Changes in health-related quality of life (EQ-5D) dimensions associated with community-based musculoskeletal physiotherapy: a multi-centre analysis

The aim of this study was to determine how community-based physiotherapy treatment for musculoskeletal injury was related to change in the five dimension scores of the EQ-5D HRQoL outcome measure. A main finding was that the pain/discomfort dimension of the EQ-5D outcome reduced across all symptom locations, but this reduction did not equate to an improvement in EQ-5D index for all patients. Therefore, during a course of physiotherapy, pain reduction often occurs but EQ-5D will not improve in total score. The anxiety/depression dimension mean score improved in patients that showed an increase in EQ-5D index of greater than 0.1, suggesting that anxiety and depression could be a key dimension and status linked to the potential to improve EQ-5D and hence HRQoL.

EQ-5D responders had higher pre-treatment pain/discomfort (2.54–2.97) than EQ-5D non-responders (1.89–2.44), but showed reduced post-treatment pain/discomfort (1.27–1.51) when compared to EQ-5D non-responders (1.76–2.23). Of all five dimensions, pain/discomfort has been reported as being the most frequently problematic dimension across a range of musculoskeletal conditions [18]. The fact that community-based physiotherapy was associated with at least small reductions in the pain/discomfort dimension in all patients, regardless of whether they reported overall improvements in HRQoL is encouraging. However, this would need to be confirmed via a randomised controlled trial using validated joint-specific pain outcomes.

The anxiety/depression dimension reduced in EQ-5D responders, with large effect magnitudes in symptom locations except the shoulder (moderate effect). EQ-5D non-responders reported no change in anxiety/depression, with trivial effect magnitudes across all symptom locations. Pre-treatment anxiety/depression dimension scores were higher in EQ-5D responders than EQ-5D non-responders. Following treatment, the anxiety/depression dimension was lowest in EQ-5D responders. These data suggest that if patients present with higher levels of anxiety/depression, they are, in fact, more likely to show improved HRQoL than those who present with less anxiety/depression. This finding contradicts the findings of Bergbom et al [19], who observed a lack of response to treatment in patients who presented with depressed mood. Previous studies have also shown high levels of depression to be linked to poor adherence to exercise rehabilitation programmes [20‐23]. Before treatment, the mean anxiety/depression score in EQ-5D non-responders ranged between 1.11 and 1.40, compared to between 1.37 and 1.78 in EQ-5D responders.

The anxiety/depression dimension scores at presentation for EQ-5D responders are still relatively low, so may not represent true clinical depression in most patients in a physiotherapy cohort compared to those reported in previous studies as showing poor outcome and poor adherence to treatment. The lower levels of depression in the EQ-5D non-responders could be considered as very low, though Shaw et al. [20] found that intervention adherence was poor in osteoarthritic patients with very low depression as well as with very high depression contrary to our findings. Further research is warranted to look at the relationship between treatment adherence and HRQoL scores at presentation.

The mobility dimension showed the greatest improvements in EQ-5D responders that received treatment for lower limb and low back conditions, showing large effect magnitudes. In EQ-5D responders receiving treatment at other symptom locations, only small to moderate improvements in the mobility dimension were observed. In EQ-5D non-responders, improvements with small effect magnitudes were observed for lower limb and wrist symptoms, and only trivial effects for trunk and upper limb symptom locations. Mobility is sometimes assessed using functional tests such as the 6-min walk test [24], which naturally make more use of the legs and trunk (load bearing) than the upper limbs (non-load bearing). The questions asked in the EQ-5D outcome specific to the mobility dimension are also specific to walking [4]. It is not surprising, therefore, that changes to the mobility dimension of the EQ-5D outcome were most associated with lower limb and lumbar pain. These findings suggest that evaluation of mobility status should only be considered in patients presenting with lower limb or low back symptoms.

The self-care dimension improved in EQ-5D responders, with moderate to large effect magnitudes observed in all symptom locations except the foot (small effect magnitude). In EQ-5D non-responders, most symptom locations showed no change in the self-care dimension, and patients receiving treatment for ankle, hip and elbow symptoms reported small declines in the self-care dimension. When completing the EQ-5D outcome, patients indicate their ability to wash themselves and to get dressed [4]. Having full range of motion at the hip and elbow will be key to washing and bending to be able to reach the feet, and ankle pain will likely impact on a patient’s ability to put on their socks and shoes. These data suggest that, for patients with ankle, hip and elbow symptoms, self-care is an important factor in whether a patient’s overall HRQoL improves.

In EQ-5D responders, improvements in the usual activities dimension (e.g. work, study, housework, family or leisure activities [4] with large magnitudes of effect reported across all symptom locations. In EQ-5D non-responders, usual activities were seen to improve with small effect magnitude for most symptom locations, with the exception of the elbow (moderate effect), foot (trivial effect) and wrist (trivial effect). This finding is similar to that discussed for the mobility dimension. Walking (the activity referred to in the mobility dimension question) could be considered one of the most prevalent “usual activities” that many patients will perform. In fact, Simpson et al [25] reported that walking was the most frequently reported physical activity amongst US adults in 2000, with 46% primarily walking for a minimum of 30 min, five times per week, as a leisure-time physical activity. This figure is likely to be slightly lower in the UK as Adams [26] reported only 34% of UK adults complete this duration/frequency of physical activity each week. As such, it could be expected that the usual activities dimension would make a similar contribution to overall HRQoL as the mobility dimension, as they are both functional dimensions within the EQ-5D outcome measure [27].

The data suggest that pain/discomfort and anxiety/depression show the greatest change in EQ-5D responders compared to EQ-5D non-responders. Interestingly, EQ-5D responders reported pre-treatment scores for the pain/discomfort dimension and the anxiety/depression dimension that were higher (worse) than in EQ-5D non-responders. These data suggest that those patients presenting with increased pain/discomfort and/or anxiety/depression may have the greatest capacity for improving their overall HRQoL. It must be noted, however, that the majority of patients reported only slight to moderate pain/discomfort and either no or slight anxiety/depression. This suggests that the EQ-5D dimensions may not be sufficiently sensitive to identify improvements in HRQoL in these patients. Herdman et al. [27] reported on the development of the EQ-5D five-level outcome (as used here) which attempted to reduce ceiling effects over the three-level version of the outcome. The current data suggest that even the five-level version of this outcome shows a ceiling effect.

The fact that many patients reported no pain/discomfort on the EQ-5D pain/discomfort dimension (i.e. score of 1), especially with the lack of improvement in overall HRQoL for many of those patients, raises the question as to whether it is economically appropriate for them to receive intensive physiotherapy. This would be in line with the approach of the STarT Back stratification tool for low back pain [28]. It could be that patients reporting no or low levels of pain/discomfort and anxiety/depression would see just as good outcomes in terms of HRQoL from receiving minimal intervention. Black [2], however, cautioned against HRQoL outcomes being used to ration care, suggesting that other factors, such as long-term outcomes of the specific disease, must also be considered. The assessment of HRQoL is suggested as being more useful in clinical decision making compared to disease-specific outcomes when it forms a part of a patient’s early evaluation [3]. However, quality of life is difficult to interpret as it is highly individual [3]. Quality of life should, therefore, be measured by the clinician in order that it is appropriately interpreted [2] and is used in a proportionate way to aid clinical goal setting [3].

It must be remembered that the individual EQ-5D dimensions are not validated as individual outcome measures. Multi-dimensional HRQoL measures are also not as effective as joint-specific outcomes in the evaluation of treatment outcome [3]. The present data point, however, to the importance of assessing pain/discomfort and anxiety/depression using a validated joint-specific outcome as part of a randomised controlled trial to assess the effectiveness of community-based physiotherapy for musculoskeletal conditions. Future studies should use validated outcomes for pain/discomfort and anxiety/depression in order to fully evaluate the effectiveness of community-based physiotherapy for musculoskeletal conditions. Whilst the EQ-5D dimensions suggest improvements over the time course of treatment, it is not possible from the current data to confidently say that this improvement in HRQoL was due to the physiotherapy treatment due to the retrospective nature of the data (i.e. no control group), and there is potential that changes in HRQoL reported could be due to the natural time course of healing. The data presented here should be used, therefore, to inform the design of a randomised controlled trial to fully evaluate the effectiveness of community-based physiotherapy. Such a trial should use specific pain/discomfort and anxiety/depression outcomes in this patient group to investigate the relevance of these dimensions to HRQoL and clinical outcomes.