Home Medicine Smartphone-based assessment of long COVID symptomatology and risk factors

Smartphone-based assessment of long COVID symptomatology and risk factors

96
0


In a recent study posted to the medRxiv* preprint server, researchers quantified the features and risk factors of post-acute coronavirus disease 2019 (COVID-19) syndrome or long COVID (LCOVID) based on the integration of survey (active) data and passive [mobile health (mHealth) wearable device] data.

Study: Presentation of long COVID and associated risk factors in a mobile health study. Image Credit: Ralf Liebhold/Shutterstock
Study: Presentation of long COVID and associated risk factors in a mobile health study. Image Credit: Ralf Liebhold/Shutterstock

Background

LCOVID refers to the persistence of COVID-19 symptoms beyond the acute phase of the disease and has been reported to affect several individuals globally. LCOVID symptomatology, prevalence, risk factors, and clinical features have largely been documented during the initial pandemic wave based on subjective data, and further investigation is required to understand LCOVID better. Smartphone app-based data provide objective health assessments, and wearable devices enable long follow-ups.

About the study

In the present study, researchers performed a mHealth analysis to assess LCOVID prevalence and symptomatology and identify risk factors incorporating active data and passive data.

PCR-confirmed or antibody test-confirmed COVID-19 (before 1 February 2022 patients) were enrolled through the mass science study app between August 2020 and May 2021 and formed the case group. For comparison, 3,600 non-COVID-19 individuals were included who formed the control group. The participants were asked to fill out survey questionnaires to obtain data on COVID-19 symptoms experienced, the status of vaccination, mental well-being and mood pertaining to the acute COVID-19 period (<4 weeks), ongoing (four weeks to 12 weeks), and post-COVID (≥12 weeks).

The mHealth metrics included heart rate (HR), physical activity (PA) or step counts, sleep, moods, and symptoms. An individual was identified to have LCOVID based on (i) persistent physiological from COVID-19 diagnosis onwards and lasting ≥12 weeks, and (ii) self-reported persistent COVID-19 symptoms for ≥12 weeks. In addition, risk factors for LCOVID development were assessed, and the Pubmed database was searched between the study’s inception and 1 July 2022 for LCOVID studies involving the use of wearable or mHealth technologies.

For LCOVD diagnosis, resting heart rate variability (RHRV) over 12 weeks post-COVID-19 infection and the historic PA values and durations of sleep obtained passively between one to two years before severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection were evaluated. Participants with persistent symptoms following COVID-19 diagnosis were also considered for LCOVID diagnosis. If ≥1 symptom was documented ≥1 time weekly for ≥12 weeks, the individual was included in the LCOVID symptom-based group (Lsymp). Logistic regression modeling was used for the analysis.

Results

A total of 1,743 individuals were considered for the final analysis, of which 44% (n=759) had completed the extended socio-demographic questionnaires. Among passive metrics, RHR values were significantly elevated among cases in comparison to controls during all three study periods. HR differences during acute COVID-19 (0.6 bpm) were lesser than in the ongoing phase (1.1 bpm) and comparable to the post-SARS-CoV-2 infection period (0.5 bpm) over 12 weeks.

The general trend in HR changes exhibited a peak in the initial week of COVID-19 diagnosis, with a subsequent drop in the following week, followed by a second long-term elevation thereafter, probably implying two sub-phases of acute COVID-19 in less than four weeks. The step count was significantly but negatively impacted during acute COVID-19 and did not significantly change thereafter. Increased sleep duration but significantly reduced sleep efficiency was observed among cases and controls across the three study time points.

All self-documented mental health measures were significantly and negatively impacted during all the study periods, with reductions in mean differences between the two groups with time. Anxiety and depression were significantly and persistently impacted on a group level following the COVID-19 diagnosis. Greater historic PA levels and average durations of heavy PA one to two years before COVID-19 diagnosis negatively correlated with LCOVID risks.

The female gender was associated positively but non-significantly with LCOVID. Among self-reported symptoms, most symptoms had the greatest counts and disease severity around the time of COVID-19 diagnosis. Fatigue was the most persistent symptom (>140 days) among moderate to severe cases, whereas breathing difficulties and cough persisted for longer periods among mild cases. Over 160 individuals (12%) reported persistent COVID-19 symptoms and were categorized as Lsymp, and the Lsymp cohort comprised elder individuals.

RHR values were persistently greater in the Lsymp cohort. In addition, the increases and decreases in the HR pattern observed in during acute COVID-19 were more prominent among the Lsymp group participants. Of interest, the Lsymp group of individuals had higher step counts with differences in the rate of reduction in steps close to the date of COVID-19 diagnosis, whereas the recovery of steps was found to be more similar. The differences in sleep durations between the Lsymp and the symptom-based short COVID group (Ssymp) showed peaks close to the date of COVID-19 diagnosis.

Age was the most prominent risk factor for LCOVID, with LCOVID risk being 6.4-fold and 6.5-fold greater among individuals aged >60 years and 50 to 60 years, respectively, compared to those aged 10 to 30 years. Of 2,144 records screened, eight studies investigated LCOVID using digital technology, of which six studies used self-reported questionnaire data to characterize LCOVID symptoms, including the ZOE COVID symptom app users. Three studies evaluated wearable sensor-based data to assess HR, sleep, and PA changes following COVID-19 diagnosis, of which two studies showed a persistent pattern of RHR elevations and reductions, lasting >4 months in a few cases.

Overall, the study findings showed that mHealth and wearable sensor-based devices could identify LCOVID presence and monitor recovery.

*Important notice

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behaviour, or treated as established information.

Previous articleBrainstorming Ideas for the Medical School Personal Statement
Next articleDe-risking regulatory submissions for medical technology companies