Although collectively dermatologists are doing more, each dermatologist may be doing less in the higher quintiles.
Due to supply sensitivity and practice differences between dermatology providers and other specialists who provide dermatology services, adding more dermatologists won't necessarily increase access to dermatology services suggests an analysis published in JAMA Dermatology.
Published online November 29, researchers led by Arash Mostaghimi, M.D., MPH, of Brigham and Women's Hospital/Harvard Medical School, undertook the study to understand why areas like Boston have too few dermatologists to meet demand even though it's among the most dermatologist-dense areas in the country.
"We were focused on the difference in dermatologic care between areas with many dermatologists versus fewer dermatologists. When there are no dermatologists around, who is taking care of skin problems?"
Using 2013 Medicare provider utilization and payment data to calculate physician density and utilization, researchers focused on the top 50 outpatient dermatology-related procedure codes (excluding dermatopathology) and associated reimbursement rates. Mostaghimi et al. also used U.S. Census Bureau data to calculate ratios of dermatologists and procedures per population.
Within Medicare, dermatologist density per 100,000 persons ranged from 5.3 in the first quintile to 54.8 in the fifth quintile. Dermatologists' most commonly reported codes included 17003, destruction of multiple premalignant growths (47%); 17000, destruction of single premalignant growth (14%); and other codes relating to biopsies and removal of benign skin lesions.
Of 44 million dermatologic procedures costing $2.76 billion in 2013, dermatologists accounted for 28 million procedures (64%) at a cost of $2.21 billion (83%). Average per-person dermatology spending ranged from $15.87 in the lowest (least dense) quartile to $92.02 in the fifth quintile. Linear regression analysis showed that for each increase of 10 dermatologists per 100,000 population, spending increased $14.81 per person (P = 0.005). Conversely, mean utilization per dermatologist fell from $2.99 in the first quintile to $1.68 in the fifth.
Although these figures must be interpreted carefully, says Dr. Mostaghimi, "It could mean that there is less efficient clinical care when you have more dermatologists; they have other responsibilities so they're in clinic less, or they may have cash-based or cosmetically oriented practices. Non-Medicare patients may account for a greater proportion of visits. Although collectively the dermatologists are doing more, each dermatologist may be doing a little bit less" in the higher quintiles.
Overall, "We found that consistently, the more dermatologists there were, the more dermatologic services were billed. That's not shocking, but what was interesting was that in our analysis, it never hit a plateau." The more dermatologists practicing in an area, the more dermatology procedures are billed under Medicare.
Therefore, "Adding more dermatologists is probably not the way to improve access, at least in a reliable long-term way. It seems as though demand consistently increases to fill the supply."
Shifting dermatologist demand directly to PCPs may not solve access issues either. Dermatologists and nurse practitioners/physician assistants perform a roughly similar mix of services, with the most commonly billed procedure being destruction of premalignant lesions (66% of procedural volume versus 48%, respectively).
Perhaps mid-level extenders' practice patterns closely mirror those of dermatologists, Dr. Mostaghimi says, because they work within dermatology practices. "In many ways, these mid-level providers may be serving as dermatology substitutes, doing a similar distribution of tasks." Conversely, PCPs and other specialists most often performed injections (89% and 74%, respectively), and only 9% of PCPs' procedures involved destruction of premalignant lesions.
Mostaghimi et al. write that although several studies have noted that PCPs perform dermatologic procedures, "Our results suggest that they may be imperfect substitutes for dermatologists. Across all quintiles, PCPs primarily bill for injections rather than for skin biopsies or destructions of premalignant lesions. They seem to be less focused on treatment and prevention of skin cancer."
The study's findings are unlikely to impact dermatologists' daily practice, says Dr. Mostaghimi, "But as organizations, institutions and healthcare systems on a larger scale try to figure out how to deal with dermatologic demand, adding dermatologists may improve access to a certain extent. But we also must rethink what do we do that adds value – what's the best way for dermatologists to spend their time, and other ways we can provide high-quality dermatologic care that's not done by dermatologists or that scales dermatology, whether it's through mid-level providers, teledermatology or other mechanisms. It seems that demand, best we can tell, it is hard to fill by just expanding the workforce."
Using Medicare data provided a comprehensive national snapshot, says Dr. Mostaghimi, though the data apply only to people over age 65. And just as the report does not answer why dermatologists in higher-density markets do more procedures, it does not answer whether simply performing more dermatologic procedures improves patient outcomes. "There's some suggestion that places that have more dermatologists have better melanoma outcomes.2,3 Maybe that's the first sign," which could provide the impetus for examining outcomes across dermatologic conditions.
Ultimately, says Dr. Mostaghimi, the study is "a jumping-off point to ask the question why is dermatology performed by dermatologists important? Do dermatologists do dermatology more efficiently and effectively than non-dermatologists? And if so, how should we think about distributing dermatologists throughout America? There are some conditions where dermatologists are critical. How do we focus on those and leverage other practitioners to take care of other skin diseases and expand our reach?"
If research ultimately reveals that high dermatologist density improves outcomes, "Our goal should be to increase the number of dermatologists everywhere in America." But if crowded markets produce no better outcomes than sparse ones, "Perhaps policy should try to de-incentivize people from coming to places with the highest density and redistribute dermatologists differently to provide more consistent and standardized dermatologic care throughout the nation."
1. Tan SY, Tsoucas D, Mostaghimi A. Association of dermatologist density with the volume and costs of dermatology procedures among Medicare beneficiaries. JAMA Dermatol. 2017 Nov 29. doi: 10.1001/jamadermatol.2017.4546.
2. Aneja S, Aneja S, Bordeaux JS. Association of increased dermatologist density with lower melanoma mortality. Arch Dermatol. 2012;148(2):174-8.