Hip fractures: are we asking the right questions?

The headline statistics for outcome after hip fracture are widely reported—it feels as though every research paper on hip fracture has an almost identical opening paragraph. Annual numbers are high: around 70,000 in UK [1], 300,000 in the US [2], 1.6 million globally [3]; 30-day mortality is somewhere between 5% and 10% and gradually improving [4]; one-year mortality is around 25–30% and not really changing [5]. Depending on editorial constraints on word counts, occasionally some surrogate of functional outcome gets a mention: discharge destination or return to home rates seem most common. But, these data tell only part of the story. Thirty-day mortality is measured largely because we can measure it, not because there is something inherently special about 30-days. It is a surrogate for overall survival, not an end in itself. Thankfully, early mortality is falling, albeit gradually, most likely due to improved processes and increased involvement of geriatricians [6]. The consequence is that quality of survival is becoming increasingly important. If 1 in 12 people dies in the month following hip fracture, then the other eleven survive; three-quarters will be alive a year later. What happens to these people? How do we know? Can we predict it? Can we improve their outcomes? To date this is a relatively poorly researched area [7]. There are probably several reasons for this. In part, there has been a lack of focus—the medical research model has been dominated by mortality, medical and surgical complications. In addition, routine data collection beyond the hospital period is difficult—more so in an older group, with high rates of cognitive impairment and changes in place of residence. Beyond that, it is less than clear how to actually assess functional outcome. Patient reported outcomes (PROMs) are becoming increasingly popular in evaluation of care quality and efficacy. Unfortunately, the shift toward patient rather than researcher or service-centred outcomes has moved further than our understanding of which PROMs to use. Haywood et al. [8] recently reviewed PROMs used in hip fracture research. They found a wide variety of instruments in use but generally poor evidence for their use. Clinimetric properties were generally poorly reported. Of particular concern to the hip fracture population, there was little evaluation of the impact of completion by proxies. The evidence seems to support general assessments such as the 36-Item Short Form Health Survey (SF-36) and the European Quality of Life-5 Dimensions (EQ-5D) and possibly the more specific Functional Impact Measure and Barthel Index (BI). Only SF-36, EQ-5D and BI have evidence for responsiveness in the hip fracture population. Of these, only EQ-5D has been recommended as part of a core outcome set for clinical trials in hip fracture. All of these measures attempt to make a subjective experience—quality of survival—into an objectively quantifiable metric. Important for research and evaluation of care, yes, but inevitably missing the lived experience of patients. Bruun-Olsen et al.’s interviews with people recovering from hip fracture make sobering reading: ‘I do feel gloomy, yes…. I know that the people around me do what they can for me, but they can’t be me’. Although I have a lot of ‘blessings’ in my life, I now struggle to ‘count my blessings’ [9]. In this issue of the journal, Shehan et al. [7] have attempted to advance our understanding of functional outcomes a little further. A vast amount of work has gone into understanding factors that influence early medically orientated outcomes such as mortality [10, 11], complications [12, 13] and discharge planning [14, 15]. Given that we know that functional outcomes vary, are there any factors that can help us understand why? Can we use this information to improve the care we offer to our patients? On one hand, the review results are a little disappointing. The quality of the underlying studies was not particularly good. Of all the potential prognostic factors, Shehan et al. could only assign a weak level of evidence for cognitive impairment and anaemia. Given the lack of good evidence for functional outcome measures perhaps this is inevitable. If that were all that the review had found, then perhaps we would have learned little. But Shehan et al. have done something that is too often lacking in systematic reviews—they have attempted to add clinical interpretation to the data. They have drafted a framework linking factors and mediating mechanisms to potential outcome. Chronic cognitive impairment may be immutable, but the mechanisms by which it influences outcome may not be. These mediating mechanisms give us, clinicians and researchers, potential levers to improve outcomes. Some will work, some will not—but we will only find out if we look. There are limitations to their review that Shehan et al. have largely identified. The assessment of functional outcome was limited to the time of hospital discharge. There were methodological reasons for this, but it does betray researchers’ obsession with collecting data when we can rather than when we should. There is a message here for the national databases increasingly being used to drive improvement outside of dedicated research trials [1]. An efficient mechanism for collecting valid measures of patient relevant outcomes is needed urgently. The linkage between factor and mediating mechanism is clearly an oversimplification—some links are not identified—cognitive impairment, delirium and pain are much more closely intertwined than implied. Other aspects seem to be missing despite being common—anxiety and fear of falling being one example. But it is a starting point and provides a framework for future research in this area. Improving hip fracture care has to address a paradox. Hip fracture populations vary little between institutions. But the people that make up the hip fracture population vary greatly. The expectations of patient, healthcare staff and families for the functional outcomes of a fit and active younger person are different to the person nearing the end of their days with very limited pre-fracture function. We need to understand how our process of care interacts with the individual patient to produce the best outcome for that individual. Functional outcome after hip fracture is poorly researched and understood. As mortality outcomes improve following hip fracture we need to understand survivorship better. Valid and reliable assessments of patient-centred outcomes need to be collected routinely in research and clinical practice. I.K.M.’s employer receives reimbursement for his time with Health Services Research Centre (HSRC). He is on the editorial board of the British Journal of Anaesthesia. He has received research grants related to hip fracture care and patient safety. He is a member of the NICE Quality Standards Group for hip fracture. He has participated in peer reviews of hospitals as part of the National Hip Fracture Database/British Orthopaedic Association process. None.