HIP fracture Supplemental Therapy to Enhance Recovery (HIPSTER): a protocol for a randomised controlled trial


Hip fractures affect 1.6 million people across the world each year.1 Hip fractures predominantly occur in older people, with the average age ranging from 80 to 83 years.2 The ageing of the population will increase the number of people at risk of hip fractures and contribute to an increasing burden of care.3–6 The incidence of hip fractures is expected to rise to 4.5 million globally by 2050.7 In Asia, it is anticipated that by 2050, the direct hospital cost of hip fractures will increase by over 50%.8 Costs are primarily attributable to prolonged hospital stays, including inpatient rehabilitation, with the average total length of stay exceeding 30 days in many jurisdictions.9

Within the Australian healthcare system, acute care provides acute medical treatment by diagnosing and surgically managing the hip fracture. Subacute care is provided once acute medical treatment is no longer required and refers to healthcare services that provide rehabilitation to improve physical function and independence, palliative care, geriatric evaluation and management or psychogeriatric treatment.10 A key strategy to optimise a patient’s functional recovery following a hip fracture is rehabilitation, including early mobilisation delivered in the acute setting.11 12 Australian National Hip Fracture Registry data (2021) indicate that 90% of patients were offered mobilisation the day after surgery and 49% of those patients achieved mobilisation on this day.13 Previous research shows that completing physical rehabilitation guided by a physiotherapist on day 1 postsurgery is associated with better functional mobility on day 5 and on hospital discharge.14 Every foot walked was associated with a 0.57% decrease in the odds of a hospital complication (p=0.04).15 Despite the known benefits of early mobilisation, patients can still expect protracted physical recovery, with data indicating only 24% of patients regained their preadmission mobility level at 120 days postfracture.12 Early and intensive rehabilitation in the acute setting could provide opportunities to optimise recovery and reduce hospital days. However, current clinical practice guidelines do not include any recommendations on the intensity of physiotherapy in the acute setting for patients following hip fracture surgery.16 17

There is some evidence that intensive physiotherapy-led rehabilitation, delivered early following hip fracture surgery, may improve patient and health system outcomes. A randomised controlled trial comparing early and intensive physiotherapy to usual care (one time per day, 7 days per week) demonstrated improvement in patient outcomes.18 Intensive physiotherapy included two additional treatments per day, one delivered by a physiotherapist and one by an allied health assistant, providing additional opportunities to mobilise, practise functional tasks and progress towards independent ambulation. Intensive physiotherapy resulted in significantly better physical function at day 5 following surgery, a reduced length of stay in the acute hospital (median 6.0 days vs 7.4 days, p=0.08) and a reduced overall hospital length of stay (inclusive of acute and subacute hospital days, median 24.4 days vs 35.0 days, p=0.01).18 This clinical trial was conducted at a single site, so there is a need to test the efficacy of this intervention across a broader range of acute care settings. There is currently no evidence regarding the health and economic impact of intensive physiotherapy following hip fracture surgery or the barriers and enablers to its implementation. Such evidence will be critical to inform international clinical practice guidelines for the care of patients following hip fracture surgery.

The aims of the HIP fracture Supplemental Therapy to Enhance Recovery (HIPSTER) multicentre randomised controlled trial are to determine the impact of intensive physiotherapy following hip fracture surgery on1 total hospital length of stay2; clinical outcomes and3 healthcare costs in the 12 months following hip fracture, compared with usual care and4 evaluate barriers and facilitators to implementation of intensive physiotherapy in hospital settings, from the perspective of patients and clinicians.

Methods and analysis

Trial design

A two-arm randomised controlled, assessor-blinded trial will be conducted in eight Australian hospitals. The selected sites deliver usual care according to the Australian Clinical Care Standard,19 admit at least 175 patients per year with an isolated hip fracture and are representative of large-sized and medium-sized hospitals delivering acute care, including public and private facilities and metropolitan and regional sites. The trial design is depicted in figure 1. The study is designed according to the Standard Protocol Items: recommendations for interventional trials guidelines20 and was prospectively registered (ACTRN12622001442796). The study will be conducted in accordance with the principles of good clinical practice (GCP) and reported according to Consolidated Standards of Reporting Trials standards.21 The primary outcome is the total hospital length of stay. Twelve months of follow-up will capture data on healthcare utilisation for economic analysis (table 1).

Figure 1
Figure 1

Study design, time points and outcomes. mILOA, modified Iowa Level of Assistance Score; EQ-5D-5L, EuroQol five-dimension five-level; ICECAP-O, ICEpop CAPability measure for older people; short FES-I, short Falls Efficacy Scale-International and MBS/PBS, Medicare Benefits Schedule/Pharmaceutical Benefits Scheme.

Table 1

Schedule of trial follow-up and procedures


Participants will be randomised 1:1 to either the intervention group (intensive physiotherapy) or the control group (usual care). Randomisation (sequence generation and allocation concealment) will be performed using a secure online randomisation service managed by an independent agency. A permuted block randomisation schedule will be used with stratification for the recruitment site.


Clinical staff delivering usual care will be blinded to group allocation using methods successfully employed in our previous trial.18 The usual care physiotherapist will aim to deliver standard physiotherapy sessions in the morning. The intervention group physiotherapist and allied health assistant will provide sessions at a separate time of day. These sessions are not documented in the medical record. An intervention log, only accessible to the intervention team, will be completed to track intervention fidelity and maintain the blinding of the usual care team. The intervention log will record the number and duration of additional treatments delivered each day. All outcomes will be measured by an assessor who is blinded to group allocation.

The success of blinding will be measured using Bang’s Blinding Index after discharge from the acute hospital, at the completion of the modified Iowa Level of Assistance Score (mILOA) assessment and at the completion of a 12-month follow-up. Usual care physiotherapists and blinded assessors will be asked to guess the participant’s group allocation. These responses will then be used to create an index scaling −1 (opposite guessing) to 1 (complete lack of blinding), with 0 representing perfect blinding.22


To be eligible for inclusion, participants must be aged ≥65 years with an isolated subcapital or intertrochanteric hip fracture suitable for surgical management. In Australia, 97% of admitted patients receive surgery.13 Patients will be excluded if the fracture is subtrochanteric or pathological, if postoperative orders do not allow for full-weightbearing or weight-bearing-as-tolerated on the operated hip (relevant to 5% of patients)23 or if they are unable to mobilise (with or without a gait aid) prior to the fracture. Patients residing in an aged care facility will also be excluded, as their length of stay is usually determined by factors such as medical readiness for discharge and facility requirements rather than functional ability.


All patients admitted with an isolated hip fracture will be screened for inclusion. Informed, written consent will be obtained by a member of the research team. If patients are unable to provide informed consent (38% of patients with hip fractures have cognitive impairment),12 consent will be requested from the person responsible for their medical decisions.18 Demographic information collected will include age, sex, usual place of residence, carer status, home set-up (stairs at home) and healthcare funding (public, private or other including compensable bodies).

Participants in both groups

Participants in both groups will receive medical and surgical management according to best practice standards.19 Recruitment will occur prior to the first physiotherapy review and will not influence the timing of surgery. Allied health interventions, aside from physiotherapy (eg, occupational therapy), will be delivered according to usual practice at each centre, with all treatments recorded.

Usual physiotherapy care

Participants in both groups will receive physiotherapy according to usual practice at the site over 7 days or until discharge from the acute hospital. The treatment provided will be individualised and include exercises, regaining independence with transfers and mobility, as well as gait retraining. Discharge planning will follow local site policies and procedures.

Intervention group

Participants in the intervention group will receive two additional daily therapy sessions. One session will be delivered by an allied health assistant to practice the achievements of the usual care session. The other session will be delivered by a physiotherapist, with the aim of progressing the functional gains achieved in the usual care session as documented in the participant’s medical record, including increasing independence with mobility, progression of gait aids (eg, frame to crutches) and increasing walking distance. The intervention will be delivered for 7 days or until discharge from the acute ward, whichever occurs sooner. A 7-day intervention period has been chosen, as this represents the median length of stay in the acute ward in Australian hospitals.12

Data collection and follow-up

Data and follow-up schedule

Healthcare utilisation (including visits to health professionals and hospitalisations) and falls will be recorded in a diary by the participant or proxy for 12 months following surgery. Participants will be telephoned monthly to encourage diary completion and capture self-reported information on healthcare utilisation and falls. Hospitalisation and use of other hospital services will be confirmed by a medical record audit, while federal government-funded healthcare and resource use will be obtained using Medicare Benefits Schedule (MBS) and Pharmaceutical Benefits Scheme (PBS) data.


Outcomes will be assessed at baseline, day 7 postoperatively (at the acute hospital or the participant’s location), 6 weeks, 120 days and 12 months.

Primary outcome

The primary outcome is the total hospital length of stay for the index admission, including both acute hospital days and subacute hospital days. The length of stay will be measured in the number of days from the date of acute hospital admission to the date of discharge from the hospital (either acute or subacute) to the participant’s home or long-term residence. We will collect this data through a medical record review at each site. We have chosen the total length of stay because any hospital stay has cost implications and because our Phase II study demonstrated reductions in both acute and subacute length of stay,18 suggesting benefits for patients across each phase of the hospitalisation. Consumer engagement also identified length of stay as an outcome that matters to patients.24 25

Secondary outcomes

Clinical outcomes will be evaluated at 7 days, 6 weeks, 120 days and 12 months postoperatively to capture both early and late recovery and obtain data for economic analyses (figure 1). All of the following measures will be obtained by an assessor who is blind to group allocation.

Modified Iowa Level of Assistance Score

The mILOA is a measure of functional mobility assessed by a physiotherapist that is valid and reliable for use in patients following trauma.26 The mILOA is responsive to change following early rehabilitation interventions in acute care27 and reflects levels of independence in transfers and mobility, a high-priority outcome for patients following hip fracture.25 28 The mILOA will be collected on day 7 or at the time of discharge from acute care, whichever is earlier.

Health-related quality of life

Two preference-based measures will be used. The EuroQol five-dimension five-level (EQ-5D-5L) is a validated generic quality of life measure which is used to estimate health benefits in terms of quality-adjusted life-years (QALYs). Australian preference weights will be used to calculate utilities. Where a participant is unable to provide a response to the EQ-5D-5L, a proxy response will be sought29 from either the person responsible for the participant’s medical decisions or a caregiver who is able to provide responses. The ICEpop CAPability measure for older people (ICECAP-O) is a generic quality of life measure which measures perceived capability in older people and has been validated in older Australians following hip fractures.30 It is recommended to incorporate capability-based measures of well-being into economic evaluations, where a focus on physical functioning alone may provide an incomplete evaluation of the benefits of an intervention.31 Concerns about falling will be measured using the short Falls Efficacy Scale-International, a high priority for patients following hip fracture.25 28 It is valid and reliable in patients following hip fracture aged ≥65 years.32

We will record the proportion of participants remaining in hospital at 6 weeks postsurgery and the proportion of participants that return to preadmission mobility (yes/no) determined by walking ability, current residence, survival and reoperation rate at 6 weeks, 120 days and 12 months. Discharge destination from the acute ward will be collected at the time of discharge through a medical record review and confirmed using data from the Australia and New Zealand Hip Fracture Registry (ANZHFR).

The number of adverse events

The number of adverse events, as defined by GCP standards and recorded in our Phase II trial,18 will include hip-related (eg, wound infection and hip fixation failure) and non-hip-related (eg, chest infection and acute kidney injury) in-hospital and out-of-hospital adverse events during the 12-month follow-up period. These will be collected via phone interviews, monthly diaries and a medical record review. A combination of fall calendars and active asking has been shown to capture falls more comprehensively than either method alone.33

Healthcare utilisation during 12-month follow-up

We will confirm healthcare utilisation from monthly diaries and phone calls, medical record reviews and MBS and PBS data. We have demonstrated that this combination of methods results in good diary completion rates34 with the identification of community-based and non-medical encounters (allied health and nursing) that are not captured in administrative data sources.35

Sample size

Sample size calculations are based on the primary outcome (total hospital length of stay for the index admission, including days in acute and subacute care). In our Phase II trial, the median total length of stay was 24.4 days for those receiving intensive physiotherapy versus 35 days for those receiving usual care (30% reduction).18 This trial will be powered conservatively to detect an effect of half this size (15% reduction, equating to a 5-day reduction in length of stay).

Length of stay data in our Phase II study were positively skewed, as is typical for this outcome. Data were therefore transformed (natural log (lnLOS)) for power calculations. Using transformed data, the mean lnLOS was 2.98 (SD 0.80) and 3.41 (0.74) for High Intensity Physiotherapy for hip fractures (HIP4Hips) and usual care, respectively. To detect a reduction in log length of stay of 0.19 (equivalent to approximately 5 days), assuming an SD of 0.8 with 80% power and a two-sided 0.05 significance level, a total of 556 participants are required (278 in each group). Allowing for 10% attrition, we will recruit a total of 620 participants.

We will also be able to detect differences in the secondary outcome (number of participants remaining in hospital at 6 weeks following admission). The proportion of participants remaining admitted at 6 weeks in our Phase II trial was 0.29 and 0.54 in HIP4Hips and usual care, respectively. Assuming a more conservative reduction (0.30 vs 0.45), with 90% power and a two-sided 0.05 significance level, a total of 428 participants are required (214 in each group) to detect a difference in the number of participants remaining admitted at 30 days.

Statistical analysis

The primary outcome (total hospital length of stay) of hospitalisation and use of other hospital services will be confirmed by a medical record review, which is expected to be positively skewed and so is likely to be log transformed. Outcomes will be analysed using linear mixed models, with the main explanatory variable being the intervention and the site fitted as a random effect. Participant characteristics that could plausibly affect the primary outcome, such as age and gender, will be considered for inclusion as covariates. The probability of remaining admitted at 6 weeks (and other data based on binary outcomes) will be analysed using logistic regression. The analysis will be by intention-to-treat. A per-protocol analysis will be undertaken, including participants who received their designated therapy for at least 4 days in the acute hospital. The designated therapy for usual care will be physiotherapy, as usually provided at that site. The designated therapy for the intervention group will be usual care physiotherapy as provided at that site, plus at least one additional therapy session. The level of significance will be set at p<0.05.

Economic evaluation

Cost comparisons

A comparison of per-person costs associated with the delivery of the intervention for the two groups. Intervention costs will include staff inputs by duration, type and resource use.

Cost-effectiveness analysis

An incremental cost-effectiveness analysis from a societal perspective will be undertaken to compare the differences in healthcare costs between the two groups with the differences in QALYs over the 12-month follow-up period.

A utility index will be calculated from the EQ-5D-5L by applying a ‘social tariff’, which enables the estimation of health benefits in terms of QALYs. The conversion to QALYs will be based on the assumption that the duration of each status is exactly one-half of the time between two measurement intervals.36 Participants who die or are censored during the follow-up period will be included in the analyses, with their total costs weighted by time and a zero value for the 12-month follow-up EQ-5D-5L score imputed to allow calculation of QALYs. Multiple imputations will be used to address missing data. Group values and differences between groups for QALYs will be assessed using general linear models, controlling for respective baseline values.

Health system costs will include Health Services Australia data (incorporating MBS and PBS) and costs associated with hospitalisations, which will be valued at standard Australian Refined Diagnosis Related Group cost weights. Where able, the provision of other healthcare services will be valued as reported by participants or at standard rates if required. Personal costs will include out-of-pocket expenses associated with access to healthcare. Group values and differences between groups for costs will be assessed using a generalised linear model with an appropriate choice of distribution.

Results for the primary analysis of cost-utility (total cost per QALY) will be expressed as bootstrapped point estimates (mean differences between groups) and plotted on the cost-effectiveness plane. A cost-effectiveness acceptability curve will be used to represent uncertainty, depicting the probability that intensive physiotherapy is cost-effective relative to usual care for a range of alternative values for willingness to pay for a QALY.

Sensitivity analyses will incorporate analyses from a health system perspective as well as an evaluation of capability-adjusted life-years using the ICECAP-O to examine the broader economic impacts of the intervention on the health system and personal costs.

Implementation evaluation

A detailed protocol for the implementation evaluation will be published separately. In brief, a theory-informed implementation evaluation will be embedded in the trial, examining barriers and facilitators to the implementation of HIPSTER. The embedded process evaluation will follow the UK Medical Research Council recommendations on process evaluation of complex frameworks.37 Offering intensive physiotherapy to patients following a hip fracture requires practice change and resourcing adjustments; thus, understanding the clinician and patient experience is critical to guide future implementation. We will also identify any changes to usual care over the course of the trial, including changes to clinician behaviours. Barriers and facilitators will be mapped to the Theoretical Domains Framework to better understand the drivers of behaviour change. The Reach Effectiveness Adoption Implementation Maintenance (RE-AIM) framework will guide implementation evaluation using a mixed-methods approach.28

These data will inform the development of an implementation toolkit that will be tested for feasibility at sites naive to HIPSTER and used to facilitate sustainable broader adoption and implementation if intensive physiotherapy is found to be effective and cost-effective.

Ethics and dissemination

This protocol (V.4, 10 March 2023) received approval from the Alfred Hospital Human Research Ethics Committee. Local research ethics committee approvals have also been obtained at sites where the National Mutual Acceptance Scheme does not apply. Local governance approval was obtained from each site prior to the commencement of recruitment and data collection. Any protocol modifications will be submitted as a formal ethics amendment and approved changes will be communicated accordingly. All participants in this trial will receive usual care, and no participant will receive less physiotherapy care as a result of study participation. Potential participants at each site may be in a dependent relationship with physiotherapy staff. For this reason, consent will be obtained by an independent researcher or health professional who is not directly involved in the patient’s clinical care. If the participant is unable to consent to surgery (eg, premorbid dementia and other medical comorbidities), the person responsible (who consented to the surgical procedure) will be approached for study consent. Our previous Phase II study18 found no difference between groups in pain scores or opioid equivalence scores, indicating that participants tolerated the intervention well. Additionally, there was no increased risk of falls or other adverse events between the two groups.18

All trial findings will be published in peer-reviewed journals and submitted to national and international conferences for presentation. The authorship of any publications will adhere to the guidelines established by the International Committee of Medical Journal Editors. Results will be disseminated to participants and carers.

Safety monitoring and reporting

Each site will collect and report all hip-related and non-hip-related adverse events according to GCP standards. Health professionals delivering the intervention will cease the intervention if clinically indicated and liaise accordingly with the medical team, as per usual care in their hospital. The trial coordinator will report all adverse events to the Data Safety Monitoring Board (DSMB). This board will consist of one orthopaedic surgeon, an academic physiotherapist and a geriatrician. A biostatistician will be included if requested by the other DSMB members. All members are independent of the investigator team. The DSMB will meet at least twice a year.

Data management

Hardcopy information will be stored in a locked filing cabinet within a locked office at each site, with access restricted to named researchers. Electronic trial data from sites will be collected and managed using an encrypted and password-protected online tool, Research Electronic Data Capture, hosted and managed by Monash University. No identifying information will be stored in the online database. Site investigators will only have access to data relating to their site.

Patient and public involvement

Consumer engagement undertaken by our group24 and others25 found that patients following orthopaedic trauma and hip fracture identify return to their previous place of residence as a critical goal and desire a shorter stay in rehabilitation. Patients perceived that the most important indicators of recovery were stable mobility without falls or fear of falls and the capability to undertake valued activities of daily living.28 While early mobilisation was reported as challenging initially, pain was perceived as manageable.28 This is consistent with the modest pain scores in our Phase II trial, with no difference between HIP4Hips and usual care.18 The trial will involve one consumer representative, who will join site representatives on the Physiotherapy Intervention Committee. They will provide feedback on study processes and review patient-facing documents such as participant information and consent forms. Dissemination of the results through ANZHFR, social media and professional avenues will ensure the findings are broadly distributed with the aim of improving care and outcomes for patients following hip fracture surgery.

Trial status

Trial recruitment began in January 2023 and is ongoing.


The ageing of the population and the growing burden of hip fractures highlight the need for new ways to optimise rehabilitation outcomes and return patients to independent living. This will be the first multicentre randomised controlled trial examining the clinical benefits and healthcare costs associated with providing intensive physiotherapy in comparison to usual care for patients following hip fracture surgery. The current study builds on our Phase II trial, which demonstrated a significant decrease in total hospital length of stay for those who received intensive physiotherapy in the first postoperative week.18 ,25. The National Institute for Health and Care Excellence Clinical Guideline for Management of Hip Fracture in Adults16 states that studies investigating the clinical and cost-effectiveness of intensive physiotherapy are a research priority.38 Patients report that regaining physical function, mobility and independence are the most important outcomes of rehabilitation following a hip fracture.25 This trial will therefore address outcomes of importance to patients, health professionals and the health system.

In order for this study to determine the true impact of intensive physiotherapy on length of stay, it is important that discharge decisions are based on the hospital’s usual process to determine readiness for discharge and are not influenced by knowledge of group allocation. Blinding health professionals providing usual care is a particular challenge in trials of physical interventions. For our Phase II trial, we developed an effective method to blind the treating clinical team to group allocation by ensuring the intensive physiotherapy intervention was delivered at a separate time of day from usual care, with no documentation in the medical record. This method of blinding was effective, with physiotherapists delivering usual care correctly identifying group allocation in only 56% of participants (kappa=0.11, minimal agreement) and no evidence of contamination of usual care.18 As a result, for our Phase III trial, we have elected to conduct a patient-randomised trial rather than a cluster-randomised trial, which would have required a larger sample size. However, the potential risk of contamination of usual care with this approach must be acknowledged. We will employ our established methods for separating the delivery of the intensive intervention from usual care in the current trial, along with blinding all outcome assessors, another critical aspect of robust trial design when testing physical interventions. We will evaluate the success of blinding using the Bangs Blinding Index.22

This trial will provide important evidence to inform future guidelines and recommendations for physiotherapy following hip fracture surgery. If intensive physiotherapy is successful in improving patient outcomes, it will inform a key change in the acute postoperative management of hip fracture, provide greater opportunity for patients to achieve what is important to them and decrease the burden on the healthcare system.

This post was originally published on https://bmjopen.bmj.com