2006 OPEN FORUM Abstracts
AN INDEPENDENT PROBABILITY ANALYIS OF VAPOTHERM RALSTONIA COLONIZATION
Wesley M. Granger, PhD, RRT, Jonathan B. Waugh PhD, RRT
University of Alabama at Birmingham, Birmingham, Alabama
Introduction: An independent probability analysis was conducted using data submitted to Vapotherm, Inc., the CDC and the FDA as part of an investigation regarding a recent FDA Preliminary Public Health Notification concerning reported Ralstonia cultured from a respiratory gas administration device at several hospitals. The purpose of this analysis was to determine the probabilities associated with reports that Vapotherm cartridges were being contaminated during manufacturing.
Methods: We were allowed independent access on a consultant basis (for the Company) to the data reported on Ralstonia cultures to the company via MDR, the CDC, and the FDA. We performed a probability analysis of reported positive Ralstonia cultures from clean, unopened cartridges and an analysis of CDC reported positive cultures of cartridges.
Results: Hospital "A" reported positive cultures from 3 new cartridges out of a box of 10. This hospital did not culture the other 7 cartridges but these were cultured by three independent labs, including the CDC and FDA, with all 7 being negative. The question is: What is the probability the hospital selected the only 3 contaminated cartridges from a box of 10 assuming that 3 of the 10 were actually contaminated? The result of this calculation was a 0.83% probability that they would select the 3 contaminated cartridges at random. Hospital "B" reported that they cultured 2 positive cartridges out of a box of 10. However, further testing by the CDC did not find contamination in an additional 67 cartridges from the same lot. If we use the contamination rate of 2/(2+67) = 0.0290 then what is the probability of the hospital getting exactly 2 contaminated samples when we sample 10 cartridges from a single lot? Binomial Probability analysis showed a probability of less than 3% (Pr = 0.0299) that they would choose exactly 2 contaminated samples out of 10 randomly selected samples when the contamination rate is 2.90%. A separate analysis of data from a CDC Spreadsheet of Cartridge cultures revealed 13 out of 54 (25%) cartridges tested positive at the CDC but these came from only 4 hospitals in 2 states and they were all opened by the hospital. Also, 10 of these 13 (77%) appear to be used cartridges and it is impossible to determine how these became contaminated.
Conclusions: (1). There is a less than 3% probability that the cartridges were contaminated during manufacture. (2). All of the cultures verified by the CDC as positive were from cartridges opened at the two hospitals and then sent to the CDC. None of the cartridges opened at the CDC or FDA tested positive (Total of 74 cartridges). (3). Analysis involving cultures of new, unopened Vapotherm cartridges, showed there is a high probability that these were contaminated with Ralstonia at the Hospital.