If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Human Care Research Team, Tokyo Metropolitan Institute of Gerontology, Tokyo, JapanDepartment of Clinical Epidemiology and Health Economics, School of Public Health, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan
To examine the associations of 3 major hospital discharge services covered under health insurance (discharge planning, rehabilitation discharge instruction, and coordination with community care) with potentially avoidable readmissions (PARs) within 30 days in older adults after rehabilitation in acute care hospitals in Tokyo, Japan.
Retrospective cohort study using a large-scale medical claims database of all Tokyo residents aged ≥75 years.
Acute care hospitals.
Patients who underwent rehabilitation and were discharged to home (N=31,247; mean age in years ± SD, 84.1±5.7) between October 2013 and July 2014.
Main Outcome Measure
Among the patients, 883 (2.9%) experienced 30-day PAR. A multivariable logistic generalized estimating equation model (with a logit link function and binominal sampling distribution) that adjusted for patient characteristics and clustering within hospitals showed that the discharge services were not significantly associated with 30-day PAR. The odds ratios were 0.962 (95% confidence interval [CI], 0.805-1.151) for discharge planning, 1.060 (95% CI, 0.916-1.227) for rehabilitation discharge instruction, and 1.118 (95% CI, 0.817-1.529) for coordination with community care. In contrast, the odds of 30-day PAR among patients with home medical care services were 1.431 times higher than those of patients without these services (P<.001), and the odds of 30-day PAR among patients with a higher number (median or higher) of rehabilitation units were 2.031 times higher than those of patients with a lower number (below median) (P<.001). Also, the odds of 30-day PAR among patients with a higher Hospital Frailty Risk Score (median or higher) were 1.252 times higher than those of patients with a lower score (below median) (P=.001).
The insurance-covered discharge services were not associated with 30-day PAR, and the development of comprehensive transitional care programs through the integration of existing discharge services may help to reduce such readmissions.
Short-term unplanned readmissions can exert heavy clinical and economic burdens on patients, providers, and payers. A United States study reported that almost one-fifth of Medicare beneficiaries were rehospitalized within 30 days after being discharged and estimated that unplanned readmissions cost Medicare a total of $17.4 billion in 2004.
Patients may require hospital rehabilitation services for functional impairment after a major illness or injury, and studies have identified such patients to be at a higher risk of short-term readmissions.
In the United States, potentially avoidable readmissions (PARs) to acute care hospitals represent an important quality indicator, and financial incentives have been introduced to reduce readmission rates.
examined the factors associated with 30-day PAR after discharge from inpatient rehabilitation facilities, but few studies have analyzed the efficacy of existing interventions in preventing readmissions. The assessment of such interventions may help to identify areas of improvement and enable the formulation of more effective preventive strategies.
Transitional care services, including discharge planning and follow-up, are increasingly used to prevent unnecessary readmissions.
The current health care payment system used by the majority of acute care hospitals in Japan is based on the diagnosis procedure combination (DPC) system, which enables case-mix adjustments for reimbursements. Because Japanese acute care hospitals, including DPC hospitals, generally provide patients with both acute care and postacute care (such as rehabilitation services) during the same hospitalization episode, the length of stay (LOS) in these hospitals tends to be longer than in acute care hospitals in the United States.
Although empirical analyses have indicated that mean LOS did decrease from 20.4 days in 2002 to 13.7 days in 2014, the 6-week readmission rates (including planned readmissions) increased substantially from 4.7% in 2002 to 11.2% in 2014.
In this way, patients in DPC hospitals may be more susceptible to PAR in Japan. To facilitate the development and improvement of interventions for reducing short-term readmissions after rehabilitation in DPC hospitals, it is important to first understand if current discharge services at these hospitals are associated with such readmissions.
To the best of our knowledge, no previous studies have been conducted on this topic.
This study aimed to examine the associations of major discharge services covered under health insurance with 30-day PAR among older adults who received rehabilitation while admitted at acute care hospitals in Tokyo, Japan. Based on prior studies on discharge services and transitional care services,
we tested the hypothesis that major discharge services covered under health insurance are associated with reduced 30-day PAR.
Database and study sample
This retrospective cohort study was conducted using a large-scale, anonymized medical claims database obtained from the Tokyo Extended Association of Medical Care for Latter-Stage Older People, which manages the health insurance program for Tokyo residents aged 75 years or older.
Accordingly, this database comprised data from all citizens aged 75 years or older living in Tokyo, Japan. The data included patient-level sociodemographic characteristics, treatments, medical facilities used, drugs prescribed, and diagnoses made during clinical encounters for the purpose of insurance claims. Diagnoses were recorded as International Classification of Diseases–10th Revision (ICD-10) codes. We obtained data from approximately 1.35 million patients aged 75 years or older who had received outpatient care, inpatient care, or home medical care from a medical institution between September 1, 2013 and August 31, 2014.
For this study, we focused on patients who had been admitted to a DPC hospital between October 1, 2013 and July 28, 2014, and had been discharged between October 4, 2013 and July 31, 2014. Patients who underwent rehabilitation during this index admission were identified and included in analysis. If patients were admitted to a DPC hospital twice or more during the study period, we defined the index admission as the first hospitalization episode with rehabilitation at a DPC hospital. As patients discharged from DPC hospitals to long-term care facilities would have more stable conditions and are unlikely to be readmitted, we excluded these patients from analysis. Patients younger than 75 years were also excluded from analysis. Moreover, patients with a short LOS (≤3d) were also excluded because they are unlikely to use discharge planning services due to lower barriers to discharge.
Outcome: 30-day PAR
The outcome was the occurrence of 30-day PAR following rehabilitation in DPC hospitals. The causes of readmissions were identified using the reported ICD-10 codes.
We defined 30-day PAR as the first unplanned readmission within 30 days after discharge to a DPC hospital due to any of the following 15 medical conditions: respiratory infection; congestive heart failure; urinary tract infection; fracture; electrolyte imbalance; constipation, fecal impaction, and obstipation; skin ulcers and cellulitis; chronic obstructive pulmonary disease and asthma; seizures; weight loss and malnutrition; anemia; diabetes; hypertension; acute renal failure; and gastroenteritis. These medical conditions were selected based on a prior study on potentially avoidable hospitalizations among residents of long-term care facilities in Japan.
Diarrhea, Clostridium difficile infections, sepsis, and altered mental status were excluded as reasons for PAR because we determined that these conditions are neither preventable nor manageable in the home care setting, as indicated in previous studies.
Independent variables of interest: discharge services
We identified the following 3 types of hospital services that are provided at discharge and covered under Japan’s health insurance system from fiscal year 2013 to fiscal year 2014: (1) discharge planning; (2) rehabilitation discharge instruction; and (3) coordination with community care.
Discharge planning aims to reduce barriers to discharge to home, institutions, or other hospitals. Nurses and medical social workers in discharge planning departments promptly identify patients with potential barriers to discharge due to medical, physical, or environmental problems; formulate an optimized discharge plan for each patient; and provide the necessary support for safe discharge. Moreover, discharge planning involves arrangements with home care agencies, institutions, or other hospitals to ensure that the patients’ and their families’ postdischarge needs are met.
Rehabilitation discharge instruction is designed to educate patients and their families on self-management to prevent functional impairment and improve independence in activities of daily living. Physical therapists, occupational therapists, and speech therapists coordinate with doctors, nurses, and medical social workers to provide this service before discharge. For example, physical therapists provide instruction on safe exercise methods and activities of daily living to patients and their families.
Coordination with community care can be subcategorized into 2 types of services: “instructions for community care at discharge” and “coordination with long-term care.” Instructions for community care at discharge promote coordination between hospital staff and community care staff to provide postdischarge instructions (such as instructions for medication safety at home) to patients and their families. Hospital doctors and nurses consult with community care doctors and nurses to share information on patient medications and activities of daily living, and hospital staff provide summaries of these consultations to community care workers.
Coordination with long-term care aims to promote coordination between hospital workers and community care managers to assess the need for formal long-term care services and formulate a long-term care plan for each patient according to their medical, physical, or environmental problems. Hospital staff (such as nurses and community care managers) may provide this service once or twice before discharge. When long-term care is determined to be necessary, staff assess the appropriate care settings that can meet both patients' and their families' needs after discharge, and provide this assessment to a care manager. In addition, staff may also assess the care needs of patients just before discharge and provide this assessment to a care manager. The care manager then formulates a postdischarge long-term care plan based on this care needs assessment.
The provision or nonprovision of each discharge service was identified using the medical claims data for the index admission. Each discharge service can be provided alone or in combination with the other services to a single patient.
We selected other variables available in medical claims data that were previously identified to be associated with providing discharge services or 30-day PAR or had been used in similar risk-adjustment models in other studies.
We also collected information from the database on other variables, such as patient sex, age groups (75-79, 80-84, 85-89, and ≥90y), and insurance copayment rate (10% for residents with a taxable income<$14,078/y and 30% for residents with a taxable income≥$14,078/y).
we counted the number of 22 chronic diseases that are common in older adults using ICD-10 codes and prescriptions of relevant therapeutic agents. The number of chronic comorbidities was divided into 4 categories (0-1, 2-3, 4-5, and ≥6 diseases) for analysis. Also, the use or nonuse of home medical care services before the index admission to a DPC hospital was identified through relevant records in the claims data. We calculated the LOS and the mean number of rehabilitation service units used per day for each patient during the index admission. A rehabilitation service unit consisted of 20 minutes of rehabilitation by a physical therapist, occupational therapist, or speech therapist. LOS and rehabilitation service units were each divided into 2 categories (LOS: <20 and ≥20d; rehabilitation service units: <0.66 and ≥0.66 units per day of hospital stay) based on their median values. Patient frailty was measured with the Hospital Frailty Risk Score (HFRS), which was developed to screen for frail patients in the United Kingdom using ICD-10 codes in claims databases.
The HFRS is an aggregate of 109 conditions that are known to be associated with frail patients, and the score ranges from 0-99. Patients with higher scores (ie, frail patients) were found to be associated with higher mortality, longer hospitalizations, and higher emergency readmission rates.
these cut-off points have yet to be validated in the Japanese health care system. We therefore used the median score to dichotomize the cohort for this analysis (<1.8 and ≥1.8). We also identified 8 categories of primary diagnoses in the index admission using the corresponding ICD-10 codes for the main causes of hospitalization, and included these as covariates in the analyses. In addition to 6 diagnostic categories of hospitalization causes identified in a previous report,
we also included diagnoses of fractures and musculoskeletal diseases that require rehabilitation services.
First, the chi-square test was used to compare the characteristics between patients with and without 30-day PAR. We then examined the association of each discharge service with 30-day PAR using a logistic regression model fitted with a generalized estimating equation (GEE) that adjusted for the covariates, other discharge services, and clustering of patients within hospitals.
The multivariable logistic GEE model used a logit link function and binomial sampling distribution. P values (2-tailed) below .05 were considered statistically significant. The odds ratios (ORs) and their 95% confidence intervals (CIs) were adjusted for the covariates, other discharge services, and clustering of patients within hospitals. All analyses were conducted using SPSS version 23.0.a
The study protocol was approved by the Ethics Committee of the Tokyo Metropolitan Geriatric Hospital and Institute of Gerontology. We performed all procedures in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects established by the Japanese government.
Figure 1 shows the flow chart of patient selection. We first identified 33,221 candidate subjects who had received rehabilitation at a DPC hospital during the study period. We excluded 1513 patients who were discharged to long-term care facilities, 230 patients aged below 75 years, and 231 patients with short LOS (≤3d). The final sample for analysis comprised 31,247 patients.
The characteristics of the patients are summarized in table 2. The overall mean age in years ± SD was 84.1±5.7, and 12,363 of the patients (39.6%) were men. A total of 883 patients (2.9%) were identified as having 30-day PAR according to our identification criteria.
Table 2Comparison of characteristics (percentages) between patients with and without PARs within 30 days (N=31,247)
Table 2 also shows the comparison of characteristics between patients with and without 30-day PAR. Significantly more patients with 30-day PAR received coordination with community care than patients without 30-day PAR (P=.008). In contrast, there were no significant differences in the use of discharge planning (P=.228) and rehabilitation discharge instruction (P=.325) between the groups. Moreover, patients with 30-day PAR had significantly higher use of home medical care before the index admission (P<.001), longer LOS (P=.775), higher mean number of rehabilitation service units (P<.001), and higher HFRS (P<.001) than patients without 30-day PAR. Table 3 presents the causes of 30-day PAR. The most common causes were respiratory infection (259 patients; 29.3%).
Table 3Breakdown of potentially avoidable readmission within 30 days (n=883)
Table 4 shows the associations of the 3 discharge services with 30-day PAR after adjusting for the covariates and patient clustering within hospitals. None of the discharge services were significantly associated with 30-day PAR: The ORs were 0.962 (95% CI, 0.805-1.151) for discharge planning, 1.060 (95% CI, 0.916-1.227) for rehabilitation discharge instruction, and 1.118 (95% CI, 0.817-1.529) for coordination with community care. In contrast, the odds of 30-day PAR among patients with home medical care services were 1.431 times higher than the odds among patients without home medical care services (P<.001), and the odds of 30-day PAR among patients with a higher number (median or higher) of rehabilitation units were 2.031 times higher than the odds among patients with a lower number (below median) (P<.001). In addition, the odds of 30-day PAR among patients with a higher HFRS (median or higher) were 1.252 times higher than the odds among patients with a lower score (below median) (P=.001).
Table 4Associations of discharge services with PARs within 30 days (N=31,247)
This retrospective cohort study examined the associations between 3 major discharge services covered under Japanese health insurance and 30-day PAR among older adults after rehabilitation in an acute care setting using a large-scale claims database that included all Tokyo residents aged 75 years or older. The 30-day PAR rate was 2.9% in DPC hospitals. Discharge planning, rehabilitation discharge instruction, and coordination with community care were not associated with 30-day PAR after adjusting for variations in patient characteristics. This indicates that these services may have little or no effect on preventing PAR in DPC hospitals.
In contrast to our findings, a previous meta-analysis of randomized controlled trials reported that discharge planning can reduce readmission rates and LOS.
In Japan, discharge planning is performed by nurses or medical social workers belonging to a specialized discharge planning department. Therefore, the lack of association between discharge planning and PAR in our analysis may be influenced by staff-related factors. It may be important to determine if these staff have the necessary time, training, and resources to formulate discharge plans that can effectively prevent 30-day PAR. Moreover, discharge planning is provided to patients with potential barriers to discharge in order to decrease LOS, which is incentivized by the health insurance system. Therefore, insurance-covered discharge planning may be more focused on reducing LOS than preventing short-term readmissions.
Rehabilitation discharge instruction and coordination with community care were also not associated with 30-day PAR in our subjects. Although previous studies have indicated that improving patient self-care and coordination with other providers can reduce 30-day readmission rates in heart failure patients,
few studies have examined the effects of these services on readmissions among older patients using rehabilitation in acute care hospitals. It has also been reported that multicomponent interventions are more effective than single-component interventions in reducing short-term readmissions.
conducted a systematic evidence review to examine if transitional care interventions reduced readmission rates among older adults with chronic diseases after being discharged to home. Their review indicated that transitional care (including discharge planning, education on self-management, follow-up, and coordination among health care providers) can reduce readmission, mortality, and acute care resource utilization.
Although Japanese hospitals do not provide insurance-covered transitional care programs with multicomponent interventions, such programs may be developed by integrating current services and including a follow-up system. Postdischarge follow-up activities under US transitional care programs include home visits and telephone calls from nurses,
but resource and workforce limitations may make it difficult for nurses in Japan to conduct these activities. Older adults with functional impairment can receive medical care (eg, nursing care and rehabilitation) at low cost under the medical insurance system or long-term care insurance system in Japan.
Because many older adults who use rehabilitation services during hospitalization also use medical care after discharge, it is important to promote continuity between hospital care and community-based care to prevent readmissions.
However, the insurance-covered coordination with community care services were not associated with 30-day PAR in this study. As these services are currently provided only during hospitalization and not after discharge, there may be inadequate communication and information sharing between hospital staff and community care staff or long-term care staff to ensure postdischarge continuity of care. Japan’s medical and long-term care insurance systems should be modified to promote communication and information sharing between hospital staff and community care staff or long-term care staff after discharge.
The finding that a higher frailty score is associated with 30-day PAR is consistent with previous studies on the association between frailty and all-cause emergency readmission.
this risk score may be applied to Japanese medical claims data to efficiently screen for patients at higher risk of 30-day readmission. Also, patients who received home medical care services before admission and underwent more intensive rehabilitation were more likely to have 30-day PAR. Under Japan’s health insurance system, patients must be physically unable to travel to receive nonemergency outpatient care in order to be eligible for the use of home medical care services. Accordingly, these patients have a higher risk of severe functional impairment. Similarly, patients with severe functional impairment may have a greater need for intensive rehabilitation than other patients. Functional impairment in older patients is associated with an elevated risk of readmissions after rehabilitation during hospitalization,
and it is therefore important to consider functional status when developing PAR prediction models.
Highly sensitive prediction models that are easy to use may provide an efficient method for identifying patients at high risk of 30-day PAR. Comprehensive geriatric assessments involve the systematic evaluation of frail older adults by teams of health professionals and typically include assessments of medical, psychiatric, functional, and social domains in order to increase a patient’s likelihood of survival and living in their own homes after an emergency hospitalization.
Therefore, data from easy-to-use, comprehensive, geriatric assessments may be linked with medical claims data to develop more accurate prediction models for short-term PAR.
A strength of this study is the use of claims data from all patients aged 75 years or older who received rehabilitation services in DPC hospitals in Tokyo, Japan. These findings are therefore representative of the population in this age group residing in Tokyo. Another strength is the quantification of 30-day PAR among older adults who received rehabilitation in acute care hospitals using a large-scale database. Furthermore, this study is the first in Japan to assess the risk of 30-day PAR to DPC hospitals among patients discharged from DPC hospitals. In contrast, previous studies using DPC databases have only assessed unplanned readmissions to the same hospital as the index admission.
This study has several limitations. First, our database did not include information on patients who were readmitted to non-DPC hospitals. DPC hospitals accounted for approximately 63% of all acute care beds in Japan in 2014, which may have led to an underestimation of 30-day PAR.
Second, this study was unable to exclude patients who died after discharge as claims data lack this information. This may have led to an underestimation of 30-day PAR. Third, we could not account for variations in disease burden due to the lack of disease severity information in our database. Because disease severity can directly affect treatment approaches and readmission rates, this variable should be considered in future analyses. Fourth, our findings may not be directly generalizable to other countries due to inherent differences in health care systems, although some predictors (such as longer index admissions) may have similar effects on PAR.
Fifth, the optimal cut-off points for the HFRS have yet to be identified in the Japanese health care system, and further analyses are required to understand the implications of this factor. Finally, we could not ascertain the quality of discharge services in each DPC hospital. However, we attempted to account for these variations by adjusting for the clustering of patients within hospitals using GEEs.
The major insurance-covered discharge services were not associated with 30-day PAR in older adults after rehabilitation in acute care hospitals in Tokyo, Japan. The development of comprehensive transitional care programs (involving components of discharge planning, education on self-management, follow-up, and coordination among health care providers) through the integration of existing insurance-covered discharge services may help to reduce 30-day PAR.
SPSS version 23; IBM Corp.
Rehospitalizations among patients in the Medicare fee-for-service program.