Prevalence of falls in the last weeks of life and relationship between falls, independence, and quality of dying: A secondary analysis of a large prospective cohort study

2.1 Study design and subjects

This study involved a secondary analysis of a large prospective cohort study conducted between January 2017 and June 2018 to investigate the dying process of patients with advanced cancer in 23 PCUs in Japan [14]. Consecutive eligible patients were enrolled if they had been newly referred to the participating PCUs during the study period, and all enrolled patients were observed until their death or six months after enrollment. All 23 institutions were asked to collect data consecutively, up to a designated number of patients of 50, 60, 70, 80, 100, 150, and 250, according to the size of the institution, without depending on high-volume centers. This study’s inclusion criteria were as follows: (1) adult patients (18 years of age or older), (2) patients diagnosed with locally advanced or metastatic cancer (including hematological neoplasms), and (3) patients admitted to PCUs. Patients who were planned to be discharged within one week or those who did not want to be enrolled were excluded. This study upheld ethical standards per the Helsinki Declaration and Japanese Ministry of Health guidelines for research involving human subjects, with written consent waived for this noninvasive observational study. It was also approved by the institutional review boards at all participating sites.

2.2 Measurements

2.3 Statistical analyses

We analyzed patients who experienced falls one month prior to death as the primary endpoint. The baseline demographic data were summarized using descriptive statistics. First, we calculated the prevalence of falls in all the patients. Additionally, we conducted a subgroup analysis for the patients with a performance status of 3 or less (i.e., excluding the patients with a performance status of 4), on the assumption that bedridden patients had no theoretical risk of falls but some patients with a performance status of 4 might recover their physical function after admission. Second, we compared the patient backgrounds between the patients who did and did not experience falls. Univariate analyses were performed using the Student’s t-test or chi-square test, as appropriate. Multiple logistic regression analyses were then performed using a forward elimination procedure. Third, we calculated the prevalence of serious complications from falls in all the patients and the patients with a performance status of 3 or less. Fourth, we determined the clinical factors associated with falls and serious complications from falls for each patient. Fifth, we calculated the percentage of activities that led to falls in the patients who experienced them. Finally, we compared the data on independence (i.e., use of indwelling urinary catheter during admission and walkability until three or seven days before death) and quality of dying measured by the GDS between the patients who did and did not fall using the Student’s t-test or chi-square test, as appropriate. Subgroup analyses were also performed in this comparison. Statistical significance was set at P <.05. All analyses were performed using SPSS version 28.0.1 (IBM Corp., Armonk, NY, USA).

3.1 Prevalence of fall

Of the total number of patients evaluated, 150 (9.2% [95% CI 7.8 to 11]) fell within 30 days prior to death; limiting the sample to patients with a performance status of 3 or less on admission, 114 (14.8% [95% CI 12 to 17]) fell within 30 days prior to death.

Patients who fell were more likely to be male, younger, had the ECOG performance status 3 on admission, had a longer length of hospital stay days and longer estimated prognosis by the primary physician on admission, and had more delirium during hospitalization (Table 1). The multiple logistic regression analysis revealed that independent determinants of falls were being male, had the ECOG performance status 3 on admission, longer estimated prognosis by the primary physician on admission, and the presence of delirium during hospitalization.

3.2 Serious complication of fall

There were five cases of serious events due to falls, three with fractures, and two with intracranial hemorrhages (Table 2). This figure corresponds to 0.31% [95% CI 0.038 to 0.57] in all patients, and 0.52% [95% CI 0.012 to 1.0] in patients with a performance status of 3 or less (Table 2). Table 3 describes the factors potentially associated with falls, fracture, and hemorrhage in each patient. Both patients with intracranial hemorrhages were elderly (79 and 91 years old); one younger patient with a fracture had bone metastases, whereas two older patients with fractures did not (81 and 83 years old). All patients received opioid medications, and 3 of 5 had an episode of delirium and received antipsychotics. None of the two patients who developed intracranial hemorrhage had risk factors of hemorrhage (i.e., severe liver dysfunction, alcohol abuse, or low platelet count).

3.3 Activities that led to falls

The behaviors that led to falls were: going to the toilet to defecate (64.7%; 97/150), picking up things in the room (12.7%; 19/150), and washing their face and hands (1.3%, 2/150) (Table 4).

3.4 Variables related to independence and quality of dying

Patients who fell were more likely to be able to walk in the last days of life (Table 5). The GDS results and indwelling urinary catheter use were not significantly associated with falls.

4.1 Prevalence of falls in palliative care settings

In this study, 150 patients (9.2%) fell within 30 days prior to death in PCUs. Limiting the sample to patients with a performance status of 3 or less, this figure slightly increased to 14.7%. These findings are consistent with those of previous studies demonstrating that 11%–18% of patients in PCUs experienced falls [1,2,22–25]. Given that the reported fall rate for patients with advanced cancer is approximately 50% (3,26), patients in PCUs were found to have a relatively lower rate. This may be because of the more serious physical condition of the PCU patients and of the medical teams supporting patients in the palliative care ward. This study confirmed that approximately 10%−20% of patients experienced falls in palliative care units.

In addition, this study explored the factors associated with falls. Patients who fell were more likely to have the following characteristics: male, an ECOG performance status of 3 on admission, longer estimated prognosis by the primary physician on admission, and presence of delirium during hospitalization. Patients who fell were more likely to have more delirium during hospitalization. Previous studies [4,22] have shown that delirium is more likely to result in falls. Delirium increases the risk of falls because it can manifest as rapid fluctuations in alertness, attention, and behavior. Moreover, the neuroleptics/antipsychotics often used to treat it may increase the risk of falls [4,27]. In addition, in this study, men were more likely to fall. Previous studies also found that men were more likely to fall [22,25]. However, one study concluded that sex is not statistically significant in relation to fall risk (4). Therefore, it seems best to use these risks as a supplement to clinical judgment to address the continually shifting symptomatology of patients in terminal decline because changes in unstable situations are frequent in hospice care, such as the ECOG performance status 3 on admission, which has a significantly higher risk of falling in this study.

4.2 Prevalence of serious cases in patients who experienced falls

There were five cases (0.31% of all patients and 0.52% of patients with performance statuses of 1−3 on admission) of serious events due to falls, three with fractures, and two with intracranial hemorrhage. These figures in the current study are lower than those reported in previous studies on serious injuries (fractures) requiring severe medical treatment in a palliative care setting (1.6%–4.1%) [1,3], in an acute care setting (6.3%), and in situations with fall prevention measures (4.1%) [5]. This may indicate that the five (0.3%) serious events might not have been caused by the palliative care setting. In addition, both patients with intracranial hemorrhage were older adults, and had no risk factors of hemorrhage. Further, a younger patient with a fracture had bone metastases and two older patients had no bone metastases. This finding suggests that intracranial hemorrhage observed is related to the cases more likely to occur in older patients (e.g., subdural hematoma). Moreover, fracture is related to bone metastases in the younger population and osteoporosis in older patients. Serious events due to falls are extremely rare but they do occur, and not only in the palliative care setting. Although prevention is theoretically important because of the potentially devastating physical and emotional consequences, special precautions simply based on the setting may be unnecessary because they do not necessarily occur frequently in palliative care settings.

4.3 Relationship between falls, independence, and the quality of dying

The most common behavioral reason for falling in terminal cancer patients was “moving to the toilet to defecate.” This means that casual guidance and measures for the patient to use the toilet may reduce the frequency of falls and preserve the dignity of terminally ill cancer patients [15–16,28–30]. In particular, humiliation and shame over the failure to use the toilet have a significant bearing on the dignity of terminally ill cancer patients [16,31], and support for toileting is an issue that cannot be separated from the association between falls and dignity [32]. However, five reasons for resisting asking for assistance to go to the toilet are: they did not want someone to help them with personal activities; if elimination was an emergency, it would be difficult to wait for help to arrive; they thought they could move to the toilet on their own without falling; the nurses were too busy to help; and they did not want to ask for assistance from a nurse who they thought was not compassionate. Even providing elimination assistance involved complex factors [16].

We considered indwelling urinary catheter use and walkability in the last three or seven days of life as indicators of independence, and the GDS as an indicator of quality of dying. These were not associated with falls. Patients who fell were more likely to be able to walk in the last three or seven days of life (Table 5). As noted above, the most common reason for falls is defecation, but neither more nor fewer urinary catheters are used by those who fall. These results represent a significant effort by healthcare providers to pay more attention to the spiritual pain of patients in hospice settings. In addition, many people can walk until the end of their lives, with or without falls, probably as a result of healthcare providers supporting them to walk on their own. Furthermore, falling was not associated with the GDS scores. In previous studies, patients who fell were more likely to have a “greater fear of losing independence,” “avoid asking for help,” and “feel uneasy about asking for help” [4,6]. Nevertheless, in this study involving healthcare providers in a hospice setting, the GDS was not influenced by the presence or absence of falls, indicating that although fall prevention is considered a priority, healthcare providers should support patients’ desires to move on their own to maintain independence, and it may be possible to maintain independence and quality of dying even for patients who have fallen.

4.4 Strengths and limitations

This study’s strengths include a large sample of patients, the high participant response rate, multicenter design, and use of validated measurement tools. However, there are some limitations. First, the same palliative physician performed all observations for each patient. Therefore, observer bias and underreported bias must be considered. However, it is unlikely that the physician would have overlooked a fall because he or she is required to medically examine the patient at the time of the fall. Second, not all possible causes of falls are considered, such as past fall history or the effects of medications [3]. Third, we considered indwelling urinary catheter use during admission and walkability in the last three or seven days of life as indicators of independence and the GDS as an indicator of quality of dying. Our study involved secondary analyses of existing data, and we recognize that these factors alone cannot measure independence and the quality of dying. Similarly, we could not obtain information on the exact time when the falls occurred from the data. Therefore, some variables may have an inverse time relationship to the outcome variable (falls). Finally, we included only patients with cancer admitted to inpatient hospices/PCUs in Japan, where interprofessional care is provided for dying patients.