Journal Club Summary

Methodology Score: 3.5/5              
Usefulness Score: 3.5/5
Friedman BW, et al.
JAMA. 2015;314(15):1572-1580.
In this pragmatic 3-arm RCT of Naproxen + placebo or cyclobenzaprine or acetaminophen/oxycodone for acute low back pain, there was no significant difference in the primary outcome of improvement in pain and functional scores at one week following discharge from the ED. The methodology of this study was quite robust, however the small sample size, lack of adherence to the investigational medications, and narrow selection criteria limit its applicability to our population.
By: Dr. David Gruber

Epi lesson:

Multiple Imputation for Missing Data
When analyzing data from a dataset, most statistical software can only include cases where data is available for each of the included predictive variables. When data is missing for one or several important predictive variables, analysts have the option to “impute” the missing data. One simple strategy is to replace the missing information with the “mean” from the rest of the available data. Multiple imputation is a more complex strategy which replaces each missing value with a set of plausible values, which includes a level of uncertainty about the right value to impute. Important considerations include if data is missing at random, and how much data is missing.
By: Dr. Christian Vaillancourt