Dermatological emergency unit, day‐care hospital and consultations in time of COVID‐19: the impact of teledermatology

Conflicts of interest

None of the authors have conflict of interest to declare (you will find for each authors the Conflict of Interest forms completed).

Funding source

None.

To the Editor,

During the first COVID‐19 pandemic wave, dermatologists were urged to postpone non‐urgent and outpatient visits, ¹ to limit COVID‐19 spreading. Teledermatology (TD) integration, through live‐and‐interactive (LI) video consultation or store‐and‐forward (SF) expertise, was raised as a potential substitute to maintain continuity of care. ² , ³ The aim of this study was to determine whether TD implementation during the pandemic could (or not) compensate for the outpatient activity predicted in the absence of COVID‐19 pandemic. It was conducted in a dermatology department of a tertiary centre, providing SF TD since 2016, notably for skin emergencies. To avoid cancelling in‐person visits, LI TD was implemented and set in March 2020. The monthly number of scheduled consultations, dermatological emergency unit (DEU) visits, LI TD consultations and SF TD requests were retrieved from January 2019 to December 2020. For each activity, an ARIMA model (Auto Regressive Integrated Moving Average) was applied to predict the evolution of a time series, as previously described. ⁴ To confirm that time series modifications were linked to the influence of COVID‐19 on the number of consultations, we used a causal inference method. The impact of COVID‐19 pandemic and lockdown on the activity of the dermatology department are shown in Fig. 1 and Table 1.

Figure 1.

Observed and predicted activity of a dermatology department, between May 2019 and December 2020. Orange is lockdown period in France (March 17 to May 5 for the first one & October 30 to December 15 for the second one).

Table 1.

Differences between observed and predicted dermatological unit visits and consultations, ARIMA and Causal inference model

	Observed (n)	ARIMA model		Causal inference model
		Predicted (IC 95%)	Difference	Predicted (IC 95%)	Difference
Dermatological emergency visits	6225	10,757 [6395; 13083]	4472	10920 [9569; 12,164]	4635
Consultations	7019	9739 [6546; 12931]	2720	9188 [8133; 10,236]	2169

During the COVID‐19 period, a sharp decrease in the number of in‐person consultations (dermatological consultations/emergencies) was observed, especially during the first lockdown. For DEU visits, the ARIMA and the causal inference models showed, respectively, a decrease of 4472 and 4635 visits between March and October 2020 (P < 0.001). Over the same period for scheduled consultations, the decrease was of 2720 and 2169 consultations, respectively, for the ARIMA and the causal inference models (P < 0.001). In parallel, LI TD and SF TD continued with an upward trend in the number of requests (high average of predictions). However, during the first wave of the pandemic (from March to July 2020), LI TD does not compensate the important drop of in‐person consultations. While from August 2020 onwards, hospital practitioners revert to their traditional habits (in‐person consultation), as evidenced by the low number of LI TD and the little difference between the in‐person consultation and in person + LI TD, curves (Fig. 1).

In the post‐lockdown period (from mid to may), number of consultations almost got back to the predicted activity (low range of the confidence interval), while for DEU, the average number of visits did not reach those predicted by the ARIMA or causal inference model, for example, 1103 visits in July vs. 1446 predicted by the ARIMA model. While remaining in the low range, the predicted activity of in‐person activity was less impacted by the second lockdown. Although an exponential increase in its use in primary care was noted, telemedicine failed to compensate the decrease in number of patients consulting during the COVID‐19 period. ⁵ Whereas SF expertise usage seemed not affected by the pandemic, patient and physician preferred in person consultation to LI when possible. COVID‐19 pandemic has delayed diagnosis and care for patients with skin cancer, while reducing access to care for all. ³ The decrease in DEU activity suggests a change in the ways of general population, perhaps due to fear of long waiting times or crowd. Determining if this loss was either due to patients fears in healthcare facilities or to other factors needs to be investigated. Unfamiliarity and lack of trust with technology tools for consultations are also possible reasons. In conclusion, while it helped substitute many in‐person consultations when necessary, TD did not to take off during the COVID‐19 pandemic. Development of TD usage remains essential to exploit its full capacities.