Unfinished Business in Bereaved Families of Patients with Advanced Cancer: Pre–Post QI Evaluation

Design and participants

We conducted pre- and post-QI anonymous bereaved family surveys at Seirei Hospice. The pre-QI survey (October–November 2020) included families of cancer patients who died between April 2018 and March 2020. ⁷ The post-QI cohort included bereaved families of patients who died at Seirei Hospice between November 2022 and January 2024. Eligibility criteria were identical across periods. In brief, the inclusion criteria were patient’s death from cancer and both patient and respondent aged ≥20 years. Exclusion criteria included inability to identify or contact family members, treatment-related deaths, cognitive or visual impairment, or severe distress anticipated. Questionnaires were mailed, and return of the questionnaire by bereaved family members constituted implied consent to participate. The study protocol was approved by the Institutional Review Board of Seirei Mikatahara General Hospital (Nos. 20-31 and 24-03).

QI project

The QI initiative consisted of a low-burden UB-focused approach integrated into usual care. A one-page pamphlet was distributed at admission, inviting families to reflect on “things they would like to do” and “things they would like to say” during the limited time remaining (Supplementary Data). The pamphlet summarized common bereaved family reflections and noted that approximately 60% of families in our prior survey reported distress related to unfinished matters. ⁷ It described how clinicians could help by (1) providing clear explanations about the anticipated clinical course and (2) facilitating communication between patients and families. Although the pamphlet included checkbox-style items, these were intended as prompts for reflection rather than for data collection, and responses were not systematically collected.

Clinicians were encouraged to maintain attention to prognostic awareness, emotional preparation, and UB-related needs throughout admission. The intervention did not mandate specific scripts or protocols; rather, clinicians responded flexibly according to each patient’s and family’s readiness. The QI initiative involved multidisciplinary hospice staff, including physicians and nurses; prior to implementation, pre-QI findings were shared and brief orientation sessions were conducted on how to use the pamphlet and support UB-related needs within usual care. During the implementation phase (July–December 2022), the pamphlet was provided to families of 126 among 178 admitted patients (71%), some of whom may have been included in the post-QI cohort; however, individual-level exposure to the pamphlet was not assessed in the bereaved family surveys.

Data collection

Both pre- and post-QI surveys collected basic background characteristics and measures on UB and related distress. Survey items were developed based on prior studies, with face validity assessed in a pilot test involving 11 bereaved individuals with hospice experience.^4,6,8–11

Measures of UB

An introductory statement explained that people who had lost an important person sometimes feel something was left “unsaid or unfinished.” UB was assessed using several measures. ⁵

Unfinished Business in Bereavement Scale, brief version

We used the brief Japanese version of Unfinished Business in Bereavement Scale (UBBS), originally developed in the United States, to assess UB among bereaved family members. ¹⁰ The Japanese version was created using the forward/backward translation method and used in our prior survey. ¹² The brief UBBS comprises two 4-item subscales: “Unfulfilled wishes” and “Unresolved conflict.” Participants rate the level of distress associated with each item on a 5-point scale (1 = not at all distressed to 5 = extremely distressed), with higher scores indicating more UB. In the present study, we used only the “Unfulfilled wishes” subscale because of its feasibility and clinical relevance in the PCU setting. ⁶

One-item UB scale and UB-related distress.

Participants were asked: “Do you on the whole have any ‘unfinished business’ concerning the final few weeks you spent with the patient?” Responses were recorded on a 7-point Likert-type scale (1 = absolutely disagree to 7 = absolutely agree). To assess UB-related distress, participants were also asked: “How distressed have you been about this issue in the past month?” This was rated on a 7-point Likert-type scale (1 = not distressed at all to 7 = extremely distressed). These items were adapted from previous UB studies, with minor modifications to response options to improve clarity^3,5,6,13 (Table 2).

OPEN IN VIEWER

Table 2.

Comparison of Unfinished Business-Related Outcomes across Pre- and Post-Quality Improvement Periods

Variables	Pre-QI period (n = 319)	Post-QI period (n = 216)	ES	p*
	Mean (SD)	Mean (SD)
UBBS, brief version (unfulfilled wishes)
Total score	12.46 (4.16)	11.24 (4.23)	0.29	0.001
Each item score
I wish we did more things together	3.19 (1.18)	2.91 (1.22)	0.24	0.007
I should have told him/her “I love you” more often	3.09 (1.17)	2.88 (1.16)	0.18	0.044
I wish I had told him/her how much s/he meant to me	3.12 (1.19)	2.80 (1.18)	0.27	0.002
I wish I would have taken my chance to say goodbye	3.08 (1.21)	2.73 (1.16)	0.29	0.001
UB, one item (any “unfinished business” concerning the final few weeks the family spent with the patient)	4.35 (1.81)	4.15 (1.94)	0.11	0.227
UB-related distress (distress about this issue in the past month)	4.73 (1.62)	4.31 (1.75)	0.25	0.005
UB-related domains of the Good Death Inventory
Good relationship with family
Total score
Good relationship with family domain, mean	4.86 (1.10)	4.77 (1.16)	0.08	0.382
Each item score
Patient spent enough time with his or her family	4.87 (1.40)	4.57 (1.54)	0.20	0.025
Patient had family to whom he or she could express feelings	4.28 (1.49)	4.29 (1.39)	−0.01	0.885
Patient had family support	5.46 (1.12)	5.46 (1.13)	−0.01	0.971
Life completion
Total score
Life completion domain, mean	4.13 (1.51)	4.50 (1.48)	−0.24	0.007
Each item score
Patient felt that his or her life was completed	4.39 (1.83)	4.73 (1.74)	−0.19	0.035
Patient felt that his or her life was fulfilling	4.46 (1.57)	4.80 (1.54)	−0.22	0.013
Patient had no regrets	3.54 (1.70)	3.92 (1.64)	−0.23	0.011
Preparation for death
Total score
Preparation for death domain, mean	4.78 (1.17)	4.71 (1.29)	0.06	0.499
Each item score
Patient was able to say what he or she wanted to dear people	4.39 (1.55)	4.62 (1.52)	−0.15	0.095
Patient met people whom he or she wanted to see	5.04 (1.47)	4.70 (1.57)	0.22	0.014
Patient felt thankful to people	4.92 (1.36)	4.81 (1.48)	0.08	0.363

Items of the UBBS brief version were graded on a five-point Likert-type scale (1: not at all distressed to 5: extremely distressed); UB, one item on a seven-point Likert-type scale (1: absolutely disagree to 7: absolutely agree); and UB-related distress on a seven-point Likert-type scale (1: not distressed at all to 7: extremely distressed).

*Item scores before and after the QI project at Seirei Hospice were compared using t-tests.

ES, effect size; UB, unfinished business; UBBS, Unfinished Business in Bereavement Scale.

UB of patients (Good Death Inventory)

We used three domains of the Good Death Inventory (GDI) that are closely related to family-perceived UB of patients: “good relationship with family,” “life completion,” and “preparation for death.” ³ As these GDI domains are phrased from the patient’s perspective, they allow indirect assessment of patient-related unfinished matters. Each domain comprises three items, rated on a 7-point scale, addressing: time spent with family, opportunities to express and receive feelings, sense of a completed and fulfilling life, absence of regrets, opportunities to say what one wanted to say, meet desired people, and feel thankful.

Statistical analyses

Descriptive statistics were used to summarize patient and family baseline characteristics. For UB-related outcomes, scores on each measure of UB and UB-related distress before and after the QI initiative were compared using independent samples t-tests, with corresponding effect sizes (ESs) reported. Given the exploratory nature of this study, two-sided p-values <0.05 were considered significant.

Comparison of UB-related outcomes across pre- and post-QI periods

Total UBBS scores (unfulfilled wishes subscale) were significantly lower in the post-QI cohort (mean, 12.46 (standard deviation = 4.16) vs. 11.24 (4.23); ES = 0.29; p = 0.001) as well as all four items (Table 2).

On the one-item UB scale, the overall levels of UB did not differ between pre- and post-QI cohorts (mean, 4.35 (1.81) vs. 4.15 (1.94); ES = 0.11; p = 0.227), whereas UB-related distresses were significantly lower in the post-QI cohort (mean, 4.73 (1.62) vs. 4.31 (1.75); ES = 0.25; p = 0.005).

Among GDI domains, life completion scores were significantly higher in the post-QI cohort (mean, 4.13 (1.51) vs. 4.50 (1.48); ES = −0.24; p = 0.007), whereas a good relationship with family (mean, 4.86 (1.10) to 4.77 (1.16); ES = 0.08; p = 0.382) and preparation for death (mean, 4.78 (1.17) to 4.71 (1.29); ES = 0.06; p = 0.499) remained stable.

Strengths and limitations

Strengths include the relatively large sample and use of validated measures, enabling pre–post comparisons in routine practice. However, recall bias is possible, patient-reported outcomes were unavailable, and data were collected at a single Japanese PCU, limiting generalizability. In addition, the intensity and fidelity of the QI activities were not systematically monitored, and individual exposure to the intervention could not be confirmed. As an observational pre–post study, causal inference is limited. Lastly, because different cohorts were compared in this pre–post design, observed differences may reflect cohort characteristics rather than intervention effects. Therefore, findings should be interpreted as exploratory and hypothesis-generating. However, the consistent reduction in distress-related outcomes and the low-burden nature of the intervention suggest potential relevance in routine palliative care. These findings support the need for hypothesis-driven interventional studies in both patients and families, rather than providing evidence of intervention effectiveness.