Quality of Life: changing the face of outcome measurements in critical care

Born from a concern about the cost and quality of healthcare, and further emphasised by an increasing awareness of the variability of clinical practice throughout different jurisdictions, interest in measuring and evaluating the effect of clinical interventions has grown consistently over the last two decades. Measuring effectiveness serves as an attempt to ensure that healthcare systems are transparent and accountable to both those who pay for them, and those who use them. While outcome measures continue to rely heavily on the use of mortality as a marker of performance, recent evidence demonstrates that both the UK and the USA have a growing interest in measuring patient function, rather than merely physiological endpoints 1. This move towards using functional outcome in performance measurement mirrors the realisation in many specialties that subjective measures of health are a worthwhile adjunct in examining the effects of treatments and interventions. Initial clinician concerns regarding their use as an outcome measure have eased, as the necessary measurement tools have been repeatedly validated in areas of medicine including oncology, rheumatology and cardiology 2-4. Indeed, the strength of these tools of subjective measurement in assessing the effectiveness of care has also led to the US Food and Drug Administration's decision to use patient reported outcomes (PROs) in the clinical evaluation of technologies 5. Further proof of their widespread acceptance is evident in clinical oncology, as formal assessments of quality of life are now a mandatory requirement of most randomised control trials 6. Nonetheless, despite increasing evidence within intensive care that both the disease process of critical illness and the treatments provided to our patients can have a substantial effect on the functional outcome of those who survive intensive care 7, randomised controlled trials continue to focus on outcome measures of survival, length of stay and duration of mechanical ventilation. In the 10 years since Wu and Gao discussed, in this journal, the merits of examining long-term outcomes of critical care, there has been an increasing emphasis on the effect of critical illness on functional outcome, yet their call for the use of subjective measures as an endpoint in clinical trials has largely gone unheeded 8. Indeed, despite the growth of PROs as a measure of effectiveness in medical care, they are rarely evident in the evaluation of individual treatments and interventions within critical care medicine. This individuality is not unexpected. The primary target of interventions within intensive care is the prevention of death, and the most costly of its interventions and technologies are those that aim to save lives, rather than improve functional status. This targeted approach is not without merit. Interventions that used survival as their only primary outcome in assessing treatment efficacy are known to have improved mortality rates within critical care 9. However, the World Health Organization reminds us that the goal of healthcare is not merely the prevention of death, but it must also ‘improve health’ 10. In response to calls for improvements in quality within intensive care, and an emphasis on performance measurement, the use of quality indicators is becoming more widespread. These quality indicators fall within two of the arms of Donebedian's model of care – structure and process 11. Consensus guidelines consider nosocomial infection, pressure sores and pulmonary embolism to be significant markers of the provision of quality care 12. Identifying these indicators represents a significant step on the pathway to realising that critical care should be about more than just the prevention of death. However, by only examining the ‘process’ and ‘structure’ aspects of care, these indicators describe the outputs of healthcare delivery, rather than the ultimate health outcome. In an era of resource constraint, this is predictable. As the unit of delivery of a service, outputs are easy to measure in terms of quantity, quality, and cost, and are a tangible measure of work performed. However, they rely heavily on the assumption that altering these outputs will be associated with an improved outcome. Health performance measurement data examining treatment differences for myocardial infarction between Ontario and New York demonstrated that increased outputs do not necessarily result in improved outcomes 13. A realistic evaluation of care requires us to examine the ultimate change in health status that is attributable to health interventions, rather than the degree of intervention itself. Therefore, to demonstrate improvements in quality within intensive care units, we need to include an assessment of changes in both health and functional outcome. While the intensive care community may be aware of the effect of critical illness on the long-term outcome of survivors, this is less likely to be identified by healthcare providers involved in their post-hospital care. In addition, the impact on families, and the resultant societal burden, are often neglected in the allocation of resources and support. Patients discharged from intensive care units are more likely to be to be unemployed or under-employed than age-related cohorts, and poor coping has been reported in 100% of survivors, and in 100% of their primary carers 14. Studies of the physical, psychological and social morbidities affecting survivors demonstrate that many are unable to return to normal care. However, despite both the burden of disease and the cost-implications for both patients and society, clinicians continue to view impairments of functional outcome and decreased quality of life as an unfortunate complication of the disease processes and treatments necessary to ensure survival. Yet, the evidence suggests otherwise. Low-cost interventions can be beneficial in improving functional outcome; for example, strict attention to sedative choice and glycaemic control are possible means of improving cognitive function without negatively affecting survival 15, 16. The failure to include quality of life as an outcome in critical care trials is particularly remarkable in the case of acute respiratory distress syndrome (ARDS). Albeit a controversial association, the effect of ARDS-related hypoxaemia on cognitive function was described in the 1990s 17. While the exact mechanism for the relationship between ARDS and cognitive function remains in dispute, what is of significance is that despite the awareness of this potential relationship, cognitive function was not included as a secondary outcome in subsequent ARDS Network (ARDSNet) trials. More recently, survivors of the Fluid and Catheter Treatment Trial (FACTT) were involved in an adjunct study to examine neuropsychological function 18. The authors determined that fluid management is a possible risk factor, but could not confirm their findings. Further retrospective examination of survivors of ARDS demonstrated that higher blood glucose levels predicted both a longer duration of mechanical ventilation and a decline in cognitive function 16. However, functional status as an outcome measure has been missing from the endpoints of the most renowned trials examining glucose control over the past 12 years 19, 20. Examining the effect of delirium on neurocognitive and neuropsychiatric function has demonstrated that it is associated with cognitive dysfunction one year after discharge 21. While recent work has reinforced the theory that delirium may be an unavoidable effect of critical illness 22, an additional body of evidence has reported that the precipitating factors for delirium may be modified by choice of sedatives, rehabilitation, noise control and sleep-promoting interventions 23, 24. However, even the relatively recent trials of dexmedetomidine failed to include long-term cognition in their measurement of outcomes 25, 26. Furthermore, no plans have been reported to do so on follow-up analysis. While trials that focus predominantly on short-term functional outcome are no doubt important in determining the degree of supportive care that will be required by patients on their immediate discharge to ward-based care, or indeed to their home environment, emphasising only the effect of these interventions on short-term function will not provide us with the necessary information regarding their ability to cope in the longer term. Part of the hesitancy surrounding the use of quality of life outcome measures in the intensive care population relates to concerns regarding the subjective nature of the measurement tools involved. Clinicians’ fears regarding their use often focus on a belief that patients are unable to evaluate their own health status appropriately. Although an imperfect science, methods to examine quality of life have been validated in both in-hospital patients and the critical care population. In cardiology, quality of life tools have shown a consistent correlation between patients’ subjective assessments of their health status and conventional clinical assessments, including exercise stress tests and New York Heart Association (NYHA) assessments 27. Furthermore, there is a growing consensus in cardiology and oncology that quality of life assessments are now the ‘gold standard’ when used as an adjunct in the evaluation of healthcare performance 28. Targeted Short Forms (SF) have been validated not only in the assessment of high-risk patients, but also in the assessment of patients whose social circumstances may prevent them from accessing care 28. Disease-specific measurement tools are of particular interest in intensive care medicine. These tools are known to detect subtle clinical changes, and both utility tools (e.g. EQ5D) and generic instruments (e.g. SF36) have been demonstrated as being effective in discrete evaluation and prediction of illness for intensive care survivors 29. These measures allow a multifaceted approach to measuring outcome, by combining functional capacity, physiological capacity, neuropsychiatric conditions, work, economic and social activity, and a subjective expectation of illness 30. Nevertheless, the use of quality of life indicators is not without controversy. Nor is the use of these tools fully validated in the assessment of cost-effectiveness. While the validity of such a form is reliably confirmed, there less consensus regarding the timing of the quality of life assessments 31. Even within Europe, differences of opinion exist regarding the optimal time to assess quality of life. Further difficulties may arise with the small size of the critical care population for comparison. As patients requiring intensive care have different admission diagnoses and exhibit varying degrees of disease progression, both generic outcome measures and disease-specific measures are required. However, these generic measures may be poorly responsive to changes in disease-specific conditions 32. Additional problems with population size occur due to loss to follow-up, which is high in these patients 33. Furthermore, selection bias within this follow-up group raises concerns that those attending discharge clinics may be a group with greater health needs 34, or that those in greatest need may be unable to access follow-up care 33. The most pressing concern relates to whether or not patient-reported outcome measures (PROMs) may result in a focus on conditions that are not amenable to intervention. In an era of resource constraint, do we risk directing scarce resources away from those treatments that improve survival? While appropriate concerns exist regarding the conceptual and methodological difficulties inherent in comparing healthcare delivery, and in how these results are communicated and used, moving towards greater accountability within healthcare will depend on accurate measurement of outcomes, and a move towards cost-effective care will require these measurements to take long-term functional outcomes into account. In 2010, an examination of the role of subjective measures as an outcome within all areas of medicine demonstrated that around 12% of industry sponsored trials, and 15% of non-sponsored trials, involved the use of PROMs 35. Ageing populations and technological advancements will continue to place an enormous strain on critical care services. Healthcare costs continue to increase globally, with an expectation that they will reach 50% of national health expenditure by 2021 in the USA alone 36. As part of an effort to control costs, while continuing to deliver care within a responsive health system, governments are turning towards value-based pricing as a potential reimbursement policy in the purchase of pharmaceuticals. Internationally, chronic disease has become widely recognised as a significant burden, in terms of both patients’ health status and costs. Those involved in the allocation of resources will be under increasing pressure to ensure that treatments and interventions will benefit patients in both the short and long term. In preparation for this, we should follow the example of clinical trials in oncology, and ensure that quality of life measures are considered as an outcome along with mortality and length of stay. Rather than focusing on the difficulties with outcome measures in critical care, we should instead be focused on developing measurement tools that accurately represent our patient population. If we fail to embrace this, critical care medicine will become isolated as other specialties within medicine move towards value-based delivery of healthcare. No external funding and no competing interests declared.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.