Utilising digital contact tracing during a pandemic to measure contact trends by risk group

Monitoring contact patterns is important for assessing impact of public-health-and-social-measures (PHSM) during a pandemic, but existing methods, such as contact surveys or mobility data, have limitations. While the COVID-19 pandemic provided impetus for population-wide adoption of digital-contact-tracing (DCT), prior studies that utilised DCT to evaluate contact patterns evaluated temporal variations, lacking granularity pertaining to impact in at-risk groups.

Population-wide DCT data was utilised to construct a cohort of SARS-CoV-2-infected community-dwelling adults (N = 544,259), evaluating temporal trends in contact patterns up to 5 days prior to positive SARS-CoV-2 test. Average contact number recorded on DCT was aggregated across four phases (baseline; tightened measures; relaxed measures; Omicron-surge). Incidence-rate-ratios (IRR) of average contacts was then compared across various phases, with the baseline phase as reference, via generalised linear models with a Poisson distribution. Results were stratified by vaccination status, comorbidities and occupation.

Over the 8-month study period, we observe a downtrend in average contacts, versus baseline. Amongst fully-vaccinated cases, average contacts fall from 5.93 at baseline to 4.68 during tightened-measures (IRR = 0.78[95%CI = 0.77-0.79]), then increase to 5.23 when measures are relaxed, albeit still below baseline (IRR = 0.86[95%CI = 0.85-0.87]), and finally drop to 3.51 during Omicron (IRR = 0.58[95%CI = 0.57-0.58]). PHSM consistently impacts non-household contact, but does not initially affect household contact, which declines only during Omicron emergence, despite no change in PHSM. Frail individuals with comorbidities have higher non-household close-contacts, as do essential/frontline workers.

DCT data can provide insight into contact patterns and enable assessment of PHSM's impact over pandemic phases.