Effectiveness of Multidisciplinary Pre-Discharge Conferences on Concordance Rate in Place of End-of-Life Care and Death: A Single-Center Retrospective Study

Abstract

Backgrounds: The pre-discharge conference (PDC) is crucial for ensuring care continuity based on patients’ preferences and goals. However, there is no quantitative on its effectiveness. We investigated the effectiveness of multidisciplinary PDC on the concordance rate between the preferred (PPEoLC) and actual place of end-of-life care and death. Methods: Overall, 551 older homebound patients (median age, 83.0 years; female, 49.4%; male, 50.6%) receiving continuous home medical care through clinics were enrolled in hospital ward admission from March 2011 to September 2018. Patient demographics, presence or absence of PDCs, statements from patients and their families regarding PPEoLC, and place of death of deceased patients were confirmed from the patients’ medical records, followed by concordance rate analyses between PPEoLC and place of death and a multivariate analysis of home mortality. We used the Mann–Whiney U test to assess attribute data, hypothesis testing to assess the difference in the population proportions, and binominal logistic regression analyses to evaluate the relationship between valuables. Results: In the conference group, the home mortality rate, patients’ and their families’ expression rates, and preferences for their home of PPEoLC were higher (p < 0.001) than those in the non-conference group. The place of death was significantly influenced by family preferences. Conclusions: PDC can affect the place of death of homebound patients, but family preferences rather than patient preferences influence the decision of the patient’s place of death. To better reflect the patient preferences, patient-centered decision support should be provided earlier in the disease process.

1. Introduction

Japan’s population is aging rapidly, with 29.0% of people aged ≥65 in the year 2023 [1]. Many older people aged ≥75 years have chronic diseases, making it difficult for them to go to the hospital when their disease worsens with impaired activities of daily living. Thus, the number of patients receiving home medical care is increasing [2]. Many of these patients wish to continue living at home as long as possible and, if possible, wish to end their lives at home [2]. A certain percentage of patients receiving home medical care are repeatedly hospitalized because of disease exacerbation of their disease and so on [3]. To continue a stable life at home after hospital discharge, the continuity of care is important [4,5]. To confirm care continuity, it is necessary to make quality discharge planning during the pre-discharge period, considering life after discharge, and to share the information among multidisciplinary staff in the hospital and community [6,7,8,9,10]. Ibrahim et al. [11] reported that intervention performed by a multidisciplinary discharge coordination team was successful in improving the hospital process to overcome barriers to patient discharge. Altfeld et al. [12] reported that patients receiving the discharge planning intervention were more likely to communicate and follow-up with their physicians after discharge.

In Japan, the multidisciplinary ‘pre-discharge conference’ (PDC), as a type of discharge planning conference, has been held to bring together hospital staff and local medical and care professionals involved in post-discharge care to share discharge plans with patients and their families and in turn promote a smooth transition to the community post-discharge. Since 2006, the PDC has been eligible for additional fees from medical insurance in Japan [13], further encouraging the movement to hold the conference. This PDC has not been held for all hospitalized patients, but the physician or nurse-in-charge of the patient considered the necessity of holding it during hospitalization. In this PDC, the discharge plan is developed on the basis of the patient’s and family’s preferences, so the practice of quality decision-making support is important [14,15,16]. However, although there are a few case reports in which patients’ preferences were met, reports on the effectiveness of PDC in decision-making support for a larger number of patients are lacking.

For patients receiving home medical care, it is important to decide where to receive treatment during their end-of-life when their disease progresses. Furthermore, considering the extent to which the preferences of patients and their families regarding the place of treatment can be fulfilled is important. Therefore, PDC, as a place to jointly create and share discharge plans, is an important activity for supporting decision-making and realizing care that is consistent with the patient’s and family’s preferences. From 2011 to 2018, we operated a home medical support ward specializing in supporting patients receiving regular physician-led home visits through neighboring clinics [17] and have held PDCs for patients judged necessary by the hospital physician or nurse-in-charge. Routinely, before performing PDC, the nurse-in-charge confirms the patient’s and family’s preferences, communicates them in advance to the local staff involved in home medical care and reflects them in the discharge plan. For this reason, determining whether the PDC plays a role in supporting patients’ decision-making and whether their expressed preferences are fulfilled is important.

As mentioned above, PDC is important for ensuring continuity of care based on patients’ preferences and goals; however, there is no quantitative data on its effectiveness. Thus, the present study aimed to investigate the relationship between holding the PDC and the patient’s and family’s expression rate of their preferred place of end-of-life care (PPEoLC), along with the degree of concordance between the expressed PPEoLC and the actual place of death. In this study, we hypothesized that there was no difference between groups that underwent PDC and those that did not.

2. Materials and Methods

2.1. Study Design

We performed a single-center, hospital ward-based retrospective observational study. The study center had a hospital ward named the “home medical care support ward”, (Obu, Aichi, Japan), which only admitted patients receiving regular physician-led home visits through neighboring clinics.

2.2. Study Population

Patients aged ≥65 years and admitted to the hospital ward between March 2011 and September 2018 were consecutively enrolled because comprehensive geriatric assessment began in the ward in March 2011, and the ward closed in September 2018. The 1269 admissions made by 551 older homebound patients in the ward were confirmed. After confirming the admission frequency for each patient, information related to the patient’s and family’s preferences, and PDCs from 718 s and subsequent admissions was integrated into the data of the 551 first admissions per patient.

2.3. Routine Multidisciplinary Pre-Discharge Conference

Figure 1 shows the routine proceedings of the multidisciplinary PDC in the hospital ward. PDCs are routinely conducted within several days before discharge. Before the day of the conference, a hospital nurse-in-charge confirms the patient’s and his/her family’s preferences for treatment and care after discharge and shares the information with the local staff, including a care manager-in-charge. On the conference day, the patient’s family and multidisciplinary staff from the hospital and local community gather to discuss treatment and care plans for the patient’s life after discharge. The patient may participate if medically possible but is not necessarily required to do so. During this discussion, the preferences of the patient and his family are confirmed once again, and at the end of the discussion, the patient’s family gives consent to the plan agreed upon by all members.

2.4. Data Collection

We collected data on patient demographics, household structure, underlying disorders, comorbidity, activities of daily living (ADLs; Barthel index (BI)), body mass index (BMI), Clinical Frailty Scale (CFS), laboratory tests at hospital admission, the number of admissions, death during the study period, home death number of the deceased patients, and the number of holding PDC. Cognitive impairment in patients was defined as a dementia diagnosis based on the medical records or long-term care insurance with a level of ‘ADL of people with dementia’ > III, which is used to assess cognitive impairment in long-term care insurance in Japan as previously reported [18]. The laboratory variables included hemoglobin (Hb), total protein (TP), albumin, serum creatinine (sCr), blood urea nitrogen (BUN), and C-reactive protein (CRP).

For the extraction of patients’ and their families’ expressions of PPEoLC, the data from 1269 admissions were sorted by patient, and one research assistant searched the medical records and meeting minutes in the electronic medical records of each admitted patient and extracted the statement contents that are related to the PPEoLC made by the patients and their family members. The contents were checked by another research assistant and finally confirmed by the first author (HM).

2.5. Follow-Up

Follow-up of all selected patients was conducted throughout the study period. A prognosis survey was conducted by mail once a year for the patients and their families. In the surviving patients in each survey, the place of residence at that time was confirmed in the questionnaire. For deceased patients, the date and place of death were confirmed by the bereaved family in the questionnaire. The requirement for obtaining consent by the patients and their families was waived because of the local Institutional Review Board’s decision. The observations for each patient were terminated upon confirmation of the patient’s death or the end of the observation period.

2.6. Statistical Analyses

Categorical variables are presented as numbers and percentages and were compared by performing a chi-squared test. Continuous variables are presented as median (interquartile range (IQR)). The attribute data were analyzed using the Mann–Whitney U test. The sample size for an unpaired sample was calculated using this test [18]. With an $\mathit{\alpha }$ value of 0.05, a power of 0.8, a two-sided test, and a moderate effect size (d = 0.5) because of the absence of prior studies, we determined that 67 participants per group were required. We used a chi-square test for categorical data, hypothesis testing for differences in the population proportions for expression rate and concordance rate, and binominal logistic regression analyses for univariate and multivariate analyses. Sample sizes for these analyses were estimated using procedures similar to those used for the Mann–Whitney U test [18]. The Univariate and multivariate analyses using the binominal regression logistic model were performed as independent tests of significance. Univariate data significant at p < 0.05 were included in the multivariate analysis along with age and sex. Several variables that exhibited multicollinearity with other variables were excluded from the multivariate analysis. The odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for all significant variables in the univariate and multivariate analyses. Two-tailed p < 0.05 was considered statistically significant.

We used IBM SPSS Statistics ver. 29 (IBM, Armonk, NY, USA) and R ver. 4.2.3. (R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/ (accessed on 28 March 2023)) for the analyses. We did not use the generated artificial intelligence while writing this article.

2.7. Research Ethics

This study involved human participants and was approved by the Institutional Review Board of the National Center for Geriatrics and Gerontology (reference number No. 675; approved on 18 September 2013). The present study complied with the principles outlined in the Declaration of Helsinki. The requirement for obtaining written informed consent from each patient was waived because of the following reasons, which were in accordance with the Japanese ethical guidelines for medical and health research involving human subjects: (1) we would use clinical information obtained in routine practice on the medical record without any risk to the participants, (2) the waiver of normal consent procedures would not adversely affect participants’ rights and welfare, and (3) the Institutional Review Board of the National Centre for Geriatrics and Gerontology approved this study, including the waiver of the written informed consent.

The authors confirm that all patient and personal identifiers have been removed or disguised to ensure that the individuals described are not identifiable and cannot be identified through the details provided.

3. Results

3.1. Comparison of Baseline Patients’ Attributes between the Conference and Non-Conference Groups

During the study period, 551 older homebound patients were admitted to the ward, resulting in 1269 admissions. Among these patients, 244 patients held a PDC more than once during hospitalization (“conference group”) and 307 patients never held a PDC (“non-conference group”). In the conference group, the average number of conferences (SD) per patient was 1.2 (0.5).

Table 1. Comparison of the baseline characteristics during the first admission between patients with and without holding pre-discharge conference.

Valuables	Total	Holding Pre-Discharge Conference	No Pre-Discharge Conference	p Value
	(n = 551)	(n = 244)	(n = 307)
Age y, median (IQR)	83.0 (13)	82.0 (13) *	84.0 (12)	0.027
Female, n, (%)	272(49.4)	116 (48.5)	156 (50.0)	0.733
Charlson Comorbidity Index, median (IQR)	2.0 (2)	2.00 (2)	2.00 (2)	0.467
BI, median (IQR) (missing data = 88)	5.0 (30)	5.0 (26)	5.0 (35)	0.245
CFS, median (IQR) (missing data = 60)	8.0 (1)	8.00 (1)	8.00 (1)	0.829
Underlying diseases
Malignant disease, n (%)	155 (28.3)	59 (24.2)	96 (31.3)	0.066
Cerebrovascular disease, n (%)	89 (16.2)	42 (17.2)	47 (15.3)	0.546
Dementia, n (%)	78 (14.2)	37 (15.2)	41 (13.4)	0.545
Neurologic disease, n (%)	71 (12.9)	44 (18.0) *	27 (8.8)	<0.001
Respiratory disease, n (%)	53 (9.4)	25 (10.2)	28 (9.1)	0.656
Heart disease, n (%)	39 (7.1)	11 (4.5) *	28 (9.1)	0.036
Bone and joint disease, n (%)	25 (4.5)	7 (2.9)	18 (5.9)	0.093
Others †, n (%)	41 (7.4)	19 (7.8)	22 (7.2)	0.783
Cognitive Impairment (missing data = 104)
Impaired, n (%)	172 (38.3)	81 (33.2)	91 (29.6)	0.669
Household structure (missing data = 2)
Living alone, n (%)	61 (11.1)	24(9.9)	37(12.1)	0.412
Living with 1 person, n (%)	151 (27.5)	61 (25.1)	90 (29.4)	0.261
Living with more than 2 persons (%)	337 (61.4)	158 (65.0)	179 (58.5)	0.119
BMI, kg/m2, median (IQR) (missing data = 73)	18.3 (4.9)	17.6 (4.5) *	18.9 (5.3)	0.012
Laboratory findings, median (IQR)
Hgb, g/dL (missing data = 11)	11.4 (2.9)	11.6 (2.7) *	11.1(3.2)	0.004
TP, g/dL (missing data = 23)	6.4 (1.0)	6.4 (1.0)	6.5 (1.1)	0.718
Albumin, g/dL (missing data = 20)	3.2 (0.8)	3.1 (0.8)	3.2 (0.8)	0.366
sCr, mg/dL (missing data = 10)	0.7 (0.6)	0.7 (0.4) *	0.8(0.5)	<0.001
BUN, mg/dL (missing data = 9)	20.0 (16)	18.00 (15) *	21.0 (18)	<0.001
CRP, mg/dL (missing data = 9)	2.3 (7.3)	2.3 (7.0)	2.4 (7.6)	0.818
The number of admissions, median (IQR)	1.0 (1)	2.0 (2) *	1.0 (1)	<0.001
The number of deaths during study period, n (%)	340 (61.7)	150 (61.5)	190(61.9)	0.921
Home death number in the deceased patients and rate (%) in total death	136 (40.0)	87 (58.0) *	49 (25.8)	<0.001

* p < 0.05 compared with data of non-conference group. BI, Barthel index; CFS, clinical frailty scale; BMI, body mass index; Hgb, hemoglobin; TP, total protein; sCr, serum creatinine; BUN, blood urea nitrogen; CRP, C-reactive protein; IQR, Interquartile range. ^† Others include digestive, renal, and skin diseases.

shows baseline patient attributes between the conference and non-conference groups. A significant difference was shown in the variables of age, underlying disease of neurological and heart diseases, BMI, laboratory findings of Hgb, sCr, and BUN, number of admissions, and home death rate in the deceased patients. Particularly, the conference group showed a significantly higher home mortality rate (p < 0.001). However, no significant difference in comorbidity, ADL, severity of frailty, and family structure was observed between the two groups.

3.2. Expression Rate of Preferred Place of End-of-Life Care

As shown in

Table 2. The expression rate of preferred place of end-of-life care (PPEoLC).

	Patient (Total = 551)			Family (Total = 551)
	Conference (+), n = 244	Conference (−), n = 307	p Value	Conference (+), n = 244	Conference (−), n = 307	p Value
PPEoLC	n (%)	n (%)		n (%)	n (%)
Home	119 (48.8)	91 (29.6)	<0.001	136 (55.7)	87 (28.3)	<0.001
Outside Home	12 (4.9)	18 (5.9)		44 (18.0)	105 (34.2)
No expression or ambiguous	113 (46.3)	198 (64.5)		64 (26.2)	115 (37.5)
Expression rate (%)	53.7 *	35.5	<0.001	73.8 *	62.5	0.005

* p < 0.05 compared with data of non-conference group.

, both patients and families in the conference group had a significantly higher rate of expressing their PPEoLC than those in the non-conference group (p < 0.001 for patient, p = 0.005 for family). Regarding PPEoLC, both patients and their families in the conference group were more likely to choose home (p < 0.001) than the non-conference group.

3.3. Patient and Family Concordance Rate of Preferred Place of End-of-Life Care and Place of Death

Table 3. (a) Patient’s concordance rate of preferred place of end-of-life care (PPEoLC) and place of death. (b) Family’s concordance rate of preferred place of end-of-life care (PPEoLC) and place of death.

	Deceased Patient with Expressed PPEoLC (Total = 145)
	Death Place of Conference (+), n = 81			Death Place of Conference (−), n = 64
PPEoLC	Home	Outside Home	Concordance Rate	Home	Outside Home	Concordance Rate
Home, n (%)	47 (100)	27 (79.4)	0.67 *	20 (95.2)	35 (81.4)	0.42
Outside Home, n (%)	0 (0)	7 (20.6)		1 (4.8)	8 (18.6)
	Deceased Patients’ Family with Expressed PPEoLC (Total = 227)
	Death Place of Conference (+), n = 113			Death Place of Conference (−), n = 114
PPEoLC	Home	Outside Home	Concordance Rate	Home	Outside Home	Concordance Rate
Home, n (%)	64 (91.4)	18 (41.9)	0.79	28 (82.4)	15 (18.8)	0.82
Outside Home, n (%)	6 (8.6)	25 (58.1)		6 (17.6)	65 (81.3)

* p < 0.05 compared with data of non-conference group. PPEoLC, preferred place of end-of-life care. PPEoLC, preferred place of end-of-life care.

a,b show the concordance rate between the deceased patients’ and their families’ PPEoLC and place of death. The number of deceased patients with expressed PPEoLC was 145, whereas the number of deceased patients’ families with expressed PPEoLC was 227. The patient’s concordance rate of PPEoLC and place of death was higher in the conference group (0.67) than in the non-conference group (0.42) (p = 0.006). Contrarily, the family’s concordance rate was high even in the non-conference group (0.82) and showed no significant difference between the conference (0.79) and non-conference groups (p = 0.60).

3.4. Factors Related to Death at Home (Table 4)

In the univariate binominal logistic analysis for death at home as a dependent factor, cerebrovascular disease, cognitive impairment, BMI, BUN, C-reactive protein, family preferences, and holding PDC as independent factors showed a significant p value (p < 0.05). In the multivariate binominal logistic analysis for death at home as a dependent factor, after adjusting for age, sex, and other factors extracted from the univariate analysis, BMI, family preferences, and holding PDC showed a significant OR, and the OR (95% confidential interval; CI) was 0.904 (0.835–0.980) for BMI (p = 0.014), 1.678 (1.155–2.438) for family preferences (p = 0.007), and 2.969 (1.682–5.239) for holding PDC (p < 0.001).

Table 4. Univariate and multivariate binominal logistic regression analyses for death at home.

	Univariate Analysis		Multivariate Analysis
	OR (95% CI)	p Value	OR (95% CI)	p Value
Age	1.018 (0.992–1.045)	0.184	1.019 (0.982–1.058)	0.312
Female	0.943 (0.610–1.456)	0.790	1.058 (0.591–1.894)	0.848
Charlson Comorbidity Index	0.934 (0.815–1.069)	0.321
BI	0.995(0.985–1.004)	0.245
CFS	1.031 (0.838–1.270)	0.771
Underlying diseases
Malignant disease	1.273 (0.811–1.998)	0.294
Cerebrovascular disease	0.537 (0.289–0.997) *	0.049	1.223 (0.558–2.678)	0.615
Dementia	0.564 (0.302–1.053)	0.072
Neurologic disease	1.370 (0.620–3.024)	0.437
Respiratory disease	1.680 (0.776–3.636)	0.188
Heart disease	1.463 (0.580–3.687)	0.420
Bone and joint disease	2.771 (0.767–10.009)	0.120
Others †	0.448 (0.166–1.207)	0.112
Cognitive Impairment
Impaired	0.464 (0.283–0.761) *	0.002	0.610 (0.326–1.139)	0.121
Household structure
Living alone	0.626 (0.288–1.361)	0.237
Living with 1 person	0.955 (0.585–1.558)	0.852
Living with more than 2 persons	1.233 (0.783–1.941)	0.367
BMI	0.922 (0.885–0.983) *	0.013	0.904 (0.835–0.980) *	0.014
Laboratory findings
Hgb	1.070 (0.970–1.181)	0.176
TP	0.806 (0.616–1.053)	0.114
Albumin	0.804 (0.570–1.135)	0.215
sCr	0.831 (0.634–1.089)	0.179
BUN	0.998 (0.977–0.999) *	0.028	0.994 (0.981–1.008)	0.419
C-reactive protein	0.962 (0.927–0.999) *	0.042	0.987 (0.943–1.032)	0.561
The number of admissions	0.905 (0.799–1.025)	0.115
Patient preferences	0.849 (0.588–1.228)	0.385	0.993 (0.595–1.657)	0.979
Family preferences	1.358 (1.029–1.790) *	0.030	1.678 (1.155–2.438) *	0.007
Holding pre-discharge conference	3.700 (2.343–5.842) *	<0.001	2.969 (1.682–5.239) *	<0.001

BI, Barthel index; CFS, clinical frailty scale; BMI, body mass index; Hgb, hemoglobin; TP, total protein; sCr, serum creatinine; BUN, blood urea nitrogen; OR, odds ratio; CI, confidence interval. ^† Others include digestive, renal, and skin diseases. * p < 0.05 in univariate and multivariate binominal logistic regression analyses.

Table 4. Univariate and multivariate binominal logistic regression analyses for death at home.

	Univariate Analysis		Multivariate Analysis
	OR (95% CI)	p Value	OR (95% CI)	p Value
Age	1.018 (0.992–1.045)	0.184	1.019 (0.982–1.058)	0.312
Female	0.943 (0.610–1.456)	0.790	1.058 (0.591–1.894)	0.848
Charlson Comorbidity Index	0.934 (0.815–1.069)	0.321
BI	0.995(0.985–1.004)	0.245
CFS	1.031 (0.838–1.270)	0.771
Underlying diseases
Malignant disease	1.273 (0.811–1.998)	0.294
Cerebrovascular disease	0.537 (0.289–0.997) *	0.049	1.223 (0.558–2.678)	0.615
Dementia	0.564 (0.302–1.053)	0.072
Neurologic disease	1.370 (0.620–3.024)	0.437
Respiratory disease	1.680 (0.776–3.636)	0.188
Heart disease	1.463 (0.580–3.687)	0.420
Bone and joint disease	2.771 (0.767–10.009)	0.120
Others †	0.448 (0.166–1.207)	0.112
Cognitive Impairment
Impaired	0.464 (0.283–0.761) *	0.002	0.610 (0.326–1.139)	0.121
Household structure
Living alone	0.626 (0.288–1.361)	0.237
Living with 1 person	0.955 (0.585–1.558)	0.852
Living with more than 2 persons	1.233 (0.783–1.941)	0.367
BMI	0.922 (0.885–0.983) *	0.013	0.904 (0.835–0.980) *	0.014
Laboratory findings
Hgb	1.070 (0.970–1.181)	0.176
TP	0.806 (0.616–1.053)	0.114
Albumin	0.804 (0.570–1.135)	0.215
sCr	0.831 (0.634–1.089)	0.179
BUN	0.998 (0.977–0.999) *	0.028	0.994 (0.981–1.008)	0.419
C-reactive protein	0.962 (0.927–0.999) *	0.042	0.987 (0.943–1.032)	0.561
The number of admissions	0.905 (0.799–1.025)	0.115
Patient preferences	0.849 (0.588–1.228)	0.385	0.993 (0.595–1.657)	0.979
Family preferences	1.358 (1.029–1.790) *	0.030	1.678 (1.155–2.438) *	0.007
Holding pre-discharge conference	3.700 (2.343–5.842) *	<0.001	2.969 (1.682–5.239) *	<0.001

BI, Barthel index; CFS, clinical frailty scale; BMI, body mass index; Hgb, hemoglobin; TP, total protein; sCr, serum creatinine; BUN, blood urea nitrogen; OR, odds ratio; CI, confidence interval. ^† Others include digestive, renal, and skin diseases. * p < 0.05 in univariate and multivariate binominal logistic regression analyses.

4. Discussion

Many people receiving home medical care in Japan prefer to wait for the end of their lives at home if possible; therefore, it is important to consider how to make their wishes come true. The present study revealed that the PDC in the functional ward that supports home care continuation is likely to influence the patient’s final place of care.

The conference group showed a higher mortality rate at home than the non-conference group (Table 1). In the conference group, the rate of patients with expressed PPEoLC and the rate of those who wanted to live at home were significantly higher (Table 2) than those in the non-conference group, suggesting that the PDC contributed to the increase in the rate of patients expressing their intention for their final place of care. Additionally, in family members, the expression rate of PPEoLC and the rate of those who wanted their loved one to live at home were significantly higher in the conference group (Table 2). In holding PDCs, the nurses in charge of the ward play an important role (Figure 1). The importance of continuing nursing care from hospital to home care has been previously reported [19]. Nagae et al. conducted a review of continuing nursing care and revealed that important activities include “a team approach aimed at the realization of patient-centered care” [19]. Additionally, Ulin et al. [20] observed that a multidisciplinary approach based on the concept of person-centered care for inpatients improved the discharge process by involving patients in planning their subsequent care and recognizing them as competent participants.

In this study, the concordance rate with the place of death was higher with family preferences than with patient preferences (Table 3a,b), and the multivariate analysis for home death as a dependent variable showed the possibility that family preferences were more associated with the home death rate than the patient preferences (Table 4). In Japan, family preference, rather than patient preference, was found as a facilitating factor for home death among older non-cancer patients receiving home medical care, although this study was a retrospective observational study conducted in a primary care clinic [21]. Although the setting and research participants are different, the study results are consistent with the fact that family members play a stronger role than the patients themselves in terms of dying at home in Japan.

In the present study, the rate of cognitive impairment complication was approximately 38% (Table 1), and it is possible that many homebound patients had challenges in expressing their preferences at the time of hospital admission. To avoid the family’s decisions that are not in accordance with the patient’s preferences, it is important to confirm the patient’s preferences from himself/herself. The present study could not assess the quality of decision support provided in relation to the PDC. However, learning decision-support skills is important, and using shared decision-making (SDM) skills reportedly can improve communication with patients and their families in the case of dementia [22,23]. Goossen et al. [22,23] reported that providing SDM training to nursing home staff can improve the quality of SDM in advanced care planning (ACP). To better reflect the patients’ PPEoLC, it is necessary not only to hold the PDC but also to provide SDM training for nursing staff who play a major role in decision-making support in the ward. Additionally, ACP implementation as early as possible, in the condition before starting home medical care, could be needed.

Recently, the importance of goal-concordant care in patient-centered care has been reported [24,25]. In Western countries, autonomy as an individual has been emphasized, but in East Asian countries, including Japan, several countries respect Confucian familism [26]. In these countries, Confucian familism insists that the family as a whole, rather than any individual member, has ontological priority and that collective decisions address matters of crucial concern [26]. The results of this study can be related to Confucian familism. Yang et al. have reported that Confucian familism weakened patient–clinician SDM in EoL care [27]. In fact, in routine PDC in Japan, family members generally participate and patients are not present. In Sweden, discharge planning conferences have been routinely held in clinical practice [8,28]. Bångsbo et al. [8] reported that achieving patient participation in these conferences is challenging, as outcomes are sometimes predetermined because of the ‘back stage’ collaboration of health professionals. This contradicts person-centered care, which emphasizes patient involvement in decision-making. Efraimsson et al. [28] also reported that decisions were often made in advance, which patients expected to accept, while institutional representatives frequently justified their actions by referring to bureaucratic practices. Despite cultural differences, as recommended by the United Nations [29], the global trend is to respect the wishes of people with disabilities. To this end, it is necessary to improve SDM and ACP education for people with disabilities, such as homebound patients in Japan. Although the concept of SDM between medical professionals and patients has not been widely introduced in Japan, the authors have reported that SDM training for healthcare professionals (HCP) changed the participants’ attitude from an HCP-centered to a patient-centered one [30]. Therefore, SDM education is important in Japan as well for practicing patient-centered care.

Strengths and Limitations

In the present study, the data were obtained from a functional ward that supports patients to continue medical care at home, and many PDCs were held for continuing home medical care. Therefore, evaluating the results based on a more homogeneous PDC was possible. The decision to hold a PDC was made by the attending physician and attending nurse, but the basis for this decision was not available in this study. Furthermore, detailed cognitive function evaluations were not performed for individual patients, and it was not possible to examine the relationship between the cognitive functional level and the rate of intention expression. Moreover, this was a retrospective observational study conducted at a single institution, and there were some missing values, which may have led to a bias in the results. There was also the possibility of the existence of unknown confounding factors.

5. Conclusions

The PDC can affect the place of death of homebound patients, but family preferences rather than patient preferences can influence the decision of the patient’s death place. To better reflect the patient preferences, patient-centered decision support needs to be provided earlier in the disease process.