Outcomes following surgery: are we measuring what really matters?

Outcome measurements following surgery have traditionally focused on objective data such as duration of hospital stay and mortality at 30 days. Recently, there has been a paradigm shift to measuring outcomes that are important to patients, such as physical activity, cognition and general well-being. These seem to be especially important to older patients, the number of which continues to rise annually 1. In this issue of Anaesthesia, Keeler et al. 2 report the effect of pre-operative intravenous (i.v.) iron compared with oral iron on quality of life (QoL) in patients with anaemia and colorectal cancer undergoing curative surgery. This was a secondary analysis of a previously published randomised controlled trial 3. Administration of i.v. iron resulted in improvements in QoL scores that met a minimum clinically important difference (MCID) across significantly more QoL domains compared with oral iron. The QoL assessments were performed using the EuroQoL 5-Dimension 5-Level (EQ-5D-5L), modified Short Form 36 (SF-36) and the Functional Assessment of Cancer Therapy-Anaemia (FACT-An) questionnaires. This is a timely addition to the growing body of literature regarding measurement of patient-reported outcomes following surgery, but we should be cautious when interpreting the improvements detailed in the study. This concept was first described by Jaeschke et al. 4 and defined as ‘the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects and excessive costs, a change in the patient's management’. In other words, the change is worthwhile to the extent that the patient would consider receiving the intervention if they had to make the choice again. The MCID is a threshold value for this change and can be useful when assessing patient-reported outcome measures (PROMs). Patient-reported outcome measures are self-reported questionnaires which are generic or condition specific and are completed by patients to ascertain perceptions of their health status. They are used to evaluate the effectiveness and safety of an intervention. There are two methods for determining MCID: anchor based; and distribution based. Anchor-based methods compare the reported change in the PROM to an external criterion (anchor). The anchor can be a laboratory test (e.g. haemoglobin (Hb)), a physiological measurement, clinician rating or a patient-reported global scale where patients can rate themselves as ‘better’, ‘the same’ or ‘worse’. The anchor should be relevant to the condition being studied, clinically acceptable and have some relationship with the PROM. Anchor methods can provide information at both group and individual level 5. Distribution-based methods, as used by Keeler et al. 2, relate to the statistical characteristics of the study sample. The observed change in the PROM is expressed as a metric such as standard error of measurement, effect size or minimally detectable change. Distribution methods provide information at an individual level 5 and are easy to generate, but have been criticised for being ‘purely statistical’, overestimating MCID, particularly with small sample sizes, and for being sample specific, that is, the MCID calculated depends on the variability of scores in that study sample only 6. Does it matter which method is used? Anchor-based methods appear preferable when calculating MCID as they attempt to integrate clinical information, but neither method is perfect and the difference between anchor-based and distribution-based methods can, in fact, be minor 6. An important limitation of both methods is that they do not consider the cost of the treatment. A more useful approach would be to present estimates using both methods in order to guide power calculations for future studies and to allow better interpretation of changes in PROMs. This has been demonstrated in orthopaedic surgery where Beard et al. analysed a PROM dataset of more than 137,000 patients to provide an array of values resulting in clinically meaningful changes to the Oxford Hip and Knee Scores 7. The choice of PROMs to be used in a study depends on four properties: objectivity, reliability, validity and responsiveness. These are crucial to any treatment decisions that are based on the findings of these questionnaires. Keeler et al. 2 have used well-validated questionnaires to assess both overall health status and, more specifically, anaemia and fatigue in cancer patients. However, clinicians may be less trusting of ‘subjective’ PROMs than they are of objective outcome measures such as exercise performance tests, duration of hospital stay or mortality. Objective measures are often perceived as being ‘hard’ or ‘real’ and are quick and easy to measure, whereas barriers to utilising PROMs include: the length of time taken to complete questionnaires; the fact that certain groups of patients may have difficulty completing them independently; and that clinicians are often unfamiliar with their interpretation and analysis 8. Randomised controlled trials often use objective measures which, although they may correlate with patient benefit and are easier to analyse statistically, ultimately do not matter to patients. Indeed, there are several examples in critical care where targeting objective measures has led to increased harm 9. From the orthopaedic literature, where the vast majority of work related to PROMs has been carried out, there is evidence that, although subjective measures do not always correlate with objective measures, patient satisfaction does 10. Modern technology may offer an opportunity to bridge this apparent discrepancy between objective and subjective measures. For example, wearable accelerometer variables (objective) have been shown to correlate well with physical activity scores in patients undergoing major surgery, although further validation is required 11. Keeler et al. 2 report a positive correlation between improvements in Hb and QoL scores, particularly for domains related to anaemia symptomatology. However, the correlation coefficients reported in their study are weak and there may be other reasons to explain this finding; for example, the unblinded study design could have led to exaggeration in reporting of PROMs. The benefits of i.v. iron administration on the pulmonary circulation, skeletal muscle function and exercise capacity have been well described and may in part have contributed to an improved QoL 12. The high oral iron dose used in the study was unlikely to be effective in the short time frame before surgery and current guidelines recommend that once daily or alternate day dosing may be more efficacious 13. The study did not specifically target patients with iron deficiency and therefore, any benefit of iron therapy may therefore either go unnoticed or be of a lesser magnitude than observed when studies include both iron-replete and iron-deplete patient. The definition of anaemia used in the study (Hb < 100 g.l−1) is also at odds with current recommendations of using an Hb cut-off of 130 g.l−1 for both sexes 13. This study also raises some other important questions. How long should the follow-up period be in such studies? Is target Hb a useful outcome measure? It is easy to measure and can be used as an external anchor for statistical comparison amongst groups, but does it matter to clinicians and patients? It is presumed to correlate with patient benefit but the evidence is conflicting and meaningful QoL improvements have still been observed in anaemic patients with cancer-related fatigue 14. Conversely, in patients with chronic kidney disease, targeting a high Hb in excess of 120 g.l−1 has not been associated with clinically meaningful improvements in QoL and safety concerns regarding a higher incidence of cardiovascular events have been noted. Current guidelines in this cohort therefore recommend a target Hb of 100–120 g.l−1 15. The PROMs we should be using will depend on what we are trying to measure. For example, in order to evaluate pain, a PROM should be considered as the gold standard tool as the patient's personal experience is the only information that matters. Studies in pain medicine commonly measure absolute differences in analgesia consumption between study groups or time to first use of opioid rescue medication. These are of little relevance to patients when compared with other outcomes such as cumulative pain scores (reflected as area under the curve) over a specified period of time, or time taken to return to usual activities. Fatigue can be assessed through PROMS, direct observation (e.g. exercise capability) and objective measures (e.g. maximal inspiratory pressure or hand-grip strength). Keeler et al. 2 used the FACT-An questionnaire to assess fatigue but other validated questionnaires include the Brief Fatigue Inventory, the Multidimensional Fatigue Inventory and the Edmonton Symptom Assessment System 16. In the absence of a ‘gold standard’ PROM, investigators have to make a trade-off between selecting an appropriate number of questionnaires and minimising questionnaire burden in order to ensure adequate return rates. Although we can question why Keeler et al. did not use an objective assessment of fatigue, evidence suggests that, even in this cohort of patients, subjective measures correlate poorly with objective measures 16. Patient-reported experience measures (PREMs) evaluate the patient's perception of the healthcare experience they have received, for example, confidence and trust in healthcare professionals, or hygiene, involvement in treatment decisions 17. This is in contrast to PROMs which measure outcomes of an intervention (Table 1). An example of a PREM is the 15-item Picker Patient Questionnaire which has been validated to provide a meaningful insight into the patient experience of healthcare 18. Picker Patient Questionnaire, CARE questionnaire, ISAS I have severe pain or discomfort (Yes/No) How often do you feel frustrated due to your breathing problems? (Not at all/A little of the time/Some of the time/Most of the time) Did doctors talk in front of you as if you weren't there? (Yes, often/Yes, sometimes/No) How good were the practitioners at showing care and compassion? (Poor/Fair/Good/Very good/Excellent) Although it is appealing to use both PROMs and PREMs, research in the peri-operative setting is limited. A number of tools are available in anaesthesia 19, but these are used infrequently and few are validated for populations of interest, for example, elderly surgical patients 1. Black et al. 20 have demonstrated a weak positive correlation between experience and outcome in patients undergoing elective hip, knee and groin hernia surgery. Similar relationships have been reported in patients with medical conditions such as acute myocardial infarction and chronic lung disease and highlight the relationship between PROMs and PREMs 21. However, it is important to note that the two do not have a causal relationship. Patients may report a good surgical outcome but have a bad experience (and vice versa) and it is perhaps in these situations where researchers and healthcare providers should concentrate their efforts. It is important that patients are involved in the design of clinical trials and selection of outcomes. Research that reflects their involvement is more likely to produce results that can be used to improve overall healthcare. A recent systematic bibliometric review of airway management, which included 1082 patient studies, highlighted that the most commonly reported outcome measures were procedure time and success rate 22. Although these are important to clinicians, their importance to patients is questionable. To try and address this disparity, funding bodies such as the National Institute for Health Research have now developed frameworks to guarantee involvement of patients and carers throughout the research process 23. The Core Outcomes Measures in Peri-operative and Anaesthetic Care – Standardised Endpoints for Perioperative Medicine initiative is a welcome step forward. This will collaborate with patients, carers, research groups and other relevant stakeholders through various methodologies including systematic reviews, expert consensus statements and Delphi processes in order to develop and harmonise outcome reporting 24. An example of what can be achieved was recently demonstrated in the NHS Coronary Revascularisation PROMs pilot study 25. The authors modified a pre-existing, validated questionnaire (the Coronary Revascularisation Outcomes Questionnaire), which is used to measure outcomes before and after coronary revascularisation and undertook a psychometric validation study across 11 hospitals in England. The modified version was reliable and valid when used as part of a large-scale PROM programme, but more importantly it was developed with patient participation. It included a much broader range of outcomes that are important to patients, such as overall satisfaction and quality of recovery following the procedure, which may have been missed when using the existing cardiac-specific PROMS. So where does this leave us? The take-home message seems to be that powering studies to detect changes in PROMs may lead to improvements in outcomes that matter to patients. We have a range of PROMs at our disposal, but challenges remain with regard to tool selection, statistical calculation of MCIDs and clinician ‘buy in’. PREMs offer the opportunity to provide an insight into the quality of care delivered during the conduct of a study and further research is needed in this area. Although there is still much work to be done on the use of outcomes that are important to patients, Keeler et al. 2 should be commended on providing us with an insight into what is possible. AS is a Trainee Fellow and CB is an Editor of Anaesthesia. AS is currently supported by an NIHR Doctoral Research Fellowship (DRF-2017-10-094). No other 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.