Use of an online questionnaire as a follow-up tool for a cohort of health personnel and patients attending a chemoradiotherapy unit in Mexico.

e13518 Background: COVID-19 detection in a timely manner in patients (pts) undergoing anticancer treatment is essential. RT-PCR test for SARS-CoV2 is the diagnostic gold standard, however it is a potentially limited resource in our setting. As an alternative, clinicians have developed symptom-based questionnaires as a screening tool for the detection of COVID-19 to optimize resources. The predictive efficacy of these tools in pts with cancer and healthcare personnel (HCP) who treat them has not been evaluated. Our objective was to describe the findings of the follow-up through an online clinical tool in our cohort, and determine its predictive performance compared against the gold standard. Methods: Data was obtained from a follow-up cohort of HCP and pts attending the chemoradiotherapy unit of a tertiary hospital designated as a COVID-19 priority facility in Mexico City . The follow-up period was from June 12 to September 30, 2020, and consisted of a 10-item clinical questionnaire (CQ) on respiratory symptoms and contact with patients diagnosed with COVID-19, collected daily electronically or by telephone . In addition, RT-PCR for SARS-CoV2 was performed every two weeks. The CQ was considered as a positive screening test if in the period between the biweekly RT-PCRs the participants had reported symptoms or contact. Results: We included 130 asymptomatic participants. 44.6% (n = 58) were HCP and 55.3 % (n = 72) were pts. Within a median follow-up of 85 days (IQR 48-103) a total of 8970 CQ were completed, 48.3% (n = 4335) were answered by HCP and 51.6% (n = 4635) by pts. 4.03% (n = 175) of CQ returned positive in HCP and 0.77% (n = 36) in pts. 634 nasopharyngeal swabs for SARS-CoV-2 RT-PCR were performed, 13.5% (n = 18) of the participants tested positive for SARS-CoV2 infection. When we evaluate within the biweekly period, the CQ for symptoms or contact was found to be a positive screening test in 12% (n = 78). If we applied the standardized definition of COVID-19 suspicious case used at our center, the CQ was found to be positive just in 3.9% (n = 25). The CQ as a general screening tool gave a sensitivity (SN) of 33.3%, specificity (SP) of 88.3%, a positive predictive value (PPV) of 7.6% and a negative predictive value (NPV) of 97.8%. Using the standardized definition, CQ gave a SN of 38.8% and a SP of 97.07%, a PPV of 28% and a NPV of 98.1%. Conclusions: Screening based on self-reporting of symptoms and contact through a questionnaire demonstrated low sensivity but high specificity in our cohort. When we applied a standardized definition of COVID-19 suspicious case, the clinical performance improved. RT-PCR testing remains as the gold standard to detect COVID-19 and should be preferred in patients undergoing anticancer treatment. Nevertheless, clinical questionnaires are an accessible tool for follow-up.