An interesting article titled “American Association of Oral and Maxillofacial Surgeons’ Anesthesia and Third Molar Extraction Benchmark Study: Rationale, Methods, and Initial Findings” written by Thomas B. Dodson and Martin L. Gonzalez appears in the 2016 Journal of Oral and Maxilofacial Surgery (vol. 74, pp. 903-910).
In 2007 the American Association of Oral and Maxillofacial Surgeons (AAOMS) Board requested a benchmarking study to assess practice patterns and outcomes of oral and maxiofacial surgeons (OMS). The study period was 12 months and began June 1, 2011 and ended May 31, 2012. The study was designed to enroll 300 OMSs randomly selected from a list of all active private practice AAOMS members. The first sample included OMSs who had enrolled as practice-based research collaborative (P-BRC) participants. The second sample included patients treated by the OMSs participating in the P-BRC.
To be eligible for inclusion in the P-BRC, the OMSs were required 1) to be in private practice, full or part-time; 2) to be delivering anesthesia services in the office-based ambulatory setting; 3) to be current AAOMS members; 4) to have Internet access; and 5) to have agreed to participate. To develop a P-BRC composed of randomly selected OMSs, a list of active AAOMS members (N = 5,966) was provided by the AAOMS. After applying the exclusion criteria, the resultant population of OMSs eligible to participate in the P-BRC was 5,455 OMSs. The sample was stratified by census region. With this distribution, a sample of 300 contained 69 OMSs from the Northeast, 62 OMSs from the Midwest, 104 OMSs from the South, and 65 OMSs from the West. A random number was generated and assigned to each OMS. The OMSs of each census region were sorted in ascending order by their assigned random numbers. Once sorted, the first 69, 62, 104, and 65 OMSs were selected from the Northeast, Midwest, South, and West regions, respectively, to achieve a P-BRC composed of 300 OMSs stratified by geographic region. The OMS were then contacted and asked to participate and methods were used to replace them if they declined.
For the P-BRC participants, the following demographic variables were collected: age, gender, board certification status, degree status (double or single degree), years in practice, member status (member, fellow, life fellow, and other, defined as candidates or provisional members or fellows), and census region. All subjects had data collected for the following variables categorized as demographic (age, gender, and American Society of Anesthesiologists status), risk factors (alcohol and tobacco use, body mass index, chronic disease status [any], airway anatomy using the Mallampati classification), subject assessment of preoperative anxiety, type of operation executed (eg, dentoalveolar), pathology, level of anesthesia (local anesthesia only, minimal, moderate, or deep sedation, or general anesthesia), types of medication used (eg, local anesthetics, narcotics), personnel (eg, delivering anesthesia, monitoring, number of personnel used per case), types and frequencies of monitoring instruments, techniques (eg, blood pressure, electrocardiogram, oximetry, and capnography), clinical outcomes (adverse events), and patient satisfaction.
Various statistics on the two samples were performed. In addition a sample size study was performed prior to selecting the surgeons to include in the study. For example, the mean age of the active P-BRC members was 50.5 years, and 93.5% were male. Active participants had statistically fewer years in practice than inactive participants (18.3 vs 22.5 years, respectively; P < .001). A total of 124 clinicians contributed 6,344 subjects to the anesthesia database. For the wisdom teeth data set, 116 clinicians contributed data from 2,978 subjects who had had 9,207 wisdom teeth removed. The actual statistics and summary of the two patients samples are contained in another publication.
Despite the stated goal, a random sample of 300 AAOMs members were not included in the study. Instead self selection had to occur. The authors state
“The nonparticipation bias resulted in a clinician study sample different from that of the desired target population of US AAOMS members in private practice or invited AAOMS members who elected not to participate.”
The authors believe that lack of enthusiasm, training required, and the expense of collecting and entering data contributed to the low participation rate.