Statistical analysis protocol for an evaluation of COVID Oximetry @home using a Regression Discontinuity Design
Links
Abstract
Coronavirus disease 2019 (COVID-19) has led to many individuals in England suffering from severe health degradation, complications, and deaths. One issue lies with people presenting to hospital with low oxygen saturation levels, often without accompanying breathlessness (known as silent hypoxia). Delays in escalating and admitting these individuals to hospital can lead to invasive treatment, prolonged hospital stay, and an increased risk of death.
Remote home monitoring models, which aim to remotely monitor COVID-19 patients at risk, have been implemented in several countries, including England, in response to COVID-19. The aim of these models is twofold:
- avoid unnecessary hospital admissions (‘appropriate care in the appropriate place’), and
- escalate cases of health deterioration earlier to avoid invasive ventilation and ICU admission.
The National Health Service (NHS) pilot of remote home monitoring models, which included the use of pulse oximeters that measure a person’s blood oxygen saturation levels, was launched in eight settings in England during the first wave of the pandemic. The pilots encompassed various models, with differing settings (eg primary care providing pre-hospital support or secondary care providing post-discharge support) and different mechanisms for triage and monitoring.