Reader Comments

Post a new comment on this article

Independent findings upon review of the FLIP study

Posted by jmasters on 14 Aug 2019 at 21:20 GMT

Independent Review Findings on “FLIP” study by Brian Hartman, PhD, ASA, Actuarial Program Director and Assistant Professor Department of Statistics, Brigham Young University, Provo, UT:
“The study is well-designed but the analysis answers the wrong question. The analysis answers the question, ‘Does this flooring reduce the number of serious fall-related injuries?’ when the question of interest is really ‘During a fall, does the flooring reduce the likelihood of a serious injury?’ These two questions have similar answers if the number of falls on both the SmartCells and the control flooring are similar, but in this case, they are quite different, with 1,009 falls on SmartCells and 898 falls on the control flooring. So rather than comparing 38 serious injuries on SmartCells to 47 serious injuries on the control floor, the researchers should have compared injury rates; 38/1009 = 0.03766 (SmartCells) to 47/898 = 0.05234 (Control). A simple z-test on the difference in those proportions gives a (one-sided) p-value of 0.06, much smaller than the p-values in the paper. They could have also assessed the difference under their preferred framework by modeling the odds that a given fall resulted in a serious injury, rather than the odds that a given resident sustains a serious injury. Setting aside any statistical tests, the rate of serious injury on the SmartCells was still 28% smaller than the rate on the control, even with all the limitations mentioned (other interventions in use e.g. high hip protector compliance, reporting challenges with high dementia diagnosis, mechanics of falls with 95% unwitnessed, some injuries not confirmed at hospital but still reported, room layout without noting role of furniture presence, most common location of injury was the head - 21.2% with lacerations being 22.4% of injuries, the sampling is from a study site with end-of-life limited mobility residents - see rebuttal paper). Additionally, the rate of minor injuries was reduced by 15%.” (Please note: Dr. Hartman was compensated for his time by SATECH, but his statement is entirely his own)

Competing interests declared: Employee of SATECH