BACKGROUND: Low research participation by primary care doctors, especially those working in the private sector, is a challenge to quality benchmarking.
METHODS: Primary care doctors were sampled through multi-stage sampling. The first stage-sampling unit was the primary care clinics, which were randomly sampled from five states in Malaysia to reflect their proportions in two strata - sector (public/private) and location (urban/rural). Strategies through endorsement, personalised invitation, face-to-face interview and non-monetary incentives were used to recruit public and private doctors. Data collection was carried out by fieldworkers through structured questionnaires.
FINDINGS: A total of 221 public and 239 private doctors participated in the study. Among the public doctors, 99.5% response rates were obtained. Among the private doctors, a 32.8% response rate was obtained. Totally, 30% of private clinics were uncontactable by telephone, and when these were excluded, the overall response rate is 46.8%. The response rate of the private clinics across the states ranges from 31.5% to 34.0%. A total of 167 answered the non-respondent questionnaire. Among the non-respondents, 77.4 % were male and 22.6% female (P = 0.011). There were 33.6% of doctors older than 65 years (P = 0.003) and 15.9% were from the state of Sarawak (P = 0.016) when compared to non-respondents. Reason for non-participation included being too busy (51.8%), not interested (32.9%), not having enough patients (9.1%) and did not find it beneficial (7.9%). Our study demonstrated the feasibility of obtaining favourable response rate in a survey involving doctors from public and private primary care settings.
METHODS: This was a multicenter prospective cohort study involving patients with cardiovascular risk factors who were undergoing major noncardiac surgery. Patients underwent home sleep apnea testing. All patients completed the STOP-Bang questionnaire. The predictive parameters of STOP-Bang scores were calculated against the apnea-hypopnea index.
RESULTS: From 4 ethnic groups 1,205 patients (666 Chinese, 161 Indian, 195 Malay, and 183 Caucasian) were included in the study. The mean BMI ranged from 25 ± 4 to 30 ± 6 kg/m² and mean age ranged from 64 ± 8 to 71 ± 10 years. For the Chinese and Indian patients, diagnostic parameters are presented using BMI threshold of 27.5 kg/m² with the area under curve to predict moderate-to-severe OSA being 0.709 (0.665-0.753) and 0.722 (0.635-0.808), respectively. For the Malay and Caucasian, diagnostic parameters are presented using BMI threshold of 35 kg/m² with the area under curve for predicting moderate-to-severe OSA being 0.645 (0.572-0.720) and 0.657 (0.578-0.736), respectively. Balancing the sensitivity and specificity, the optimal STOP-Bang thresholds for the Chinese, Indian, Malay, and Caucasian groups were determined to be 4 or greater.
CONCLUSIONS: For predicting moderate-to-severe OSA, we recommend BMI threshold of 27.5 kg/m² for Chinese and Indian patients and 35 kg/m² for Malay and Caucasian patients. The optimal STOP-Bang threshold for the Chinese, Indian, Malay and Caucasian groups is 4 or greater.
CLINICAL TRIAL REGISTRATION: Registry: ClinicalTrials.gov; Name: Postoperative Vascular Events in Unrecognized Obstructive Sleep Apnea; URL: https://clinicaltrials.gov/ct2/show/study/NCT01494181; Identifier: NCT01494181.