The worldwide prevalence of falls is 5186 per 100 000 people (James et al, 2020) and is one of the leading causes of injury for older adults (Berková and Berka, 2018). Falls are associated with a substantial increase in risk of mortality (McMahon et al, 2012) and are the second leading cause of unintentional death worldwide (WHO, 2021). A single fall can lead to reduced social participation (Pin and Spini, 2016), functional capacity (Laybourne et al, 2008) and self-confidence (Schoene et al, 2019). These broad impacts can lead to disability and long-term stays in nursing homes (McMahon et al, 2012).
As well as the individual impact, falls in older adults represent a substantial burden to the NHS. Between April 2020 and March 2021, the rate of fall-related emergency admissions in England for adults over 65 years was 2023 per 100 000 people (Public Health England, 2021). Injuries by falls account for over 4 million hospital bed days in England alone, and cost the NHS around £2.3 billion each year (Treml et al, 2010; National Institute for Health and Care Excellence (NICE), 2013a). Owing to this substantial individual burden and the subsequent financial ramifications, there is a need to identify interventions that are not just successful at reducing the risk of falls, but also cost-effective. This need forms part of a wider healthcare resource allocation decision-making process that balances healthcare systems’ limited resources, individual and collective health-related priorities, and the importance of cost-effectiveness evidence in facilitating the future roll out of the interventions. Previous reviews in this area have identified that exercise fall prevention programmes for older people are effective in reducing the rate of falls (Palmer et al, 2020).
A systematic review by Winser et al (2020) examines the cost-effectiveness of exercise-based programmes for falls prevention in older people. This review adds to the existing evidence base (Davis et al, 2010; Olij et al, 2018) and contributes to informing resource allocation decisions in budget-constrained healthcare systems.
Aim of commentary
This commentary aims to critically appraise the methods used within the review by Winser et al (2020) and expand upon the findings in the context of clinical practice.
Methods
The Methods and Results described in this and in the next section represent a summary of those presented in the original article (Winser et al, 2020). A robust multi-database search was carried out from inception to February 2019. Additional screening of reference lists of all included studies was undertaken. Randomised controlled trials (RCTs) that examined the cost-effectiveness of exercise-based prevention programmes to prevent falls in older adults (≥60 years old) were included. Studies that were not published in English, conference abstracts only or study protocols were excluded. Title and abstract screening were undertaken by a single reviewer. Full paper screening, data extraction and assessment of bias (Physiotherapy Evidence Database (PEDro scale) and Quality of Health Economic Studies (QHES) scale) were undertaken by two reviewers with arbitration by a third reviewer. The main outcomes of focus were the cost per quality-adjusted life year (QALY), incremental cost-effectiveness ratio (ICER), the benefit-to-cost ratio, and the incremental cost per fall prevented.
Results
After duplicate removal, 1033 studies were screened, from which 12 RCTs were identified. Using the PEDro scale, five RCTs were judged to be of good quality and seven of poor quality. Of the 12 included studies, eight had a high-quality economic evaluation as assessed using the QHES scale (Ofman et al, 2003). However, of the four studies that reported the interventions to be cost-effective, three had a poor methodological quality and one had both poor methodological and economic evaluation quality.
Studies were separated into two groups of exercise only (n=7) and multifactorial interventions (n=5). The multifaceted interventions included exercise plus risk management, occupational therapy and environmental modifications. The included studies adopted either a healthcare system (n=7) or a societal perspective (n=5). Cost-utility (CUA) (Robinson, 1993) and cost-effectiveness analyses (CEA) (Rutigliano, 1995) were the most common analytical frameworks (used by 10 studies). The most common primary outcome measures were the QALY in CUAs (n=4) and the number of falls during follow-up in CEAs (n=8). Of the 12 studies, four used the ICER to report the results, and only two studies had a duration longer than 12 months.
In total, seven studies reported the interventions to be cost-effective or potentially cost-effective (n=4 and n=3 respectively), and, the remaining five studies reported the interventions to not be cost-effective. Among the cost-effective or potentially cost-effective interventions, four were based on exercise-only and three were multifactorial. Based on the seven exercise only and multi-faceted studies classified to be cost-effective, a range of specific possible moderating factors were identified: these were exercise duration of 30 to 60 minutes, with frequency of once or twice weekly and minimal intervention lasting six months.
All the costs were converted by the authors into US Dollars (USD) 2018. For cost-effective exercise-only interventions, the cost of implementation ranged from 64.4 USD (McLean et al, 2015) to 533.5 USD (Robertson et al, 2001) per person. For cost-effective multifactorial interventions, the cost of implementation was estimated as 905 USD per person in one study (Rizzo et al, 1996). The incremental cost per fall prevented was estimated as 840 USD for an exercise-only intervention in one study (McLean et al, 2015) and 8824 USD for a multifactorial intervention in another study (Rizzo et al,1996).
Commentary
Using a modified version of the Joanna Briggs Institute Critical Appraisal tool for systematic reviews, this review scored 7 out of the 9 criteria (Aromataris et al, 2015). The two criteria that were not achieved were issues regarding methods of synthesis and the assessment and addressing of publication bias. Additionally, only one reviewer undertook screening of the study title and abstract, which may lead to relevant studies being excluded from the review. Furthermore, in the presentation of the results on moderating factors, there was incorrect referencing where studies deemed to be non-cost-effective were identified to be cost-effective (Iliffe et al, 2014). This review is of moderate quality and may provide a comprehensive summary of the results of the available studies that address the question of interest.
Caution needs to be applied when interpreting the cost-effectiveness findings, as issues have emerged. As recognised by Winser et al (2020), the included studies differ in terms of the perspective adopted (societal vs healthcare system), geographical contexts (such as North America vs Europe), comparators assessed (for example, exercise-only vs multifactorial interventions), type of costs included (such as direct vs indirect costs), outcomes analysed (for example, QALY vs falls averted) and methodological choices (for example, CUAs vs CEAs and whether discounting was applied or not). This clinical and methodological heterogeneity is not appropriately examined in the review. For example, minimal clinically important differences for specific outcomes (such as, how many averted falls are clinically significant) and clear-cut cost-effectiveness thresholds (such as, what is an acceptable cost per fall averted) are not discussed.
Additionally, the reporting of the results lacks clarity in some parts. For instance, in Table 5, which summarises the cost and cost-effectiveness of all included studies, the currencies and price years used in the included studies are not reported throughout, nor are the results from the sensitivity analysis. Moreover, the authors’ own calculations of the ICERs do not appear consistent (adjusted vs unadjusted incremental QALYs are used for different studies), and their interpretations seem at times different from the conclusions of the included studies (Davis et al, 2011a).
Furthermore, and perhaps most importantly, any demonstration of cost-effectiveness needs to be weighed against the quality of the underlying economic evaluation, as also recognised by the authors of the systematic review (Winser et al, 2020). In this sense, only three studies seemed to show cost-effectiveness (Davis et al, 2011a; 2011b) or potential cost-effectiveness (Isaranuwatchai et al, 2017) of the interventions and were also assessed as good quality by the QHES scale. However, two of these studies also showed poor methodological quality in terms of the PEDro scale (Davis et al, 2011a; 2011b), which reiterates how cautiously the cost-effectiveness findings need to be interpreted. Therefore, as the authors conclude (Winser et al, 2020), there is no clear-cut message as to which exercise-based fall prevention programme intervention is most cost-effective for preventing falls in older people, and the overall cost-effectiveness is still unclear. Subsequently, further research is required in this area to make any definitive decisions for informing the commissioning of these programmes. However, current NICE (2013b) guidelines recommend that community dwelling older adults who present with a history or potential to fall should be offered a multifaceted falls risk assessment. Where possible, this should be provided by a healthcare professional with appropriate skills and experience as part of a falls prevention service (NICE, 2013b).
Further research (economic)
Economic evaluations investigating the cost-effectiveness of interventions preventing falls in older people are needed to enrich the evidence base. New studies will clarify which interventions are more cost-effective, and which components are key determinants of cost-effectiveness. These economic evaluations should compare a wide range of alternative interventions across multiple settings with long-term study durations, potentially adopting modelling techniques to integrate analyses based on primary data. Also, as the authors of the review highlighted, studies set in developing and underdeveloped countries are lacking. Moreover, common metrics of assessment need to be adopted. When clinical outcome measures enter the estimation of a cost-effectiveness outcome (for example, cost per fall averted), clear thresholds are then needed to determine cost-effectiveness. The use of QALYs represents a positive example in this sense since, at least in certain jurisdictions such as the UK (NICE, 2013a), a specific maximum acceptable ICER on the cost per QALY gained is set to delimit the cost-effectiveness region. Alongside new primary studies, there is a need to update this systematic review (whose searches were run until February 2019) and synthesise newer and older findings, while avoiding the identified shortcomings.
Key points
- There is no clear-cut message as to which exercise-based falls prevention programme intervention is most cost-effective in preventing falls in older people
- The overall cost-effectiveness of exercise-based falls prevention programmes is unclear
- Wherever possible, exercise-based falls prevention programmes should be undertaken on the basis of robust cost-effectiveness analyses/clinical audits
CPD reflective questions
- What are the sources of heterogeneity in the included studies?
- What outcomes were investigated in terms of cost-effectiveness?
- Is there a clear-cut indication of what a cost-effective intervention to prevent falls in older people would look like?