This retrospective study examined total care, healthcare, and medical care facilities for the elderly, as defined by Japan’s long-term care insurance system. Total care facilities provide care for elderly people who will require permanent care, and for whom providing in-home care would be complicated by significant physical or mental decline. In contrast, healthcare facilities provide rehabilitation care after acute medical care, which is designed to enable residents to return home.

We randomly selected 6920 facilities from the LTCFs registered with the Welfare, Health, and Medical Care Information Network of the Welfare and Medical Service Agency. A questionnaire was mailed to the nurse managers or facility officers at the 1402 facilities that agreed to participate in this study.

We collected information on long-term care services at the facilities during the fiscal year 2012-2013. Based on the Donabedian framework, the questionnaire was composed of three parts: structural indicators, process indicators, and outcome indicators [11]. This study’s structural indicators consisted of 12 items, including facility organization, personnel allocation, and standard care system characteristics of the facility. Process indicators consisted of 26 items, including the frequency with which risk assessments were conducted, conference enforcement status, and patient safety strategy implementation. Outcome indicators consisted of 7 items, including improving or maintaining the required level of care, improving or maintaining independence in daily living, improving or maintaining cognition in daily living, and improving or maintaining levels of incontinence in the year preceding data collection. Adverse events were documented as the frequency (per 100 residents in one facility) of falls or behavioral problems, such as harm to self or others, aimless wandering, or going out without permission. The definition and formula used to calculate the outcome indicators are summarized in Table 1.

The facility manager or nursing manager completed the questionnaire. We asked the facility manager for data on the 12 structural indices of the LTCFs and

the nursing managers for data on process and outcome indicators. We asked whether care was provided for the 26 process indicators.

We calculated seven outcomes per 100 residents for each facility during the year preceding data collection as follows: maintaining or improving the required care level, independence in daily living, cognitive function in daily living, incontinence, the occurrence of tumbles and falls, behavioral problems such as harm to self or others, and the occurrence of aimless wandering or going out without permission. We compared the seven outcomes for each facility type using chi-square tests and Mann-Whitney U tests. We constructed multivariate linear regression models using a stepwise selection method to examine the relationship of each of the seven outcomes to the 12 structural indicators and the 26 process indicators.

Facility factors’ influences were analyzed after statistical adjustment for facility type. Analyses were carried out using SPSS ver.24.0 for Windows, with the threshold of significance set at p < 0.05.

This study was conducted with the approval of the ethics committee of the Graduate School of Nursing of Nagoya City University (12024-2). Survey respondents were provided with a written explanation of the purpose of the study, informed that participation was voluntary, and advised that the data from all participating facilities would be published collectively, so that individual facilities would not be identifiable. Returning a completed questionnaire was considered to represent implied consent to participate in this study.

Completed questionnaires were returned from 1402 facilities, with a response rate of 20.3%. Questionnaires with more than an 80% completion rate regarding the seven outcome indicators were accepted as valid responses; based on this criterion, 1067 facilities (76.1%) provided valid responses. Of these, 541 (50.7%) were total care facilities for the elderly, 324 (30.4%) were healthcare facilities for the elderly, and 202 (18.9%) were medical care facilities for the elderly. Table 2 lists the characteristics of participating facilities.

The average number of residents per facility was 75.3, the average age of residents was 87.8 years, and 74.3% were women. Japan’s long-term care insurance system has five levels of classification which are used to indicate the level of care required by residents. Based on this classification system, an average of 4.2% of facility residents were Level 1 residents who required partial care for some aspects of ADL; 16.4% were Level 2 residents and required a low level of care; 24.2% were Level 3 residents and required a moderate level of care; 26.8% were

Level 4 residents and required a high level of care; and 28.4% were Level 5 residents and required the highest level of care.

In one year, 86.5% of the total care facilities, 83.3% of the healthcare facilities, and 91.1% of the medical care facilities maintained or improved their levels of care needs for their residents. Independence in daily living was maintained or improved in 86.0%, 88.5%, and 89.6% of the facilities, respectively. Cognitive levels in daily living were maintained or improved in 91.4%, 88.1%, and 96.2% of the facilities, respectively. There were significant differences between these indicators by facility type. Instances of incontinence were maintained or improved in 91.5%, 91.1%, and 92.5% of the facilities, respectively, and there were no significant differences among the facility types.

Adverse events, such as the occurrence of falls, were measured over a one-year period; per100 residents, there were 120.2 reports of falls per a facility in total care facilities, 163.3 in healthcare facilities, and 47.0 in medical care facilities. The rate of behavioral problems was reported as 24.1, 11.7, and 13.3 per 100 residents, respectively. The rate of incidents of aimless wandering and going out without permission was reported as 11.6, 7.0, and 5.3 per 100 residents, respectively. There were significant differences among the facility types for all three indicators.

Table 3 displays structural and process indicators by facility type, number of care staff on-site, and the standard care systems in place to ensure patient safety. During the day, the average number of residents per care staff member was 2.5 in total care facilities, 3.7 in healthcare facilities, and 3.8 in medical care facilities. During the day, the average number of residents per nurse was 18.4, 9.3, and 6.9, respectively.

More than 80% of the care staff in all facility groups had processes of: persuading incontinent residents to use training pants or pads, allocating a staff member to watch over residents while they were bathing, and paying attention to how staff members positioned wheelchairs and bed stoppers.

Fewer than 50% of the care staff in all facility group sengaged in the processes of: understanding the resident’s cognitive condition using the Hospital Anxiety and Depression Scale (HADS) and the Mini-Mental State Examination test (MMSE) every three months; routinely modifying clothes to accommodate an individual’s disabilities; using protective devices (e.g., hip protectors and cushioning mats) for residents at high risk of falls; and occasionally restricting the range of activities for residents who were likely to wander aimlessly.

Table 4 lists the associations between process indicators and outcome indicators. Maintained or improved levels of long-term care needs were not related to structural or process indicators. Maintained or improved levels of independence in daily living were related to measuring ADL using indexes of FIM (functional independence measure) and BI (Barthel index) every three months (β = 0.20, p < 0.001), holding care conferences regarding the maintenance and enhancement of ADL and cognitive condition every three months (β = 0.18, p < 0.001), and providing individual residents who can go out with opportunities to have a walk or an outing (β = 0.09, p = 0.004). Maintained or improved cognitive levels in daily living showed a relationship to measuring ADL using indexes of FIM and BI every three months (β = 0.14, p < 0.001), measuring mental health by using HADS and MMSE every three months (β = 0.15, p < 0.001), holding care conferences for the maintenance and enhancement of ADL and cognitive condition every three months (β = 0.09, p < 0.001), and reviewing oral medicines (e.g., sleeping tablets and psychotropic drugs) used by residents at high risk of falls (β

= 0.10, p = 0.001). Maintained or improved management of incontinence was related to measuring ADL using indexes of FIM and BI every three months (β = 0.11, p < 0.001), keeping an incontinence diary for each resident (β = 0.14, p < 0.001), and holding care conferences for behavior enhancement of incontinence problems every three months (β = 0.06, p = 0.002).

The three adverse events—the number of falls, behavioral problems, and aimless wandering or going out without permission—all showed a similar relationship with the number of residents per care staff member (β = 0.19, β = 0.16, β = 0.20, respectively, p < 0.001). Additionally, the three adverse events were positively related to factors that restricted residents’ behavior: occasionally using a waist belt or a wheelchair table to prevent residents from slipping off wheelchairs, putting mittens or gloves on residents to prevent them from removing tubes and damaging their skin, and occasionally restricting the range of activities of residents who wander aimlessly.