Elina Naydenova, CEO and Co-Founder of DigitalHealth.London Accelerator alumni company, Feebris, explains the importance of capturing oxygen saturation measurements and shares how Feebris is supporting patient home-monitoring.
Why is oxygen saturation so important?
Oxygen saturation (SpO2) is a critical vital sign that measures the proportion of oxygenated blood in the body. Normal oxygen saturation levels vary with age, sex, and posture, but it is generally accepted that a range of 94 – 98% is considered healthy . A sudden drop in oxygen saturation may be the first evidence clinicians have of an acute illness such as pneumonia or pulmonary embolism, or an exacerbation of a chronic condition such as heart failure, asthma, or chronic obstructive circulatory disease (COPD). Furthermore, in recent times, pulse oximeters have been used for COVID-19 patients to self-monitor their SpO2 at home, allowing them to be discharged from hospital in a safe manner (even with mild to moderate disease) [2,3]. In the UK, SpO2 is one of the six core physiological parameters that is used in determining a patient’s aggregate National Early Warning Score (NEWS), which is used to detect and respond to clinical deterioration in adult patients.
How is oxygen saturation normally captured?
Pulse oximeter devices can be used to calculate SpO2 in primary and community care settings. Since the technology makes continual readings of a patient’s SpO2 at a high frequency, in some high dependency or secondary care units, bedside monitors automatically aggregate these to acquire a single value for the patient’s true SpO2. Otherwise, the continual measurements from the device must be carefully monitored, aggregated, and manually charted. Several studies have investigated differences in automatically acquired versus manually charted SpO2 measurements in hospital settings, evidencing different sources of bias and error that can arise from the latter [4-7].
What are the challenges with using pulse oximeters at home?
There are limited studies investigating the clinically reliability of pulse oximetry data in community settings. However, issues with pulse oximeters (even medically certified ones) being unable to calibrate to different skin types or underlying physiologies came into focus last year, with the NHS issuing a warning against the reliability of readings in patients with darker skin [8-10]. Yet beyond this there are additional factors that can dramatically impact the reliability of a reading, notably low perfusion (reduced peripheral blood flow, common in elderly patients), noise artifacts (which can be caused by ambient light, nail varnish, or unclean skin), or patient motion (common amongst elderly patient suffering with dementia) .
How is Feebris solving the problem through AI augmentation?
The Feebris mobile application is designed to guide a non-clinical user in capturing clinically reliable data in a home setting. Equipped with advanced AI, the app augments the behaviour of the user by automatically identifying sources of error and helping them address any issues in real time. Coupled with a pulse oximeter, the Feebris AI-guided system evaluates the quality of the underlying pulse oximetry signal, identifies where the most reliable parts of the measurement are and generates a clinically reliable aggregate.
In designing this algorithm, we worked with a number of clinical experts, such as anaesthetists, who are experienced in interpreting pulse oximetry signals and extracting clinically reliable insights from them. We designed the algorithm to mirror that advanced clinical judgement and guarantee that any data captured outside the hospital is of equivalent quality.
As the NHS is incorporating more decentralised programs of care, such as virtual wards and pulse oximetry @Home, it is crucial that we ensure the technology used preserves quality and safety and does not waste precious clinical time, attending to poor data.
The support we’ve received during our time on the DigitalHealth.London Accelerator has been pivotal in helping us clarify our value proposition and ensure it meets the needs of the NHS. This new feature of our platform aims to do just that.
 O’Driscoll BR, Howard LS, Earis J, et al. BTS Guideline for Oxygen Use in Adults in Healthcare and Emergency Settings. Thorax 2017; 72: i1 – i90.
 Greenhalgh T, Knight M, Inada-Kim M, Fulop N J, Leach J, Vindrola-Padros C et al. Remote management of covid-19 using home pulse oximetry and virtual ward support. BMJ2021; 372.
 Orla O’Carroll et al. Remote monitoring of oxygen saturation in individuals with COVID-19 pneumonia. European Respiratory Journal. Aug 2020; 56 (2).
 Vawdrey DK, Gardner RM, Evans RS, Orme JF Jr, Clemmer TP, Greenway L, Drews FA. Assessing data quality in manual entry of ventilator settings. J Am Med Inform Assoc. 2007 May-Jun; 14(3): 295-303.
 Friesdorf W, Konichezky S, Grob-Alltag F, Fattroth A, Schwilk B. Data quality of bedside monitoring in an intensive care unit. International Journal of Clinical Monitoring and Computing. 1994; 11: 123-128.
 Ward N. The accuracy of clinical information systems. Journal of Critical Care. 2004 Dec; 19(4): 221-225.
 Levy M. Computers in the intensive care unit. Journal of Critical Care. 2004 Dec; 19(4): 199-200.
 Feiner JR, Severinghaus JW, Bickler PE. Dark skin decreases the accuracy of pulse oximeters at low oxygen saturation: the effects of oximeter probe type and gender. Anesth Analg. 2007 Dec; 105(6 suppl): s18-23
 Bickler P, Feiner J, Severinghaus J. Effects of Skin Pigmentation on Pulse Oximeter Accuracy at Low Saturation. Anesthesiology. Apr 2005; 102: 715–719.
 NHS, COVID Oximetry @home and COVID virtual wards https://www.england.nhs.uk/nhs-at-home/faqs-for-for-covid-virtual-wards-and-covid-oximetry-home/.
 Mark N, Lyubin A, Gerasi R, Ofir D, Tsur A, Chen J, Bader T. Comparison of the Effects of Motion and Environment Conditions on Accuracy of Handheld and Finger-Based Pulse Oximeters. Military Medicine. Jan-Feb 2021; 186 (1 suppl): 465–472.
DigitalHealth.London is delighted to publish blogs by the NHS staff and digital health companies we support through our programmes, as well as sector thought-leaders, experts and academics. Any opinions expressed within blogs published on our website are those of the author and not necessarily held by DigitalHealth.London. For more information, or if you would like to write a blog for our website, please email info@digitalHealth.london.