Given the public nature of these data, no institutional review board approval was necessary for this study. The study group was composed of all NBA athletes who had sustained an acute Jones fracture during competition in an NBA game over a time period between the 1994-1995 and 2012-2013 seasons. Subjects were identified using a comprehensive online injury database (http://www.prosportstransactions.com ). Results were cross-referenced with team press releases, online injury reports, and player profiles to confirm diagnosis. This method has been validated in multiple studies.2 (link),3 ,7 (link),8 ,14 (link),16 (link) Effort was made to differentiate between stress fractures of the fifth metatarsal and acute Jones fractures. Access to an official NBA-controlled injury database was not available.
Thirty-seven players with acute Jones fractures were identified. Of these, 26 players with appropriate complete statistical performance data were included in the study. Return to play was defined as returning to at least 1 game of equal level of competition the year after injury. A 1-to-1 matched control group was selected based on “similarity scores” provided by a comprehensive online database (http://www.basketballreference.com ). This score identifies players whose careers are most similar according to performance data and seasons of play. Regarding position, backcourt players were designated as “guards,” while frontcourt players were classified as “forwards.” Controls were additionally selected according to position, and all efforts were made to select controls without a significant injury history.
The index year was defined as the season in which the player sustained the Jones fracture. The index year for controls was matched to the age of the matched-subject at the time of injury. Study parameters included the season before and after injury, as well as averaged data over 3 seasons before and after injury, if available. Method of treatment and instances of recurrence and/or reoperation were recorded. Demographic data, including height, weight, and body mass index (BMI), were collected. Performance data were recorded before and after injury and included games missed, minutes per game, assists per 36 minutes played, rebounds per 36 minutes played, steals per 36 minutes played, blocks per 36 minutes played, and points per game. Statistics were evaluated on a per minute basis to help control for an athlete’s playing time.
Additionally, NBA player efficiency rating (PER) data were collected on all subjects and controls. PER is a novel statistical method that involves the summation of a player’s positive statistical contributions and then subtracting negative measures.9 Yearly, a PER of 15 is designated to represent an average NBA player. PER is a comprehensive statistic that accounts for variables such as a player’s playing time as well as their team’s pace and style and allows for standardized comparisons between players. It has been previously used as a variable in similar studies.1 (link),2 (link),20 (link)As no previous studies have investigated the variability in PER with regard to Jones fractures, this pilot study used all available injuries to establish the PER means and variances for this injury to allow for future studies. Unfortunately, no sample size estimate was performed. Descriptive statistics were performed to describe the cohorts and their demographics. Univariate analysis was performed to compare the cohorts and the outcomes of the subjects (continuous variables were evaluated using Student t tests and categorical data were evaluated using the Fisher exact test). Additionally, multivariate regression analysis was performed to identify which factors were associated with performance (as measured by PER) after the injury while controlling for age, height, BMI, seasons in the NBA, minutes played per game prior to the injury, and PER 1 year prior to the injury as predictor variables. All statistical analyses were performed using SPSS version 19.0 software (IBM Corp), and significance level was set at P < 0.05.
Thirty-seven players with acute Jones fractures were identified. Of these, 26 players with appropriate complete statistical performance data were included in the study. Return to play was defined as returning to at least 1 game of equal level of competition the year after injury. A 1-to-1 matched control group was selected based on “similarity scores” provided by a comprehensive online database (
The index year was defined as the season in which the player sustained the Jones fracture. The index year for controls was matched to the age of the matched-subject at the time of injury. Study parameters included the season before and after injury, as well as averaged data over 3 seasons before and after injury, if available. Method of treatment and instances of recurrence and/or reoperation were recorded. Demographic data, including height, weight, and body mass index (BMI), were collected. Performance data were recorded before and after injury and included games missed, minutes per game, assists per 36 minutes played, rebounds per 36 minutes played, steals per 36 minutes played, blocks per 36 minutes played, and points per game. Statistics were evaluated on a per minute basis to help control for an athlete’s playing time.
Additionally, NBA player efficiency rating (PER) data were collected on all subjects and controls. PER is a novel statistical method that involves the summation of a player’s positive statistical contributions and then subtracting negative measures.9 Yearly, a PER of 15 is designated to represent an average NBA player. PER is a comprehensive statistic that accounts for variables such as a player’s playing time as well as their team’s pace and style and allows for standardized comparisons between players. It has been previously used as a variable in similar studies.1 (link),2 (link),20 (link)As no previous studies have investigated the variability in PER with regard to Jones fractures, this pilot study used all available injuries to establish the PER means and variances for this injury to allow for future studies. Unfortunately, no sample size estimate was performed. Descriptive statistics were performed to describe the cohorts and their demographics. Univariate analysis was performed to compare the cohorts and the outcomes of the subjects (continuous variables were evaluated using Student t tests and categorical data were evaluated using the Fisher exact test). Additionally, multivariate regression analysis was performed to identify which factors were associated with performance (as measured by PER) after the injury while controlling for age, height, BMI, seasons in the NBA, minutes played per game prior to the injury, and PER 1 year prior to the injury as predictor variables. All statistical analyses were performed using SPSS version 19.0 software (IBM Corp), and significance level was set at P < 0.05.
