A Statistical Review Based on the Research Dealing with Patellofemoral Pain in Young Dancers: A One-Year Follow Up - The 5th Congress of Exercise and Sport Sciences

Background: Patellofemoral pain (PFP) is one of the most frequent injuries in young dancers. A previous study by Steinberg et al. (2017) tried to evaluate the prevalence and incidence of PFP in young dancers in a special dance program at school, and to characterize body morphology and anatomical alignment of those dancers who suffered from PFP. They screened 7th grade children and monitored them the following year (8^th grade). 89.6% of the dancers who participated in the research in the first year continued participation in the second year. Those seven participants who did not participate the second year left the program.

Aims: The aim of this study is to present the statistical advantages and the weaknesses of this research.

Results: In the first year (7^th grade) the prevalence of PFP was 62.7% (50.0%-74.2%) while in the next year the prevalence was 85.0% (73.4%-92.9%). Yet, this cannot say anything about incidence because we cannot detect the number of new case of PFP, or recovery rate from PFP of the dancers.

The advantages of the follow up is the possibility to answer the following questions:

What is the rate of new cases (incidence)?
What is the rate of recovery?

The number of new cases of PFP (an estimate of incidence) from which none recovered was 36.7% (25.3%-49.3%).

Major statistical issues: A known problem that arises when dealing with leg injuries is how to display the injuries? By foot or by dancer? In the literature researchers often present injuries per leg (in our research it means120 legs versus 60 dancers). Also, how do we account for pain? Is the pain specifically connected to the injured leg? In this case, the proper answer is to deal with each leg, but the legs are not independent, so how do we address this problem?

Conclusion: Statistics do not always have the answers to the questions raised by the researchers, but knowing and understanding what researchers are looking for makes it much easier to find the right statistical procedures.