Comparing antibiotic prescription rates between in-person and telemedicine visits.

To the Editor: The COVID-19 pandemic has necessitated an expansion of telehealth. However, telehealth encounters may contribute to diagnostic uncertainty and alter clinician practice. For instance, studies of direct-to-consumer telehealth visits found that these encounters were associated with higher antibiotic prescription rates and lower concordance with guideline recommendations. ,

To examine the potential impact of increased telehealth use in dermatology, oral antibiotic prescription rates at a large academic health system (Department of Dermatology, University of Pennsylvania) were compared between prepandemic in-person visits (from April 12, 2016, to March 15, 2020), intrapandemic in-person visits (after March 15, 2020), and intrapandemic telehealth encounters (after March 15, 2020), stratified by new-patient visits (NPV) and return-patient visits (RPV).

Encounters for patients aged ≥18 years and having diagnoses commonly associated with antibiotic prescriptions were included: “abscess” (L02.9, L02.91-L02.93), “acne” (L70.0-L70.9), “atopic dermatitis” (L20.0, L20.81-L20.89, L20.9), “hidradenitis suppurativa” (L73.2), and “impetigo” (L01.00-L01.09, L01.1, L08.0, B08.02). Logistic regression, adjusted for race, sex, age, and insurance type, was used to compare the odds of receiving antibiotics between visit types. Statistical analyses were performed with R. This study was deemed exempt by the University of Pennsylvania institutional review board.

Among 7940 in-person and 1067 telehealth encounters associated with a diagnosis of interest, younger patients (17-40 years of age) were more likely to use telehealth than those 40-65 or over 65 (Table I ). Among NPVs for acne, antibiotics were more likely to be prescribed at telehealth encounters (15.8% of visits) than at prepandemic in-person (7.2%) or intrapandemic in-person (4.4%) visits. In a post hoc analysis, isotretinoin prescriptions were also more likely to be prescribed at NPV telehealth encounters (5.8% of visits) than at prepandemic in-person NPV visits (0.9%; P < .001). Among RPVs for acne, antibiotics were less likely to be prescribed via telehealth (4.2% of visits) than at in-person visits (9.7% and 7.0%). Isotretinoin prescriptions in RPVs were more likely to be prescribed at intrapandemic telehealth encounters (41.8% of visits) than at prepandemic in-person encounters (12.0%; P < .001). No other significant differences were noted (Table II ).

Table I

Patient demographics and multivariable analyses for in-person and telehealth for individuals seen for diagnoses of interest between April 13, 2016, and April 14, 2021

Variable	Type	Overall	In-person	Telehealth	Odds ratio (CI)	P value
		9007	7940	1067
Race	Asian	745 (8.3)	644 (8.1)	101 (9.5)	Reference	Reference
	Black	2603 (28.9)	2307 (29.1)	296 (27.7)	0.85 (0.66-1.09)	.2
	Hispanic Latino/Black	78 (0.9)	65 (0.8)	13 (1.2)	1.25 (0.66-2.36)	.49
	Hispanic Latino/White	205 (2.3)	160 (2.0)	45 (4.2)	1.77 (1.20-2.62)	.004
	Multiracial	93 (1.0)	77 (1.0)	16 (1.5)	1.29 (0.72-2.30)	.39
	Unknown/other	689 (7.6)	597 (7.5)	92 (8.6)	1.00 (0.74-1.36)	.98
	White	4594 (51.0)	4090 (51.5)	504 (47.2)	0.82 (0.65-1.04)	.1
Sex	Female	6581 (73.1)	5781 (72.8)	800 (75.0)	Reference	Reference
	Male	2426 (26.9)	2159 (27.2)	267 (25.0)	0.91 (0.79 - 1.06)	.25
Age group (years)	17-40	6173 (68.5)	5374 (67.7)	799 (74.9)	Reference	Reference
	40-65	2135 (23.7)	1924 (24.2)	211 (19.8)	0.74 (0.63-0.88)	.0004
	65-104	699 (7.8)	642 (8.1)	57 (5.3)	0.57 (0.39-0.82)	.003
Insurance	Commercial	5536 (61.5)	4869 (61.3)	667 (62.5)	Reference	Reference
	Medicaid	2083 (23.1)	1813 (22.8)	270 (25.3)	1.06 (0.90-1.25)	.49
	Medicare	793 (8.8)	715 (9.0)	78 (7.3)	1.17 (0.84-1.63)	.34
	Other	179 (2.0)	154 (1.9)	25 (2.3)	0.74 (0.55-0.99)	.046
	Self pay	12 (0.1)	12 (0.2)	0 (0.0)	-	-

Bolded text indicates statistical significance at P < .05 level.

Table II

Antibiotic prescription rates in prepandemic in-person visits, intrapandemic in-person visits, and intrapandemic telehealth encounters, by visit type

Diagnosis	Prepandemic in-person encounters with antibiotics (% total)	Intrapandemic in-person encounters with antibiotics (% total)	Intrapandemic telehealth encounters with antibiotics (% total)	Prepandemic in-person vs intrapandemic telehealth encounters odds ratio, adjusted (95% CI)	Prepandemic in-person vs intrapandemic telehealth encounters P value	Intrapandemic in-person vs intrapandemic telehealth encounters odds ratio, adjusted (95% CI)	Intrapandemic in-person vs intrapandemic telehealth encounters P value
New-patient visits
Abscess	7 (41.2)	3 (60)	0 (0)	-	-	-	-
Acne	90 (7.2)	4 (4.4)	19 (15.8)	0.41 (0.24-0.70)	.001	0.24 (0.08-0.75)	.013
Atopic dermatitis	1 (0.3)	0 (0)	0 (0)	-	-	-	-
Hidradenitis suppurativa	36 (39.1)	1 (20)	1 (20)	2.57 (0.28- 23.94)	1.00	1.00 (0.045-22.18)	1.00
Impetigo	1 (6.3)	0 (0)	0 (0)	-	-	-	-
Return-patient visits
Abscess	18 (42.9)	5 (41.7)	0 (0)	-	-	-	-
Acne	391 (9.7)	21 (7.0)	36 (4.2)	1.73 (1.22-2.46)	.002	1.21 (0.69-2.10)	.51
Atopic dermatitis	11 (1.4)	4 (4.3)	3 (2.3)	0.60 (0.16- 2.16)	.43	1.85 (0.40-8.48)	.43
Hidradenitis suppurativa	127 (23.8)	21 (33.3)	18 (27.7)	0.81 (0.46- 1.45)	.49	1.31 (0.61-2.78)	.49
Impetigo	18 (42.9)	3 (18.8)	3 (33.3)	0.23 (0.05-1.04)	.06	0.46 (0.07-2.99)	.42

Bolded text indicates statistical significance at P < .05 level.

Overall, there do not appear to be significant differences in antibiotic prescriptions between telehealth encounters and in-person encounters for these common dermatologic conditions. Although there was an increase in antibiotic prescriptions at telehealth NPVs for acne, this may be related to patients presenting with greater acne severity during the pandemic, given that there was a corresponding increase in isotretinoin prescriptions for telehealth NPVs. This finding is in line with recent work highlighting higher antibiotic and spironolactone prescription rates among acne patients seen via teledermatology relative to those seen in-person. The lower antibiotic prescription rates in telehealth RPVs for acne may be related to an increase in telehealth visits for isotretinoin follow-up during the pandemic in the setting of reduced iPLEDGE requirements.

The similar prescription rates across visit types suggest that telehealth is not associated with inappropriate antibiotic prescribing practices among dermatologists. Although it is unclear why this finding differs from studies of other specialties, one potential explanation is that dermatologists may be more comfortable with telehealth, given the visual nature of dermatologic practice.

This study is limited by its single center design and future studies are needed to evaluate whether these findings generalize to other populations. In addition, while the demographics of the telehealth and in-person cohorts were relatively comparable, there may be differences in disease severity between groups.

Conflicts of interest

None disclosed.