Telemedical management of symptomatic COVID-19 outpatients

Discussion

In our randomised clinical trial, we allocated symptomatic non-hospitalised COVID-19 patients at risk for an adverse course of disease to either receive standard management or a smartwatch-based telemedicine care. Regarding the primary end-point of hospitalisation or unscheduled utilisation of outpatient emergency medical service, we found no difference between treatment arms. Yet, we observed that patients receiving telemedicine care less frequently utilised an outpatient-based emergency medical service.

In the beginning of the SARS-CoV-2 pandemic, the prospect of COVID-19 disease was a threat to individual health, particularly in high-risk patients for an adverse course [12, 13]. With the advent of vaccines [14–16] and therapeutic options [17], combined with less virulent SARS-CoV-2 variants [18, 19], the threat to individual health faded. Yet, the increasing number of infected patients increased pressure on the capacity of healthcare systems worldwide [20]. Today, many patients seek medical attention for non-COVID-19-related health issues while being actively infected with SARS-CoV-2. Throughout all phases of the pandemic, and to the present day, the provision of care to COVID-19 patients constitutes a burden to healthcare systems and their providers, and in fact high-risk patients such as those who are immunocompromised remain especially vulnerable to active COVID-19. For all others, care for vaccinated COVID-19 patients mostly shifted from inpatient and intensive to outpatient care as the number of hospitalisations decreased [21].

Telemedicine care combining access to experienced physicians with the availability of objective disease-relevant parameters like oxygen saturation or an ECG promised to alleviate the pressure on healthcare providers and resources. Based on initial data from the Robert Koch Institute in May 2020, we anticipated a 20% incidence rate of the primary end-point in our target population at risk for severe COVID-19. We further anticipated that telemedicine care could reduce this rate to 10%. However, with the described attenuated pandemic kinetics, and despite recruitment of all planned participants, we only observed a 9.6% incidence of our primary end-point in our control group. As a point estimate, we found a 28% relative reduction of the primary end-point in our intervention group, but this difference did not reach statistical significance.

Considering the increasing importance of outpatient-based COVID-19 care, we observed that our telemedicine concept reduced the unscheduled utilisation of outpatient-based emergency medical care by 56%. This effect was driven by patients seeking the attention of outpatient on-call physicians, whereas the utilisation of emergency ambulances providing immediate transport to an emergency room was comparable across treatment groups. This indicates that the telemedicine concept of COVID-SMART might be most suited to coordinate care and reduce in-person medical contacts of infectious individuals in a subacute, less symptomatic context.

Facing the risk of infection transmission, the need to reduce in-person contacts in the COVID-19 pandemic has been recognised early on. Telemedicine care thereby promised a good balance between maintaining a high standard of care and a reduction of contacts. It remains an important question whether the concept of telemedicine would withstand scientific evaluation. Telemedicine for COVID-19 care has been examined previously; recent reviews report various attempts to use smartwatches and wearables, mostly for symptom monitoring and contact tracing [22]. Only few studies have demonstrated that telemedicine in COVID-19 is practicable also for continuous monitoring of objective health measures like heart rate and oxygen saturation. A study by Wurzer et al. [9] showed that provision of an in-ear device for such 24/7 surveillance transmitted to a telemedicine team of physicians is feasible. Yet, randomised controlled trials demonstrating the benefit of telemedicine in COVID-19 are sparse. One of the few existing randomised controlled studies investigated the addition of oxygen saturation measurements to a text message-based telemedicine concept [11]. As a result, the authors found that such an addition provided no significant benefit. Importantly, the reported telemedicine concept itself had only been evaluated based on a retrospective, observational analysis [10, 23]. Hence, our COVID-SMART study may be the first randomised clinical trial evaluating the benefit of telemedicine care to COVID-19 patients. This is emphasised particularly as we investigated the clinically relevant composite primary end-point of hospitalisation or any unscheduled utilisation of a hospital-based or outpatient-based emergency medical service. Against the expectations raised that telemedicine may improve the level of medical care while reducing personal contacts, the primary result of our study was negative.

A number of factors may have contributed to the neutral findings and the dynamic of the pandemic may have played the most important role. At the beginning of the pandemic, the menace of SARS-CoV-2 was highest, resulting in the most threatening burden on healthcare systems and the strictest restriction measures [24]. Yet, during this phase, the absolute numbers of COVID-19 patients in ambulatory care were low [25], limiting a speedy study enrolment. In contrast, as the pandemic evolved with more virulent SARS-CoV-2 variants, higher absolute numbers of infected individuals were registered [26]. However, the more benign course of COVID-19 combined with increasing vaccination rates led to a reduced threat to personal health and fading associated fear, both negatively impacting the willingness to participate in our study as well as the occurrence of our primary end-point.

Patients recruited into our study were affected by a high risk of an adverse course of COVID-19 as defined by the inclusion criteria. One might speculate that such an at-risk cohort is more ready to seek emergency medical attention despite the provision of telemedicine care due to the severity of symptoms. This is further reflected by the neutral findings of our secondary end-point of utilisation of an emergency medical rescue and ambulance service. Against this background it is interesting to note that we did observe a difference in our secondary end-point of utilisation of an outpatient-based emergency medical service for less acute circumstances. Although this finding may serve only as hypothesis-generating, one might postulate that telemedicine care may be of highest usefulness in less symptomatic patients. However, facing the negative result of our study and considering the high effort for the provision of 24/7 on-call physicians, a systematic, clinical implementation of our concept to date seems inappropriate.

Interestingly, the rate of contacts with the participants’ primary care physicians during the study period did not reach statistical significance but numerically more control group patients contacted their primary care physician. Furthermore, we noted a high acceptance of our innovative telemedical concept with 90% of patients in the intervention group stating that they would use telemedical care in the light of active COVID-19 again. This finding is well in line with previous analyses that showed a sufficient patient satisfaction with telehealth providers during the COVID-19 pandemic [27].

A few limitations need to be considered when interpreting our data. First, we enrolled voluntary participants in domestic isolation. Given German data protection regulations, unlike in previous studies [11], no direct contact via the healthcare provider was possible. Instead, intense communication efforts were required to inform our target population about our study. This rendered it possible for all SARS-CoV-2 infected at-risk individuals in our catchment area to participate irrespective of their insurance status or healthcare provider. Yet, it may have introduced some degree of selection bias favouring more health-literate individuals. The requirement of a personal smartphone may have further accentuated selection bias. We aimed to reduce such bias by using a smartwatch compatible with any current smartphone and common operating system. We also acknowledge that our study was conducted open-label at a single centre in Germany only, which may limit the transferability to other, possibly less developed healthcare systems. Furthermore, our results may not be fully transferable to other risk groups for COVID-19, such as immunocompromised patients.

In conclusion, we conducted one of the first and largest prospective randomised clinical trials investigating the benefit of an innovative smartphone-based telemedicine care concept in COVID-19 patients at risk for an adverse course of disease. Despite such an at-risk cohort, our telemedicine care concept provided no significant benefit to our patients regarding our primary end-point. Interestingly, non-acute outpatient-based emergency medical contacts were reduced in the intervention group. Future studies may need to target the benefit of telemedicine in lower risk or less severely affected individuals.