METHOD

We examined the contents of the four journals using criteria modified from the appraisal component of the READER method.³ The study was restricted to original research articles published in these journals in the five year period 1996–2000 inclusive. We classified the studies according to the method. The first group comprised randomised controlled trials. A second group of quasi-experimental designs was divided into three categories: cohort studies, case-control studies, and single case repeated measures design. We included a further group of observational/descriptive studies and finally a group of case reports and studies using other methods. We compared the journals according to the number of studies in each group using κ² analysis.

RESULTS

We identified 1051 original articles and classified them into the various groups (table 1). Randomised controlled trials comprised 10% or less of all original research articles, with a slightly greater proportion of cohort studies (10.5%). Observational/descriptive studies were the most commonly published study design, and the British Journal of Sports Medicine contained more case reports. In total, cohort, case-control, and single case study designs comprised nearly 46% of published work, across all four journals, with observational/descriptive and case studies making up the remaining 43%. We compared the journals according to the percentages of each study type, and there was a highly significant difference in their contents (p<0.0001) because of the high proportion of case studies in the British Journal of Sports Medicine and fewer case studies in Medicine and Science in Sports and Exercise. When the journals were compared by categorising the better quality methods together (randomised control trial, case-control, and cohort studies), the difference between the journals was not significant (p = 0.09), although there are a greater percentage of better quality articles in Medicine and Science in Sports and Exercise (37.5%) than in Physical Therapy (28.2%).

DISCUSSION

Overall, the proportion of randomised controlled trials seems low (9.5%), but we can compare the proportion of such studies published in sports medicine with other specialties, albeit within slightly different time periods. In a study of three UK primary care journals, the British Journal of General Practice, Family Practice, and the British Medical Journal, more than 50% of studies published in a five year period were either qualitative studies or surveys of attitude and opinion. Overall, just 6% of studies were randomised controlled trials, although in the British Medical Journal the proportion was greater at 16% (42).⁷ Similarly, the proportion of randomised controlled trials published in US family medicine⁸ is relatively small at 3.4%. In a review of research published in nine general surgical journals, 46% were case series, with only 7% randomised controlled trials.⁹ A study of six community health journals found 4% randomised controlled trials and suggested that 42% of the remaining trials could have used a randomised study design.¹⁰ In a study of seven leading rheumatological journals, almost half of the published studies were descriptive, with 16% classified as randomised controlled trials.¹¹ Therefore the proportion of studies identified in sports medicine in which high quality methods were used compares well with other medical specialities, in particular, general practice and surgery, but there were more high quality studies in rheumatology journals. Although we compared these journals directly, this may be unfair as contents are determined by both editorial policy and the quality of material submitted.

Notwithstanding these factors, the overall pattern of publication type seems remarkably stable and can be seen to reflect the overall quality of sports medicine research. Most encouraging is the finding that the proportion of randomised controlled trials published in the British Journal of Sports Medicine has increased from 3% to 7% in the two five year periods compared.

The importance of developing a high quality research base within sports medicine is twofold. Primarily, it will provide patients with diagnostic tests that are more reliable and interventions that are more efficacious. Furthermore undertaking and publishing high quality research studies will enhance the reputation of sport and exercise medicine as an academic discipline.¹² Attracting funding for clinical research can be difficult, and trials of clinical interventions are complex and time consuming. The nature of sports medicine may also attract its own particular set of problems. Highly motivated athletes, for example, may be unhappy to be allocated to a control group so that recruitment of elite athletes, in particular, into randomised controlled trials may be problematic. Furthermore, sports medicine as a discipline relates to people across the sporting spectrum, from the recreational athlete to the highly paid professional. These population groups may differ greatly in the nature and intensity of physical activity and injury, and findings in one group may not be directly applicable to others.