi.org/10.1093/eurpub/ckaa166.513
Yip W, Fu H, Chen AT, Zhai T, Jian W, Xu R, et al. 10 years of health-care reform in China: progress and gaps in Universal Health Coverage. Lancet. 2019; 394(10204): 1192–204. DOI: https://doi.org/10.1016/S0140-6736(19)32136-1
Wei X, Yang N, Gao Y, Wong SY, Wong MC, Wang J, et al. Comparison of three models of ownership of community health centres in China: a qualitative study. Journal of health services research and policy. 2015; 20(3): 162–9. DOI: https://doi.org/10.1177/1355819615579700
World Bank Group, World Health Organization, Ministry of FinancecPRC, National Health and Family Planning Commission PRC, Ministry of Human Resources and Social Security PRC. Deepening Health Reform in China: Building High-Quality and Value-Based Service Delivery; 2016. Available from: https://openknowledge.worldbank.org/handle/10986/24720.
Ma X, Wang H, Yang L, Shi L, Liu X. Realigning the incentive system for China’s primary healthcare providers. British Medical Journal. 2019; 365: l2406. DOI: https://doi.org/10.1136/bmj.l2406
Medical Reform Office of the State Council PRC, National Health commission PRC, National Development and Reform Commission PRC, Ministry of Civil Affairs PRC, Ministry of Finance PRC, Ministry of Human Resources and Social Security PRC, et al. Notice on the issuance of guidelines on the promotion of contract services for family doctors; 2016. Available from: http://www.gov.cn/xinwen/2016-06/06/content_5079984.htm.
National Health commission PRC, National Administration of Traditional Chinese Medicine PRC. Guiding Opinions on Standardizing the Management of Contract Services for Family Doctors; 2018. Available from: http://yzs.satcm.gov.cn/zhengcewenjian/2018-10-10/8052.html.
National Health commission PRC. The National Health Commission has announced 10 key grassroots health work in 2021; 2021. Available from: https://www.cn-healthcare.com/articlewm/20210306/content-1196038.html.
Ye T, Sun X, Tang W, Miao Y, Zhang Y, Zhang L. Effect of continuity of care on health-related quality of life in adult patients with hypertension: a cohort study in China. BMC health services research. 2016; 16(1): 674. DOI: https://doi.org/10.1186/s12913-016-1673-2
Johnson JA, Stoskopf C, Shi L. Comparative health systems: a global perspective (2nd edition). Burlington: Jones & Bartlett Learning; 2017.
National Bureau of Statistics PRC. China Statistical Yearbook. Beijing: China Statistical Publishing House; 2020. Available from: http://www.stats.gov.cn/tjsj/ndsj/2020/indexch.htm.
Yuhuan Municipal Bureau of Statistics PRC. The Statistical Bulletin of Yuhuan on National Economic and Social Development 2019; 2020. Available from: http://www.yuhuan.gov.cn/art/2020/4/27/art_1229302012_3157382.html.
Hussey PS, Schneider EC, Rudin RS, Fox DS, Lai J, Pollack CE. Continuity and the costs of care for chronic disease. JAMA internal medicine. 2014; 174(5): 742–8. DOI: https://doi.org/10.1001/jamainternmed.2014.245
Saultz JW. Defining and measuring interpersonal continuity of care. Annals of Family Medicine. 2003; 1(3): 134–43. DOI: https://doi.org/10.1370/afm.23
Bice TW, Boxerman SB. A quantitative measure of continuity of care. Medical Care. 1977; 15(4): 347–9. DOI: https://doi.org/10.1097/00005650-197704000-00010
Eriksson EA, Mattsson LG. Quantitative measurement of continuity of care. Measures in use and an alternative approach. Medical Care. 1983; 21(9): 858–75. DOI: https://doi.org/10.1097/00005650-198309000-00003
Breslau N, Haug MR. Service delivery structure and continuity of care: a case study of a pediatric practice in process of reorganization. Journal of health and social behavior. 1976; 17(4): 339–52. DOI: https://doi.org/10.2307/2136712
Steinwachs DM. Measuring provider continuity in ambulatory care: an assessment of alternative approaches. Medical Care. 1979; 17(6): 551–65. DOI: https://doi.org/10.1097/00005650-197906000-00001
Barker I, Steventon A, Deeny SR. Association between continuity of care in general practice and hospital admissions for ambulatory care sensitive conditions: cross sectional study of routinely collected, person level data. British Medical Journa. 2017; 356: j84. DOI: https://doi.org/10.1136/bmj.j84
Swanson JO, Vogt V, Sundmacher L, Hagen TP, Moger TA. Continuity of care and its effect on readmissions for COPD patients: A comparative study of Norway and Germany. Health Policy. 2018; 122(7): 737–45. DOI: https://doi.org/10.1016/j.healthpol.2018.05.013
Kristjansson E, Hogg W, Dahrouge S, Tuna M, Mayo-Bruinsma L, Gebremichael G. Predictors of relational continuity in primary care: patient, provider and practice factors. BMC Fam Pract. 2013; 14: 72. DOI: https://doi.org/10.1186/1471-2296-14-72
Nutting PA, Goodwin MA, Flocke SA, Zyzanski SJ, Stange KC. Continuity of primary care: to whom does it matter and when? Annals of Family Medicine. 2003; 1(3): 149–55. DOI: https://doi.org/10.1370/afm.63
Hyer JM, Paredes AZ, Tsilimigras DI, Azap R, White S, Ejaz A, et al. Preoperative continuity of care and its relationship with cost of hepatopancreatic surgery. Surgery. 2020; 168(5): 809–15. DOI: https://doi.org/10.1016/j.surg.2020.05.020
Jung B, Cho KH, Lee DH, Kim S. The effects of continuity of care on hospital utilization in patients with knee osteoarthritis: analysis of Nationwide insurance data. BMC health services research. 2018; 18(1): 152. DOI: https://doi.org/10.1186/s12913-018-2951-y
Liang C, Mei J, Liang Y, Hu R, Li L, Kuang L. The effects of gatekeeping on the quality of primary care in Guangdong Province, China: a cross-sectional study using primary care assessment tool-adult edition. BMC family practice. 2019; 20(1): 93. DOI: https://doi.org/10.1186/s12875-019-0982-z
Velasco Garrido M, Zentner A, Busse R. The effects of gatekeeping: a systematic review of the literature. Scandinavian journal of primary health care. 2011; 29(1): 28–31. DOI: https://doi.org/10.3109/02813432.2010.537015
Sun J, Lyu S. The effect of medical insurance on catastrophic health expenditure: evidence from China. Cost effectiveness and resource allocation: C/E, 2020; 18: 10. DOI: https://doi.org/10.1186/s12962-020-00206-y
Ma M, Li Y, Wang N, Wu Q, Shan L, Jiao M, et al. Does the medical insurance system really achieved the effect of poverty alleviation for the middle-aged and elderly people in China? Characteristics of vulnerable groups and failure links. BMC Public Health. 2020; 20(1): 435. DOI: https://doi.org/10.1186/s12889-020-08554-3
Zhao Y, Oldenburg B, Mahal A, Lin Y, Tang S, Liu X. Trends and socio-economic disparities in catastrophic health expenditure and health impoverishment in China: 2010 to 2016. Tropical medicine & international health: TM & IH. 2020; 25(2): 236–47. DOI: https://doi.org/10.1111/tmi.13344
Li Y, Wu Q, Xu L, Legge D, Hao Y, Gao L, et al. Factors affecting catastrophic health expenditure and impoverishment from medical expenses in China: policy implications of universal health insurance. Bulletin of the World Health Organization. 2012; 90(9): 664–71. DOI: https://doi.org/10.2471/BLT.12.102178
Li X, Lu J, Hu S, Cheng KK, De Maeseneer J, Meng Q, et al. The primary health-care system in China. Lancet. 2017; 390(10112): 2584–94. DOI: https://doi.org/10.1016/S0140-6736(17)33109-4
Wong WCW, Jiang S, Ong JJ, Peng M, Wan E, Zhu S, et al. Bridging the Gaps Between Patients and Primary Care in China: A Nationwide Representative Survey. The Annals of Family Medicine. 2017; 15(3): 237. DOI: https://doi.org/10.1370/afm.2034
Zhang Y, Wang Q, Jiang T, Wang J. Equity and efficiency of primary health care resource allocation in mainland China. International journal for equity in health. 2018; 17(1): 140. DOI: https://doi.org/10.1186/s12939-018-0851-8
Pollack CE, Hussey PS, Rudin RS, Fox DS, Lai J, Schneider EC. Measuring Care Continuity: A Comparison of Claims-based Methods. Medical Care, 2016; 54(5): e30–4. DOI: https://doi.org/10.1097/MLR.0000000000000018