Evaluation of the Implementation of a Multidisciplinary Fast Track Program for Geriatric Acute Hip Fracture Patients at a University Hospital in Resource-Limited Settings

Background: Hip fractures are common among frail, older people and associated with multiple adverse outcomes, including death. Timely and appropriate care by a multidisciplinary team may improve outcomes. Implementing a team to jointly deliver the service in resource-limited settings is challenging, particularly on the effectiveness of patient outcomes. Methods: A retrospective cohort study to compare outcomes of hip fracture patients aged 65 or older admitted at Siriraj hospital pre- and post-implementation of the Fast-track program. The primary outcome was the incidence of medical complications. The secondary outcomes were time to surgery, factors related to the occurrence of various complications, in-hospital mortality, and mortality at month 3, month 6 and month 12 after the operation. Results: 302 patients were enrolled from the Siriraj hospital’s database between October 2016 and October 2018; 151 patients in each group with a mean age of 80 years were analyzed. Clinical parameters were similar between groups except the Fast-track group comprising more patients with dementia (37.1% VS 23.8%, p<0.012). In the Fast-track group, there was a signicantly higher proportion of patients underwent surgery within 72-hours (80.3% VS 44.7%, p<0.001) and the length of stay was signicantly shorter (11 days (8-17) VS 13 days (9-18), p=0.017). There was no signicant difference in medical complications and mortality. Stratied analysis by dementia status showed a trend in delirium reduction in both demented and non-demented groups, and a pressure injury reduction among patients with dementia after the program was implemented but there was no statistical signicance. Conclusions: The implementation of a multidisciplinary team for hip fracture patients is feasible in resource-limited settings. In the Fast-track system, time to surgery was reduced and the length of stay was shortened. Other outcome benets were not shown, which may be due to incomplete uptake of all involved disciplines. variables were analyzed by using Chi-square test. Fisher’s exact test was used for categorical data with a count of less than 5. All continuous data were tested for normality. Independent sample t-test and Mann-Whitney U test were used to compare continuous variables, depending on the data distribution. The pvalue < 0.05 was considered statistical signicance. Statistical analysis was performed by using SPSS for Windows version 18 software.


Background
Hip fractures are common injuries that result in loss of function, reduction in quality of life, an increase in morbidity and mortality in older people 1,2 . Globally, more than 4.5 million patients per year suffer from medical complications due to improper management of hip fractures. As society ages, post-hip fracture morbidity may affect up to 21 million persons in 2050 [3][4][5][6] . Male gender, older age, and multiple comorbidities are associated with an increased risk of death within the rst year after a hip fracture 7

. In
Thailand, there is an increase in the incidence of hip fractures in people aged 65 years old and older, which substantially increased short term and long term mortality.
Most hip fractures are more likely to occur in frail older people with several geriatric syndromes such as functional impairment, malnutrition and dementia 8, 9 . Consequently, those patients with complex medical comorbidities require more attention prior to undergoing operation. Moreover, they are more likely to be at risk of postoperative complications and prolonged hospital stays with their pre-morbid complexity.
After the existing literature with limited evidence from limited resource settings had been reviewed [27][28][29][30] , a multidisciplinary program titled "Acute Geriatric Hip Fracture: Fast Track in Siriraj Hospital" was initiated for older patients with hip fractures in September 2017. Although the primary target was to reduce the time to surgery, the ultimate goal was to improve patients' outcomes. The purpose of this study was to evaluate patient outcomes before and after implementation of the Fast Track for the Acute Hip Fracture program. The primary outcome was the incidence of medical complications. The secondary outcomes were time to surgery, causes of delayed surgery, short-term and long-term morbidity and mortality.

Study Design and Data Collection
All patients with hip fractures admitted to the Department of Orthopedics, Siriraj Hospital, Thailand, from October 2016 to October 2018 were identi ed retrospectively from the Siriraj hospital's database and the Orthopedic department's database. The hip fracture patients aged 65 years or older admitted and had a complete medical record were selected to be the subjects for the study. Patients who underwent elective surgery were excluded, and then the included population were divided into the PRE-fast track group and the Fast-track one according to the time of the program implementation.
Medical records were retrieved according to the ICD-9,10 (International Classi cation of Diseases 9-10th Revision) diagnostic codes for hip fracture (820.0-820.9 and S720-S722). Among 905 medical records identi ed, 803 were patients aged ≥ 65 years (Fig. 1). After having initially reviewed, medical records were sorted, according to admission number (AN) for both the PRE-fast track group and the Fast-track one until the sample size target was met.
All relevant medical records were reviewed to identify patients' medical comorbidities, premorbid functional status, interventions and complications occurring during hospitalization. With respect to delirium, the detection of delirium was performed by a clinical trainee in geriatric medicine. The patient was determined to have delirium if the medical record contained any documentation representing awareness of the syndrome, progress notes describing delirium or confusion, notes attempting to identify the causes of delirium, or notes describing any treatment to control delirium symptoms. Discharge summaries were reviewed to identify any documentation of the signs and delirium symptoms, as well.
Beginning in July 2016, all hip fracture patients admitted to orthopedic wards were recruited in the Siriraj Fracture Liaison Service (FLS). Patients in the FLS registry were followed from hospital admission until discharge and subsequent visits. Information regarding mortality and functional outcomes was obtained through electronic hospital records and data from the FLS registry.

PRE-fast track program
The orthopedic trauma team was responsible for the standard hip fracture treatment including pain control, basic preoperative assessment and scheduling surgery time. Surgery would be performed according to the availability of operative rooms and surgeons' regular work schedule. Consultation with the on-call medical teams and the geriatric team was available on request. Physical therapists assessed the patients when they were admitted to the orthopedic ward, but there was no standard protocol on mobilization or postoperative medical management (Table 1).

Siriraj Hip Fracture Fast track program (Fast-track program)
The Fast-track model is a multidisciplinary team of medical specialists and allied health teams including orthopedists, geriatricians, anesthesiologists, physical therapists and nurse coordinators. The Fast-track care commences on admission to the hospital. Acute pain service (APS), operated by anesthesiologists, aims to optimize pain control within the rst 24 hours by providing femoral nerve catheter blockade, and then customizing pain medications. Applying the orthogeriatric model of care as a framework, the geriatric team manages the patients within the rst 24 hours of admission until the patients were discharged from the hospital. Surgery is scheduled as quickly as possible, and spinal anesthesia is the preferred method. The Fast-track program aims to have all patients in the surgical operating room within 72 hours of admission. The program also aims to prevent medical complications (delirium, urinary tract infection, pressure injury, stroke, pulmonary embolism/deep vein thrombosis, myocardial infarction), provide nutritional counseling and supplements, achieve at least a sitting position on the rst postoperative day, and begin discharge planning early in the course of care (Table 1).
Results 302 patients were enrolled from the Siriraj hospital's database between October 2016 and October 2018; 151 patients in each group. The mean age of the total population was 80 years, and 43 (28.5%) of subjects were ≥ 85 years of age. Gender distribution, comorbidities, and the Charlson Comorbidity Index

Outcomes
The primary outcome was the incidence of medical complications. The secondary outcomes were the proportion of achieving the 72-hour time-to-surgery target, causes of delayed surgery, hospital mortality, mortality at month 3, month 6, month 12 after the surgery and function status (Barthel index) at postoperative day 4, month 3 and month 12.

Statistical analysis
For the comparison between the patient outcomes of the PRE-fast track program and those of the Fasttrack program, a sample size calculation was conducted by assuming the incidence of delirium in the conventional group of 53% and 37% in the intervention group 31 . With 80% power and a 5%, 2-sided level of signi cance, the estimated sample size was 151 subjects per group.
Baseline characteristics and related factors were analyzed by using descriptive statistics. Categorical variables were analyzed by using Chi-square test. Fisher's exact test was used for categorical data with a count of less than 5. All continuous data were tested for normality. Independent sample t-test and Mann-Whitney U test were used to compare continuous variables, depending on the data distribution. The pvalue < 0.05 was considered statistical signi cance. Statistical analysis was performed by using SPSS for Windows version 18 software.
(CCI) between both groups were not statistically different. There was a higher proportion of dementia in the Fast-track group (37.1%) compared to the PRE-fast track group (23.8%) (p = 0.012). The mean BMI for patients in the PRE-fast track group was 21.9 ± 4.14 and 22.0 ± 3.56 kg/m 2 in the Fast-track group (p = 0.783) in Table 2. There was no difference between groups in the hematocrit level, white blood cell count and the serum albumin level. More than 80% of patients had the inadequate vitamin D level in both groups.   On the 3rd post-operative day, more than 50% of patients were able to stand or walk in the Fast-track group compared to only 31% of the PRE-fast track group in Table 4. The proportion of participants with delirium at any point postoperatively was similar in both groups. The percentage of delirium in the PRE-fast track group and the Fast-track group were 34.0% and 31.0% respectively (p = 0.583) in Table 5. Other complications were similar in both groups. Importantly, there was a higher proportion of dementia patients in the Fast-track group that may have contributed to clinical outcomes. Therefore, a strati ed analysis by dementia status was performed, which revealed a nonsigni cant trend toward reduced delirium after implementing the Fast-track program among patients with dementia in Table 6.    The secondary outcomes were summarized in Table 7. The length of stay in the Fast-track group was signi cantly shorter (11(8-17) VS 13(9-18), p = 0.017). However, there was no signi cant difference between hospital mortality and long term mortality. Most patients in both groups were discharged to home. The readmission rates in both groups were all similar. The information about the patient's function (Barthel index) at day 4, month 3 and month 12 after the operation was collected. The Barthel index of patients in both groups were subsequently improved after discharging to home, but there was no difference in the Barthel index between group in Table 7.

Discussion
This study has demonstrated the outcome of implementing a multidisciplinary team for caring of hip fracture older patients in a large referral center university hospital in resource limited settings. The implementation was successful for accelerating the operation time and reducing length of stay. However, bene ts on patient-centered outcomes were not demonstrated in this analysis contrast to several existing evidences 13,30 . This nding might stem from several factors.
Outcomes including in-hospital complications after implementation of the Siriraj Hip fracture Fast Track System were analyzed. There was no statistical difference in the incidence of delirium between pre-and post-implementation of the program. Published studies reporting reductions in post-operative medical complications, delirium and in-hospital mortality mainly occurred in settings with routine geriatric consultation 10,12,13,17,19,21,31,32 . In Siriraj hospital, geriatric consultation was already common practice (96%) before the Fast-track program was implemented. This may have contributed to the inability to detect important differences following implementation of the program. Moreover, the overall incidence of post-operative delirium in the study was approximately 32%, lower than the 45% rate reported by a previous study in a similar Thai population 33 and lower than rates reported by studies in other countries 10,[34][35][36][37] . This might indicate that the existing care in the PRE-fast track era was already at a standard level of care for delirium reduction.
The prevalence of dementia in this study was signi cantly higher in the Fast-track group (37.1% VS 23.8%, p < 0.012) and higher than that reported by earlier studies 12,31,33,36 . Dementia has a signi cant impact on post-operative complications including delirium 36, 38 , so a post-hoc analysis according to dementia status was performed. The analysis revealed a non-signi cant trend of delirium reduction for both demented group and non-demented one after the program was implemented. The incidence of pressure injury was also lower in the demented population in the Fast-track group. This makes clinical sense because delirium and pressure injury may result from inadequate pain control and prolonged immobilization. The Fast-track program was designed to prevent these undesirable symptoms.
The incidences of pneumonia, urinary tract infection, and pressure injury were higher than those in the previous studies 13,17,39 and were not reduced after the program had been implemented. Early ambulation is considered one of the most valuable postoperative strategies to reduce postoperative pneumonia and pressure injury 26, 40 . Early ambulation was more common in the Fast-track group, but the statistically signi cant reductions in post-surgical complications were not observed. Barriers could possibly stem from the attitudes of involved healthcare staff, availability of therapists, or the patients' condition. A more comprehensive plan of barrier reduction from all involved parties with more practical strategies to apply the program might improve engagement and outcomes.
Moreover, during the rst few months of the Fast-track program, the Acute Pain Service (APS) was not fully operational and some patients might not obtain the optimal pain control. Complications such as delirium and inadequate rehabilitation were also more common before the APS was fully implemented.
Hip fractures in elderly people reduce short-term mobility and long-term functional outcomes 41, 42 43 . The functional outcomes was measured by using the Barthel index. Although, there were no signi cant difference in the Barthel index for all comparisons between groups but there was a trend toward better function in the Fast-track group. Moreover, most patients achieved improved functional outcomes over time and the majority reached full mobility at 1 year, which has been better results than other studies 44 . Discharge destination might contribute to mobility status, and most of the patients in this study were discharged back to home. In other studies, most of the patients were discharged to institutional care.
Debate on time to surgery for hip fracture surgery remains, not for the bene t of expedited surgery but rather the optimal time of how fast the time should be 23,45,46 . Most studies appears to indicate the 'optimal time' at 48 hours following the cutoff points utilized in meta-analyses 46

Strengths And Limitations
Our study has several strengths. Data for this study was collected from several sources, including a prospective registry (FLS). Data regarding related medical complications and possible interventions were explored to identify the gap of practice. The selection of consecutive cases would reduce the selection bias for this study. Limitations of the study include its retrospective design, which may have resulted in the under ascertainment of delirium; nevertheless, several measurements were taken to ensure that all cases would be identi ed.

Conclusion
We demonstrated the feasibility of implementing comprehensive multidisciplinary care team approach in a large referral center, teaching hospital with complex administrative structures in resource-limited  Figure 1