Epidemiology and outcomes of dermatology in-patient consultations in a midwestern U.S. university hospital
- Author(s): Davila, Manuel;
- Christenson, Leslie J;
- Sontheimer, Richard D
- et al.
Published Web Locationhttps://doi.org/10.5070/D364h8j3kz
Epidemiology and outcomes of dermatology in-patient consultations in a Midwestern U.S. university hospital1. Private Practice of Dermatology, Pasedena, Texas. Manuel.email@example.com
Manuel Davila MD1, Leslie J Christenson MD2, Richard D Sontheimer MD3
Dermatology Online Journal 16 (2): 12
2. Division of Dermatologic Surgery, The McFarland Clinic, Ames, Iowa. firstname.lastname@example.org
3. Department of Dermatology, University of Utah School of Medicine, Salt Lake City, Utah. email@example.com
CONTEXT: A paucity of data exists concerning the utilization of in-patient dermatologic consultations. Previous studies on this subject have indicated a knowledge deficit of primary care providers with regard to common dermatoses, prompting a need for more effective teaching mechanisms in this area. OBJECTIVE: To identify dermatologic conditions in the in-patient setting that are frequently misdiagnosed by non-dermatologists in order to improve future patient care and cost reduction through physician education. DESIGN: Retrospective chart review of 271 consecutive dermatologic consultations from primary ward teams between January 20, 1998, and May 19, 1999. SETTING: Non-dermatology in-patient services at a Midwestern state-supported university hospital system in the U.S. PATIENTS: Patients hospitalized on non-dermatology wards with skin problems prompting a formal dermatologic consultation. INTERVENTIONS: None. MAIN OUTCOME MEASURE: Prevalence of dermatologic conditions that are most frequently misdiagnosed on non-dermatology in-patient services. RESULTS: Seventy-six percent of the dermatologic consults were requested by Internal Medicine, Surgery, and Psychiatry departments. Common skin conditions accounted for a large majority of dermatologic consultations including: dermatitis (21.0%) and drug eruption (10.0%). The primary ward team submitted a correct dermatologic diagnosis in only 23.9 percent of cases. Dermatology consultation resulted in a change in or addition to treatment in 77 percent of patients. CONCLUSIONS: Our results suggest that common skin conditions account for a large majority of dermatologic consultations in a University hospital setting. Modern hospital ward teams continue to struggle with accurately recognizing and appropriately managing common skin problems resulting in inappropriate treatment, wasted resources, and prolonged hospitalization. Increasing medical student and house staff knowledge and experience in the diagnosis and management of common skin disorders could help address this problem.
Skin disease is highly prevalent in the United States. It is one of the top 15 groups of medical conditions for which prevalence and health care spending increased the most between 1987 and 2000, with approximately 1 of 3 people in the United States with a skin disease at any given time . The Burden of Skin Diseases Report indicated that 22 skin diseases ranging from melanoma to acne represented a healthcare cost of $39.3 billion in 2004 dollars . The majority of patients with skin disease in the U.S. are managed by physicians not formally trained as dermatologists . This is especially true of inflammatory dermatoses, which pose a particularly challenging area of dermatology (Table 1). There has traditionally been minimal emphasis on education in skin disease in US medical school curricula. Some medical schools in the US have no requirements for a formal clinical rotation on their dermatology service. Others may require at most two weeks on such a rotation. In addition, most medical schools typically offer less than 20 hours of formal dermatology lectures over a four-year curriculum. Thus, it is not surprising that patients hospitalized on non-dermatology in-patient services are frequently found to have skin problems and present as a source of confusion for their admitting physicians.
Over a three decade career in academic dermatology, one of us (R.D.S.) has gained the impression that a limited set of common dermatologic problems (e.g., inflammatory cutaneous drug hypersensitivity reactions, eczematous dermatitis, infections) results in a disproportionately large percentage of requests for dermatologic consultations for patients hospitalized in the two academic health care centers. This impression is supported by the published experience of others . This, plus the growing dermatologist manpower shortage in the United States, would appear to present a need for a new approach to medical education concerning in-patient skin disease for both the sake of improved patient care as well as cost savings.
To examine the scope of this problem, we undertook a retrospective analysis of utilization of dermatology consultations by non-dermatology in-patient services and identified skin conditions associated with these consultation requests.
Informed Consent. Medical record review described in this study was approved by the Institutional Review Board of the University of Iowa College of Medicine.
Patients/Setting. In-patients on non-dermatology services at the University of Iowa Hospitals and Clinics having a skin finding/condition that prompted the primary care in-patient team to seek a formal dermatologic consultation.
Study Design. A retrospective chart review of 271 consecutive dermatologic consultations was performed. These consultation requests were received over a 16-month period between January 20th, 1998 and May 19th, 1999 by the University of Iowa Hospitals and Clinics (UIHC) Department of Dermatology for inpatients on non-dermatology wards at UIHC and the Childrens Hospital of Iowa. The following variables were collected and entered into a Microsoft Access database: patient demographics, past medical history, patient medications, patient allergies, reason for admission, preliminary diagnosis or differential diagnosis, preliminary treatment (if any), reason for consultation, laboratory testing obtained by the consulting dermatologists, final dermatologic diagnosis, and recommended treatment by the dermatology consultant (if any).
Data analysis. We calculated the frequency of dermatologic consultations requested by the various hospital clinical departments. We also calculated the frequency of the primary diagnoses by the non-dermatology teams and compared it to the final diagnoses of the dermatology consult service. Standard mathematical formulas were applied to summarize the data in the Microsoft Access database.
Patient demographics. The age and gender distribution of the study population can be found in Table 2. 87 percent of the dermatologic consultations were for patients over 18 years of age while 56 percent were for individuals over 45 years of age. The male:female ratio was 146 (54%):125 (46%).
Hospital services requesting dermatologic consultations. The various hospital medical services (syn. “ward team”) that requested dermatologic consultations are listed in Table 3. A similar overall source pattern for in-patient dermatologic consultations at the University of Miami was observed by Falanga et al. . Two-hundred six (76%) of the dermatologic consultations included in the analysis at UIHC were requested by three departments - Internal Medicine, Surgery, and Psychiatry. Compared to the University of Miami, there were relatively more consultation requests from the Surgery and Psychiatry Services at UIHC and relatively fewer requests from the Pediatrics Service. Fewer pediatric dermatologic consultation requests at UIHC could have resulted from the fact that there was no board-certified pediatric dermatologist on the dermatology faculty at UIHC.
Annual hospital admission rates of the consulting departments tended to correspond to the rates of dermatologic consultations they requested. Internal medicine and surgery constituted the largest percentage of annual hospital admissions and dermatologic consultations. However, if the percentage of annual hospital admissions per department referred for dermatologic consultation are compared, psychiatry referred the largest percentage of its patients (1.8%) followed by neurology (1.2%), medicine (1%), and ICU (1%). Pediatrics referred only 0.5% of their annual admissions for dermatologic consultation.
Reasons for hospital admissions. Table 4 presents a comparison of the reasons for admission of the patients in the present study to that of Falanga et al. . Except for a considerably higher percentage of AIDS at the University of Miami, the patterns are similar.
Clinical information listed on the consultation request form. Our analysis revealed that 52 percent of the time the primary ward team demonstrated the ability to generate dermatologic differential diagnoses for the skin findings in question. The ward team was given credit for this ability if they proposed two or more possible diagnoses on the consultation sheet accompanying the patient. Forty-eight percent of the time the consultation had only vague descriptions of the skin lesions in question with no differential diagnosis.
Skin disorders responsible for the consultations. The final diagnosis determined by the consulting dermatology staff served as the gold standard in this study. The following common dermatoses, as listed in Table 5 accounted for a large majority of dermatologic consultations: dermatitis (21.0%), drug eruption (10.0%), superficial dermatophyte infections (5.0%), viral infections (4.6%), folliculitis (3.5%), acne (or acniform eruptions) (3.5%), bacterial infections (3.1%), candidiasis (2.9%), cellulites (2.6%), and seborrheic keratosis (2.3%). Other diagnosis, which were seen in consultation with a frequency of 6 cases or less include: No lesions, trauma, vasculitis, nevi, intertrigo, xerosis, dystrophic nail, actinic keratosis, excoriations, abscess, post-inflammatory hyperpigmentation, hemangioma, icthyosis vugaris, hyperkeratosis, calciphylaxis, septic emboli, cholesterol emboli, adenocarcinoma, urticaria, erythema nodosum, transient neonatal pustular melanosis, ulceration, psoriasis, miliaria, pruritus, acrochordons, epidermoid cyst, angioimmunoblastic lymphadenopathy, alopecia, dermatomyositis, Wegner's granulomatosis, bullous pemphigoid, bowenoid papulosis, Bowen's disease, bacillary angiomatosis, bullae, chondrodermatitis nodularis helices, insect bites, pyogenic granuloma, keratosis lichenoides chronica, keratosis pilaris, lentigines, leukemia cutis, discoid lupus erythematosus, mongolian spot, pyoderma gangrenosum, pigmented purpura, pediculosis pubis, petechia, polymorphous light eruption, porphyria cutanea tarda, frostbite, dystrophic epidermolysis bullosa, lip erosion erythema ab igne, sarcoidosis, scurvy, hidradenitis suppuritiva, sarcoma, and extravasation injury.
Accuracy of submitted diagnoses on dermatologic consultation request form. Overall, the consulting ward teams provided an accurate preliminary diagnosis in 23.9 percent of the cases. This is exceedingly low in comparison to a mean 48 percent diagnostic accuracy of nondermatologists reported in the study by Falanga et al. It is still lower than the ~60 percent accuracy in the out-patient based studies on non-dermatologist diagnosis accuracy. The percent of correct diagnosis by consulting departments is listed in Table 6. Internal medicine, psychiatry and pediatrics had the highest percent correct, while neurology and obstetrics and gynecology had the lowest percent correct. One must take into account that different services obtain consults for various reasons and regarding cases of varying complexities, which may skew this data. The percentage of time that the consulting team was able to correctly diagnose the most common diagnosis seen in consultation is listed in Table 5.
Laboratory tests performed by the Dermatology Consultation Service. In many of the dermatologic consultation work-ups, laboratory tests were used to aid in diagnosis. The Dermatology Consultation Service performed 132 tests during the 271 consultations (Table 7). Fifty-four skin biopsies were performed. Forty-six (85.6%) of them resulted in a definitive diagnosis. Sixty-one percent of all other lab tests performed resulted in a positive result aiding in determining the diagnosis. Negative results were also beneficial in ruling out possible diagnoses in the differential, however it was not possible to quantify the role of negative tests in this study.
Impact of dermatologic consultation on patient management. Dermatology consultation resulted in a change in or addition to treatment in 77.7 percent of patients. A change is treatment was defined as the initiation or addition of a new topical or oral medication, the discontinuation of a previous medication and the addition of wound care instructions. The most common change in treatment was the addition of a topical corticosteroid or emollient. Table 8 lists the top ten most common changes in treatment recommend by the consulting dermatologic team.
Our study showed that the mean accuracy of the submitted diagnosis by the ward teams was 23.9 percent. A change in treatment based on dermatologic consultation occurred in 77.7 percent of the cases reviewed. The high rate of misdiagnosis or null diagnosis by ward teams could be expected to result in unacceptably high morbidity and mortality rates for the affected patients.
This study also raises the question of whether dermatologic consults may be suboptimally utilized resulting in an increased cost of health care to the patient secondary to prolonged hospital stay and ineffective treatment trials. This data supports the need for increased training of nondermatologists enabling them to better recognize and treat inflammatory and other common dermatoses.
Our findings were somewhat different from those of a previous study . This discrepancy may be accounted for by the fact that the study by Falanga et al. was prospective, allowing consult requesting physicians to make a more concentrated effort in providing a differential diagnosis, because they were in essence being evaluated.
Basic data presented in 1984  on inpatient dermatologic consultations raised a concern about a lack of diagnostic ability of “ward teams” with regard to common dermatoses. The need for an in-patient dermatology instructional program to provide ward teams with a better sense of when and how to utilize a dermatologic consultation was identified.
In the report “National Program for Dermatology - To Conquer Disability due to Skin Disease Through Patient Service, Education, and Research ,” very similar goals to those of Sheretz were proposed. Unfortunately, neither of these proposals or goals have been met as demonstrated in the present study and previous studies .
Ward teams in our hospital rarely attempted to perform simple diagnostic laboratory tests on their patients with skin problems. In part, this may be due to regulations limiting access to the materials needed to perform these tests on an inpatient unit or due to a lack of education on how and when to perform these tests properly. It is also clear from this data that use of common dermatologic laboratory tests such as potassium hydroxide preparations for dermatophyte infection and Tzanck smears for Herpes simplex/zoster infections should be included in the medical school dermatology curriculum. However, the majority of final dermatologic diagnoses were made by the Dermatology Consultation Service by visual examination alone. This emphasizes the importance of the ability to recognize common dermatoses by non-dermatologist physicians who care for patients in a hospital setting.
There are few other studies that evaluate the ability of non-dermatologists to diagnose and treat cutaneous disease, as well as the ability to use a dermatologic consultation appropriately in the inpatient setting. However, there have been numerous studies evaluating these issues in the outpatient setting. The outpatient based studies showed that primary care providers were able to identify common dermatoses approximately 40 - 60 percent of the time. This is very similar to that seen in the inpatient setting. Inadequacies in dermatologic knowledge most likely stem from the fact that medical students typically receive less than 20 hours of formal dermatologic training during four years of medical school. Despite this, physicians in pediatrics, family medicine, and internal medicine were responsible for seeing 42.4 percent of the 65.7 million skin disease visits in 1997 .
One survey  demonstrated the large percentage of dermatologic concerns seen and treated by internists despite their lack of training. Fifty percent of internists reported 10-30 percent of their patients presented with a cutaneous concern as their chief complaint. Of all the internists surveyed, 89 percent attempted treatment of the cutaneous concerns presented by this patient population and 85 percent of these internists had received < 1 month of dermatologic training in or after medical school. There was no correlation between the internist's ability in dermatology and the number of patients encountered for cutaneous concerns and treatment. There was a correlation between the amount of training the internist received and their ability in dermatology.
Often it is taken for granted that dermatologic concerns are incidental, and although perhaps uncomfortable and unattractive, of little consequence to the patient's general health. A South African study emphasized the considerable interface between cutaneous and systemic disease and the importance of the dermatologic diagnosis to the overall care of the patient . Five hundred inpatients on non-dermatologic primary teams demonstrated skin conditions related to the presenting illness in 50 percent of cases. These skin conditions contributed substantially to the diagnosis of the systemic illness in 36.6 percent of cases. Interestingly, although general dermatologic problems were seen in 20 - 25 percent of the patient population presenting to primary care or general practitioners in that region, only an equivalent of 0.6 percent of patients were referred to dermatologists in this study.
Our study confirmed that in the inpatient setting dermatologic consultation was frequently called upon for evaluation of drug eruptions. It is the experience of the authors that early signs and symptoms of delayed drug hypersensitivity syndromes (drug-induced delayed multiorgan hypersensitivity syndrome [DIDMOHS] ; drug rash with eosinophilia and systemic symptoms [DRESS] ) are often unnoticed by the primary team. This results in significant morbidity and mortality that could have been avoided if appropriate diagnosis and treatment had been provided in a timely manner. Other diagnoses such as purpura fulminans or acute graft-versus-host disease also have cutaneous signs associated with life threatening systemic conditions.
Evaluation of costs related to misdiagnosis was not a specific goal of this study. However, there is published data projecting cost-inefficiences resulting from delivery of dermatologic care by primary care providers. Many of the initial studies comparing ability of non-dermatologists to that of dermatologists were done out of “defense” at a time of concern for the future of dermatology with the advent of managed care. Such studies reviewed many ineffective care practices resulting from the inability of non-dermatologists to diagnose common dermatoses. The time efficiency of non-dermatologists vs dermatologists was evaluated  by examining the average duration of the physician-patient encounter. Dermatology visits were 12-20 percent shorter than primary physician visits for the same concern. Initial evaluation by a dermatologist was projected to possibly lead to decreased costs in a health care system or capitated environment, but not definitively shown.
An outpatient study by Clark and Rietschel evaluated calculations on the accuracy of diagnosis, treatment and referral patterns, procedures, office fees, number and costs of prescriptions and follow-up patterns and charges of family physicians vs dermatologist. The authors found that the cost of appropriate care rendered by the dermatologist was less than that by family physicians, although this did not reach statistical significance. Higher prescription costs and professional fees by the dermatologists were offset by increased laboratory testing, treatment of incorrect diagnosis and increased referrals by family physicians. Indirect costs were not included in this study but may have resulted in favoring the dermatologist more. Feldman et al.  showed that therapeutically incorrect treatment would have been administered by residents in his study 50 percent of the time and by attending physicians 43 percent of the time. Additionally, the internal medicine residents and faculty were more likely to request skin biopsies than dermatologists. The authors extrapolated from their data that 5 million inappropriate biopsies would be performed during the 30 million outpatient visits seen by nondermatologist at a cost of more than $765 million. No studies to date have looked at the cost-inefficiencies accrued secondary to the primary ward team's inability to care for common dermatoses in the inpatient setting. As stated, our study did not address cost-efficiency directly. Of note, in the year 2000, the average cost of stay for one day on a medical/surgical non-monitored floor at our university hospital was $1056. The average cost for an inpatient dermatology consultation at our institution was approximately $200. We would expect that these costs could be lessened with increased knowledge of the primary ward team with regard to the dermatoses contributing to the patients hospital stay.
Falanga et al. concluded in their study that although common dermatologic diseases were often not recognized or were misdiagnosed by nondermatologists, resulting in inappropriate treatment, to remedy this would require continued dermatologic training for the generalist. This was then stated to be an impractical solution. We disagree, for primary care physicians should be instructed on recognition and treatment of common dermatoses for their patient's health and welfare, as well as to lessen the economic burden of health care costs.
Since completion of our study in the year 2000, no similar reports based upon dermatology inpatient consultations in United States hospitals have appeared in the PubMed/Medline indexed literature. However, a similar study was reported from Ireland in 2008 by Ahmad & Ramsay . The results of that study were for the most part similar to ours considering differences between the U.S. and European health care systems. The five most common skin conditions identified in 703 Dermatology inpatient consultations were in descending order: skin infections, atopic eczema, psoriasis, drug rash, and urticaria. The higher prevalence of psoriasis and urticaria diagnosis seen in this study likely reflects the fact that European health care systems still support hospital admission primarily for skin disease problems. This is reflected by the fact that 53 percent of the 703 inpatient dermatology consultations examined by Ahmad & Ramsay were on patients who had been admitted because of acute skin failure (defined as “a state of total dysfunction of the skin resulting from different dermatological conditions”) . Institution of the Diagnosis-Related Groups policy by Medicare in the U.S. in the 1970s as a component of its prospective payments program functionally eliminated hospitalizations in the U.S. primarily for common skin disorders such as psoriasis and eczema. Similar to our study, Ahmad & Ramsay concluded stating “Much work remains to be done on dermatologic under-graduate and post graduate education for trainees as over one third of referrals had no diagnosis on the referral form.”
Another approach to addressing this challenge would be to support the development of a dermatology hospitalist career path . Hospital-based dermatologists would have the training and experience necessary to efficiently deal with inpatient skin disorders. The ideal pool of such subspecialists would be those who have completed a combined dermatology-internal medicine residency training program. However, as this combined training program is a relatively new one, there are currently few such dual trained individuals available in the United States.
It is unfortunate that U.S. medical schools provide so little training in a category of disease that is so common in every day medical practice. Based upon the rate of observation of mismanagement of skin disease by general physicians, the patient morbidity and economic burden of this educational failing is likely to be enormous. However, if educational policy within medical schools should change suddenly to correct this problem, academic departments of dermatology, traditionally the smallest departments in medical schools, would be overwhelmed with the increased teaching responsibility. Therefore, new educational mechanisms need to be envisioned to correct this problem. Taking advantage of the enhanced teaching productivity made possible by modern electronic information technology could be one such approach.
It is our feeling that more effective computer-based educational tools universally available to medical students, residents, fellows, and attending faculty on ward teams could target knowledge deficits concerning common dermatoses. The use of such educational tools in real time when questions about in-patient skin changes first arise could positively impact in-patient outcomes and conserve financial resources (e.g., length of hospital stay). An interactive, case-based, Internet-accessible teaching program based on well-illustrated case examples of common dermatoses could be one such tool. The “Skin Disease Finder” Internet open access tool made available by Logical Images as a component of its VisualDxHealth program is one example of recent progress that has been made in this area that might be adaptable to in-patient ward teams and hospitalists.
In agreement with a prior study by others, common skin conditions, especially inflammatory dermatoses, account for a large majority of dermatologic consultations in a University hospital setting. Difficulty in recognizing and treating such common dermatoses extends to all medical disciplines throughout the hospital. This knowledge deficit stems largely from limited formal education in dermatology during medical school. Dermatologic conditions encountered in the inpatient setting have been shown to contribute to the diagnosis and management of the patient's systemic illness requiring their hospital admission. Therefore, a lack of knowledge in this area may lead to unnecessary increase in morbidity and mortality as well as cost of patient care. It is our belief that primary ward teams should have better training and support in this area. Computer based educational tools concerning common dermatoses that are universally available in real time in an in-patient setting may achieve this goal in an effective manner
ACKNOWLEDGMENTS: At the time of that this study was performed, Dr. Sontheimer's contributions to this project was supported in part by the John S. Strauss Chair in Dermatology at the University of Iowa Hospitals and Clinics.
References1. Bickers DR, Lim HW, Margolis D, Weinstock MA, Goodman C, Faulkner E, Gould C, Gemmen E, Dall T. The burden of skin diseases: 2004 a joint project of the American Academy of Dermatology Association and the Society for Investigative Dermatology. J Am Acad Dermatol 2006; 55: 490-500. [PubMed]
2. Stern RS, Nelson C. The diminishing role of the dermatologist in the office-based care of cutaneous diseases. J Am Acad Dermatol 1993; 29: 773-777. [PubMed]
3. Falanga V, Schachner LA, Rae V, Ceballos PI, Gonzalez A, Liang G, Banks R. Dermatologic consultations in the hospital setting. Arch Dermatol 1994; 130: 1022-1025. [PubMed]
4. Sherertz EF. Inpatient dermatology consultations at a medical center. Arch Dermatol 1984; 120: 1137. [PubMed]
5. Feldman SR, Fleischer AB, Jr., Williford PM, White R, Byington R. Increasing utilization of dermatologists by managed care: an analysis of the National Ambulatory Medical Care Survey, 1990-1994. J Am Acad Dermatol 1997; 37: 784-788. [PubMed]
6. Thompson TT, Feldman SR, Fleischer AB, Jr. Only 33% of visits for skin disease in the US in 1995 were to dermatologists: is decreasing the number of dermatologists the appropriate response? Dermatol Online J 1998; 4: 3. [PubMed]
7. Kirsner RS, Federman DG. Managed care: the dermatologist as a primary care provider. J Am Acad Dermatol 1995; 33: 535-537. [PubMed]
8. Jessop S, McKenzie R, Milne J, Rapp S, Sobey G. Pattern of admissions to a tertiary dermatology unit in South Africa. Int J Dermatol 2002; 41: 568-570. [PubMed ]
9. Sontheimer RD, Houpt KR. DIDMOHS: A proposed consensus nomenclature for the drug-induced delayed multiorgan hypersensitivity syndrome. Arch Dermatol 1998; 134: 874-875. [PubMed]
10. Revuz J, Valeyrie-Allanore L. Drug Reactions. In: Bolognia JL, Jorrizo JJ, Rapini RP (eds.), Dermatology. London: Mosby; 2003.
11. Feldman SR, Fleischer AB, Jr., Young AC, Williford PM. Time-efficiency of nondermatologists compared with dermatologists in the care of skin disease. J Am Acad Dermatol 1999; 40: 194-199. [PubMed]
12. Ahmad K, Ramsay B. Analysis of inpatient dermatologic referrals: insight into the educational needs of trainee doctors. Ir J Med Sci 2008.
13. Fox LP. Inpatient dermatology. Semin Cutan Med Surg 2007; 26: 131-132. [PubMed]
© 2010 Dermatology Online Journal