Addressing obesity in the management of knee and hip osteoarthritis – weighing in from an economic perspective

Obesity poses one of the greatest contemporary global public health challenges. Worldwide, the number of overweight or obese individuals has more than doubled since 1980 [1]. With rising prevalence rates in adult populations, the consequential growth in obesity-related chronic disease, including osteoarthritis (OA), is inevitable. Globally, the prevalence of OA, particularly of the large weight-bearing joints such as the knee and hip, is also predicted to grow [2], spurred on as a result of an ageing population but also an ageing population that is getting heavier [2].

There is well established evidence associating obesity (BMI ≥ 30 kg/m² [3]) and OA, with obesity being identified as a major but modifiable risk factor for both OA disease incidence and progression [4‐6]. The issue with obesity and OA is that the two conditions often coincide, working synergistically to perpetuate poor function and a greater likelihood of sedentary lifestyles which inevitably lead to higher levels of disability and a reduction in quality of life [7]. Furthermore, the need for total joint arthroplasty (TJA) surgery for the treatment of severe OA in knees and hips is more likely to arise in the obese [8] and earlier in life [9]. Changulani et al. 2007 reported that patients with a BMI of >35 kg/m² who were treated by knee replacement were, on average, 13 years younger than their normal weight counterparts; the authors alluded to the potential implications of younger age for lifetime management of OA [9]. Wang et al. 2013 also highlighted that weight gain and persistent excess weight from early adulthood increased the risk of TJA of the hip and knee for OA [10].

Knee and hip OA is one of the leading causes of global disability and was ranked as the 11th highest contributor to global disability in the most recent Global Burden of Disease study in 2010 [2]. As such, OA as a disease results in large indirect costs to society, mostly driven by impacts on productivity [11]. These indirect costs, when coupled with rapid growth in direct healthcare related costs associated with management of the disease and specifically TJA as a major cost driver [12], mean that the economic burden is huge. Furthermore, it is likely to be exacerbated by the aforementioned growing obesity and OA prevalence. Despite TJA being a very effective and cost-effective treatment option for end stage hip and knee OA, 2013 data from the Organisation for Economic Co-operation and Development (OECD) highlights continuing growth in the number of procedures being carried out with faster growth in knee replacements (TKA) particularly in countries with higher rates of overweight and obesity such as Australia, USA and the UK. It has been estimated that in 2007 in Australia, allocated health care expenditure on OA alone was approximately AUD 2 billion [13]. However, given that obesity rates are likely to continue to rise, this will potentially lead to substantial increases in the prevalence of OA and an even greater burden on health care expenditure [13].

There is now a substantial body of evidence focusing on the relationship between obesity and OA from a variety of perspectives. This includes investigation of the causal relationships of obesity and OA through biomechanical [14], physiological [15] and inflammatory mechanisms [16] and quantification of the impact of obesity on OA outcomes [14]. Attention has also been afforded to the benefits of weight loss in OA populations including a systematic review which concluded that a 10 % reduction in body weight is likely to have positive clinically meaningful effects on OA symptoms such as pain and disability [17]. A recent review focused on addressing obesity in knee OA, identified 9 randomised controlled trials of weight loss interventions in people with knee OA. It concluded that there are several strategies likely to be successful in the management of knee OA [18]. However, the evidence base is far from definitive in terms of what is the most efficacious type of intervention, who it should specifically target and when along the treatment pathway it should be offered. There are the usual challenges of whether obesity interventions can achieve long term weight maintenance and prevent weight regain [19], particularly in a population where OA symptoms may hinder physical activity efforts. There is also some concern that interventions that target weight loss alone may lead to muscle weakness and some bone density loss [20], meaning that an intervention that targets both weight loss and appropriate physical activity simultaneously is likely to be more favourable. This is reflected in current treatment guidelines recommending that weight maintenance and exercise in combination are suitable for managing OA symptoms [21, 22], although, in reality, it is often difficult to determine compliance with these recommendations.

Despite this apparent research momentum, it is not clear what the contribution of research from a health economics perspective has been in terms of supporting the use of obesity interventions in the prevention or management of OA. The discipline of health economics is fundamentally concerned with the allocation of scarce healthcare resources in an environment of competing demands with the goal of maximising a society’s welfare in the process [23]. Health economists typically carry out two different types of studies designed to achieve different purposes. Firstly, they describe and predict the economic impact of a disease or risk factor in order to quantify its cost burden (commonly known as cost of illness studies); secondly, they evaluate the incremental costs and benefits of alternative options to current practice (cost-effectiveness studies) [24] to inform resource allocation decisions. The latter essentially addresses the main objective of economics per se, that being efficiency, or maximising the benefit from available resources.

The goals of this study are two-fold. Firstly, it aims to identify and review the published economic evidence that is focused on describing and quantifying the economic impact of obesity in OA populations. Secondly, it reviews economic evaluations of interventions designed to explicitly address obesity in OA populations in order to prevent the onset of OA or improve OA clinical outcomes and potentially health related quality of life (HrQoL), to determine which interventions represent value for money. In doing so, the study also highlight gaps in the literature and identifies future priorities for this burgeoning area of research.

Figure 1 identifies all steps in the search process and the resulting studies extracted. In the first instance, 2048 citations were retrieved from searched databases (1997) and other sources (51). After removal of duplicates and irrelevant citations, 215 abstracts of candidate articles were screened of which 71 full text articles were assessed for eligibility. The search yielded 12 costing studies (Table 1) and one full economic evaluation (Table 2).

Despite a growing body of literature highlighting the importance of addressing obesity in OA populations, this review highlights the current paucity of economic evaluation research conducted in this area. Therefore, it remains unclear as to whether it is efficient to address obesity explicitly for the prevention or management of OA, how to best do it, which population to target, and when along the disease spectrum this should occur.

Nearly all studies identified were cost burden studies in TJA populations conducted to assess whether obesity leads to higher healthcare sector costs. It should also be noted that all of identified the studies related to OA in knees and hips, despite the search of OA being much broader. All of the costing studies adopted the narrow perspective of the individual healthcare setting from which the main cost data were derived; therefore, they did not account for costs associated with the whole continuum of care for TJA and only provided a snapshot of the potential economic impact compared to if a societal approach had been taken. With greater streamlining of acute health care clinical pathways for joint replacement recipients, some of the traditional post-operative costs have shifted away from the acute care setting and onto inpatient rehabilitation, outpatient, community and ambulatory care settings as well as patients and families themselves. Ideally, resource use should be tracked throughout the whole episode of care, using database linkage to gain a broader understanding of the costs associated with the presence of obesity in TJA populations. A Canadian study by Tarride et al [44] which tracked total healthcare resource use in both hospital and ambulatory care settings (though not OA attributable costs per se), demonstrated that healthcare costs per year to OA patients were nearly double those of non OA control participants. A subsequent finding that was not the primary objective of the study was that costs were more than triple for obese OA patients compared to non-obese controls; this highlights that the co-existence of obesity and OA possibly augments the excess financial burden of OA.

Not all of the costing studies reviewed concluded that higher costs in TJA populations were due to obesity per se. However, the more recent studies with stronger methodologies trended towards showing significant differences in healthcare costs between the severely obese compared to other BMI categories, which suggests that there are potential health care cost savings to be realised if weight loss in the severely obese is addressed prior to TJA surgery. Yet focusing only on the cost side of the equation simply identifies a potential problem, but does not address issues of technical (“how to do”) or allocative efficiency (“what to do”) in terms of choosing alternate healthcare programme options; these can only be addressed by economic evaluation studies. Similarly, establishing and subsequently investing in interventions with proven efficacy alone without consideration of the cost implications will inevitably drive healthcare spending upwards [45]. For example, there is increasing interest in the use of laparoscopic adjustable gastric banding surgery, for severely obese (BMI ≥ 35 kg/m ²⁾ to achieve rapid weight loss prior to TKA [46] given some evidence of negative health outcomes post TKA for these patients such as higher rates of infection [47], higher rates of revision [47], and worse pain and function post operatively [48]. Lap band surgery is highly efficacious [49, 50] and cost-effective [50, 51] as a weight loss intervention in the severely obese but it is also a very costly intervention. It remains to be seen whether the incremental cost of adding this procedure prior to TJA in this population is outweighed by the incremental benefits achieved through cost offsets and improved health outcomes including health related quality of life gains immediately post-surgery and over longer time periods. It is certainly worthy of further investigation however, given evidence of the association of pre-operative obesity predicting additional weight gain post TJA [52] which is only likely to further precipitate the ill effects of OA and obesity in combination.

Only one economic evaluation was identified in this review [38]. This showed that despite a small but growing number of efficacy trials that have been conducted to assess the impact of interventions to address obesity in OA populations [18], published cost-effectiveness studies are still lacking as in many areas of obesity research. The Sevick et al study, provided some evidence that weight loss and exercise in combination was most cost-effective in improving pain and function in overweight and obese patients with moderate knee OA compared to only exercise or weight loss independently. Yet the study provides no indication of the intervention’s longer term cost-effectiveness beyond the 18 month trial duration. It would be of particular interest to a range of stakeholders (clinicians, funders and patients) to know if the intervention was able to slow disease progression, control symptoms and maintain function for longer in persons with obesity and OA and subsequently delay their demand for more intensive and costly healthcare that coincides with higher levels of disability. Longer term cost-effectiveness studies however are reliant on either longer duration efficacy studies being conducted or, alternately, extrapolation of outcomes through modelling; the latter can be very informative but are dependent on assumptions especially in relation to maintenance of program effect. It would also be interesting to see how changes in OA symptoms as a result of an intervention manifests in terms of the quality of life impact using a utility-based HRQoL measure that can be used in economic evaluation to produce a cost per Quality Adjusted Life Year (QALY). QALYs provide a common generic measure of benefit taking into account both mortality and morbidity impacts and importantly, within this context, may capture benefit from weight loss and exercise that extends beyond that of the effect on immediate OA related outcomes. This metric also facilitates comparison of the cost effectiveness of interventions within and across disease areas. Even if studies have not incorporated a generic utility based HRQoL measure, there are reliable mapping techniques available. For example mapping techniques have been used to map disease specific measures such as the WOMAC to HRQoL scores on measures such as the EQ-5D [53] and the Assessment of Quality of Life (AQoL) [54] for Australian populations, although there are some limitations to the applicability of such mapping exercises [54].

With only one published cost-effectiveness study to inform policy and practice in this area, it is evident that there is a need for more economic evaluations of interventions to be conducted. This is inherently reliant on a growing evidence base of the efficacy of such interventions which ideally would incorporate economic evaluation alongside such efforts. Modeling could then be utilised to predict longer term cost-effectiveness over the life course and OA disease progression; this would require explicit statements of the assumptions about maintenance of program effect and their testing in sensitivity analysis. Economic evaluation in addressing obesity in OA should ideally adopt a societal perspective given that both costs and benefits may be incurred beyond the heath-care sector and where possible, an incremental cost per QALY should be reported to provide comparison across disease areas.

The notion of intervening to alter disease progression, delay, avoid or reduce complications post TJA by addressing obesity explicitly is, prima facie, likely to have both downstream economic and health benefits to healthcare systems, individuals and society as a whole. In practice, guidelines identify the need to focus on obesity in the management of OA, however it is evident that there is a lack of data both from an efficacy and cost-effectiveness perspective to support and to guide to whom, how and when this should be conducted and whether it should be explicitly incorporated into clinical pathways and public health practice. More research is therefore warranted in order to guide clinicians, funders, decision makers and patients alike.

Overall, this review has demonstrated that while the presence of obesity in OA populations, and in particular severe obesity, has been shown to have significant health and economic impacts, there is a dearth of evidence from an economic perspective to guide resource decisions about whether addressing obesity explicitly is considered value for money.

AQoL, assessment of quality of life; AUD, Australian dollars; CEA, cost-effectiveness analysis; HRQoL, health related quality of life; ICD-9 M, international classification of diseases – 9^th revision; LOS, length of stay; NIS, Nnationwide inpatient sample; OA, osteoarthritis; OECD, organisation for economic cooperation and development; THA, total hip arthroplasty; TJA, total joint arthroplasty; TKA, total knee arthroplasty; USD, United States dollars; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index