Author ORCID Identifier

http://orcid.org/0000-0002-7220-3291

Date Available

4-17-2017

Year of Publication

2017

Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation

College

Public Health

Department/School/Program

Epidemiology and Biostatistics

First Advisor

Dr. Wayne T. Sanderson

Second Advisor

Dr. Mary Kay Rayens

Abstract

An observational study was used to illustrate the application of time to event analysis methods to return to play; a secondary data analysis of athlete injury data from the High School RIOTM Injury Surveillance System (ISS) database was conducted. National Athletic Trainers’ Association (NATA)-certified athletic trainers from approximately 100 high schools in the US enroll their school in the system and complete the online “Exposure Report Form” for reportable injuries each week. New lateral ankle sprains and single-ligament knee injuries experienced by high school athletes during regularly scheduled participation in school-sanctioned sports for seven academic years (2005-2006 through 2011-2012) were analyzed. Field and court sport athletes (football, boys/girls soccer, volleyball, wrestling, basketball, baseball, and softball) were considered as these athletes were more likely to suffer lateral ankle or knee ligament sprains.

Detailed guidance was provided to assist athletic trainers and sports medicine researchers with understanding the appropriate data structure and programming statements required for time to return to play (T-RTP) analysis and the methodology appropriate for analyzing discrete time RTP categories. A data example was presented using lateral ankle sprain information to demonstrate how the life-table is useful for generating directly applicable information on expected T-RTP, and a discrete logistic regression model for this example highlights the relationship between severity of injury and T-RTP. Coding statements and life-table output were detailed for the LIFETEST procedure in SAS; SPSS instructions for generating life-tables were documented. The PHREG procedure in SAS using the TIES=DISCRETE option was presented to generate the discrete logistic regression model. An alternative method for computing hazard odds ratios was discussed to reduce computing time for large datasets with high numbers of tied event times using a pseudo dataset and the LOGISTIC procedure.

For 1st and 2nd degree lateral ankle sprains, the probability of RTP was highest 10-21 days after injury. For 3rd degree lateral ankle sprain, the probability of RTP was highest at least four weeks after injury. Gender had a marginal effect on RTP; male athletes were 18% more likely to return to play than female athletes. There was a significant interaction effect on RTP between time interval of return and ankle sprain severity. Athletes who experienced a 1st degree sprain were 458% more likely to RTP in 1-2 days than athletes who experienced a 3rd degree sprain, and 2nd degree sprains were 259% more likely to RTP in 1-2 days than 3rd degree sprains. In general, 1st and 2nd degree LAS were more likely to return than 3rd degree sprains in the three weeks after injury.

Regardless of which knee ligament was injured, athletes had a very small chance of RTP within two weeks of injury. Athletes injuring the ACL any time during the season had only a 1 in 3 chance of returning before the end of the season. RTP probabilities increase slightly for PCL, LCL, and MCL injuries after two weeks. Athletes suffering a single-ligament knee sprain during competition were 25% less likely to RTP before the end of the season than athletes injured during practice. Gender did not have a significant effect on RTP. There was a significant interaction effect on RTP between time interval of return and injured knee ligament. Athletes who experienced ACL sprain were 78% less likely to RTP in 1-2 days than athletes with MCL sprain, 81% less likely to return in 3-6 days, 91% less likely to return in 7-21 days, and 74% less likely to return 4 weeks after injury. Athletes who experienced LCL sprains were 213% more likely to return in 1-2 days than athletes with MCL sprain, 73% more likely to return in 3-6 days, and 103% more likely to return in 7-9 days.

The literature on return to play has been largely descriptive in nature, and time to event analysis methodology has not been heavily utilized. The applied methods paper presented here provides sports medicine researchers with direction to apply the methodology and interpret the results. The findings suggest that ankle sprain severity has the strongest impact on RTP timelines. ACL sprains have the longest RTP times and athletes are not likely to return during the season; athletes who suffer MCL sprains could potentially return during the season, but can expect to be out a minimum of three weeks. These RTP probability estimates are directly applicable for use by coaches, athletic trainers, and other members of the sports medicine team as they help provide reasonable expectations for return time following injury and allow for more accurate RTP planning.

Digital Object Identifier (DOI)

https://doi.org/10.13023/ETD.2017.085

Download

Share

COinS