Methodology: 3/5
Usefulness:2/5
Chinnock B, et al. Acad Emerg Med. 2021 Jun;28(6):612-620. doi: 10.1111/acem.14213.
Question and Methods: This prospective observational cohort trial sought to assess if self-obtained vaginal swabs were not inferior to provider-performed endocervical swab in the ED diagnosis of chlamydia (CT) and gonorrhea (NG) using a rapid NAAT.
Findings: SOVS was not inferior and had a 95% (CI =88-99%) for detection of NG/CT when compared to PPES.
Qualitative SOVS (Self obtained) felt acceptable in 93% of pts and 75% preferred SOVS to PPES (provider obtained)
Limitations: Single center study. Prospective cohort study with a convenient sample. This test is not performed at TOH.
Interpretation: This study alone does not provide conclusive evidence that SOVS are non-inferior to PPES and cannot be used to change management locally. May consider the use of SOVS if you have access to rapid NAAT and you are only concerned about ruling out NG/CT
Thank you for the review of our group’s article. Your final interpretation of the study was that the study was not conclusive and cannot be used to change management. I would have to disagree with this interpretation. First, this study is just 1 of numerous studies that have universally shown SOVS to be noninferior, in many cases with a trend towards superiority, compared to standard provider–performed endocervical sampling (PPES). Second, while it is true that the lower limit of the 95% confidence interval for sensitivity of SOVS was below 90%, we addressed this in the last part of the Limitations section of the paper. It is true that if you use an absolutely strict definition of the lower limit of the 95% CI being above 90% as the only criterion of clinically relevant noninferiority, then we did not demonstrate noninferiority. However, we think the sum of many parts demonstrated a clinically practical noninferiority for the following reasons:
– The overall number was above what we set for the noninferiority margin
– The gold standard itself is flawed, which we acknowledged in the first part of the Limitations section. There are several previous studies in non-ED populations that demonstrate SOVS is noninferior to PPES, and even with a trend towards getting better performance. In truth, that is also what we found with this study. The Venn diagrams of Figure 2 demonstrate this the best. While gonorrhea detection was the same, SOVS detected 5 cases of chlamydia that PPES did not, while PPES detected only 3 cases of chlamydia that SOVS did not. If we ran the same analysis on PPES using SOVS as the gold standard, it’s sensitivity would be worse than SOVS (although not likely clinically significant worse). You could call those SOVS positives in which PPES was negative as false positives (which we did for the purposes of reporting), but the reality is that every study using any NAAT swab has always shown specificities in the 99% area, so these “false positives” are very unlikely to be false positives- this means that, if anything, our study demonstrated SOVS was minimally better than PPES. The ideal way that we wanted to do this study was with 3 different testing methods (ideally another PPES Aptima swab) so that we could call SOVS a false negative or false positive only if both of the other 2 methods of collection/swab disagreed with it. However, we did not have funding for this unfortunately.
I really appreciate you reviewing new studies like this. Based on this study, and a significant body of studies showing SOVS to be noninferior to PPES, we have instituted doing self–sampling very early in the provider–at–triage model that we have, and we are often getting results back before patient can even get into the room for a pelvic exam. Our ultimate hope is we try to work on improving the process is to significantly decrease under–and overtreatment of GC (studies show undertreatment rates as high as 40% and overtreatment as high as 80% in ED) by getting results as early as possible (doing current study examining this).