Concerns about the Validation of the Berlin Questionnaire and American Society of Anesthesiologist Checklist as Screening Tools for Obstructive Sleep Apnea in Surgical Patients
Bibliographic record
Abstract
We appreciate the questions from Drs. Perez Valdivieso and Bes-Rastrollo regarding our article.1When we prepared the manuscript, we were considering publishing postoperative complications as a separate paper. In that way we could have presented data on postoperative complications more comprehensively.We agree with the doctors that it would be more accurate to state that the STOP questionnaire and American Society of Anesthesiologists (ASA) checklist identified the patients with higher incidence of postoperative respiratory complications. Since the odds ratio was calculated based on the incidence of total postoperative complications, it is not conflicting to the above statement that the 95% CI of the odds ratio presented in table 71for the STOP questionnaire and ASA checklist included the null value.To further evaluate the predictive value of different apnea-hypopnea index (AHI) cutoffs, high risk score on the STOP questionnaire, STOP-Bang scoring model, Berlin questionnaire, and ASA checklist, we did multivariate logistic regressions on the potential risk factors for total postoperative complications and respiratory complications. The analysis was carried out with the procedure LOGISTIC from the SAS statistical package (SAS Institute Inc., Cary, NC). The candidate factors were selected to enter the model through the stepwise method. The P value for an effect to enter and stay in the model was 0.1. In models, AHI > 5, AHI > 15, AHI > 30, STOP questionnaire high risk, STOP-Bang scoring model high risk, Berlin questionnaire high risk, or ASA checklist high risk was respectively combined with age (> 50), sex (male), and preexisting conditions (hypertension, gastroesophageal reflux disease, diabetes, and asthma) as candidate risk factors. The result suggested that AHI > 5, AHI > 15, or STOP-Bang high risk were, respectively, significant predictors for total postoperative complications and respiratory complications, with P < 0.05 and a 95% CI of an odds ratio excluding 1. The score of high risk on the STOP questionnaire or ASA checklist was a significant predictor for postoperative respiratory complications. The other predictive factor retained in final models was gastroesophageal reflux disease, with P = 0.0776 and odds ratio = 1.828 (95% CI: 0.931–3.592).Of the 211 patients, 44 had an AHI > 30. Compared with the patients with an AHI ≤ 30, this group of patients had a significantly higher percentage of men (75% vs . 46%, P = 0.0005), bigger neck circumference (42 ± 8 vs . 38.±4 cm, P = 0.0035) and higher prevalence of hypertension (61% vs . 39%, P = 0.0132). They did show a higher rate of total postoperative complication (25% vs . 22.2%), severe desaturation (18% vs . 9%), intensive care unit admission (11% vs . 5%), and prolonged oxygen therapy (18% vs . 10%). However, the differences were not statistically significant. There were several possible explanations why we did not see the significantly increased incidence of postoperative complication in this group of patients. The first is the awareness of obstructive sleep apnea by anesthesiologists and surgeons, because of the requirement of our institutional research ethics board to inform anesthesiologists and surgeons if patients had an AHI > 30. The patients with AHI > 30 from the hospital which automatically monitor a patients in the intensive care unit for first night if the patient had a AHI > 30 showed a lower rate of postoperative complication (21% vs . 28%) and increased prolonged oxygen therapy (21% vs . 16%), as compared with the patients with AHI > 30 who were from the other hospital, although the difference is not significant. The second possible reason is that the sample size was too small.As we stated in the original paper,2there was a self-selection of patients involved in the process of conducting the study. Because of the difficulty to arrange a sleep study before surgery, and the stress the patients faced before surgery, it was almost impossible to avoid self-selection for this kind of study.*University of Toronto, Toronto Western Hospital, University Health Network, Toronto, Ontario, Canada. frances.chung@uhn.on.ca
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.337 | 0.593 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.006 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".