Article Text
PDF
Abstract
Objective This study assessed whether patients with potentially preventable emergency admissions had limited access to outpatient care immediately before admission and whether they received appropriate outpatient care during their outpatient visits.
Design Retrospective observational study.
Setting Linked outpatient and inpatient care records obtained from a nationwide claims database in Japan.
Participants Patients who experienced emergency admissions for ambulatory care-sensitive conditions between April 2005 and March 2020. Patient and regional characteristics were examined to assess the types of patients who faced difficulties with outpatient visits and receiving outpatient care related to the disease that resulted in admissions (hereafter referred to as admission-related outpatient care).
Main outcome measures (1) Whether patients had an outpatient visit during the 2 weeks preceding admission and (2) whether patients received admission-related outpatient care during the 2 weeks before admission.
Results This study included 18 449 emergency admissions for ambulatory care-sensitive conditions, representing 16.3% (18 449/113 669) of all emergency admissions in our data. Among patients with emergency admissions for ambulatory care-sensitive conditions, 37.4% did not have an outpatient visit within the 2 weeks preceding admission and 29.9% did not receive admission-related outpatient care despite having an outpatient visit. In total, 67.4% did not receive admission-related outpatient care during the 2 weeks preceding admission. Patients in their 40s and 50s were less likely to have outpatient visits and receive admission-related outpatient care before admission. No evidence associates regional characteristics with outpatient visits and receiving admission-related outpatient care before admission.
Conclusion Most patients who underwent emergency admissions for ambulatory care-sensitive conditions did not have an outpatient visit or receive admission-related outpatient care, despite having an outpatient visit immediately before admission. Our findings suggest that emergency admissions may be prevented by improving access to timely and effective outpatient care.
- Emergency Service, Hospital
- EPIDEMIOLOGIC STUDIES
- Observational Study
- Quality in health care
- Health Services Accessibility
Data availability statement
No data are available. No additional data is available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
- Emergency Service, Hospital
- EPIDEMIOLOGIC STUDIES
- Observational Study
- Quality in health care
- Health Services Accessibility
STRENGTHS AND LIMITATIONS OF THIS STUDY
We investigated how patients with emergency admissions for ambulatory care-sensitive conditions were managed in outpatient settings before admission.
We analysed a nationwide claims database from April 2005 to March 2020, with approximately 12 million enrolees in Japan.
We used multivariable logistic regression to analyse how patient and regional characteristics relate to the likelihood of having an outpatient visit and receiving admission-related outpatient care in the 2 weeks before admission.
This study used simulation analysis to compare outpatient visit patterns before emergency hospitalisation with the typical patterns.
Our study has certain limitations. In the absence of detailed clinical data in insurance claim records, we were unable to evaluate the clinical effectiveness of outpatient care for individual patients. Furthermore, our data were limited to employees of large Japanese companies and their families, excluding low-income individuals.
Introduction
Health systems in many developed countries have focused on preventing emergency hospital admissions, which are usually undesirable for patients, expensive and disrupt the provision of other healthcare services.1 A popular approach to reducing emergency hospital admissions is improving access to outpatient care,2 3 since a substantial proportion of hospital admissions may be prevented through timely and effective interventions in the outpatient setting.4–6 For example, the accountable care organisations in the USA and the government of the UK have adopted this approach to reduce the demand for emergency medicine.2 3 7 8
Despite efforts to prevent emergency admissions through outpatient care, recent evidence shows that improving access (eg, increasing open hours or reducing financial barriers) has achieved limited or mixed effects on reducing emergency admissions.9–13 This raises important questions about whether patients with potentially preventable emergency admissions face limited access to outpatient care and whether they receive appropriate outpatient care if they have an outpatient visit. Little is known about whether these patients had an outpatient visit before admission and how they were treated. Existing studies on ambulatory care-sensitive conditions (ACSC), defined as conditions for which hospitalisation could be avoided through timely and effective outpatient care,2 14 15 and found that 10–20% of emergency admissions were ACSC admissions (eg, admissions for congestive heart failure, chronic obstructive pulmonary disease and asthma).4–6 However, beyond interview research with a limited sample size,16 no study has examined whether and how patients with emergency ACSC admissions were treated immediately before admission in the outpatient settings. This lack of research makes it difficult for health policymakers to distinguish whether patients lacked an outpatient visit or received inadequate care before admission. It is important to distinguish which possibility is correct because it shows where opportunities for reduction exist and have different policy implications (ie, improving access to outpatient care in the former case may be effective, but it may not be effective in the latter case). We conducted a simulation analysis to assess whether outpatient visits before emergency admission differed from the usual pattern.
We used a nationwide claims database spanning between April 2005 and March 2020, with approximately 12 million enrolees in Japan. We examined: (1) whether patients who experienced emergency ACSC admissions had an outpatient visit in the immediate 2 weeks before admission and (2) whether those patients had received outpatient care related to the condition that resulted in their admissions (hereafter referred to as admission-related outpatient care) immediately before admission. Additionally, we investigated whether having an outpatient visit and receiving admission-related outpatient care immediately before their admission varied according to patient and regional characteristics. Japan’s public health insurance system provided an appropriate opportunity for examining whether and how patients were treated before admission. This system covers a wide array of medical services, including inpatient and outpatient services. This inpatient and outpatient claims data can be linked at the individual level, allowing us to obtain detailed information regarding the care provided to individual patients.
Methods
Health system in Japan
In Japan, public health insurance is compulsory for all residents. Employees and their dependants are covered by employment-based public insurance, while those not covered by employment-based public insurance are enrolled in community-based public insurance, except for individuals receiving public assistance. The government sets uniform benefit packages and fee schedules, irrespective of the insurance type. These benefit packages encompass various services, including inpatient and outpatient care, prescription medications and basic dental services. Additionally, patients in Japan have the freedom to choose and visit any providers, which means that there is no gatekeeper system for the first visit.
Data sources
We used a nationwide claims database from JMDC, a private company collecting and analysing anonymised administrative claims data from employee-based public health insurance for large companies in Japan. As of April 2021, this database included approximately 12 million enrolees (employees and their families), making it one of the largest claims databases in Japan. It excludes individuals aged 75 and above, as well as those on social aid because they are not covered by employee-based public health insurance. It includes information on enrolees’ basic demographics (such as gender, birth year and month and the number of family members), all healthcare services provided, and associated health expenditure. We obtained the enrolees’ basic demographic information and details of all inpatient and outpatient care services from the claims database, including dates of inpatient admissions and outpatient visits, diagnoses for inpatient stays and outpatient visits, recorded using the International Classification of Diseases, the 10th revision (ICD-10), and associated health expenditures. We linked the claims database with official statistics from the Ministry of Internal Affairs and Communications in Japan, using hospital addresses, to obtain information on regional characteristics.17
Patient population
We analysed patients who were admitted for emergency ACSC using the claims database between April 2005 and March 2020. We identified ACSC admissions based on the recorded diagnoses for admissions. Although different definitions of ACSC exist in research,14 we considered admissions for 21 conditions as ACSC admissions, following the recent literature (including admissions for congestive heart failure, chronic obstructive pulmonary disease and asthma; see online supplemental eTable 1).2 18 We examined patients who were not hospitalised 30 days before emergency ACSC admission to assess whether and how patients are treated in the outpatient settings before the emergency ACSC admission. Additionally, patients were required to have at least 1 year of enrolment in the employee-based public health insurance before the emergency ACSC admission. We excluded a small number of patients whose basic demographic information was not recorded (online supplemental eFigure 1).
Supplemental material
Outpatient care before emergency admission
Our primary concerns were (1) whether patients had an outpatient visit 2 weeks before admission and (2) whether patients received admission-related outpatient care during the same period, since some patients might not have received admission-related outpatient care even despite having an outpatient visit. Using a similar approach used for identifying ACSC admissions, we identified admission-related outpatient care based on the recorded diagnoses of outpatient care claims (online supplemental eTable 1). For instance, when patients were hospitalised for diabetes-related complications, the admission-related outpatient care was outpatient care for diabetes.
Statistical analysis
We first calculated the percentage of patients with emergency ACSC admissions who did not have an outpatient visit and did not receive admission-related outpatient care 2 weeks before admission. Moreover, we calculated the percentage of patients who had an outpatient visit and received admission-related outpatient care each day during the 2 weeks before admission. We conducted the simulation analysis as follows. We randomly selected a ‘pseudoadmission-date’ within 1 year before the emergency ACSC admission for each patient. We then estimated the percentage of patients having a daily outpatient visit during the 2 weeks before the ‘pseudo-admission-dates’ in 100 simulation trials. We then compared this simulated average pattern to the actual pattern before the genuine emergency ACSC admissions. Subsequently, we calculated the percentage of patients receiving admission-related outpatient care on each day during the 2 weeks before admission among patients who had an outpatient visit.
Moreover, we analysed whether outpatient visits and admission-related outpatient care during the 2 weeks before admission varied according to patient and regional characteristics using a multivariable logistic regression model to determine which patients face difficulty in having an outpatient visit and receiving admission-related outpatient care. Patient characteristics included age (<20, 20–29, 30–39, 40–49, 50–59 or ≥60), gender, the number of family members, breadwinner status and health expenditure in the previous year. Regional characteristics included physician density, regional average income and population density to extract data for analysing their association with outpatient visits. After fitting the logistic regression models, we calculated the predicted percentage of not having an outpatient visit and not receiving admission-related outpatient care for each patient group (such as age and gender) using the predictive margins.19 As a sensitivity analysis, we analysed whether having an outpatient visit and receiving admission-related outpatient care during 1 week before admission varied according to patient and regional characteristics.
We used Amazon Athena (Athena engine V.2) for data preparation as well as R V.4.0.5 (R Foundation) and Stata V.16.1 (Stata Corp) for statistical analyses.
Patient and public involvement
As our data were anonymised, the patients were not involved in setting the research question or the outcome measures, nor were they involved in developing plans for the design or implementation of the study. The patients did not advise on the interpretation or writing up of results. We cannot disseminate the results of our study to participants directly because our data were anonymised. We plan to disseminate the results of our study to the public through institutional press releases and ensuing news articles.
Results
Outpatient care immediately before emergency admission
Our sample included 18 449 emergency admissions for ACSC, representing 16.3% (18 449/113 669) of all emergency admissions in our data. The average age of patients with emergency ACSC admissions was 42.6 years, with 37.9% of them being female (online supplemental eTable 2). Angina, convulsions and epilepsy, congestive heart failure, pneumonia and asthma accounted for over half (52.4%) of emergency ACSC admissions (online supplemental eTable 3).
Among patients with emergency ACSC admissions, 37.4% (6918/18 449) did not have an outpatient visit 2 weeks before admission and 29.9% (5513/18 449) did not receive admission-related outpatient care despite having an outpatient visit (table 1). In total, 67.4% (12 431/18 449) of the patients did not receive admission-related outpatient care during the 2 weeks before admission.
Number of admissions and outpatient care during the 2 weeks before admission
An outpatient visit before emergency admission differed from the usual pattern. We did not find any clear difference between the actual utilisation pattern and the simulated pattern 8–14 days before admission. However, more patients received outpatient care 1–7 days before admission (figure 1), indicating that more patients sought outpatient care immediately before admission. Moreover, the proportion of patients who received admission-related outpatient care increased from 7 days before admission (online supplemental eFigure 2). Among patients who had an outpatient visit, the proportion of patients who received admission-related outpatient care gradually increased. For instance, 33.0% of patients received admission-related outpatient care 14 days before admission and 60.0% received admission-related outpatient care 1 day before admission.
Pattern of having an outpatient visit before emergency admission compared with the usual pattern generated through simulation. The red line shows the proportion of patients with an outpatient visit before admission. We simulated whether this pattern differed from the usual before emergency admissions. For each patient, we randomly generated a ‘pseudoadmission-date’ for each patient within a year before their emergency ACSC admission. We then analysed the percentage of patients with an outpatient visit in the 2 weeks before these ‘pseudoadmission-dates”’ in 100 simulation trials. The blue line shows these results, with error bars showing 95% CIs.
Patient and regional characteristics
Patient characteristics, such as age, gender and health expenditures in the previous year, were associated with the absence of an outpatient visit during 2 weeks before admission (table 2). The predicted probabilities of not having an outpatient visit before admission were 31.5% for patients aged 19 and younger, 36.4% for patients aged 20–29, 36.0% for patients aged 30–39, 40.8% for patients aged 40–49, 41.2% for patients aged 50–59 and 36.1% for patients aged 60 and older (table 2 and figure 2). Patients aged 40–49 and 50–59 were less likely to have an outpatient visit before admission compared with patients aged 20–29, and this difference was statistically significant. Moreover, predicted probabilities of not having an outpatient visit before admission were 34.3% (reference) for female patients and 39.3% (OR, 0.80; 95% CI, 0.74 to 0.87; p<0.001) for male patients (table 2 and online supplemental eFigure 3). No evidence indicating that other patient characteristics, including the number of family members and patient’s breadwinner status, as well as regional characteristics were associated with having an outpatient visit before admission (table 2).
Association between patient and regional characteristics and not having an outpatient visit during the 2 weeks before admission
Predicted probabilities of not having an outpatient visit and not receiving admission-related outpatient care during 2 weeks before admission by patient age. The blue bars represent the predicted probabilities of not having an outpatient visit. The red bars represent the predicted probabilities of not receiving admission-related outpatient care. The predicted probabilities were calculated by adjusting patient and regional characteristics, as well as hospital fixed effects. The error bars show the 95% CIs.
Patient age and health expenditures in the previous year were associated with not receiving admission-related outpatient care during the 2 weeks before admission (table 3). The predicted probabilities of not receiving admission-related outpatient care before admission were 63.6% for patients aged 19 and younger, 60.9% for patients aged 20–29, 64.1% for patients aged 30–39, 68.9% for patients aged 40–49, 69.8% for patients aged 50–59 and 69.5% for patients aged 60 and older (table 3 and figure 2). Patients aged 40 and older were less likely to receive admission-related outpatient care before admission compared with patients aged 20–29, and this difference was statistically significant. Little evidence indicating that other patient characteristics, including gender, number of family members and patient’s breadwinner status, as well as regional characteristics, were associated with receiving admission-related outpatient care before admission (table 3). Our findings were qualitatively unchanged when we assessed whether and how patients were treated 1 week rather than 2 weeks before admission (online supplemental eTables 4 and 5).
Association between patient and regional characteristics and not receiving admission-related outpatient care during the two weeks before admission
Discussion
Using the nationwide claims data from approximately 12 million enrolees in Japan, the majority of patients who underwent emergency admissions for ACSC did not have an outpatient visit or did not receive admission-related outpatient care despite having an outpatient visit immediately before admission. Our findings revealed three types of emergency admission patients, each accounting for approximately the same proportion. The first group consists of patients who did not have an outpatient visit before admission. The second group had an outpatient visit but did not receive admission-related care. The final group had an outpatient visit and received admission-related care. For the first group, improving access to outpatient care may be effective in reducing emergency admissions, although not for all cases. However, for the other groups, simply improving access to outpatient care is unlikely to reduce emergency admissions, and enhancing the effectiveness of outpatient care may be more effective in reducing admissions. This suggests that a combined approach may be necessary when considering strategies to reduce admissions. Moreover, middle-aged patients were less likely to have an outpatient visit and receive admission-related outpatient care immediately before admission. Previous literature has often focused on children and older individuals to reduce ACSC admissions.5 However, our findings suggest that middle-aged adults represent a promising target for prevention.
Our estimates indicate that approximately one in three patients did not have an outpatient visit immediately before emergency ACSC admission. While previous studies have reported an association between prior outpatient visits and factors such as regional physician density or income, our findings did not demonstrate significant evidence supporting this relationship. However, a correlation with age was observed in our data. The results concerning physician density or income may have been influenced by Japan’s universal health coverage, which keeps costs low while maintaining equity.20 21 It has been noted that middle-aged and older adults tend to delay seeking medical treatment and advice,22 23 and a similar tendency was observed in this study. Individuals in their 40s and 50s, who are at higher risk for work-related illnesses,24 may face challenges in accessing outpatient care. Given that this age group often plays a crucial role in the workforce, it is essential to consider targeted interventions for this demographic from both a medical and societal perspective.
Next, we discussed why approximately one in three patients did not receive admission-related outpatient care despite having had an outpatient visit immediately before admission. This may be influenced by Japan’s healthcare system. In Japan, there are two mechanisms for providing delayed admission-related specialty care. First, patients can consult a specialist directly of their choice. Patients often visit specialists; however, in doing so, they may sometimes choose an inappropriate specialty for their condition. Second, doctors may be unable to determine the suitability of hospitalisation for patients with medical conditions outside their own field of expertise. In such cases, it may take longer for patients to visit the appropriate specialist, potentially leading to a deterioration in their condition to the point where hospitalisation is required. It may be necessary to establish a system similar to that of the coordinators or general practitioners in other countries who assign patients to appropriate specialists or outpatient clinics.
That emergency ACSC admissions accounted for 16.3% of all emergency admissions in our data, which is consistent with previous studies reporting that 10–20% of emergency admissions were ACSC admissions.4–6 However, to the best of our knowledge, no study has examined whether and how patients with emergency ACSC admissions were treated immediately before admission in the outpatient setting, using large-scale data. We examined whether these patients had an outpatient visit immediately before admission and received admission-related outpatient care using nationwide claims data.
Our study has certain limitations. First, as with any studies on ACSC admissions, our findings may be sensitive to the definitions of ACSC. Although there are various definitions of ACSC, we selected a core set of conditions following the recent literature. Second, we were unable to examine whether the outpatient care received by individual patients was clinically effective because of the lack of detailed clinical information in the claims data. Due to the lack of clinical information, verifying the presence of unpredictable acute exacerbations was impossible. For example, while it is known that appropriate daily management of asthma can reduce emergency admissions,25 these unpredictable exacerbations are not entirely preventable and may be present in a certain number within our dataset. Third, our study shows that there is potential for preventing emergency admissions but does not assess how to realise it. Further research is warranted to understand how to improve access to outpatient care and the quality of outpatient care, focusing on patients at a high risk of experiencing potentially preventable emergency admissions. Fourth, in Japan, individuals aged 75 years and older are covered by the Late-Stage Elderly Healthcare System, so this important age group, which represents a large share of hospital admissions, was not included in our data. Finally, considering our study was focused on employees of large companies and their families in Japan, our findings may not apply to those with low incomes (such as individuals on social aid) and populations of other countries.
Based on nationwide claims data from Japan, we concluded that most patients who underwent emergency admissions for ACSC did not have an outpatient visit or did not receive admission-related outpatient care immediately before admission. Our findings suggest that it is necessary to specifically improve access for middle-aged individuals. Additionally, rather than simply increasing access, a combined approach may be required—one that establishes an effective outpatient system to ensure timely visits for admission-related outpatient care.
Data availability statement
No data are available. No additional data is available.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by the Institutional Review Board of the Graduate School of Business Administration, Keio University, on 8 July 2021. Patient consent was not necessary as the claims database from JMDC is anonymised.
Acknowledgments
We thank Takashi Tanaka for providing administrative support.
References
Footnotes
Contributors RN and HK conceived and designed the study, and they also drafted and finalised the manuscript. RN, HK, TM, TN and RG contributed to the interpretation of the data and critically revised the manuscript for essential intellectual content. TM performed the statistical analysis. HK is the guarantor of this work.
Funding JMDC Inc. provided support. The funding source played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.