“Research highlights” is a weekly round-up of research papers appearing in the print BMJ. We start off with this week’s research questions, before providing more detail on some individual research papers and accompanying articles.
- Which treatment for heavy menstrual bleeding do women find least troublesome: hysterectomy; endometrial destruction; or Mirena?
- Are surrogate outcomes appropriately reported and discussed in randomised trials published in general medical journals?
- Are composite outcomes appropriately defined, reported, and interpreted in reports of randomised trials?
- Can we predict which people with psychological distress are at increased risk of dependency on welfare benefits for long term sickness?
- Is substantive criticism in letters to the BMJ’s editor adequately addressed by authors of research papers?
Dissatisfaction with treatment for menorrhagia
This is a bumper BMJ for anyone interested in research methods. Even this clinically focused study by Lee Middleton and colleagues was selected for the BMJ mainly because of its methodological approach: it was the first systematic review on this topic to include an individual patient data meta-analysis. As Richard Riley and colleagues explained earlier this year, obtaining and synthesising the raw individual level data from multiple related studies has both statistical and clinical advantages over using study level data; for example, it increases the power to detect differential treatment effects among trial participants (BMJ 2010;340:c221).
Middleton and colleagues’ study found that women with heavy periods were less dissatisfied with hysterectomy than endometrial destruction and that there wasn’t enough evidence to compare the levonorgestrel releasing intrauterine system (Mirena) definitively against the other interventions. In a linked editorial Jane Clark says these findings support the status quo and calls for more head to head trials with longer follow-up and adequate economic evaluations.
Reporting outcomes and interventions in randomised controlled trials
A randomised controlled trial should provide the strongest form of evidence for evaluating one or more interventions. But that holds only if the trial design, and the way it’s reported, truly answer the research question.
Jeppe Lerche la Cour and colleagues’ study of all trials published in 2005 and 2006 in the Journal of the American Medical Association, New England Journal of Medicine, Lancet, BMJ, Annals of Internal Medicine, and PLoS Medicine found that only 57% of the trials clearly reported using surrogate outcomes and only 35% discussed the validity of those surrogate measures. These low rates are worrying, because researchers should use a surrogate outcome (“a laboratory measurement or a physical sign used as a substitute for a clinically meaningful outcome”) only if it measures directly how a patient feels, functions, or survives. Indeed, doubly worrying, given that the Food and Drug Administration and other drug regulators frequently approve new drugs on the basis of trials using surrogates.
Gloria Cordoba and colleagues scrutinised all parallel group randomised clinical trials published in 2008 that had a composite primary outcome. Three quarters were about cardiovascular disease and almost two thirds were industry funded, and only one of the 40 included trial reports provided a good rationale for the choice of components. Using a composite outcome often allows a smaller sample size and savings in money and time. But the authors explain the pitfalls too: for instance “suppose a drug leads to a large reduction in a composite outcome of ‘death or chest pain.’ This finding could mean that the drug resulted in fewer deaths and less chest pain. But it is also possible that the composite was driven entirely by a reduction in chest pain with no change, or even an increase, in death.”
In an article for the Research Methods and Reporting section, Paul Glasziou and colleagues argue that the results of thousands of trials are never acted on because their published reports do not describe the interventions in enough detail. They spell out how to improve the reporting of trials at all stages, from the trial protocol through to systematic reviews and, finally, to patients’ websites.
Dependency on sickness benefits
Will Whittaker and colleagues analysed the annual British Household Panel Surveys of 1991-2007 and the Scottish Health Surveys in 1995, 1998, and 2003 to see if these routine datasets might help predict who would become a long term claimant of incapacity benefits for those chronically unfit for work. Mild mental health problems and frequent visits to general practitioners predicted dependence on these welfare payments up to three years beforehand.
Fruit and vegetable intake and incidence of type 2 diabetes
This systematic review and meta-analysis of six cohort studies suggests that increasing daily intake of green leafy vegetables could significantly reduce the risk of type 2 diabetes (doi:10.1136/bmj.c4229). Evidence is growing all the time on the role of nutrition in diabetes. For example, the recent Lifestyle Over and Above Drugs in Diabetes trial (BMJ 2010;341:c3337) found that intensive dietary advice could improve glycaemic control and anthropometric measures in patients with diabetes, and this week’s letter by Conor Magee and colleagues highlights the contribution of bariatric surgery, arguably another form of nutritional intervention.
Research online: For these and other new research articles see www.bmj.com/channels/research.dtl