Supplemental data. Detailed sample size considerations
Interim analyses and sample size considerations
To test for difference with 5% limit for type I error, we needed 1,000 patients, hereby obtaining a power of 80% to detect a 7.5% absolute risk reduction (ARR), assuming mortality in the standard-of-care-only group of 25.0%. The initial sample size estimation was based on our previous cohort study (S1), and on other cohort studies estimating the importance of timely and appropriate antimicrobial therapy (S2,3).
The standard-of-care-only group mortality was estimated from unpublished data from the participating sites and a cohort study (S1). The projected absolute risk reduction was estimated from studies assessing the frequency of inadequate antimicrobials and the mortality rate differences according to use of adequate versus inadequate antimicrobials support this estimate. Further, unpublished data from our cohort study supported this: The 30-day mortality rate was 44% for those who after the first ‘alert-procalcitonin’ had one or more subsequent ‘alert-procalcitonin’ over the next 3 days (indicating inadequate therapy for infection), but only 27% for those who did not encounter such a secondary ‘alert-procalcitonin’ situation. Taking into account that only 73% had at least one ‘alert-procalcitonin’ (allowing for intervention) and assuming that an intervention would only be successful in preventing further progression and procalcitonin increase in two out of three patients, the expected absolute risk reduction was set at 7.5%.
A data and safety monitoring board (DSMB) reviewed data for every 250 recruited patients, and an O´Brien-Fleming boundary was applied at the planned interim analyses (S4). At 950 patients recruited, the DSMB recommended to enlarge the target sample size to 1,200 patients. This was based on interim results, revealing a higher 28-day mortality rate than expected for all the patients so far recruited (31%). Additionally, there was a trend for improved survival in the procalcitonin-group for patients recruited between the 2nd and 3rd interim analysis, indicating a possible effect of the strategy after an initial “learning phase” in conducting interventions. The PASS steering committee – blinded to the arguments for the recommendation - endorsed it.
Reference for table S2: Ref: S5 below.
Reference List
S1. Jensen JU, Heslet L, Jensen TH, Espersen K, Steffensen P, Tvede M. Procalcitonin increase in early identification of critically ill patients at high risk of mortality. Crit Care Med. 2006;34(10):2596-602.
S2. Ibrahim EH, Sherman G, Ward S, Fraser VJ, Kollef MH. The influence of inadequate antimicrobial treatment of bloodstream infections on patient outcomes in the ICU setting. Chest 2000;118(1):146-55.
S3. Leibovici L, Shraga I, Drucker M, Konigsberger H, Samra Z, Pitlik SD. The benefit of appropriate empirical antibiotic treatment in patients with bloodstream infection. J.Intern.Med. 1998;244(5):379-86.
S4. O'Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics 1979;35(3):549-56.
S5. Vibeke Frøkjær Jensen. DANMAP 2008 - Use of antimicrobial agents and occurrence of antimicrobial resistance in bacteria from food animals, foods and humans in Denmark. 1-9-2009.