The impact of dementia on rehabilitation outcomes following hip fracture

Abstract Objective To compare clinical outcomes between patients for whom their participation in inpatient rehabilitation was and was not impacted by dementia through matching patients reporting dementia (dementia group) with those not reporting dementia (non‐dementia group). Methods Prospectively collected data held by the Australasian Rehabilitation Outcome Centre (AROC) were analyzed for patients aged 65 years or older receiving inpatient rehabilitation in public hospitals in Australia following a hip fracture and discharged between July 1, 2014, and June 30, 2019. Patients reported as having dementia impacting their rehabilitation program were matched to patients not reporting dementia based on age, admission motor Functional Independence Measure (FIM) score, and accommodation prior to rehabilitation. The matched cohorts were compared in relation to clinical outcomes (motor and cognitive FIM improvement, FIM efficiency, length of stay, and discharge destination) following participation in hospital‐based rehabilitation using univariate analysis. Results Patients with dementia had significantly lower cognitive FIM scores on commencing rehabilitation (17.6 and 26.9, respectively, P < 0.001) and their median length of stay was 2 days shorter than those without dementia (21 and 23 days, respectively, P < 0.001). Relative change in FIM score and FIM efficiency (per week) were lower in the dementia group [relative FIM score change of dementia vs non‐dementia, respectively, 26.2% vs. 44.0% (P < 0.001) and FIM efficiency, 6.5 vs. 8.9 (P < 0.001)]. Discharge destination between the two groups was statistically different, with 35.7% of patients with dementia being discharged to residential aged care facilities (RACFs) compared to 21.7% of those without dementia (P < 0.001). More patients with dementia had carers in their private residence in the post‐rehabilitation phase, 82.2% vs. 57.6% (P < 0.001). Conclusion Patients with dementia who sustain a fractured hip benefit from inpatient rehabilitation, although their clinical outcomes are not as good as those without dementia. FIM change and FIM efficiency were lower in the dementia group. Length of stay in the hospital for patients with dementia was shorter due to earlier recognition for the need for placement in either an RACF or at home with carer support. The need for placement in an RACF or carer support in a private residence was significantly greater in the dementia group.


| INTRODUC TI ON
Dementia describes a syndrome characterized by gradual impairment of brain functions. It affects memory, thinking, orientation, comprehension, calculation, learning capacity, language, and judgment. 1 Worldwide, around 50 million people have dementia and almost 10 million new cases are diagnosed yearly. The total number of people with dementia is projected to reach 82 million in 2030 and 152 million in 2050. 1,2 In Australia, an estimated 472,000 Australians had dementia in 2021. 3 In 2020, dementia was estimated to cost Australia more than $15 billion, with a cost estimate of $18.7 billion in 2023, increasing to $36.8 billion by 2056. 4 Hip fractures contribute to both mortality and morbidity in the elderly. Dementia is associated with a hip fracture hospitalization rate 2.5 times that of adults without dementia. 5 In Australia, between 2015 and 2016, there were 50,900 episodes of hip fractures that required hospital care for both new hip fractures and management and repair of previous fractures. These hospitalizations accounted for 579,000 bed days and required 206,300 procedures or interventions. 6 A meta-analysis revealed that women sustaining a hip fracture had a fivefold increase and men almost an eightfold increase in relative likelihood of death within the first 3 months as compared with age-and sex-matched controls. 7 The CHANCES project, a follow-up of 122,808 participants from eight cohorts in Europe and the United States for a mean of 12.6 years reported 4273 incident hip fractures and 27,999 deaths, with both increased short-and long-term all-cause mortality for both sexes. 8 The changing demographic of dementia worldwide will continue to impact on hip fracture rehabilitation. Fragility fractures are relatively common in the elderly, particularly among those with dementia. Studies exploring the predictors of poor clinical outcomes and level of function on discharge for those who underwent rehabilitation in subacute settings are limited. A recent Cochrane review of hip fracture management by Smith et al suggested that models of enhanced rehabilitation and care used in the included trials show benefits over usual care for preventing delirium and reducing the length of stay (LOS) for people with dementia. However, the level of evidence of these results is low as data were available from only seven trials with a total of 555 participants. 9 This highlights the importance of further research to define optimal rehabilitation and supportive care for people with dementia following hip fracture surgery.

| AIM OF THE S TUDY
There were two objectives of this study. First, we described the patient profiles on admission to inpatient rehabilitation for those who did and did not report dementia impacting their rehabilitation.
Second, we compared patient profiles and clinical outcomes for patients who reported dementia impacting on their participation in rehabilitation (dementia group) to a matched cohort of those who did not report dementia (non-dementia group).

| Study design
This study used prospectively collected episodes of inpatient rehabilitation routinely submitted to the Australasian Rehabilitation Outcomes Centre (AROC) for benchmarking. The AROC is the national rehabilitation medicine integrated outcomes center of Australia and New Zealand. It is a joint initiative of the Australian rehabilitation sector (providers, funders, regulators, and consumers) 10,11 and collects data on over 95% of inpatient rehabilitation services across both countries as members.

| S TUDY P OPUL ATION
Patients receiving inpatient rehabilitation following a fractured neck of femur, who were aged 65 years or older and were treated in an Australian public hospital between July 1, 2014, and June 30, 2019, were included in this study.

| Identifying patients with dementia
The AROC dataset includes comorbidities, defined as pre-existing illness/impairments, not part of the principal presenting condition, impacting the rehabilitation process. Dementia is one such comorbidity.

| Data collection
The AROC data collection includes: • On admission only: age, sex, and date of fracture.
• On admission and discharge: accommodation, carer status, services received, and functional status.

| Length of stay
An inpatient rehabilitation episode of care is the date that the patient's care is transferred to a rehabilitation physician or physician with interest in rehabilitation, and it is recorded in the medical record that the K E Y W O R D S dementia, hip fracture, rehabilitation rehabilitation team has commenced the rehabilitation program/provision of care. 12 Episode end date records the date that the patient completes their rehabilitation episode; it defines the end of rehabilitation episode and is the date of which the LOS concludes. An inpatient rehabilitation episode of care ends when the patient is discharged from the rehabilitation unit and/or the care type is changed from rehabilitation to acute or some other subacute (maintenance/palliative care).

| Statistical Analysis
To create the matched cohort, patients with dementia (dementia group) were matched by age, motor functional status on admission, and residence at the time of fracture with patients not reporting dementia (non-dementia group). Four passes were required to obtain the full matched cohort ( Table 1), with subsequent passes being less rigid. Most dementia episodes (96%) found a match in the first pass, giving a match that was within 1 year of age and one point of motor admission FIM score and came from the same prior accommodation.
Univariate analyses were used to describe group differences; the chi-square test was used with categorical variables, and the t test was used with continuous variables. A P value < 0.05 was regarded as statistically significant.
The Australian National Subacute and Non Acute Patient (AN-SNAP) is a case-mix classification that categorizes rehabilitation, palliative care, geriatric evaluation and management, psychogeriatric care, and non-acute care. The rehabilitation branch classifies patients based on impairment, level of function on admission, and age. 13,14 In Australia, AN-SNAP is used for funding and case-mix adjustment of clinical outcomes. Version 4 AN-SNAP was used in this study.

| Complete Cohort
During the study period 20,905 episodes of care were admitted to inpatient rehabilitation units following a fractured neck of femur. Of these, 1901 episodes reported dementia impacted participation in rehabilitation. Table 2 describes the patient profiles on admission to inpatient rehabilitation for those who did and did not report dementia impacting their rehabilitation.
Most patients were admitted to the rehabilitation unit within 2 weeks since the initial injury with 67.3% in the demen-

TA B L E 1 Matching details
TA B L E 2 Patient profiles on admission to inpatient rehabilitation for those who did and did not report dementia impacting their rehabilitation using the complete cohort

| Matched Cohort
After controlling for age, motor function, and prior accommodation by matching, there was no difference between the two groups in the timing of the rehabilitation unit admission since the initial insult (P = 0.074; Table 3). Admissions from a private residence showed both higher needs for carer assistance and support services (71.0% and 43.9%, respectively) for the dementia cohort compared with the non-dementia cohort (47.8% and 37.1%, respectively). Further, mean admission cognitive FIM was significantly lower in the dementia group compared with the non-dementia group (17.6 and 26.9 points, respectively, P < 0.001).
In comparing clinical outcomes ( Table 4) Patients with dementia on admission stayed 2 days less in the rehabilitation ward compared to those without dementia (21 and 23 days, respectively, P < 0.001). Patients reporting dementia were statistically more likely to be discharged to an RACF (35.7% and 21.7%, respectively, P < 0.001; Figure 2). Patients with dementia were more likely to require carer support in their private residence in the post-rehabilitation phase (82.2% and 57.6%, respectively, P < 0.001; Figure 3).
Both groups showed increased needs for services postrehabilitation of almost 30% compared to admission (73.3% and 68.2%, respectively, P = 0.045). Allied health care and social support increased in both groups, being greater for the non-dementia group.
Personal care support increased about 20% in both groups, whereas support for meals declined. A higher proportion of patients with dementia (5.8% on admission and 21.4% on discharge) needed case management support in the post-rehabilitation phase than those without dementia (5.7% on admission and 16.2% on discharge).

| DISCUSS ION
Dementia is a common comorbidity in older people admitted to acute hospitals and is associated with poor outcomes. 15

TA B L E 2 (Continued)
TA B L E 3 Patient profiles on admission to inpatient rehabilitation for those who did and did not report dementia impacting their rehabilitation using the matched cohort

TA B L E 3 (Continued)
TA B L E 4 Comparison of clinical outcomes among patients who did and did not report dementia impacting on their participation in rehabilitation using a matched cohort

| Strengths and Limitations
This study is a large retrospective analysis of rehabilitation out- A limitation of the study was the inability to delineate the severity of dementia and therefore its impact on rehabilitation outcomes. Nondementia group could have undiagnosed cognitive impairment but did not impact on rehabilitation at the time of admission and therefore not recorded. Although FIM scoring has both motor and cognitive components, FIM scoring is heavily weighted for motor functioning.

| CON CLUS ION
With the rising incidence of dementia and hip fractures associated with a globally aging population, providing optimal inpatient rehabilitation and post-discharge support services for this vulnerable group is essential to maximize functional gains and reduce long-term health costs. Despite the increased discharge to an RACF for the dementia group, we demonstrated that improvements in both motor and cognitive function are likely to have a beneficial effect on long-term management. By identifying functional and cognitive limitations and assistance prior to admission, strategic points of intervention across the trajectory of recovery can be introduced thereby improving clinical outcomes. Future research is needed to delineate the impact of severity of dementia on rehabilitation outcome as the disease progresses.
F I G U R E 2 Accommodation at discharge from inpatient rehabilitation among a matched cohort of patients who did and did not report dementia impacting on their participation. RACF, residential care facilities.