An exploratory analysis recently published in the Journal of the American Academy of Dermatology examines whether it is possible to classify dermatologists and internists into different patterns of prescribing behavior for patients with acne.
“Prior research has highlighted that prescribing for acne may not be aligned with guideline recommendations, including the overuse of oral antibiotics and lack of use of concomitant topical medications such as topical retinoids,” the study’s corresponding author,, of the department of dermatology at the University of Pennsylvania, Philadelphia, said in an interview.
“In addition, there is substantial variation in prescribing practices among clinicians.. By identifying such groups, it would facilitate future qualitative interviews to understand factors that might contribute to clinicians having certain prescribing patterns, which could help guide implementation science work to better align practices with evidence and guidelines.”
For the, which appeared online on March 1, Dr. Barbieri and colleague , professor of dermatology and epidemiology at the University of Pennsylvania, evaluated all clinical encounters associated with an ICD-9 or ICD-10 code for acne that occurred in the university’s departments of dermatology and internal medicine between Jan. 1, 2011, and Dec. 31, 2019. They used a machine-learning method known as k-means clustering to cluster clinicians based on their relative use of acne medications, as well as the ratio of spironolactone versus tetracycline use among female patients and stratified their analyses by specialty.
Of the 116 dermatologists included in the analysis, the researchers identified three clusters. The first cluster included 17 dermatologists (14.7%) and was characterized by low use of topical retinoids, high use of oral tetracycline, and low use of spironolactone, compared with oral antibiotics, among women with acne. Physicians in this cluster were more likely to be male and to have more years in practice.
The second cluster included 46 dermatologists (39.6%) and was marked by high use of spironolactone and low use of isotretinoin. The third cluster included 53 dermatologists (45.7%) and was characterized by high use of topical retinoids and frequent use of systemic medications.
Of the 86 internists included in the study, the researchers identified three clusters. The first cluster included 39 internists (45.4%) and was characterized by low use of topical retinoids, high use of oral tetracycline, and limited use of spironolactone. The second cluster included 34 internists (39.5%) and was marked by low use of topical retinoids and systemic medications. The third cluster included 13 clinicians (15.1%), most of whom were nurse practitioners, physician assistants, and other advanced practice providers. This cluster was characterized by high use of topical retinoids and relatively high use of spironolactone.
“There are likely opportunities to improve the use of topical retinoids by internists caring for patients with acne, since these are a first-line treatment option that may be underutilized by internists,” Dr. Barbieri said in the interview. “Future work is needed to identify underlying factors associated with different prescribing phenotypes among both dermatologists and internists. By understanding these factors, we can develop implementation science efforts to align prescribing behavior with best practices based on the guidelines and available evidence.”
He acknowledged certain limitations of the analysis, including its single-center design and the lack of data on patient characteristics. “Future studies are needed to examine whether our results generalize to other settings,” he said.
Dr. Barbieri disclosed that he receives partial salary support through a Pfizer Fellowship in Dermatology Patient Oriented Research grant to the Trustees of the University of Pennsylvania. The authors had no other disclosures.