Predictors of walking ability at discharge post proximal femoral fracture in an acute care hospital

Abstract

[Purpose] To identify factors immediately post surgical treatment of proximal femoral fractures that can predict walking ability at discharge from an acute care hospital within one month. [Participants and Methods] A total of 122 patients >50 years, admitted to our institution with a fractured proximal femur, were able to walk independently or with a cane or walker prior to admission, and were prescribed physical therapy, post-surgery, were evaluated. Each patient was assigned to one of two groups: those who could walk with a cane or independently at discharge (walking group; n=35) and those who could not walk with the assistance of a cane at discharge (non-walking group; n=87). Multivariate regression analysis was performed to identify predictors of walking ability at discharge. [Results] Age, sex, pre-injury Functional Independence Measure motor (FIMm) score, and pre-injury Functional Independence Measure cognitive (FIMc) score were significantly related to walking ability. In the multivariate regression analysis, the FIMc score was a significant predictor of walking ability. [Conclusion] A detailed assessment of cognitive function on admission may enable predictions regarding the feasibility of direct discharge and the provision of appropriate rehabilitation.

INTRODUCTION

Proximal femoral fractures are among the most common types of fractures in elderly people and are associated with decreased bone mineral density. Many patients with proximal femoral fractures are unable to regain their pre-fracture status of mobility and daily activities¹⁾. It is estimated that by 2050, 4.5 million people will have hip fractures worldwide²⁾. Osteoporotic fractures, such as proximal femoral fractures, are associated with increased mortality, disability, and severe long-term decreases in function^3,4,5). For these reasons, proximal femoral fractures are typically treated with surgical intervention in the majority of cases, with early rehabilitation.

Despite advancements in postoperative rehabilitation facilities, many patients with proximal femoral fractures continue to face significant challenges in regaining their pre-fracture mobility and independence¹⁾. Approximately 20% of patients moving from their own home transfer to institution-based care provision following proximal femoral fracture⁶⁾. Other reports suggested that less than 60% of these patients return to their own place of residence and even fewer reach their pre-fracture status^{7, 8)}. Thus, some individuals do not return to their pre-fracture status despite surgery and adequate rehabilitation.

Proximal femoral fractures are not only a common reason for admission to an orthopedic trauma ward, but also represent a life-changing event that significantly impacts patients’ quality of life and health economics. In Japan, the Diagnosis Procedure Combination (DPC) system is in place, and facilities performing surgeries should avoid long-term hospitalization. To stay within the DPC II timeframe, it is desirable to keep the hospitalization period within 22 days. Previous reports suggested that the average length of hospital stay exceeds one month, which is the duration from admission to an acute care hospital to discharge from a rehabilitation hospital^{9, 10)}. Because 22 days is not sufficient for returning to the patient’s pre-fracture status, it is common for acute care hospitals to initiate transfer arrangements to rehabilitation hospitals with rehabilitation wards after surgery. It is preferable to initiate the transfer as soon as the transfer arrangements have been finalized. If it is determined that returning to the patient’s own place of residence or to the nursing home where the patient lived before the injury is possible through rehabilitation during the acute care hospitalization, priority is given to discharge to their own place of residence. There is a significant variation in individual capabilities in terms of walking ability at the time of discharge from the acute care hospital, making it difficult to predict activities of daily living (ADL) at discharge during this short period. If we could predict walking ability at the time of discharge from acute care hospitals immediately after surgery, it would allow us to determine whether the patient can be discharged directly or needs to be transferred immediately after surgery, making goal-setting for discharge easier.

A few previous studies have investigated the factors related to the possibility of home discharge^{6, 11)}. Nanjayan et al. suggested that increased age, male gender, accompaniment for outdoor mobilization, increasing ASA grade, abbreviated mental test score (AMTS) <6, delay greater than 36 hours for medical reasons were predictors of a change in discharge destination⁶⁾. Salar et al. identified younger age, female gender, AMTS, absence of certain co-morbidities, cohabiting, walking independently outdoors, no use of walking aids, no assistance required with basic ADL, and intracapsular fracture as predictors of home discharge¹¹⁾. However, the duration of hospital stay was unclear in these previous reports. There were few previous reports on the short-term recovery of walking ability after surgery for proximal femoral fractures. If walking ability at the time of discharge from the acute care hospital could be predicted from the immediate postoperative period, it would allow for tailored rehabilitation based on individual capabilities and enable prompt transfer arrangements for patients who are likely to have difficulty with walking from the early stages of hospitalization. The purpose of this study was to identify factors immediately after surgery that can predict walking ability at the time of discharge from an acute care hospital within one month.

PARTICIPANTS AND METHODS

Ethics approval for this retrospective study was provided by the institutional review board of our institution. The opt-out method was implemented after detailed information was provided to patients, with an opportunity to decline participation. This approach was deemed ethically appropriate. Between July 2022 and April 2024, 162 consecutive patients above the age of 50 years admitted to our institution with a fractured proximal femur, who were able to walk independently or with a cane or walker prior to admission, and were prescribed physical therapy after surgery, were identified for inclusion in this study. Proximal femoral fractures are reported to occur 2–3 times more frequently in individuals aged 50 and above¹²⁾. For those under 50, the majority of injuries are caused by high-energy trauma¹³⁾. It has been noted that women experience a rapid decline in bone density after menopause, while in men, the risk of proximal femoral fractures increases from the age of 50¹⁴⁾. Based on these previous studies, we focused on individuals aged 50 and above as the target population. Patients who did not undergo surgery, had fractures at multiple locations, had high-energy trauma injuries, and had missing data were excluded from this study. High-energy trauma was defined as fractures resulting from traffic accidents or falls from significant heights, which were excluded due to their potential to confound results. Patients with missing data were excluded from the analysis, and no imputation methods were applied. This approach was taken to minimize potential biases caused by missing data.

A total of 122 patients (age 84.5 ± 9.4 years, height 150.5 ± 9.0 cm, weight, 46.8 ± 8.6 kg) were finally enrolled and subsequently divided into two groups: one that could walk with a cane or independently at the time of discharge (the walking group; n=35), and the other that did not achieve cane-assisted walking at the time of discharge (the non-walking group; n=87). Walking ability was assessed using Functional Ambulation Categories, and the walking group included patients with a score of 2 or higher. The mean length of hospital stay was less than one month.

Information was collected retrospectively from the patients’ medical records. This study is retrospective in nature, and as such, is subject to potential selection bias and confounding factors that were not measured or accounted for. These limitations should be considered when interpreting the results. Data included age, gender, body mass index (BMI), pre-injury Functional Independence Measure (FIM) scores, including FIM Motor (FIMm) and FIM Cognitive (FIMc) scores, living alone prior to injury, walking aids used prior to injury (no aids/with aids), number of medications, pre-injury co-morbidities (hypertension/diabetes mellitus/cardiovascular disease/cerebrovascular disease/respiratory disease/renal disease/gastrointestinal disease/malignancy/rheumatoid arthritis), preoperative hemoglobin level, location of fracture (femoral neck/intertrochanteric), method of operation (open reduction and internal fixation (ORIF)/bipolar hip arthroplasty (BHA) or total hip arthroplasty (THA)), time between injury and surgery, and Geriatric Nutritional Risk Index (GNRI) in the first week after surgery. These data were compared between the two groups. Moreover, Multivariate regression analysis was performed on preoperative factors, such as age, gender, BMI, pre-injury FIMm and FIMc scores, living alone prior to injury, walking aids used prior to injury, number of medications, and preoperative hemoglobin level. In addition, the rates of patients discharged either to their original living environment or transferred to a rehabilitation hospital were investigated.

Statistical analysis was performed using EZR software (Saitama Medical Center, Jichi Medical University, Saitama, Japan). The results are shown as means and standard deviation. Statistical comparisons were conducted using Student’s t-test for continuous variables. Categorical variables were analyzed by the χ² test. Multivariate regression analysis was performed to identify predictors significantly related to walking ability at hospital discharge. Prior to performing multivariate regression analysis, assumptions such as linearity, multicollinearity, and normality of residuals were tested and confirmed to be valid. Differences were considered significant at p<0.05.

RESULTS

In this study population, 72.1% resided at own home preoperatively, whereas 27.9% were living in senior care facilities, residential care facilities, or nursing homes. Of those who lived at home, 14.8% were able to return to their original homes, 1.1% moved to a nursing facility, and the remaining 84.1% were transferred to another institution. Of the patients who lived in a facility, 23.5% were able to return to their original facility, whereas 76.5% were transferred to another institution (Fig. 1).

Fig. 1.

A, Pre-injury living situation. B, Discharge location of patients living in their own home before injury. C, Discharge location of patients living in a senior care facility, residential care facility for the elderly, or nursing home before injury.

In the walking group, 37.1% returned directly to their home, whereas the remaining 62.9% were transferred to another institution. In the non-walking group, none were discharged to home: 9.2% were discharged to a facility, and 90.8% were transferred to another institution (Fig. 2).

Fig. 2.

Discharge location for patients admitted following proximal femoral fracture. A, walking group. B, non-walking group.

The baseline characteristics of the walking and non-walking groups are shown in Table 1. Patients who could achieve walking ability with a cane or independently were significantly younger (p=0.004), male (p=0.046), and had high FIM scores (FIM p=0.013, FIMm p=0.020, FIMc p=0.001). Comparisons between groups for age, FIM scores, and other variables were made using Student’s t-test for continuous variables, and the χ² test for categorical variables.

Table 1. Baseline characteristics of the two groups.

	Waking group (n=35)	Non-walking group (n=87)	p-value
Age (years)	80.7 ± 9.9	86.1 ± 8.8	**
Gender
Male	11 (31.4%)	12 (13.8%)
Female	24 (68.6%)	75 (86.2%)	*
BMI (kg/m2)	20.6 ± 3.2	20.6 ± 3.0
Pre-injury FIM score (points)	113.9 ± 22.3	101.5 ± 25.2	*
FIMm	82.5 ± 14.1	74.1 ± 18.8	*
FIMc	32.4 ± 5.6	27.4 ± 8.4	**
Living alone prior to injury
Yes	8 (22.9%)	22 (25.3%)
No	27 (77.1%)	65 (74.7%)
Walking aids used prior to injury
No aids	24 (68.6%)	48 (55.2%)
With aids	11 (31.4%)	39 (44.8%)
Number of medications	7.7 ± 4.0	8.2 ± 4.5
Pre-injury co-morbidity
Hypertension	15 (42.9%)	51 (58.6%)
Diabetes mellitus	12 (34.3%)	22 (25.3%)
Cardiovascular disease	6 (17.1%)	31 (35.6%)
Cerebrovascular disease	9 (25.7%)	21 (24.1%)
Respiratory disease	3 (8.6%)	7 (8.0%)
Renal disease	2 (5.7%)	6 (6.9%)
Gastrointestinal disease	12 (34.3%)	14 (16.1%)
Malignancy	7 (20.0%)	10 (11.5%)
Rheumatoid arthritis	1 (2.9%)	1 (1.1%)
Preoperative hemoglobin level (g/dL)	11.6 ± 1.9	11.2 ± 1.7
Location of fracture
Femoral neck	21 (60.0%)	50 (57.5%)
Intertrochanteric	14 (40.0%)	37 (42.5%)
Method of operation
ORIF	19 (54.3%)	36 (41.4%)
BHA or THA	16 (45.7%)	51 (58.6%)	0.274
Time between injury and surgery (hours)	82.5 ± 67.2	96.6 ± 84.8	0.385
GNRI one week after surgery (points)	95.8 ± 11.7	93.4 ± 10.1	0.261

*p<0.05, **p<0.01.

BMI: body mass index; FIM: functional independence measure; FIMm: functional independence measure motor; FIMc: functional independence measure cognitive; ORIF: open reduction and internal fixation; BHA: bipolar hip arthroplasty; THA: total hip arthroplasty; GNRI: geriatric nutritional risk index.

The result of the multivariate regression analysis is shown in Table 2. On multivariate regression analysis, the FIMc score was a significant predictor (p=0.020).

Table 2. Multivariate regression analysis for predictors of walking ability at discharge.

	Multivariate analysis
	OR	95% CI	p-value
Age	1.0	0.92–1.01
Gender	2.8	0.95–8.32
BMI (kg/m2)	0.9	0.81–1.10
Pre-injury FIMm score (points)	1.0	0.96–1.04
Pre-injury FIMc score (points)	1.1	1.02–1.27	*
Living alone prior to injury	0.7	0.25–2.02
Walking aids used prior to injury	1.5	0.53–3.95
Number of medications	0.9	0.85–1.04
Preoperative hemoglobin level (g/dL)	1.0	0.75–1.30

*p<0.05.

OR: odds ratio; CI: confidence interval; BMI: body mass index; FIM: functional independence measure; FIMm: functional independence measure motor; FIMc: functional independence measure cognitive; ORIF: open reduction and internal fixation; BHA: bipolar hip arthroplasty; THA: total hip arthroplasty; GNRI: geriatric nutritional risk index.

DISCUSSION

Proximal femoral fractures are among the most prevalent fractures in elderly individuals, significantly impacting their prognosis and postoperative ADL^{15, 16)}. Postoperative complications are common, and regaining walking ability can be particularly challenging¹⁷⁾. Therefore, patients with proximal femoral fractures often face significant challenges in being discharged to their own homes. Rates of home discharge after proximal femoral fractures are reported to range from 13% to 66% in previous reports^{9, 11, 18)}. Chow et al. reported that 66.1% were discharged home at the time of follow-up, but only 19.4% of these patients were directly discharged to home, and the other patients were transferred to a rehabilitation hospital⁹⁾. Ferris et al. reported that 29% of patients were discharged directly home from an acute care hospital¹⁸⁾. In the present study, only 14.8% of individuals who lived in their own home were discharged directly to home, and only 37.1% of patients in the walking group were discharged directly to home. Moreover, those who did not regain walking ability did not return directly to their own homes. More than 90% of patients in the non-walking group were transferred to another institution. Based on these results, it can be inferred that walking ability is essential for discharge. In practice, considering the safety after discharge and the need for outpatient rehabilitation until walking ability is regained, being able to walk at the time of discharge is a key condition. Thus, even after surgery for proximal femoral fractures, it is challenging for patients to regain their pre-fracture level of ADL in acute care hospitals, and walking ability may be very important for direct discharge to home. In addition, the presence of various complications in the elderly population likely hampers rehabilitation progress, thereby reducing the possibility of direct discharge to home. Therefore, we believe that it is essential to perform rehabilitation with clearly defined, individualized goals tailored to each patient’s needs. Due to the limited and brief length of stay in acute care hospitals such as ours, predicting postoperative ADL in the immediate postoperative period is extremely difficult. However, if it is possible to somewhat predict whether a patient will achieve the walking ability necessary for direct discharge from an acute care hospital at the time immediately after surgery, it would be highly beneficial for determining whether to focus rehabilitation efforts towards discharge or transfer from the early postoperative stage. By assessing the preoperative status, for patients presumed to be capable of home discharge, we propose a rehabilitation protocol aiming for them to achieve a level of walking within one month, without requiring a transfer to another hospital. Furthermore, support for home discharge, such as installing handrails or adjusting beds, would be provided immediately after surgery. For patients presumed to be incapable of home discharge, adjustments could be made in collaboration among families, a support hospital and our hospital to facilitate eventual home discharge. It is speculated that such an approach could lead to the possibility of efficient management starting immediately after surgery. However, reference standards or benchmarks are necessary, and further research is required to establish these.

Some previous reports suggested predictors of whether patients could be discharged or needed to be transferred. Deakin et al. suggested that lack of physical independence, increased age, male gender, and sustaining a fracture while an inpatient were all predictors of discharge to an alternative location⁷⁾. Chow et al. showed that Self-Efficacy for Exercise (SEE) score ≥53, female gender, ≥8 physiotherapy sessions, Mini-Mental State Examination (MMSE) score ≥17, and caregiver availability were significant predictors of home discharge⁹⁾. Nanjayan et al. reported that increased age, male gender, accompaniment for outdoor mobilization, increasing ASA grade, and delay greater than 36 hours for medical reasons were predictors of a change in discharge destination⁶⁾. Moreover, in a study focusing on walking ability at discharge, Nagai and Okawa reported that 78.9% of patients were able to walk with an assistive device, including cane walker, and that age, bone and joint disease, and dementia were significant factors for failure to walk after proximal femoral fracture surgery¹⁰⁾. In the present study, younger age, male gender, and better preoperative physical and cognitive functions as significant factors associated with the acquisition of walking ability at discharge in a comparison between the walking and non-walking groups. The present study’s results differed partially from those of previous reports in that men were more likely to achieve walking ability at discharge, and the rate of patients able to walk at discharge was lower. This is likely because walking conditions were observed for an average of less than one month. In a previous study, Deakin et al. reported that the length of hospital stay varied depending on discharge location, and that patients who lived in their own residence prior to injury required more than 30 days to be discharged to a residential home or nursing home⁷⁾. Chow et al. stated that the average postoperative lengths of stay at an acute hospital and a rehabilitation hospital were 8.6 and 31.2 days, respectively⁹⁾. Nagai et al. reported that the mean length of hospital stay was 45 days¹⁰⁾. The mean length of stay was less than one month in the present study, shorter than in previous studies. Regarding the differences from previous studies, we believe that men had an advantage because of greater muscle strength than women in the shorter period and suggest that walking ability was not as fully achieved within an average of one month as reported in previous studies. Chow et al. reported that the presence of a caregiver influences walking ability at discharge. However, they do not address differences in background or the quality of caregivers⁹⁾. These factors may potentially explain the discrepancies observed in our results. Although, due to the limited sample size in this study, further investigation is required.

Based on the results of the present study, cognitive function was identified as a predictor for achieving walking ability with a cane or better at the time of discharge from an acute care hospital on multivariate analysis. Ryder et al. reported that cognitive status was one of the most significant patient-level characteristics predictive of discharge destination from the acute ward¹⁹⁾. There may be a close relationship between physical function and cognitive function. It has been pointed out that decline in physical capacity often co-occurs with decline in cognition²⁰⁾. Elderly individuals with lower physical capabilities tend to perform worse on cognitive assessments, and vice versa^{21, 22)}. Varma et al. have suggested that research on mobility should incorporate cognitive measures and that consideration should be given to cognition and mobility as an intertwined subject for research²³⁾. Some previous reports showed that lower extremity function also partially explains the association between executive function, a higher older cognitive function, and life-space mobility, a correlate of self reported 2-km walking ability^{24, 25)}. Siltanen et al. suggested that older people who concurrently manifest poorer lower extremity function and cognitive performance have the highest odds for walking difficulty²⁶⁾. From the results of previous reports, cognitive function is considered to be one of the very important factors for physical function, such as walking ability. Assessing cognitive function upon admission may face several challenges, such as the potential decrease in accuracy due to advanced age, the limited time available for assessment in emergency admissions, and the possibility of obtaining inaccurate information due to pain from fractures. Based on these considerations, it is necessary to utilize cognitive function assessments.

The findings of the present study were consistent with those of previous research^{9, 10)}. The reason for this finding may be that the narrow support base of a cane requires sufficient balance, necessitating the recognition of unstable environments, which relies on cognitive function. In addition, the ability to manage objects is essential due to the risk of dropping or tipping the cane, and the understanding of pain onset during weight-bearing activities is also required from a cognitive perspective. Therefore, it can be inferred that understanding one’s own physical capabilities and fall risks is crucial, making cognitive function an important predictor.

There are two major advantages of predicting walking ability at discharge in the immediate postoperative period. First, it enables us to design rehabilitation programs based on predicted walking ability at discharge from the immediate postoperative period. For patients expected to regain the ability to walk, rehabilitation can focus on gait training. For those anticipated to remain non-ambulatory, an extended rehabilitation plan that allows time for transfer arrangements can be proposed, enabling each patient to participate in a feasible rehabilitation process. However, we have not investigated the criteria to a sufficient extent. Therefore, further research is necessary, including the collection of more cases and the use of various cognitive function assessments such as AMTS and MMSE. The second advantage is that transfer coordination can be streamlined from the early postoperative stage. If walking ability at discharge is expected to be limited, early discussions involving family members can facilitate smooth transitions to suitable rehabilitation facilities. This proactive approach is expected to help reduce the workload on healthcare staff and contribute to lower healthcare costs.

The present study has several limitations. First, the sample number was small. Although significant differences were found in some comparisons between the two groups, the sample size might have been inadequate to detect significant differences in others. Second, the number of independent variables in the multivariate analysis was limited. Adding more variables related to patient background characteristics may provide a more detailed understanding. An increase in the number of cases is necessary to achieve this. Third, the possibility of bias due to retrospective data collection is suggested. We think that randomized controlled trials and prospective studies focusing on specific populations are necessary. Forth, the potential impact of gender was unclear in this study. From our results, male patients may have advantage about walking ability at the time of discharge from our hospital, although there is the discrepancy between our result and previous studies. We believe that further research is necessary to elucidate this. Fifth, we assessed walking ability at discharge, however, other aspects, such as recovery in ADL or independence in other aspects of life, were not evaluated. Since these abilities are essential after discharge, we will consider conducting research to include these as evaluation targets in the future. Future studies should be conducted using randomized prospective designs to address this limitation. Furthermore, the exclusion of patients with missing data may have also influenced the study results.

In conclusion, of patients with proximal femoral fractures who underwent surgical treatment and were ambulatory prior to surgery, only 28.7% regained walking ability at the time of discharge from our acute care hospital. Furthermore, of these patients, a mere 37.1% were discharged directly to their homes. Postoperative ambulatory status at discharge appeared to be significantly affected by preoperative cognitive function. Detailed assessment of cognitive function at admission may enable predictions regarding the feasibility of direct discharge. The use of cognitive function assessment as a standard protocol upon admission for patients with proximal femoral fractures is considered highly beneficial. Although further investigation is warranted, accurate prediction could facilitate the development of efficient rehabilitation protocols aimed at direct discharge, the creation of individualized rehabilitation plans, and the seamless coordination of transfer arrangements from the early stages of hospitalization.

Conflict of interest

There are none.