2015 ISAKOS Biennial Congress ePoster #1130

A Non-Response Analysis of 2-Year Data in the Swedish National Knee Ligament Register

Jesper Kraus Schmitz, MD, PhD, Malmö SWEDEN
John Reinholdsson, MD, Stockholm SWEDEN
Magnus Forssblad, Stockholm SWEDEN
Anders Stalman, MD, PhD, associate professor, Saltsjobaden, Sweden SWEDEN
CLINTEC, Karolinska Institutet, Stockholm, SWEDEN

FDA Status Not Applicable

Summary: The results indicate that the information in the Swedish National Knee Ligament Register is valid but there is room for improvement in the information given to the patient in order to increase response rate.

Abstract:

Introduction

In Sweden 90% of the anterior cruciate ligament surgeries are registered in The Swedish National Knee Ligament Registry (SKLR). Since the response rate at the two-year follow-up is 50 % percent the validity of register data is questionable and it is necessary to make an analysis of the non-response group. Our aim was to study differences between the response and non-response groups with respect to baseline-data, demographic and outcome variables at two-year follow-up. Our second aim was to, in a questionnaire, examine reasons for drop out in order to be able to improve data collection in the future.

Methods

All 3,589 patients in the SKLR that had undergone anterior cruciate ligament surgery in 2010 were included. Responders (n=1865) and non-responders (n=1723) at the two-year follow-up survey were identified and assessed for potential differences in demographics (age, gender) and baseline data (time between injury and surgery, concomitant meniscus and/or cartilage injury, type of graft, primary/revision surgery, healthcare region, activity at time of injury). KOOS/EQ5D questionnaires were sent to the non-responders together with a non-response survey asking questions about reasons for drop out.

Results

The mean age was significantly higher for responders (27.8 years) than non-responders (25.9 years) (p < 0.001). Women had a significant higher rate of responders (62.8%) than men (44.4%) even after correction for age (p<0.001). Regional differences in the response rate were found but not statistically significant. Alpine/telemark-skiing was the only activity at time of injury that showed significantly higher rate of responders (62.5%) compared to non-responders (37.5%) (p<0.001). No differences of EQ5D at two-year follow-up were found between the groups. The change of KOOS from 0-2 years showed statistical significant difference in the subscale pain with 7.3 in the response group compared to 4.2 in the non-response group (p<0.05) and the subscale Quality of Life with a difference of 26.3 and 22.6 respectively (p<0.05). The non-response questionnaire showed shortcomings in patient information regarding the importance of the SKLR.

Conclusion

A mean difference in age of two years is too small to be considered clinically relevant. Women in our material had a significantly higher rate of response which is also consistent to findings in other studies. No influence in the different analysis could be attributed to the patients’ gender. The KOOS-analysis showed statistical significant difference in two subscales. The minimal clinically important change (MIC) for KOOS is considered 8-10 thus we conclude that the differences in KOOS are too small to be clinical relevant. The non-response survey showed that there is need for improvement in the way information is given to the patients about the SKLR in order to increase the response rate and decrease regional differences.