Intended for healthcare professionals

Endgames Statistical Question

Poisson regression

BMJ 2014; 349 doi: https://doi.org/10.1136/bmj.g6150 (Published 10 October 2014) Cite this as: BMJ 2014;349:g6150
  1. Philip Sedgwick, reader in medical statistics and medical education1
  1. 1Institute for Medical and Biomedical Education, St George’s, University of London, London, UK
  1. p.sedgwick{at}sgul.ac.uk

Researchers investigated whether a neuromuscular training programme was effective in preventing non-contact leg injuries in female floorball players. The programme was designed to enhance players’ motor skills and body control, as well as to activate and prepare their neuromuscular system for sports specific manoeuvres. A cluster randomised controlled study design was used. Participants were 457 players (mean age 24 years), recruited from 28 top level female floorball teams in Finland. Teams were allocated to treatment using cluster random allocation with 14 teams allocated to the intervention (256 players) and 14 to the control (201 players). Teams in the control group were asked to do their usual training during the study period. Clubs were followed up for one league season (six months).1

The primary outcome was the occurrence of acute non-contact injuries of the legs during training or play. In total, 32 327 scheduled hours of training and play were recorded for the intervention group (mean number per player 126.3 hours (standard deviation 32.5)), compared with 25 019 hours for the control group (124.5 (30.8)). Over the season, 20 acute non-contact leg injuries were reported for the intervention group, compared with 52 injuries for the control group. Some players may have experienced more than one injury. The injury incidence rate per 1000 hours of training and play was 0.65 (95% confidence interval 0.37 to 1.13) in the intervention group and 2.08 (1.58 to 2.72) in the control group. The unadjusted injury incidence rate ratio for the intervention compared with the control was 0.31 (0.17 to 0.58). Poisson regression was used to adjust the injury incidence rate ratio for potential confounding (including age, body mass index, floorball experience, playing position, number of orthopaedic operations, and league). The adjusted injury incidence rate ratio was 0.34 (0.20 to 0.57).

Which of the following statements, …

View Full Text

Log in

Log in through your institution

Subscribe

* For online subscription