Strategies to improve end-of-life decision-making and palliative care following hip fracture in frail older adults: a scoping review

Abstract

Background

Although surgery is the gold standard following a hip fracture, the potential for rehabilitation and survival rates are low in frail older patients. Some patients may derive more benefit from palliative care. The objectives of this review were to identify the available strategies to improve end-of-life decision-making and palliative care for frail patients with hip fractures and to synthetise their level of support.

Methods

We conducted a scoping review of the scientific and grey literature, searching seven databases and websites of associations. We included all study designs, expert opinion articles and clinical practice guidelines (CPGs). Data were synthetised according to the Approach to Patient with Limited Life Expectancy and Hip Fracture framework. The number of research items and their level of evidence were tabulated for each of the recommended strategies.

Results

Of the 10 591 items identified, 34 were eligible. The majority of included articles were original research studies (n = 15). Half of the articles and CPGs focused on intervention categories (55%) such as goals of care discussion and comfort care, followed by factors to consider in the end-of-life decision-making process (25%) and prognosis assessments (20%), mainly through the estimation of life expectancy. The level of evidence for these strategies remains low, given the limited number of prospective studies supporting them.

Conclusions

This scoping review highlighted that end-of-life care in frail older patients with a hip fracture remains understudied. The strategies identified could be prioritised for future research to improve the well-being of the target population while promoting sustainable resource management.

Key points

• Non-surgical management of hip fractures in frail older adults is a treatment option and should be discussed with patient-proxy.

• Assessment of life expectancy and quality of life are central elements to consider when determining whether surgery is indicated.

• Pain management and mobilisation are key aspects of care for patients with hip fractures who decide not to undergo surgery.

Background

Falls are the leading cause of hospitalisation among adults aged 65 years and over [1]. Osteoporosis weakens bone, making it more fragile and more likely to break on impact, and is among the pathologies that increases the risk of fracture following a fall in older patients [2]. As such, hip fracture is the admitting diagnosis in over 33% of hospitalisations following a fall in older adults [3, 4]. With the ageing population, it is estimated that the annual number of hip fractures will rise from 1.7 to 6.3 million worldwide by 2050 [5].

Although surgery is the treatment of choice for hip fracture repair and is available in many healthcare settings, it is not recommended for all patients [6–8]. The risks associated with surgery are greater in older adults with pre-existing comorbidities, as they may affect their recovery and survival prognosis [7, 9, 10]. For example, older adults with neurocognitive disorders (NCDs) are at greater risk of developing postoperative delirium, anaesthesia-related complications and mobility impairments [7, 11]. They are also more likely to experience a cognitive decline following surgery [7, 12]. A frailty syndrome can also add to existing comorbidities and adversely affect prognosis [9, 13, 14].

Postoperative complications and frailty syndrome are reported to significantly increase mortality in hip fracture patients [10, 15]. Mortality rates for older adults are estimated at 13% at 30 days, 21% at 6 months and 30% at 1 year after surgery [15–18]. Mortality rates after 6 months increase to 36% in previously institutionalised elders and to 55% in patients suffering from NCD [19]. Furthermore, among patients who undergo surgery and survive, only 50% will regain their initial level of autonomy [7, 20–22]. As such, nearly 20% of older adults who have undergone hip surgery will have to be relocated to long-term care [20, 21].

Considering the high mortality rate among frail older patients, as well as their low level of recovery following hip fracture, non-surgical (i.e. conservative) treatment and management by a palliative care team may be warranted in this population [7, 9, 19, 23]. Non-operative management combined with palliative care would provide comfort-first for targeted patients, as well as psychological, social and spiritual support towards a humane and dignified end-of-life [7, 14, 24]. However, to date, there has been no knowledge synthesis on what needs to be considered in the provision of palliative care for frail older adults with hip fractures and significant comorbidities. Therefore, the aims of this scoping review were to describe strategies for improving end-of-life decision-making and palliative care in older adults with hip fractures and to synthetise their level of empirical evidence.

Methods

We carried out a scoping review in accordance with the recommendations from the Joanna Briggs Institute [25]. This manuscript was written in accordance with the reporting guidelines from the Preferred Reporting Items for Systematic Reviews and Meta-analysis Protocols—Extension for Scoping Reviews (PRISMA-ScR) [26]. This review was also guided by the Approach to Patient with Limited Life Expectancy and Hip Fracture framework [23]. This approach makes it possible to determine which type of treatment is indicated, based on the patient’s prognosis after a hip fracture. In this regard, patients whose life expectancy is estimated in days or months should be assessed in terms of key factors (e.g. life goals, surgical risk, pain) when deciding whether or not to offer non-surgical treatment and palliative care. The protocol was registered in Open Science Framework (https://doi.org/10.17605/OSF.IO/WVZ3D).

A collaborative approach, involving multiple knowledge users (orthopaedic surgeons, geriatricians, physicians specialised in palliative care medicine, nurses) working in acute care and long-term care settings, was used. We defined end-of-life care as decisions made at the end of life that may have a potentially life-shortening effect and that are embedded-in and fostered by clinical, ethical, sociocultural, religious, political and economic concerns [27]. Palliative care was defined as health services dedicated to the relief of serious health-related suffering, be it physical, psychological, social or spiritual [28]. The threshold to determine whether studies focused on older adults was ≥65 years of age, which is the benchmark most commonly used in a context of hip fracture [29].

Search strategy and eligibility criteria

We developed a search strategy in collaboration with a health research information specialist and experts in orthopaedics and palliative care. The following databases were systematically searched: MEDLINE (Ovid), EMBASE (Ovid), CINHAL (EBSCO), PsycINFO (Ovid), Cochrane Central Register of Controlled Trials (CENTRAL), Web of Sciences (Clarivate) for scientific literature and ProQuest for grey literature from their inception to August 2023. The search strategy was validated according to the Peer Review of Electronic Search Strategy [30]. The Medline search strategy is outline in Supplement 1. We included all types of study designs and expert opinion articles.

In addition, we consulted 38 websites of orthopaedic associations and organisations publishing guidelines for improving orthopaedic care (Supplement 2) of high-income countries to identify clinical practice guidelines (CPGs) with recommendations on end-of-life decision-making and palliative care following a hip fracture. The list of associations and organisations was drawn up with knowledge users participating in the review who have in-depth knowledge of the orthopaedic ecosystem and by consulting articles. High-income countries, as defined by the World Bank definitions [31], were targeted, taking into account the differences in the access to specialised treatments that may exist with middle-income and low-income countries. Research protocols and conference abstracts were excluded. No language or publication year limits were applied. We used an open-access online translator (https://www.deepl.com/translator) for articles and CGPs not written in English or French [32].

Selection of eligible items

We used EndNote 20 (V.20.5, Clarivate, Philadelphia, 2023) to collect citations, remove duplicates and facilitate screening of titles and abstracts. Automatic and manual sorting was performed to remove duplicates. Pairs of reviewers (A.T., V.D.; M.B., M.A.G.) independently screened titles, abstracts and full-texts and two reviewers (A.T., M.B.) searched professional association websites to find relevant CPGs. All reviewers used a systematic and iterative approach to pilot test the process on a random selection of 200 eligible citations and five professional association websites for CPGs until acceptable agreement was achieved (kappa > 0.8) [33]. A third reviewer (S.P.) was consulted to resolve any disagreements. Reference lists of included publications were also manually reviewed to identify any additional eligible studies or CPGs.

Data charting

Two data extraction forms were created: one for studies and expert opinion articles and another for CPGs. These forms were pilot tested with the first five publications and with two CPGs; then, the data were extracted independently by two separate reviewers (A.T., M.B.). To optimise inter-rater reliability, regular meetings were held between the reviewers to check agreement until a consensus was reached (kappa > 0.8) [33]. The following information was extracted from scientific articles: authors and year, objectives, population, design, interventions (i.e. strategies), comparators, clinical outcomes and results (quantitative and qualitative). For each recommendation within CPGs, we extracted information on the characteristics of included CPGs: title and year, country, organisation, target users, population and quality of evidence. If important information was missing or unclear, we requested it by sending up to three emails to the first and last authors.

Data synthesis

We used the Approach to Patient with Limited Life Expectancy and Hip Fracture framework to synthetise data. This framework includes the following categories: (i) prognosis assessment, (ii) factors to consider in the end-of-life decision-making process (i.e. are life goals being met?, is pain well controlled?, is surgical risk high?, is the patient ambulatory?, is the quality of life poor?) and (iii) related interventions (e.g. goal-of-care discussions, advanced directives for medical decisions, comfort care, palliative care consultation) [23].

We synthesised the level of evidence for each strategy by classifying studies according to an adaptation of the Oxford Center for Evidence-based Medicine (OCEBM) [34]: randomised controlled trials (RCTs) or systematic reviews of RCTs (I), prospective studies or systematic reviews of RCTs and cohort studies (II), retrospective cohort studies, case–control studies, cross-sectional and case series studies (III) and expert consensus and others (e.g. case studies, editorials) (IV) [35]. CPG recommendations were also classified according to their level of empirical support (e.g. CPG recommendations supported by expert consensus were grouped with level IV studies) [35]. We then calculated the number of studies and CPG recommendations per adapted OCEBM category to illustrate the importance and evidence for each strategy [35]. The extraction and categorisation process were also reviewed by two other team members (SP, EB) with the relevant expertise to ensure accuracy.

Results

Literature search and selection

A total of 31 articles [7, 9, 14, 23, 24, 36–62] and 3 CPGs [63–65] were included (Figure 1). The main reasons for exclusion were wrong study population (n = 29), wrong target of care or lack of strategy suggestion (n = 6), conference abstract (n = 3) or data were unavailable after consulting the authors three times (n = 2) (Supplement 3).

Figure 1.

Flow diagram of included studies.

Study characteristics

As presented in Table 1, the majority of included articles were original research studies (n = 15), with prospective cohort studies (n = 6) [50, 53, 55, 56, 58, 61] and retrospective cohort studies (n = 4) [9, 38, 43, 45] being the most frequent. Where specified, the mean age of patients was always over 80 years. Quantitative studies had sample sizes ranging from 7 to 606 (mean = 157, median = 150) and qualitative studies from 17 to 45 (mean = 35, median = 35). Case studies (n = 3) [39, 41, 46], literature reviews (n = 6) [7, 14, 23, 24, 42, 51] and editorials (n = 7) [36, 44, 47, 52, 54, 60, 62] were also included when they recommended strategies. A summary of the characteristics of the original studies and clinical outcomes are presented in Supplement 4.

Table 1.

Characteristics of included studies and guidelines.

Characteristics	n (%) Total n = 34
Country
United States of America	15 (44.1)
Europe	16 (47.1)
Canada	2 (5.8)
North Africa	1 (2.9)
Type of publication
Original research article	15 (44.1)
Editorial	7 (20.6)
Narrative review	6 (17.6)
Case study	3 (8.8)
Guidelines	3 (8.8)
Design of original studies
Prospective cohort	6 (40.0)
Retrospective cohort	4 (26.6)
Qualitative	3 (20.0)
Case control	1 (6.7)
Survey of health professionals	1 (6.7)

The majority of studies and CPGs included were from the United States (44%) [14, 23, 24, 37–39, 43–47, 49, 52, 53] and Europe (47%) [9, 36, 40, 42, 51, 54–58, 60–65]. Strategies extracted from CPGs were exclusively based on expert consensus. The articles and CPGs were published between 1977 and 2023, with 50% published since 2020 (Figure 2) [24, 40, 43, 51–58].

Figure 2.

Years of publication of included items.

Strategies to improve end-of-life decision-making and palliative care

The eligible strategies, classified according to the categories from the selected framework, and grouped according to their level of evidence, are presented in Figure 3. The strategies most commonly recommended fell within the intervention category (n = 84, 55%), followed by those on factors to consider in the end-of-life decision-making process (n = 38, 25%) and those on prognosis assessment (n = 31, 20%).

Figure 3.

Evidence synthesis on end-of-life decision-making and palliative care strategies. *The level of evidence for each strategy was synthetised by classifying research items according to an adaptation of the Oxford Center for Evidence-based Medicine (OCEBM): RCTs or systematic reviews of RCTs (I), prospective studies or systematic reviews of RCTs and cohort studies (II), retrospective cohort studies, case–control studies, cross-sectional and case series studies (III) and expert consensus and others (e.g. case studies, editorials) and by calculating the number of items per adapted OCEBM category. NHFS, Nottingham Hip Fracture Score; PPS, Palliative Performance Scale.

Prognosis assessment

Half of the included articles and CPGs in this category concerned life expectancy estimates and had the highest level of evidence with one prospective study [50], two retrospective studies [9, 45] and one survey [40]. Tools such as the Nottingham Hip Fracture (NHF) [66] were used to estimate life expectancy, as well as predictions from a multidisciplinary team [9]. Assessment of functional status, cognitive function and frailty were strategies associated with a lower level of evidence, as they were mainly found in expert opinion articles. Functional status was assessed with the Instrumental Activities of Daily Living (IADL) scale [67] and the Activities of Daily Living (ADL) index [68], while frailty could be measured with the Clinical Frailty Score (CFS) or Rockwood Frailty Scale [69]. Finally, cognitive functions were measured with various tools such as the Mini Mental Score (MMSE) [70] and the Montreal Cognitive Assessment (MoCA) [71].

Factors to consider in the end-of-life decision-making process

When conservative management did not provide acceptable pain relief and mobility for basic care, surgery was recommended in half of the publications in this category. Nonetheless, only qualitative studies [37, 57] and expert opinions [7, 23, 24, 42, 44, 47, 52, 54, 60, 63, 64] supported this strategy. Deciding whether to perform surgery or not in view of goals of care and quality of life had a higher level of evidence, with two prospective studies [53, 56] that assessed patient quality of life.

Surgical risk assessment and whether to consider surgery solely to improve mobility was only reported in lower-quality studies (III–IV) [7, 23, 37, 39, 43, 44, 51, 52, 54, 62, 64]. For surgical risk assessment, strategies focused on evaluating patient comorbidities [23, 37, 44, 51, 52, 54], assessing pre-fracture functional status [39, 51] and using validated tools such as the Charlson Comorbidity Index (CCI) [72], the ASA Physical Status Classification System [73] or the Almelo Hip Fracture Score (AHFS) [7, 9, 43, 64, 74].

Interventions

Nearly half of the publications in this category focused on goals of care discussions. Involving a multidisciplinary team in these discussions was the most commonly cited strategy and was supported by the highest level of evidence, including three prospective studies [53, 56, 58], one retrospective study [43], two qualitative studies [37, 57] and nine expert opinion items [23, 24, 36, 39, 41, 52, 60, 62, 64]. The importance of discussing goals of care early on was also reported in publications with a lower level of evidence (IV) [23, 36, 37, 40, 51, 52, 54, 60, 62, 63].

A third of included articles and CPGs addressed comfort care. The most frequently reported strategy involved pain relief and mobilisation protocols, with 2 prospective studies [56, 61], 1 practice survey [40] and 10 expert opinion articles [14, 23, 24, 42, 44, 46, 47, 52, 54, 62]. In one of the prospective studies, a multimodal pain management approach consisting of opioids, breathing exercises and mobilisation techniques was used [56]. Training nursing staff on how to assess and manage pain in nonoperative patients and the use of nerve blocks as a pain relief strategy were also only reported in expert opinion articles and in one case–control study, respectively [24, 42, 44, 49, 52, 54, 57, 63, 64].

Finally, one-fifth of the publications in the intervention category discussed early palliative care team involvement. This is meant to help address end-of-life issues and support the family in the decision-making process, according to one prospective study [61] and other publications with a lower level of evidence (III–IV) [7, 23, 24, 37, 39, 40, 42, 44, 52, 62].

Discussion

The aims of this review were to outline strategies to improve end-of-life decision-making and palliative care in older adult patients with hip fractures, and to synthetise their level of empirical support. Most of the publications meeting the inclusion criteria were published over the last 3 years. The level of evidence for the strategies described remains low, given the limited number of prospective studies identified for each of them. One of the most empirically supported strategies is the estimation of life expectancy using tools assessing various patient dimensions, followed by quality-of-life assessments, and the early involvement of a multidisciplinary team to discuss goals of care and pain relief with a multimodal approach. Assessing life expectancy based on functional status or frailty alone, and deciding whether to offer surgery to relieve pain or optimise mobilisation, was mainly reported in expert opinion articles.

Estimated life expectancy is often assessed in combination with mortality prediction models [75]. In line with the results of this review, which support an assessment of the patient's overall condition, these models include characteristics such as age, sex and comorbidities, as well as pre-fracture living conditions and mobility [75]. The NHF Score (NHFS) [76] and the Holt model [77] are the two tools with the highest discriminative ability and calibration to predict mortality in patients at both low and high risk of 30-day and 8-year mortality following a hip fracture [75]. Nonetheless, despite showing consistent predictive abilities across studies, the fact remains that a large proportion of patients deemed at high risk of mortality are still alive at 30 days (i.e. 78.5%) and even at 1 year (i.e. 59.5%) post-fracture [75], thus limiting the usefulness of current predictive models when discussing end-of-life care in older patients with hip fractures. Nevertheless, it is worth mentioning that the mortality rates could have been lower in studies assessing predictive models because they were conducted in patients who underwent surgery and active treatments [75]. In this regard, a recent study found an 80% mortality rate at 30 days post-fracture in non-surgically treated frail older adults (body mass index <18.5, severe comorbidities and mobility problems) living in nursing homes prior to the fracture [56].

Assessing quality of life was one of the key elements identified in this review, to determine whether a patient should undergo surgery or not. However, the validity and inter-rater reliability scores of the tools commonly used in adults (e.g. EQ-5D-5L) have performed poorly in patients with moderate to severe NCD [78]. For these patients, the Quality of Life for People with Dementia [79] and the Quality of Life in Late-Stage Dementia have shown good psychometric properties [80, 81]. Therefore, these tools should be considered in the end-of-life decision-making process for patients with significant cognitive impairment following hip fracture.

In fact, the assessment of life expectancy and quality of life could be integrated into a shared decision-making process involving the interdisciplinary team, patients and families, as highlighted by the findings of this review. This approach offers support to patients and/or their family while they review treatment options with healthcare professionals, to make an informed decision [82, 83]. Patients, their families and healthcare professionals report a high degree of satisfaction with this approach, particularly when the patient suffers from NCD [84]. Given the challenges involved in gathering all the information needed to decide whether or not to operate on the patient at the time of the fracture, it is recommended for front-line healthcare professionals to initiate discussions on goals of care with frail, older adults and their families even before the injury occurs [52, 57, 59]. Following these discussions, directives on whether to hospitalise or not following a hip fracture could also be included in the patient’s medical record [56, 57].

Furthermore, two main issues when caring for hip fracture patients who decide not to undergo surgery were identified in this review: pain management and mobilisation. As reported in a large multicenter study [56] included in this review, greater pain was observed in non-operated patients than in operated patients when assessed with a validated tool for older adults experiencing NCD (Pain Assessment Checklist for Seniors with Limited Ability to Communicate [PACSLAC] [85]), up to several months after their fracture. Similarly, more than half the caregivers questioned about the quality of the period leading up to death in hip fracture patients rated symptom control, including pain, as poor or average [61]. Therefore, improvements in palliative care, which aims to avoid suffering by prioritising adequate pain assessment and treatment tailored to the patient's needs [28], appear to be required in frail patients with hip fracture.

Implications for future research

This review identified several potential research avenues to better guide multidisciplinary teams and improve the well-being of older adults with hip fractures and of their families. In this respect, studies on predictive mortality models and on their implementation that focus on non-operated patients are warranted to facilitate a shared decision-making process leading to palliative care. An approach to palliative care that adequately assesses and manages pain, while ensuring that patients with non-operated hip fractures can receive basic care without sacrificing their comfort, also remains to be designed and tested.

Study strengths and limitations

This review described strategies in a developing clinical field, namely choosing between surgery and palliative care in frail older adults with hip fractures. A rigorous and exhaustive method was used to identify relevant articles, as well as to synthetise strategies and their level of empirical support. This review was also carried out in collaboration with a multidisciplinary team, routinely involved with older adults suffering from hip fractures who are offered palliative care, which increased the likelihood of relevant data interpretation and recommendations for research avenues.

This review also has limitations. First, we did not analyse the effectiveness of the strategies and it is therefore not possible to draw clear conclusions regarding their potential clinical impact. However, the majority of studies included in this review, and the small number of studies included for each strategy, limited this kind of analysis. Second, our search strategy did not include keywords on clinical practice guidelines. It is therefore possible that some were not identified. However, a search of professional association websites remains the optimal way to identify relevant CPGs on a specific theme [86]. Likewise, given the low level of evidence for the strategies identified, it is unlikely that additional clinical practice guidelines would have changed the results of this review.

Conclusions

This review identified several strategies to consider in the end-of-life decision-making process and palliative care in frail older adults with a hip fracture. Nonetheless, the ideal approach for patients with limited life expectancy and a hip fracture remains largely understudied. The various strategies identified in this review could guide future research efforts, aiming to improve the well-being of the target population and ensuring a better allocation of resources regarding specialised surgical treatments.

Declaration of Conflicts of Interests:

S.P. received grants from Taketa Pharmaceutical Company. E.B. is on the advisory board of Pendopharm. S.P. and E.B. are on the Editorial board of the Orthopaedic & Traumatology: Surgery & Research journal.

Declaration of Sources of Funding:

A.T. received funding for this study from the following organisations: CHU de Québec-Université Laval Research Center, Faculty of Nursing of Université Laval, Fondation pour l’avancement et la recherche en orthopédie du Québec and Réseau québécois de recherche en soins palliatifs et de fin de vie. These funders played no role in the design, execution, analysis and interpretation of data, nor in the writing of the study.