Paper plus: Bowling injuries in cricketers

Domhnall MacAuley takes you through a study aboutthe epidemiology of cricketing injuries

This month's paper is PL Gregory, ME Batt, WA Wallace. Is risk of fast bowling injury in cricketers greatest in those who bowl most? A cohort of young English fast bowlers. Br J Sports Med2004;38:125-8.

A harmless summer pastime, endless evenings, warm sun on your back, and the smack of ball on willow—hardly the background for major trauma? But, at the highest level, sport is not for relaxation. It is intense, professional, and incredibly serious. Competition is a welcome break from hours of training, relentless preparation, and years of development. The pattern of cricket injury is markedly different from that of contact sport, but for the players, the impact of injury is just as important. Identifying and recording injury is the first step in developing strategies in prevention, monitoring treatment, altering technique, and modifying rules. Because sport is so accessible, on television and on the back pages of our newspapers, and so much a part of daily life, we may assume that recording sports injury is easy. It is not. Coding when, why, how, technical variation, anatomical location, predisposing factors, and even weather conditions are critical to the epidemiology of injury. So, although we are familiar with sport, we know much less about the nature of sports injury. This paper is interesting in a cricketing context, but it can also help us understand the difficulties of recording injury, underline the importance of accurate data collection, and illustrate how epidemiology has a real relevance, even to sport.

TONY MARSHALL/EMPICS

Alamgir Sheriyar bowling for Kent

Why do the trial?

There has been some concern that fast bowling causes injury and the English and Wales Cricket Board (ECB) issued draft directives limiting the number of balls bowled by young fast bowlers. In this paper, the authors identify two objectives: firstly, to determine if the incidence of bowling injury varies with number of balls bowled, and, secondly, to see if young fast bowlers are exceeding the ECB directives.

Let us explore how we could do such a study. Looking at one team for one season will tell us little about injury. Teams vary, and the time is too short. For a study of sporting injury to be representative it should cover a range of teams and standards with sufficient numbers to be meaningful and over a period of time that includes seasonal variation. Looking at how we record injury, it becomes even more complex because we should record the anatomical location, the type of injury, associated factors, the duration, and the treatment. All these features must be recorded consistently. Asking players retrospectively is fraught with inaccuracy, so we should design a prospective study. Before we start, we need to agree standards and protocols so that everyone records injury information accurately and consistently. Defining the pattern of injury is much more complex that we might have thought. Ultimately our design would be a comprehensive prospective cohort study, similar to any study of the epidemiology of a medical condition.

The researchers defined their research question quite clearly. To answer this question in this study there are four key parameters: they must define their sample, record information on injury, measure the volume of bowling, and decide on the duration of the study.

CLIVE FEATHERSTONE

Every ballcounts

The researchers defined their sample as including all fast bowlers attending three English County Cricket Club centres of excellence. There were 70 players of mean age 15.3 years old. There were no hard criteria for selection of this sample and the players studied were selected simply because their coaches thought they were likely to bowl for their county. But, how do you define the “fast” in fast bowling? Again, there were no objective criteria although the researchers state that the ECB defines a fast bowler as one for whom the wicket keeper stands back.

Researchers contacted the players regularly by telephone for six months and asked them to complete a structured questionnaire. By telephoning they could ensure good follow up, but the quality of information you can get by telephone is limited. The researchers defined injury simply as a painful or disabling condition, and they coded these injuries into four grades based on the pain. This injury data is rather vague, but the researchers followed up injured patients by contacting their physiotherapist or doctor.

Measuring the volume of bowling seems easy. Counting every ball bowled would be ideal but unrealistic. So, they tried to guess the number bowled by estimating the number of balls bowled in a match and in practice. They knew the number of overs each player bowled in a match and they estimated the number of balls in each practice session. Although this gives a precise number, it may not be entirely accurate. The researchers recorded this information for six months comprising three months of winter preparation and three months of the competitive season.

The researchers identified 23 new bowling injuries in 70 players and thus estimated the incidence to be 32.8 per 100 fast bowlers. But, we are most interested in the injury rate relative to the how much they bowled. Ideally, we would look at a number of different measures of exposure, the volume of bowling, intensity, frequency, and so on. This information was not available, and the researchers simply used four groups according to the estimated number of balls bowled. To compare differences between the groups they used a weighted χ2 test. Unexpectedly, perhaps, there was no statistical difference in injury rate between those who bowled fewer than 1000 balls and those who bowled more than 3000 balls.

It is too easy to be critical of such studies. The researchers set out to find out the answer to an important question, and they systematically defined their sample, selection criteria, record of injury, and the number of balls bowled. Simple to say; difficult to do.

Looking critically at the study, note that the sample is small. The researchers do include an estimate of sample size, but it seems to have been based on a large expected difference in injury incidence between fast bowlers and others (50 injuries v 10 injuries). This seems ambitious in the context of published work they cite in the introduction. Their definition of a fast bowler does not meet objective criteria. Ideally you would use more reproducible measurements to define a “fast bowler” which, in this study, is based simply on coaches' subjective assessments. The period of follow up of six months is relatively short and did not include a full season. They argue that this corresponds to three months of the winter training period and three months of summer competition. But if injuries are predominantly due to overuse, fewer injuries would occur in the first three months of the competitive period than in the latter part of the season. You could also argue that the researchers did not accurately count the number of balls bowled. Although their estimate gives us a numerical denominator, it may not, necessarily be entirely accurate. It includes their best guess of the number of balls bowled in the nets.

Identifying injury was difficult. The best method would have been that each player keep an accurate prospective diary with players examined to assess and code the injury. They could also have also used a more systematic means of recording factors associated with injury. Telephone records are of limited accuracy without formal validation.

It is also interesting to note that those who spent less time in the nets tended to be of a lower standard. If there had been a difference in injury between groups it would have been difficult to disentangle the relationship and say categorically that injuries were related to bowling rates rather than poor technique or slower speed of delivery.

They have done a useful piece of work but it does not really tell us anything. The main difficulty is that the study is underpowered. The study did not, unfortunately, achieve what it set out to do. Sports medicine research is not an easy option. For many medical students who are interested in sport, it might come as a terrible blow to realise that one of the key factors in good sports medicine research is to know your epidemiology.