Fall risk in older adults hospitalized with tumours: Contributing factors and prediction model

Abstract Aim Rates vary widely across hospitals globally and typically range from 3 to 11 falls per 1000 bed days and as 7–11 in Affiliated Hospital of Nantong University. This study determined to explore contributing factors and poor prognosis of fall in elderly tumour patients in China. Design A cross‐sectional study. Methods 161 older adults were invited to participate in this study and completed a self‐reported questionnaire, took blood tests, and received the exam of musculoskeletal ultrasound. Results Among 161 patients, falls occurred in 41 cases, accounting for 24.8%. 51.6% of older adults suffered from intermediate‐to‐high risk of falls. Fall history, reduced self‐care ability, sleep disturbance, hearing impairment, hyperkyphosis, chronic disease, platelet count, and the thickness of left muscle rectus femoris (LF‐MLT), and left cross‐sectional area (LF‐CSA) were all contributing factors of fall, and higher risk of fall indicating lower quality of life. A fall prediction model was established in this study based on above contributing factors with good prediction efficiency (AUC = 0.920). Patient or public contribution The patient volunteers participated in this study and provided valuable data for the final analysis and the acquisition of conclusion.

who underwent surgery, radiation or chemotherapy experienced at least one fall during hospitalization or at home within 2-3 months, whereas 51.4% of older cancer patients experienced at least two falls (Dotan et al., 2021). It is significantly higher than the reported fall rate (33.3%) for older adults worldwide (Sattar et al., 2021). The consequences of a fall in a cancer patient are not limited to injury, delayed recovery and hospitalization, but also more complications, unpredictable disease progression, more care need and ultimately poor prognosis .
Studies indicated that only 10% of older adults hospitalized with tumours who fell in the last 6 months received attention, assessment or intervention from healthcare professionals in hospital (Dotan et al., 2021). Therefore, we explored the influencing factors of fall in older adults hospitalized with tumours and constructed a prediction model.

| Participants
The current cross-sectional study enrolled 161 older adults hospitalized with tumours recruited from Affiliated Hospital of Nantong University between May and December 2021. Patients aged ≥60 years, able to communicate, walk independently (or with an auxiliary tool) and receive cancer treatment without palliative or hospice care were invited in this study. The study of nursing excluded older inpatients with undiagnosed tumours, physical disability and missing clinical biochemical index data.

| Demographic and clinical characteristics
A questionnaire was used to collect participants' demographic and clinical information, including sex, age, smoking and drinking status, exercise status, tumour category, fall history, upper arm circumference, leg circumference, hyperkyphosis, status of medication and chronic diseases, and so on.

| Comprehensive geriatric assessment
The risk of falls in older adults hospitalized with tumours was assessed using the Morse Fall Scale (MFS) with total score ranging from 0 to 125, and higher scores indicate higher risk of falling. Chinese version of MFS, in which low risk of fall (score < 25), moderate risk (score of 25-45) and high risk (score > 45) can be divided, was used in this study (Huang et al., 2021).
The psychological and cognitive state were evaluated using 7-item generalized anxiety disorder (GAD-7) scale for anxiety, 15-item geriatric depression scale (GDS-15) for depression and Montreal Cognitive Assessment (MoCA) for cognition. Each item in the GAD-7 is scored from 0 to 3, with the total scores ranging between 0 and 21 (Terlizzi & Villarroel, 2020). The maximum total score of GDS-15 is 15, and higher scores indicate severer depression (Zhang et al., 2020). The MoCA score ≥26 indicates normal cognitive state and score of ≤25 indicates impairment, while a point is added as a normal demarcation score if the education period is >12 years (Cersonsky et al., 2022). Sleep quality was evaluated by the Pittsburgh Sleep Quality Index (PSQI), the total score of which ranged from 0 to 21 (Orskov & Norup, 2022). The higher the PSQI score, the poorer sleep quality, and PSQI score >5 is indicative of sleep disorder (Zhou et al., 2020).
Physical functional status was evaluated by Barthel Index (BI) for self-care ability, Short Physical Performance Battery (SPPB) and Timed Up and Go (TUG). Total score of BI ranging from 0 to 100 and higher score indicates better self-care ability (McGill et al., 2022).
The SPPB was used to measure balance, gait, strength and endurance, with a total SPPB score ranging from 0 to 12, and lower score indicating poorer physical function . The TUG test is used to evaluate functional mobility and physical performance in older people based on the time it takes to rise from a chair, walk 3 metres, and then turn and walk back to sit in the chair (Hendriks et al., 2022). The frailty index was used to assess frailty and was calculated by counting the number of deficits, with higher score indicating more severe status of frailty (Atkins et al., 2021). QoL (Wang et al., 2022). The Memorial University of Newfoundland Scale of Happiness (MUNSH) was used to measure the subjective well-being of older adults. The MUNSH contains 24 items, and total scores range from −24 to +24 points (Zhang et al., 2022). The Life Satisfaction Index-A (LSIA) was used to measure perceived life satisfaction. The total score of LSIA ranges from 0 to 20, and higher score indicates better life satisfaction (Aydogdu, 2023 #1753).

| Clinical features
The patient's blood biochemical and other laboratory indicators are also collected of this study, including white blood cells (WBC), red blood cells (RBC), platelets (PLT) counts and haemoglobin, and cancer marker (carcinoembryonic antigen [CEA]).

| Musculoskeletal ultrasound
A GE LOgiQ E9 ultrasound machine (GE Corp) fitted with a 15-MHz linear array ultrasound probe was used to measure thickness of muscle rectus femoris, cross-sectional area and thickness of medial femoral muscle rectus femoris on both sides of the patient's thigh.
The following parameters were measured for muscle mass: LF-MLT (the thickness of left muscle rectus femoris), LVI-MLT (the thickness of the left vastus intermedius muscle), LF-CSA (left cross-sectional area), RF-MLT (the thickness of right muscle rectus femoris), RVI-MLT (the thickness of the right vastus intermedius muscle) and RF-CSA (right cross-sectional area) (Jahanandish et al., 2021).

| Statistical analysis
All data were analysed by SPSS 22.0 (IBM Corp). The normality of distribution for continuous data was assessed by the Kolmogorov-Smirnov test. Numerical variables were expressed as mean ± standard deviation or median with interquartile range, while categorical variables were presented as percentage values.
The Mann-Whitney U-test, chi-square test or double-tailed t-test were used to analyse the statistical data. The prediction model was built after selecting all significant variables (p < 0.05) for a stepwise binary logistic regression analysis. ROC curves were used to analyse the sensitivity and specificity of all variables in predicting fall risk and to validate the prediction model. p < 0.05 was considered statistically significant.

| Differences in sociodemographic and clinical features according to fall risk
Fall risk was associated with age, fall history, fall times, mean grip strength, multiple medications, hyperkyphosis, impaired eyesight, hearing impairment and had other chronic diseases. However, no statistically significant association was found between fall risk and tumour category (Table 1).

| Association between fall risk and comprehensive geriatric assessment scores
Fall risk was associated with physical functional status, sleep disturbance, frailty and nutritional status (p < 0.05) ( Table 2).

| Associations of fall risk with tumour-related blood indices and muscle mass
Older adults hospitalized with tumours with an intermediate-to-high risk of fall had lower RBC counts, lower haemoglobin levels, higher PLT count and lower muscle mass than those with low risk of fall (p < 0.05; Table 3).

| Independent risk factors for falls.
We handled each dimension score as the dependent variable and baseline factors impacting fall risk as independent variables, using binary logistic regression analysis. The statistically significant factors are presented in Table 4.

| Fall-risk prediction model in older patients with tumour
According to the ROC curve analysis, high PLT count (

| Association of fall risk with QoL, happiness and life satisfaction
Fall risk of older patients with tumour showed significant side effects on patients' QoL, happiness and life satisfaction (Table 5).

| DISCUSS ION
This nursing study, which was conducted at an institution in eastern China, revealed that 51.6% of older adults hospitalized with tumours were at an intermediate-to-high risk of falls and 24.8% had fall history. We also confirmed that fall risk was evenly distributed across both sexes. Fall history, hyperkyphosis, hearing impairment, chronic diseases, impaired self-care ability, sleep disturbance, higher  side an impact on falls and that fall rates are positively correlated with chronic diseases of Chinese. Alenazi et al., (2023) identified the association between falls and chronic diseases, including hypertension and neuropathy. The above data are consistent with the result in this study.
Older patients with cancer and impaired self-care ability were reported at a higher risk of falls (Goineau et al., 2018). It is probably because older cancer patients are more prone to frailty, which impacts their self-care ability. Poor sleep quality, resulting in insufficient nighttime sleep duration and daytime dysfunctions, was also important risk factor for falls in older adults (Sattar et al., 2021) (Lee, Chung, & Kim, 2021). Increasing sleep duration, while reducing daytime dysfunctions and sleep disturbances, could mitigate unintentional falls (S. . In this study, self-care ability and sleep disturbance also were revealed to be associated with fall risk among. TA B L E 2 Univariate analysis between comprehensive geriatric assessment scores and fall risk. This study demonstrated that the muscle mass, including thickness and cross-sectional area of the lower limb muscle, was significantly associated with fall risk. Sai et al., (2021) showed that the thickness of quadriceps femoris muscle was significantly correlated with fall injuries, with optimal cut-off values of 3.37 cm and 3.54 cm for men and women, respectively. In our study, the average muscle thickness of older cancer patients from China was approximately 7-9 mm, which is lower than the 3-4 cm reported for older people in Europe and the United States. Based on these findings, it is speculated that the muscle of older tumour patients has been adversely affected by the disease and then faced more risk of fall.
According to the result in this study, higher fall risk resulted in lower QoL, lesser happiness and poor life satisfaction, which is consistent with the report of Silva et al., (2021). The feeling of lonely and unaccompanied made them more prone to falls. Therefore, it is necessary to improve the social connections of older adults within family, hospital and communities, preventing social isolation, ultimately reducing the risk of falls.

ACK N O WLE D G E M ENTS
We would like to acknowledge the support given by all the funders of this study in ensuring its successful undertaking. We also thank all the support of all the nursing administrators who participated in this research. We thank Bullet Edits Limited for the linguistic editing and proofreading of the manuscript.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available upon request by contact with corresponding author.

E TH I C S S TATEM ENT
This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of the Affiliated Hospital of Nantong University (No.