Between May 2013 and March 2014, 222 patients were randomized to the group allowed remove their dressings and shower at 48 hours and 222 to the group permitted to shower only after the original dressing and the sutures were removed in clinic. There were 4 (1.8%) superficial surgical site infections in the early shower group and 6 (2.7%) in the late shower group, an insignificant difference with p = 0.751.
The authors concluded that clean and clean-contaminated wounds can be safely showered 48 hours after surgery, and early postoperative showering may increase patient satisfaction.
I have always been an advocate of early showering after surgery. Wounds properly closed will be bridged by epithelium within 48 hours. Tap water is relatively sterile or we couldn't drink it. Many studies have shown that even irrigating open wounds with tap water instead of sterile saline does not lead to more infections. [Links here and here.]
Much as I would like to believe the Annals study, I can’t because it is probably underpowered to show a difference between the two groups.
Here is a nice definition of statistical power from a website called effectsizeFAQ.com:
“In plain English, statistical power is the likelihood that a study will detect an effect when there is an effect there to be detected. If statistical power is high, the probability of making a Type II error, or concluding there is no effect when, in fact, there is one, goes down.”
To their credit, the authors did try to estimate the sample sizes they would need by doing a power calculation. They knew that the wound infection rate for the cases they intended to enroll was about 1%. The problem is they estimated that showering at 48 hours would result in a wound infection rate of 5%. That seems very high to me for the types of cases included in their investigation—thyroid, lung, inguinal hernia and skin tumors.
If they had hypothesized that early showering would merely triple the rate of wound infections from 1% to 3%, they would have needed at least 1536 patients in each arm of the study. Then if there was no difference, one could conclude that early showering truly does not cause more wound infections.
Even if the known incidence of wound infection was much larger, say 5%, and the rate of infection with showering was presumed to be doubled (10%), to have enough power a study would need 434 patients in each arm.
Many websites provide calculators for determining the appropriate sample sizes to detect with a reasonable degree of certainty whether one intervention is better than another. Anyone thinking about doing a prospective randomized trial should realistically estimate the expected difference and calculate the power.
Whenever you read a negative study, the first question to ask is, “Was the study adequately powered to avoid a type II error?”