Comparison of Dermatologist Density Between Urban and Rural Counties in the United States

Single Model Deep Learning on Imbalanced Small Datasets for Skin Lesion Classification

Deep convolutional neural network (DCNN) models have been widely explored for skin disease diagnosis and some of them have achieved the diagnostic outcomes comparable or even superior to those of dermatologists. However, broad implementation of DCNN in skin disease detection is hindered by small size and data imbalance of the publically accessible skin lesion datasets. This paper proposes a novel single-model based strategy for classification of skin lesions on small and imbalanced datasets. First, various DCNNs are trained on different small and imbalanced datasets to verify that the models with moderate complexity outperform the larger models. Second, regularization DropOut and DropBlock are added to reduce overfitting and a Modified RandAugment augmentation strategy is proposed to deal with the defects of sample underrepresentation in the small dataset. Finally, a novel Multi-Weighted New Loss (MWNL) function and an end-to-end cumulative learning strategy (CLS) are introduced to overcome the challenge of uneven sample size and classification difficulty and to reduce the impact of abnormal samples on training. By combining Modified RandAugment, MWNL and CLS, our single DCNN model method achieved the classification accuracy comparable or superior to those of multiple ensembling models on different dermoscopic image datasets. Our study shows that this method is able to achieve a high classification performance at a low cost of computational resources and inference time, potentially suitable to implement in mobile devices for automated screening of skin lesions and many other malignancies in low resource settings.

Disparities in access for melanoma screening by region, specialty, and insurance: A cross-sectional audit study

Background

Early detection of melanoma is critical for positive outcomes. However, access for the diagnosis of melanoma remains problematic for segments of the general population.

Objective

To compare the rates of dermatology and family medicine practitioner acceptances for a public insurance (Medicaid) versus private insurance (Anthem Blue Cross) and clinic wait times for an appointment for a changing pigmented skin lesion concerning melanoma in rural and urban regions in California.

Methods

Cross-sectional audit study between June 2017 and March 2019; scripted phone calls were made to dermatology and family medicine practices (FMPs).

Results

Family medicine and dermatology practices in both regions had significantly decreased acceptance of Medicaid. Dermatology practices had 11.3% to 13.0% Medicaid acceptance rates that were less than FMP rates of 28% to 36%. In both regions, FMP wait times were 2.4- to 3.2-fold longer for public versus private insurance; there were little differences in wait times for the 2 insurance types in dermatology practices, in both regions.

Limitations

Assessment of only 2 regions in the state of California.

Conclusion

Delays at FMPs and insurance types limit access to melanoma screening in California for underserved segments of the general population, which has implications for melanoma outcomes and health policy.

Dermatology workforce over a decade in Saudi Arabia: demographics, distributions, and future challenges

BACKGROUND

The dermatology workforce is an important topic, as many countries are facing an undersupply of dermatologists, while some are expecting a surplus. Therefore, we conducted this study to identify the current dermatology workforce demographics in Saudi Arabia (SA) and the changes in such demographics over the last 10 years to identify future workforce-related challenges.

METHODS

This study was conducted in SA, and it included all the practicing dermatologists in the country over the last decade (2010-2020). The number of practicing dermatologists, their gender, their nationality, and dermatology residency candidates and graduates were obtained from the Saudi Commission for Health Specialties (SCFHS). The geographic distribution of dermatologists was obtained from the Ministry of Health Statistical Yearbook 2018.

RESULTS

As of September 2020, there were 2678 practicing dermatologists in SA at a ratio of 7.82 dermatologists per 100 000 people. Of the 2678 dermatologists, only 24.8% were Saudis. The Saudi dermatologist ratio has been almost constant over 10 years, ranging from 1.3 to 1.9 per 100 000 people. Of all Saudi dermatologists, 42% were female. The number of residents who graduated from the residency program was not consistent for each year and ranged from 4 to 25. The number of dermatologists varied by region, with 9.2 in Riyadh and 3.4 in Najran per 100 000 people.

CONCLUSIONS

The results of our study revealed that a quarter of dermatologists in SA are Saudis. In addition, the number of non-Saudi dermatologists has increased in the last 10 years, while the number of Saudi dermatologists to the population has remained almost constant. There is also a geographic maldistribution of dermatologists, with urban areas having a higher number of dermatologists than rural areas. We encourage local studies that can elucidate the factors influencing the workforce, such as the dermatologist appointment waiting time, dermatologists' working hours, and the geographic maldistribution of dermatologists in the country.

Usefulness of Smartphones in Dermatology: A US-Based Review

(1) Background: As smartphones have become more widely used, they have become an appealing tool for health-related functions. For dermatology alone, hundreds of applications (apps) are available to download for both patients and providers. (2) Methods: The Google Play Store and Apple App Store were searched from the United States using dermatology-related terms. Apps were categorized based on description, and the number of reviews, download cost, target audience, and use of AI were recorded. The top apps from each category by number of reviews were reported. Additionally, literature on the benefits and limitations of using smartphones for dermatology were reviewed. (3) Results: A total of 632 apps were included in the study: 395 (62.5%) were marketed towards patients, 203 (32.1%) towards providers, and 34 (5.4%) towards both; 265 (41.9%) were available only on the Google Play Store, 146 (23.1%) only on the Apple App Store, and 221 (35.0%) were available on both; and 595 (94.1%) were free to download and 37 (5.9%) had a cost to download, ranging from USD 0.99 to USD 349.99 (median USD 37.49). A total of 99 apps (15.7%) reported the use of artificial intelligence. (4) Conclusions: Although there are many benefits of using smartphones for dermatology, lack of regulation and high-quality evidence supporting the efficacy and accuracy of apps hinders their potential.

General and Vulnerable Population’s Satisfaction With the Healthcare System in Urban and Rural Areas: Findings From the European Social Survey

Introduction: Access to the healthcare system when patients are vulnerable and living outside metropolitan areas can be challenging. Our objective was to explore healthcare system satisfaction of urban and rural inhabitants depending on financial and health vulnerabilities.

Methods: Repeated cross-sectional data from 353,523 European citizens (2002–2016). Multivariable associations between rural areas, vulnerability factors and satisfaction with the healthcare system were assessed with linear mixed regressions and adjusted with sociodemographic and control factors.

Results: In unadjusted analysis, the people who lived in houses in the countryside and those who lived in the suburbs were the most satisfied with the healthcare system. In the adjusted model, residents living in big cities had the highest satisfaction. Financial and health vulnerabilities were associated with less satisfaction with the healthcare system, with a different effect according to the area of residence: the presence of health vulnerability was more negatively correlated with the healthcare system satisfaction of big city inhabitants, whereas financial vulnerability was more negatively correlated with the satisfaction of those living in countryside homes.

Conclusion: Vulnerable residents, depending on their area of residence, may require special attention to increase their satisfaction with the healthcare system.

Exploring convolutional neural networks with transfer learning for diagnosing Lyme disease from skin lesion images

BACKGROUND AND OBJECTIVE

Lyme disease which is one of the most common infectious vector-borne diseases manifests itself in most cases with erythema migrans (EM) skin lesions. Recent studies show that convolutional neural networks (CNNs) perform well to identify skin lesions from images. Lightweight CNN based pre-scanner applications for resource-constrained mobile devices can help users with early diagnosis of Lyme disease and prevent the transition to a severe late form thanks to appropriate antibiotic therapy. Also, resource-intensive CNN based robust computer applications can assist non-expert practitioners with an accurate diagnosis. The main objective of this study is to extensively analyze the effectiveness of CNNs for diagnosing Lyme disease from images and to find out the best CNN architectures considering resource constraints.

METHODS

First, we created an EM dataset with the help of expert dermatologists from Clermont-Ferrand University Hospital Center of France. Second, we benchmarked this dataset for twenty-three CNN architectures customized from VGG, ResNet, DenseNet, MobileNet, Xception, NASNet, and EfficientNet architectures in terms of predictive performance, computational complexity, and statistical significance. Third, to improve the performance of the CNNs, we used custom transfer learning from ImageNet pre-trained models as well as pre-trained the CNNs with the skin lesion dataset HAM10000. Fourth, for model explainability, we utilized Gradient-weighted Class Activation Mapping to visualize the regions of input that are significant to the CNNs for making predictions. Fifth, we provided guidelines for model selection based on predictive performance and computational complexity.

RESULTS

Customized ResNet50 architecture gave the best classification accuracy of 84.42% ±1.36, AUC of 0.9189±0.0115, precision of 83.1%±2.49, sensitivity of 87.93%±1.47, and specificity of 80.65%±3.59. A lightweight model customized from EfficientNetB0 also performed well with an accuracy of 83.13%±1.2, AUC of 0.9094±0.0129, precision of 82.83%±1.75, sensitivity of 85.21% ±3.91, and specificity of 80.89%±2.95. All the trained models are publicly available at https://dappem.limos.fr/download.html, which can be used by others for transfer learning and building pre-scanners for Lyme disease.

CONCLUSION

Our study confirmed the effectiveness of even some lightweight CNNs for building Lyme disease pre-scanner mobile applications to assist people with an initial self-assessment and referring them to expert dermatologist for further diagnosis.

Factors Influencing Career Decisions and Satisfaction Among Newly Practicing Ophthalmologists

PURPOSE

To describe the career choices of newly practicing ophthalmologists and explore factors influencing career decisions and satisfaction.

METHODS

A cross-sectional study was conducted using data from an electronic survey of ophthalmologists who completed training within the prior 5 years. The survey included questions about demographic information, medical education, current practice, factors affecting career choices, and career satisfaction. Statistical comparisons were made based on gender, type of practice, subspecialty training, and practice area.

RESULTS

Surveys were completed by 696 (32%) newly practicing ophthalmologists, including 276 (40%) women, 179 (29%) academicians, and 465 (67%) subspecialists. A higher proportion of female respondents entered academics than male respondents (36% vs 26%, P = .009). Female and male respondents pursued fellowship training with similar frequency (64% vs 68%, P = .32), but men were more likely to seek vitreoretinal fellowships (30% vs 11%, P < .001) and women were more likely to undertake fellowships in pediatric ophthalmology (21% vs 8%, P < .001), uveitis (10% vs 2%, P = .002), and neuro-ophthalmology (6% vs 2%, P = .042). A total of 514 (83%) respondents reported being happy with work life.

CONCLUSIONS

The career choices of newly practicing ophthalmologists differ based on gender, type of practice, subspecialty training, and practice area. Many factors affect career decisions, and they have varying influence on subgroups within ophthalmology. Ophthalmologists have high levels of career satisfaction. This information may prove useful when developing workforce strategies to meet future eye care needs. NOTE: Publication of this article is sponsored by the American Ophthalmological Society.

Trends in Patch Testing in the Medicare Part B Fee-for-Service Population

BACKGROUND

Patch testing is a vital component of the workup for allergic contact dermatitis. There are limited data on changes of patch testing use among Medicare providers, as well as patch testing reimbursement rates.

OBJECTIVE

The aim of the study was to evaluate trends in the use of patch testing among various Medicare providers and Medicare patch testing reimbursement.

DESIGN

A longitudinal analysis of patch testing claims was performed with the Medicare Part B Physician/Supplier Procedure Summary files from 2010 to 2018. The primary outcomes were the total number and change in the number of submitted patch testing services from 2010 to 2018 by 3 provider groups: dermatology physicians, nondermatology physicians, and nonphysician providers. Secondary outcome measures included Medicare reimbursement amounts and changes in reimbursement amounts for patch test services (total and per 1000 enrollees) from 2010 to 2018 for the 3 provider groups, as well as per patch test service.

RESULTS

From 2010 to 2018, submitted patch testing services per 1000 enrollees grew by 89.0%. The annual trend estimate for submitted services relative to 2010 was +10.1% (95% confidence interval [CI] = 8.1 to 12.0) for physicians and +34.1% (95% CI = 32.1 to 36.0) for nonphysician providers (physician assistants and nurse practitioners). Among physicians, the annual trend estimate for submitted services was +5.1% (95% CI = -11.3 to 21.5) for dermatologists and +31.40% (95% CI = 15.00 to 47.81) for allergists.

CONCLUSIONS

Patch testing increased in the US Medicare population from 2010 to 2018, and this increase was largely driven by nonphysician providers and allergists.

Geographic distribution and characteristics of the pediatric dermatology workforce in the United States

BACKGROUND/OBJECTIVES

Up to 30% of pediatric primary care visits include a cutaneous complaint, yet the pediatric dermatology workforce has historically been too small to provide adequate specialized care. This study assesses the geographic distribution of pediatric dermatologists to determine physician-to-patient ratios, analyzes urban-rural disparities, and determines post-fellowship migration patterns.

METHODS

Board-certified pediatric dermatologists were identified using the Society for Pediatric Dermatology's public database, and their demographics and credentials were subsequently verified by an online search. Analysis included physician density per 100 000 children for each state and region, along with geographic distribution for rural and urban areas, based on the United States Census Bureau's definitions. The distances between practice locations and the American Board of Dermatology-approved Pediatric Dermatology fellowship training sites were reviewed.

RESULTS

An estimated 336 board-certified pediatric dermatologists currently work in the United States with 76.8% being women and 71.1% practicing within 50 miles of the nearest fellowship program. 96.4% are located in urban areas and 3.6% in rural areas with an average ratio of 0.54 and 0.09 per 100 000 children, respectively. The average ratio of pediatric dermatologists in the United States was 0.46 per 100 000 children. On average (standard deviation), there are 6.6 (8.8) pediatric dermatologists per state but with 7 states having zero.

CONCLUSIONS

The demand for pediatric dermatologists continues to outpace the current physician availability with a disparity between urban and rural areas. Further awareness and emphasis on training and recruitment of additional pediatric dermatologists are essential to addressing this important issue.

Oculofacial plastic surgeon distribution by county in the United States, 2021

PURPOSE

To characterize the number of oculofacial plastic surgeons (OPS) per county in the United States (U.S.).

METHODS

The 2021 public databases of the American Society of Ophthalmic Plastic and Reconstructive Surgery and the American Academy of Ophthalmology were used to identify all OPS in the U.S. Surgeon practice location was used to determine per capita physician density by county.

RESULTS

A total of 1184 OPS in the U.S. were identified. Three hundred forty-eight counties were served by at least one OPS whereas 2795 counties (89%), and two states, North Dakota and Wyoming, had no OPS. The average ratio of OPS to 100,000 population was 0.3572 (1 per 279,955). Of the counties with at least one OPS, the average was 0.5860 surgeons per 100,000 population (1 per 170,648), ranging from 0.0705 (1 per 1,418,440) to 11.26 (1 per 8,881) per 100,000. The counties with the greatest OPS density were Pitkin County, CO (1 per 8,881), San Juan County, WA (1 per 17,580), and Montour County, PA (1 per 18,231). Counties with the lowest density of those with at least one OPS were Bronx County, NY (1 per 1,418,238), San Bernardino County, Ca (1 per 1,090,037), and Gwinnett County, GA (1 per 936,329). The counties with the most OPS were Los Angeles County, CA (46), New York County, NY (38), and Cook County, IL (25).

CONCLUSIONS

Geographic disparities in OPS distribution exist in the U.S. Future investigations of OPS supply according to population and other characteristics for demand may be useful.

Melanoma classification using light-Fields with morlet scattering transform and CNN: Surface depth as a valuable tool to increase detection rate

Medical image classification through learning-based approaches has been increasingly used, namely in the discrimination of melanoma. However, for skin lesion classification in general, such methods commonly rely on dermoscopic or other 2D-macro RGB images. This work proposes to exploit beyond conventional 2D image characteristics, by considering a third dimension (depth) that characterises the skin surface rugosity, which can be obtained from light-field images, such as those available in the SKINL2 dataset. To achieve this goal, a processing pipeline was deployed using a morlet scattering transform and a CNN model, allowing to perform a comparison between using 2D information, only 3D information, or both. Results show that discrimination between Melanoma and Nevus reaches an accuracy of 84.00, 74.00 or 94.00% when using only 2D, only 3D, or both, respectively. An increase of 14.29pp in sensitivity and 8.33pp in specificity is achieved when expanding beyond conventional 2D information by also using depth. When discriminating between Melanoma and all other types of lesions (a further imbalanced setting), an increase of 28.57pp in sensitivity and decrease of 1.19pp in specificity is achieved for the same test conditions. Overall the results of this work demonstrate significant improvements over conventional approaches.

Teledermatology in the Control of Skin Neglected Tropical Diseases: A Systematic Review

Introduction

Neglected tropical diseases (NTDs) include a group of about 20 illnesses that have garnered relatively little attention, despite their ability to inflict significant suffering and disability. Skin neglected tropical diseases (sNTDs) are a subset of NTDs that present with cutaneous manifestations and are well known for their ability to generate stigma and promote poverty. Teledermatology (TD) represents a potential method to control sNTDs.

Objective

We sought to analyze the potential for TD to ease the burden of sNTDs.

Methods

We performed a systematic literature search using the Texas Medical Center Library One Search, which scans 167 databases, including Embase, PubMed, and Scopus. We included all original investigations published after 2011 that assessed the impact of TD intervention in the control of one or more sNTDs. We excluded studies not written in English and studies that did not perform any outcome analyses.

Results

Twenty studies met our search criteria, and 18 expressed positive attitudes towards TD. Overall, we found that TD may be a sustainable, cost-effective strategy for expanding access to care for individuals afflicted with sNTDs. However, poor image quality, lack of access to further diagnostic tests, and ethical, legal, and cultural issues pose as barriers to TD utilization.

Conclusion

TD may be helpful in achieving control of sNTDs but has its limitations. An integrated approach, which employs TD in conjunction with other strategies, represents a realistic path for alleviating sNTDs.

Gender differences in dermatologist practice locations in the United States: A cross-sectional analysis of current gender gaps

Background

The percentage of female dermatologists has increased from 6.9% in 1970 to 48.9% in 2017. Despite the changing gender composition of the dermatologist workforce, it is unknown whether there are gender-based differences in dermatology practice locations.

Objective

This study aimed to characterize gender-based differences in dermatology practice locations across the United States.

Methods

A cross-sectional study of all dermatologists in the 2020 Centers for Medicare and Medicaid Services Physician Compare Database was performed. The number of self-identified female dermatologists and total dermatologists in each county and state was tabulated, and Spearman's correlation coefficients between county-level demographic and socioeconomic characteristics and female practices were calculated.

Results

Among 11,911 dermatologists, 5945 (49.9%) self-identified as female and 5966 (50.1%) as male. Of the 1052 counties with a dermatologist, 291 (27.7%) had no female dermatologist and 149 (14.2%) had no male dermatologist. The percentage of female dermatologists in each state ranged from 18.4% to 62.2%. Female dermatologists practiced more in areas with a higher percentage of democratic voters (r = +0.22) and higher median household income (r = +0.18), and less in rural counties (r = -0.18) or counties with higher uninsured rates (r = -0.11).

Conclusion

Female dermatologists remain significantly underrepresented in some regions in the United States, particularly in the Mountain states and rural counties. As women continue entering the dermatologist workforce, these results can inform workforce planning strategies to improve the distribution and accessibility of dermatologists across the United States.

Trends in office visits and treatment for urticaria in children in the United States, 1998-2016

BACKGROUND/OBJECTIVES

Urticaria is a common condition with an estimated prevalence of up to 23% in the pediatric population. Studies characterizing visits and treatments for urticaria in the pediatric population are unavailable. Understanding visit and treatment trends for urticaria in the pediatric population may help inform care for patients with urticaria.

METHODS

A total of 108 278 outpatient records from the National Ambulatory Medical Care Survey representing 3.4 billion visits by patients of age 18 and younger were analyzed. This study included the calendar years 1998 through 2016.

RESULTS

Pediatricians saw the largest proportion of all visits (52.7%). Male and female patients accounted for approximately equal proportions of all visits for urticaria. There was a slight male predominance in visits to pediatricians (53.7%), whereas dermatologists saw female patients more frequently (63.3%). Most visits for urticaria were by non-Hispanic (78.1%) and White (78.2%) patients. H1 antihistamines were the most commonly prescribed treatment (70.3%), whereas topical corticosteroids were prescribed least frequently (4.9%). Topical corticosteroids were most frequently prescribed by dermatologists (7.7%). Non-H1 antihistamine and non-corticosteroid therapy were prescribed in 9.7% of all visits and in 4.5% of visits to pediatricians. Most visits for urticaria were to physicians in metropolitan areas (88.8%). Pediatricians saw the highest number of non-metropolitan area visits (56.3%).

CONCLUSIONS

H1 antihistamines were the most commonly used therapy (70%), consistent with established treatment guidelines. Male and female pediatric patients present equally often for urticaria, but sex differences were seen with visit frequencies to certain specialties.

The policy dimensions, regulatory landscape, and market characteristics of teledermatology in the United States

The COVID-19 pandemic has spurred healthcare systems across the world to rapidly redesign their models of care delivery. As such, this pandemic has accelerated the adoption of teledermatology in the United States. However, it remains unknown whether this momentum will be maintained after the pandemic. The future of teledermatology in the United States will be significantly influenced by a complex set of policy, legal, and regulatory frameworks. An understanding of these frameworks will help dermatologists more effectively adopt and implement teledermatology platforms. In this article, we review the current state of teledermatology in the United States, including policy dimensions, the regulatory landscape, market characteristics, and future directions.

Changes in dermatology practice characteristics in the United States from 2012 to 2017

Background

Dermatology practice has recently seen multiple changes. A better understanding of trends pertaining to dermatology practice setups is necessary.

Objective

To analyze the recent changes in dermatology practice in terms of geography, practice size, and gender distribution as well as to analyze the availability of dermatologists based on zip codes’ income levels.

Methods

This was a cross-sectional study. We extracted data on the sex and billing addresses of dermatologists from Medicare provider utilization and payment data for 2012 and 2017. We used 2017 tax returns data to calculate the poverty level for each zip code.

Results

Between 2012 and 2017, the number of solo practitioners decreased, while that of dermatologists working in large groups increased. The southern region experienced the largest changes. The male-to-female ratio decreased. Dermatology practices mainly comprised mixed genders, with a higher proportion of all-male groups versus that of all-female groups, but this difference decreased over time. In the northeastern and western regions, more than one third of dermatologists were located in the wealthiest zip codes.

Limitations

The Medicare data may not be exhaustively representative of the dermatology workforce, and the zip codes of 489 dermatologists’ billing addresses were missing in the tax return dataset.

Conclusions

These findings provide an understanding of the recent changes pertaining to dermatology practice setups and of the substantial health care disparities based on geographic distribution.

Teledermatology During COVID-19: An Updated Review

The coronavirus disease 2019 (COVID-19) pandemic has fundamentally transformed the landscape of providing dermatologic care. In an age of lockdowns and social distancing, teledermatology (TD) has emerged as a powerful tool to deliver remote care. Here, we review literature on TD use during the pandemic to evaluate the positives and negatives of TD implementation. We especially consider the reception of TD in underserved communities and the developing world as well as the ethico-legal challenges wrought by the burgeoning utilization of this new paradigm of care. The potential of TD to occupy a more prominent role in dermatologic care in a post-COVID-19 world is also discussed.

Employment and Migration Patterns of Recent Micrographic Surgery and Dermatologic Oncology Fellowship Graduates

BACKGROUND

There is currently an uneven distribution of the Mohs surgery workforce, with a dearth in nonurban areas. The relationship between training and employment locations of Mohs surgeons has not been studied.

OBJECTIVE

To analyze the employment location of recent micrographic surgery and dermatologic oncology (MSDO) graduates in reference to residency and fellowship locations.

MATERIALS AND METHODS

Data collection of training and practice locations of 421 MSDO fellowship graduates from 2012 to 2017.

RESULTS

Thirty-two percent (n = 136) and 53% (n = 225) of MSDO fellowship graduates' first employment locations were located within 10 and 100 miles of either their residency or fellowship locations. Ninety-six percent of graduates were employed in a metropolitan location. Female graduates worked closer to training sites than male graduates, with 62% (n = 124) versus 46% (n = 102) working within 100 miles of either training site. Analysis by fellowship census region showed that graduates clustered around training sites in all regions, with 45% to 60% of graduates working within 100 miles of either training site.

CONCLUSION

The majority of graduates chose to work in a metropolitan area. Training location strongly predicts employment location. More than 45% of graduates in any region worked within 100 miles of their residency or fellowship site, and a sizeable portion worked within 10 miles.

Teledermatology Addressing Disparities in Health Care Access: a Review

Purpose of Review

Dermatologists have been at the forefront of researching telemedicine to expand access to care. The current COVID-19 pandemic has prompted even greater expansion and implementation of teledermatology. This review discusses the research examining the potential impact of teledermatology addressing disparities in care.

Recent Findings

Teledermatology appears to increase access to dermatology given expanded means to deliver care. Specifically, recent studies have found increased access among Medicaid-insured, resource-poor urban and rural, and elderly populations. Teledermatology implementation also facilitates education among providers at different levels of training. Still, as some patients have inconsistent access to the required technology, increased reliance on telemedicine may also potentially increase disparities for some populations.

Summary

Teledermatology may serve to reduce disparities in health care access in many underserved and marginalized communities. Future research should continue to study implementation, especially given the expansion during the COVID-19 pandemic. Ultimately, teledermatology may play an important role in ensuring equitable care access for all.

Deep learning based classification of facial dermatological disorders

Common properties of dermatological diseases are mostly lesions with abnormal pattern and skin color (usually redness). Therefore, dermatology is one of the most appropriate areas in medicine for automated diagnosis from images using pattern recognition techniques to provide accurate, objective, early diagnosis and interventions. Also, automated techniques provide diagnosis without depending on location and time. In addition, the number of patients in dermatology departments and costs of dermatologist visits can be reduced. Therefore, in this work, an automated method is proposed to classify dermatological diseases from color digital photographs. Efficiency of the proposed approach is provided by 2 stages. In the 1st stage, lesions are detected and extracted by using a variational level set technique after noise reduction and intensity normalization steps. In the 2nd stage, lesions are classified using a pre-trained DenseNet201 architecture with an efficient loss function. In this study, five common facial dermatological diseases are handled since they also cause anxiety, depression and even suicide death. The main contributions provided by this work can be identified as follows: (i) A comprehensive survey about the state-of-the-art works on classifications of dermatological diseases using deep learning; (ii) A new fully automated lesion detection and segmentation based on level sets; (iii) A new adaptive, hybrid and non-symmetric loss function; (iv) Using a pre-trained DenseNet201 structure with the new loss function to classify skin lesions; (v) Comparative evaluations of ten convolutional networks for skin lesion classification. Experimental results indicate that the proposed approach can classify lesions with high performance (95.24% accuracy).

Who sees you matters: a population study examining topical corticosteroid prescribing patterns between primary care providers and dermatologists for atopic dermatitis

BACKGROUND

Many patients with atopic dermatitis seek care from both primary care physicians and dermatologists. However, little is known regarding topical corticosteroid prescribing patterns among these specialties.

OBJECTIVE

We sought to determine if differences exist in topical corticosteroid prescribing patterns among dermatologists, family medicine physicians, and internal medicine physicians.

METHODS

We conducted a population-based, cross-sectional analysis using data from the U.S. National Ambulatory Medical Care Survey from 2006 to 2016.

RESULTS

Compared to dermatologists, internal medicine physicians were 22 times less likely to prescribe a topical corticosteroid for atopic dermatitis (52.2% versus 5.1%, p = .001; adjusted OR 0.045, 95%CI 0.007-0.277). There was not a statistically significant difference in the rate of topical corticosteroid prescriptions for atopic dermatitis between family medicine physicians and dermatologists (39.1% vs. 52.2%, p = .27; adjusted OR 0.468, 95%CI 0.174-1.257). Family medicine physicians had a higher rate of prescribing topical corticosteroids for atopic dermatitis than internal medicine physicians (39.1% vs. 5.1%, p = .002).

LIMITATIONS

Severity of atopic dermatitis was not assessed.

CONCLUSIONS

Atopic dermatitis patients seen by internal medicine physicians are much less likely to receive topical corticosteroid prescriptions as compared to those seen by dermatologists.

National Trends in the United States Eye Care Workforce from 1995 to 2017

PURPOSE

To describe temporal and geographic trends in the US eye care workforce.

DESIGN

Cross-sectional study.

METHODS

We obtained data from the 2017 Area Health Resources File. The main outcomes were ophthalmologist and optometrist density, as defined as the number of providers per 100,000 individuals, the ratio of ophthalmologists ≥55 years of age to those <55 years of age, and county characteristics associated with the availability of an ophthalmologist.

RESULTS

From 1995 to 2017, the national ophthalmologist density decreased from 6.30 to 5.68 ophthalmologists per 100,000 individuals. Although rural counties experienced a mean annual increase in ophthalmologist density by 2.26%, they still had a lower mean ophthalmologist density (0.58/100,000 individuals) compared with nonmetropolitan (2.19/100,000 individuals) and metropolitan counties (6.29/100,000 individuals) in 2017. The ratio of older to younger ophthalmologists increased from 0.37 in 1995 to 0.82 in 2017, with the greatest ratio increase occurring in rural counties (0.29 to 1.90). The presence of an ophthalmologist was significantly associated with a greater proportion of individuals with a college degree and health insurance, and more developed health care infrastructure. From 1990 to 2017, the density of optometrists increased from 11.06 to 16.16 optometrists per 100,000 individuals.

CONCLUSIONS

Over the last 2 decades, the national density of ophthalmologists has decreased and the workforce has aged. In contrast, the density of optometrists has increased. Rural counties continue to have a disproportionately lower supply of eye care providers, although some growth has occurred. Given the rising ratio of optometrists to ophthalmologists, it is of interest for future work to determine how the optometrist workforce can best complement potential shortages of ophthalmologists.

PAD-UFES-20: A skin lesion dataset composed of patient data and clinical images collected from smartphones

Over the past few years, different Computer-Aided Diagnosis (CAD) systems have been proposed to tackle skin lesion analysis. Most of these systems work only for dermoscopy images since there is a strong lack of public clinical images archive available to evaluate the aforementioned CAD systems. To fill this gap, we release a skin lesion benchmark composed of clinical images collected from smartphone devices and a set of patient clinical data containing up to 21 features. The dataset consists of 1373 patients, 1641 skin lesions, and 2298 images for six different diagnostics: three skin diseases and three skin cancers. In total, 58.4% of the skin lesions are biopsy-proven, including 100% of the skin cancers. By releasing this benchmark, we aim to support future research and the development of new tools to assist clinicians to detect skin cancer.

Deep Learning for Diagnostic Binary Classification of Multiple-Lesion Skin Diseases

Background: Diagnosis of skin diseases is often challenging and computer-aided diagnostic tools are urgently needed to underpin decision making.

Objective: To develop a convolutional neural network model to classify clinically relevant selected multiple-lesion skin diseases, this in accordance to the STARD guidelines.

Methods: This was an image-based retrospective study using multi-task learning for binary classification. A VGG-16 model was trained on 16,543 non-standardized images. Image data was distributed in training set (80%), validation set (10%), and test set (10%). All images were collected from a clinical database of a Danish population attending one dermatological department. Included was patients categorized with ICD-10 codes related to acne, rosacea, psoriasis, eczema, and cutaneous t-cell lymphoma.

Results: Acne was distinguished from rosacea with a sensitivity of 85.42% CI 72.24–93.93% and a specificity of 89.53% CI 83.97–93.68%, cutaneous t-cell lymphoma was distinguished from eczema with a sensitivity of 74.29% CI 67.82–80.05% and a specificity of 84.09% CI 80.83–86.99%, and psoriasis from eczema with a sensitivity of 81.79% CI 78.51–84.76% and a specificity of 73.57% CI 69.76–77.13%. All results were based on the test set.

Conclusion: The performance rates reported were equal or superior to those reported for general practitioners with dermatological training, indicating that computer-aided diagnostic models based on convolutional neural network may potentially be employed for diagnosing multiple-lesion skin diseases.

The impact of demographics, socioeconomics, and health care access on melanoma outcomes

Disparities in melanoma care exist in the United States. Disparities in provider type, patient demographics, place of residence, insurance status, socioeconomic status, race/ethnicity, and age impact melanoma outcomes. Melanomas detected by dermatologists are thinner, at an earlier stage, and have better survival outcomes compared with detection by primary care providers or patients. Lower socioeconomic status, race/ethnicity, and place of residence are associated with decreased access to or use of dermatologists, or both, and more advanced melanomas at diagnosis. Additionally, uninsured and publicly insured individuals are more likely to present with late-stage melanomas, resulting in worse outcomes. This review provides a comprehensive overview of how structural and patient-level characteristics influence melanoma outcomes in order to inform clinical care and health care policy as it relates to addressing gaps in melanoma care.

A deep learning system for differential diagnosis of skin diseases

Skin conditions affect 1.9 billion people. Because of a shortage of dermatologists, most cases are seen instead by general practitioners with lower diagnostic accuracy. We present a deep learning system (DLS) to provide a differential diagnosis of skin conditions using 16,114 de-identified cases (photographs and clinical data) from a teledermatology practice serving 17 sites. The DLS distinguishes between 26 common skin conditions, representing 80% of cases seen in primary care, while also providing a secondary prediction covering 419 skin conditions. On 963 validation cases, where a rotating panel of three board-certified dermatologists defined the reference standard, the DLS was non-inferior to six other dermatologists and superior to six primary care physicians (PCPs) and six nurse practitioners (NPs) (top-1 accuracy: 0.66 DLS, 0.63 dermatologists, 0.44 PCPs and 0.40 NPs). These results highlight the potential of the DLS to assist general practitioners in diagnosing skin conditions.

Utilization, cost, and provider trends in patch testing among Medicare beneficiaries in the United States from 2012 to 2017

BACKGROUND

Patch testing is the best diagnostic test for allergic contact dermatitis. However, there is presently a lack of data on the test's geographic availability and the characteristics of the providers offering this test across the United States.

OBJECTIVE

To evaluate the geographic variation in the availability of patch testing for the Medicare population and to characterize the temporal trends of patch testing cost, use, and provider specialty from 2012 to 2017.

METHODS

Retrospective cohort study of the Medicare Provider Utilization and Payment Data from 2012 to 2017.

RESULTS

As of 2017, patch testing was available in 20.3% of metropolitan counties and in 1.1% of nonmetropolitan counties. From 2012 to 2017 in metropolitan regions, the number of beneficiaries tested by dermatologists grew by an average annual growth rate of 1.84%, whereas those tested by allergists grew by an average annual growth rate of 20.31%. Most providers that averaged use of 80 or more patches per beneficiary were dermatologists (76.3%).

LIMITATIONS

Analysis was restricted to Medicare Part B claims; data were unavailable on individuals with commercial insurance.

CONCLUSIONS

Most of the increase in patch testing utilization from 2012 to 2017 has been in metropolitan regions. Although growth was especially prominent among allergists in metropolitan counties, the majority of providers performing comprehensive patch testing were dermatologists.