Henry J. McQuay

Evidence-based medicine (EBM) has, over the past 20 years, made us all more critical in our thinking about the efficacy and safety of interventions. This is evident in the higher standards of our spoken and written work, formal and informal, and in our approach to the subject. The downside has been the coincidence of the squeeze on healthcare funding with the emergence of the EBM ideas – it has been all too easy to misuse the tools of EBM to deny patients access to treatment, and this, together with the off-putting political correctness of the EBM approach in some quarters, has made clinicians uneasy. Clinicians have to make decisions about therapy for the individual patient. Ideally this is guided by the best available evidence and their experience. EBM can guide one as to the population efficacy and safety of a particular intervention, but as we all know few patients are average. That guidance can only be given if there is adequate evidence, and the difficulty with guidance about symptom control is often the paucity of evidence of sufficient quality to yield credible guidance. Palliative care is often a ‘complex intervention’, and here EBM struggles to untangle which components, if any, of the complex interventions are important. The trial and review methodologies for complex intervention are wanting. Tom Chalmers, a grandfather of the EBM movement, argued late in his career that the most important function of the EBM approach was to frame the research agenda. This we think is correct. The process of systematic review of a topic throws up the deficits in trial methods and the lacunae in the data, and this then can show the way forward.

Background: Population mean changes are difficult to use in clinical practice. Responder analysis may be better, but needs validating for level of response and treatment duration. A consensus group has defined what constitutes minimal, moderate, and substantial benefit based on pain intensity and Patient Global Impression of Change scores. Methods: We obtained individual patient data from four randomised double blind trials of pregabalin in fibromyalgia lasting eight to 14 weeks. We calculated response for all efficacy outcomes using any improvement (≥ 0%), minimal improvement (≥ 15%), moderate improvement (≥ 30%), substantial improvement (≥ 50%), and extensive improvement (≥ 70%), with numbers needed to treat (NNT) for pregabalin 300 mg, 450 mg, and 600 mg daily compared with placebo. Results: Information from 2,757 patients was available. Pain intensity and sleep interference showed reductions with increasing level of response, a significant difference between pregabalin and placebo, and a trend towards lower (better) NNTs at higher doses. Maximum response rates occurred at 4-6 weeks for higher levels of response, and were constant thereafter. NNTs (with 95% confidence intervals) for ≥ 50% improvement in pain intensity compared with placebo after 12 weeks were 22 (11 to 870) for pregabalin 300 mg, 16 (9.3 to 59) for pregabalin 450 mg, and 13 (8.1 to 31) for pregabalin 600 mg daily. NNTs for ≥ 50% improvement in sleep interference compared with placebo after 12 weeks were 13 (8.2 to 30) for pregabalin 300 mg, 8.4 (6.0 to 14) for pregabalin 450 mg, and 8.4 (6.1 to 14) for pregabalin 600 mg. Other outcomes had fewer respondents at higher response levels, but generally did not discriminate between pregabalin and placebo, or show any dose response. Shorter duration and use of 'any improvement' over-estimated treatment effect compared with longer duration and higher levels of response. Conclusions: Responder analysis is useful in fibromyalgia, particularly for pain and sleep outcomes. Some fibromyalgia patients treated with pregabalin experience a moderate or substantial pain response that is consistent over time. Short trials using 'any improvement' as an outcome overestimate treatment effects.

Some people who suffer an upper gastrointestinal bleed or perforation die. The mortality rate was estimated at 12% in studies published before 1997, but a systematic survey of more recent data is needed. Better treatment is likely to have reduced mortality. An estimate of mortality is helpful in explaining to patients the risks of therapy, especially with NSAIDs. A systematic review of studies published before 1997, and between 1997 and 2008. Any study architecture was acceptable if it reported on cases who died from any cause of upper gastrointestinal bleed or perforation. Analyses were conducted separately for all cases, and those prescribed NSAID or aspirin. Information was available for 61,067 cases (81% published since 1997) of whom 5,001 died. The mortality rate in all cases fell significantly, from 11.6% (95% confidence interval, 11.0 to 12.2) in pre-1997 studies to 7.4% (7.2 to 7.6) in those published since 1997. In 5,526 patients taking NSAID or aspirin, mortality increased, from 14.7% (13.6 to 15.8) before 1997 to 20.9% (18.8 to 22.9) since 1997. Upper gastrointestinal bleed or perforation still carries a finite risk of death. Differences in study architecture, population characteristics, risk factors, definition of mortality, and reporting of outcomes impose limitations on interpreting effect size. Data published since 1997 suggest that mortality in patients suffering from an upper gastrointestinal bleed or perforation has fallen to 1 in 13 overall, but remains higher at about 1 in 5 in those exposed to NSAID or aspirin.

Perioperative analgesic strategies are frequently tested using analgesic consumption as an outcome measure. This outcome measure is intuitive and superficially attractive, but has not been evaluated rigorously. Flaws in its use may be one explanation of continuing controversies surrounding the efficacy of analgesic strategies tested by this method. We contend that the analgesic consumption outcome measure is valid only when treatment groups achieve similar pain scores. A meta-analysis of perioperative gabapentin was used to test this hypothesis. Eighteen trials were identified, which were of sufficient methodological quality to include in the analysis. Trials reporting similar pain scores in treatment groups were classified as Category A and dissimilar scores as Category B. There were seven Category A trials: four reported reduced analgesic consumption with gabapentin compared with placebo, at one or more time points, and three found no difference. There were 11 Category B trials, all of which reported a decrease in analgesic consumption with gabapentin compared with placebo, at one or more time points. Analgesic consumption after gabapentin was similar for different postoperative analgesics. Sedation, dizziness, and vomiting were significant problems in pooled analysis. Analysis according to similarity of pain scores did not clarify whether perioperative gabapentin is useful in perioperative care. More rigorous examination of analgesic consumption as an outcome measure is needed, to establish whether achieving similar pain scores is as important as this paper claims and to determine those features of the analgesic delivery system, adverse effects, and other factors which may interfere with analgesic consumption as an outcome measure.

The burden of chronic pain for those who have it and their families is substantial, says Henry McQuay, and these patients deserve better. Dawn Stacey and colleagues (doi: 10.1136/bmj.39520.701748.94) describe an example of quality improvement in practice for one group of people with chronic pain, those with osteoarthritis

Introduction: Osteoarthritis trials usually report average changes in visual analogue scale (VAS) pain, and examine the difference between treatment and placebo. We investigated whether dichotomous responder analysis provides a more informative interpretation of drug efficacy. Methods: Merck supplied the number of patients who, by 6 weeks, had achieved pain relief compared with a baseline of 0% or more, 10% or more, 20% or more, and so on at equal intervals up to 90% or more. These different levels of pain relief were used to distinguish different definitions of responders, for example at least 50% pain relief from baseline. Numbers and percentages of patients achieving each level were identified. Information was sought from a dose–response trial over 6 weeks in osteoarthritis using placebo and using etoricoxib at 5, 10, 30 and 60 mg daily. Results: With placebo, the proportions of patients achieving at least 20%, 50% and 70% pain relief over baseline at 6 weeks were 30%, 11% and 2%. With 60 mg etoricoxib the equivalent percentages were 74%, 49% and 29%. The numbers needed to treat for 30 mg and 60 mg etoricoxib to produce at least 50% pain relief at 6 weeks compared with placebo were 4.2 (95% confidence interval 3.8 to 8.6) and 2.6 (2.0 to 3.9), respectively. Levels of pain relief of 50% and above discriminated best between different doses of etoricoxib. Conclusion: Responder analysis seemed to be more sensitive than examination of average changes in VAS pain scores. Validation would require calculations to be performed on a set of trials using individual patient data not available in publications.

Background: Communicating risk is difficult. Although different methods have been proposed – using numbers, words, pictures or combinations – none has been extensively tested. We used electronic and bibliographic searches to review evidence concerning risk perception and presentation. People tend to underestimate common risk and overestimate rare risk; they respond to risks primarily on the basis of emotion rather than facts, seem to be risk averse when faced with medical interventions, and want information on even the rarest of adverse events. Methods: We identified observational studies (primarily in the form of meta-analyses) with information on individual non-steroidal anti-inflammatory drug (NSAID) or selective cyclooxygenase-2 inhibitor (coxib) use and relative risk of gastrointestinal bleed or cardiovascular event, the background rate of events in the absence of NSAID or coxib, and the likelihood of death from an event. Using this information we present the outcome of additional risk of death from gastrointestinal bleed and cardiovascular event for individual NSAIDs and coxibs alongside information about death from other causes in a series of perspective scales. Results: The literature on communicating risk to patients is limited. There are problems with literacy, numeracy and the human tendency to overestimate rare risk and underestimate common risk. There is inconsistency in how people translate between numbers and words. We present a method of communicating information about serious risks using the common outcome of death, using pictures, numbers and words, and contextualising the information. The use of this method for gastrointestinal and cardiovascular harm with NSAIDs and coxibs shows differences between individual NSAIDs and coxibs. Conclusion: Although contextualised risk information can be provided on two possible adverse events, many other possible adverse events with potential serious consequences were omitted. Patients and professionals want much information about risks of medical interventions but we do not know how best to meet expectations. The impact of contextualised information remains to be tested.

Introduction: Faecal blood loss has been measured using autologous erythrocytes labelled with radioactive chromium for several decades, using generally similar methods. We conducted a systematic review of studies employing this technology to determine the degree of blood loss associated with use of aspirin, nonsteroidal anti-inflammatory drugs (NSAIDs) and cyclo-oxygenase-2 selective inhibitors (coxibs). Methods: A systematic search of PubMed and the Cochrane Library (to December 2006) was conducted to identify randomized trials in which treatment with aspirin, NSAIDs, or coxibs was continued for at least 7 days, and with at least 7 days of washout for crossover trials. Rates of faecal blood loss associated with these agents were determined in the randomized trials identified. Comparators were placebo, active, or no treatment. Outcomes of interest were mean daily faecal blood loss, and the number or proportion of individuals recording faecal blood above 5 ml/day and above 10 ml/day. Results: Forty-five reports of 47 trials were included, including 1,162 individuals, mostly healthy volunteers and predominantly young men. Only 136 patients (as opposed to healthy volunteers; 12%) were included, and these were mostly older people with an arthritic condition. Most NSAIDs and low-dose (325 mg) aspirin resulted in a small average increase in faecal blood loss of 1 to 2 ml/day from about 0.5 ml/day at baseline. Aspirin at full anti-inflammatory doses resulted in much higher average levels of blood loss of about 5 ml/day. Some individuals lost much more blood than average, at least for some of the time, with 5% of those taking NSAIDs having daily blood loss of 5 ml or more and 1% having daily blood loss of 10 ml or more; rates of daily blood loss of 5 ml/day or 10 ml/day were 31% and 10%, respectively, for aspirin at daily doses of 1,800 mg or greater. Conclusion: At baseline, or with placebo, faecal blood loss is measured at 1 ml/day or below. With low-dose aspirin and some NSAIDs, average values may be two to four times this, and anti-inflammatory doses of aspirin result in much higher average losses. A small proportion of individuals respond to aspirin or NSAIDs with much higher faecal blood loss of above 5 ml/day or 10 ml/day. There are significant limitations regarding the quality and validity of reporting of these studies, such as limited size and inclusion of inappropriate participants. The potential for blood loss and consequent anaemia requires more study.

Ageing populations will impact on healthcare provision, especially since extra years are not necessarily spent in good health. It is important to identify and understand the significance of common medical problems in older people. Anaemia may be one such problem. We report on the prevalence of anaemia in cohorts of elderly people in the general population. The presence of anaemia is associated with a worse prognosis for both morbidity and mortality. Electronic searching and reference lists of published reports were used to identify studies that reported on prevalence of anaemia in cohorts of at least 100 individuals predominantly aged 65 years and over living in developed countries, together with criteria used to define anaemia. Studies of anaemia prevalence in specific disease groups or published before 1980 were excluded. Prevalence data for the entire cohort, for men and women separately and for different age bands were extracted. Forty-five studies contributed data. Thirty-four studies (n = 85,409) used WHO criteria to define anaemia. The weighted mean prevalence was 17% (3–50%) overall, and 12% (3–25%) in studies based in the community (27, n = 69,975), 47% (31–50%) in nursing homes (3, n = 1481), and 40% (40–72%) in hospital admissions (4, n = 13,953). Anaemia prevalence increased with age, was slightly higher in men than women, and was higher in black people than white. Most individuals classified as anaemic using WHO criteria were only mildly anaemic. Anaemia, as defined by WHO criteria, is common in older people living in the community and particularly common in nursing home residents and hospital admissions. Predicted demographic changes underline the need to understand more about anaemia in older people.

Differences between gastrointestinal and cardiovascular effects of traditional NSAID or cyclooxygenase-2 selective inhibitor (coxib) are affected by drug, dose, duration, outcome definition, and patient gastrointestinal and cardiovascular risk factors. We calculated the absolute risk for each effect. We sought studies with large amounts of information to calculate annualised rates for clearly defined gastrointestinal (complicated upper gastrointestinal perforations, ulcers, or bleeds, but not symptomatic or endoscopic ulcers) and serious cardiovascular outcomes (antiplatelet trial collaborators – APTC – outcome of fatal or nonfatal myocardial infarction or stroke, or vascular death). Meta-analyses and large randomised trials specifically analysing serious gastrointestinal bleeding or cardiovascular events occurring with five different coxibs had appropriate data. In total there were 439 complicated upper gastrointestinal events in 49,006 patient years of exposure and 948 serious cardiovascular events in 99,400 patient years of exposure. Complicated gastrointestinal events occurred less frequently with coxibs than NSAIDs; serious cardiovascular events occurred at approximately equal rates. For each coxib, the reduction in complicated upper gastrointestinal events was numerically greater than any increase in APTC events. In the overall comparison, for every 1000 patients treated for a year with coxib rather than NSAID, there would be eight fewer complicated upper gastrointestinal events, but one more fatal or nonfatal heart attack or stroke. Three coxib-NSAID comparisons had sufficient numbers of events for individual comparisons. For every 1000 patients treated for a year with celecoxib rather than an NSAID there would be 12 fewer upper gastrointestinal complications, and two fewer fatal or nonfatal heart attacks or strokes. For rofecoxib there would be six fewer upper gastrointestinal complications, but three more fatal or nonfatal heart attacks or strokes. For lumiracoxib there would be eight fewer upper gastrointestinal complications, but one more fatal or nonfatal heart attack or stroke. Calculating annualised event rates for gastrointestinal and cardiovascular harm shows that while complicated gastrointestinal events occur more frequently with NSAIDs than coxibs, serious cardiovascular events occur at approximately equal rates. For each coxib, the reduction in complicated upper gastrointestinal events was numerically greater than any increase in APTC events.