- Open Access
Erratum to: The effectiveness of substitution of hospital ward care from medical doctors to physician assistants: a study protocol
BMC Health Services Researchvolume 16, Article number: 115 (2016)
The original article was published in BMC Health Services Research 2014 14:43
Unfortunately, the original version of this article  contained an error in the text. The correction of this error and also an adjusted sample size calculation is detailed below.
Year of authorization of PAs
After publication of our study protocol, we noticed an error at the fifth bullet on page 2 . We described that since January 2013 PAs are authorized to indicate and perform predefined medical procedures and subscribe medication without supervision. January 2013 should however be January 2012 .
Adjusted sample size calculation
In the original study protocol we described a sample size calculation in which an average length of hospital stay (LoHS) of 7 days and a standard deviation of 6 days was used. These numbers were based on a study of Borghans et al, in which the LoHS was presented of all patients who were admitted at 69 hospitals in the Netherlands during one year . This concerned all possible medical specialisms. However, we included the following specialisms in our study population: general surgery, pulmonology, gastroenterology, cardiology, orthopedics and otolaryngology (ENT). The medical specialisms with relatively high LoHS (f.e. cardiothoracic surgery, geriatrics, dermatology) and relatively low LoHS (f.e. ophthalmology, plastic surgery, gynecology) were not represented . This composition prompted us to recalculate the required sample size. Instead of a LoHS of 7 days and a SD of 6, a LoHS of 6 days and a SD of 4.8 days was used, which better fitted with our study population. All other parameters remained the same. Taking into account an expected drop out of maximum 2 matched pairs, 34 wards (17 in each arm) with each 100 patients are required. In case of no drop out, 50 patients per ward are sufficient to detect a significant difference in LoHS, with an expected 20 % reduction in LoHS, alpha 5 %, power 80 % and ICC 0.06.
As a consequence of the matched controlled study design, the SD in our study population might be smaller than the above mentioned SD of 4.8, but we are unable to provide reliable estimates .
Timmermans MJC, Van Vught AJ, Wensing M, Laurant MGH. The effectiveness of substitution of hospital ward care from medical doctors to physician assistants: a study protocol. BMC Health Serv Res. 2014;14:43.
Dutch Association of Physician Assistants; 2013. http://www.napa.nl (date 29 May 2013)
Borghans I, Heijink R, Kool T, Lagoe RJ, Westert G. Benchmarking and reducing length of stay in Dutch hospitals. BMC Health Serv Res. 2008;8:220.
Bouwman B. Ligduurmonitor Nederlandse Ziekenhuizen. Arnhem: Coppa Consultancy; 2010.
Stuart EA. Matching methods for causal inference: A review and a look forward. Stat Sci. 2010;25(1):1–21.
We would like to thank Dr. S. Teerenstra (Statistician, Radboud university medical center, Department of Health Evidence) for his assistance with the recalculation of the power.
The online version of the original article can be found under doi:10.1186/1472-6963-14-43.