Article Text

Download PDFPDF

413 Methodological considerations for quantifying prior injury history to study risk of injury recurrence
  1. Mackenzie Herzog1,2,3,
  2. Steve Marshall2,3,
  3. Nancy Dreyer1,2,
  4. Christina Mack1,2
  1. 1IQVIA Injury Surveillance and Analytics, Durham, USA
  2. 2University of North Carolina at Chapel Hill Department of Epidemiology, Chapel Hill, USA
  3. 3University of North Carolina at Chapel Hill Department of Exercise and Sport Science, Chapel Hill, USA


Background History of injury is an important predictor of future injury risk. Accurately assessing injury history can be challenging as it is influenced by the period of time used for the ‘lookback window’.

Objective Compare two lookback window definitions to quantify injury history and assess impact of this definition on the association between history of ankle sprain and incident game ankle sprain: 1) ’1-year lookback period’ limited to a history of ankle sprain within the past year and requiring ≥1 year of NBA participation, and 2 )‘all-comers lookback period’ including any history of ankle sprain regardless of the timing of the prior sprain or the number of years of prior participation.

Design Retrospective Cohort Study

Setting U.S. National Basketball Association (NBA)

Participants All NBA players who participated in at least one game from 2013–14 through 2016–17.

Assessment of Risk Factors History of ankle sprain included prior game and non-game ankle sprains.

Main Outcome Measurements Game ankle sprains were obtained from the audited NBA electronic medical record, which is standardized across all 30 teams.

Results Across this 4-season study, 554 game ankle sprains were reported among 946 players and 122,010 player-games. Using the primary definition, players with a history of ankle sprain in the past year were 1.41 (95% CI 1.13, 1.74) times as likely to sustain an incident game ankle sprain, relative to players with no history of ankle sprain. These results are consistent with prior research and pathobiology of ankle sprain. In contrast, the ‘all-comers lookback period’ definition led to an entirely different conclusion with a null result (IRR=1.01, 95% CI 0.80, 1.27).

Conclusions In this analysis, accounting both for a defined exposure time via a fixed 1-year lookback period and for the recency of the prior sprain(s) yielded robust and interpretable results.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.