Association of Dermatology Consultations With Patient Care Outcomes in Hospitalized Patients With Inflammatory Skin Diseases

Key Points

Question  What is the association of dermatology consultations with outcomes of hospitalized patients with skin diseases?

Findings  In this study of a single medical center, dermatology consultations in patients with a diagnosis of inflammatory skin conditions were associated with decreased 1-year readmissions and adjusted hospital length of stay.

Meaning  The associated improvement of dermatology consultations with outcomes of hospitalized patients with skin diseases provides a rationale for expanding the role of dermatology consultative services in the hospital setting.

Importance  The value of inpatient dermatology consultations has traditionally been demonstrated with frequency in changes of diagnosis and management; however, the impact of dermatology consultations on metrics such as hospital length of stay and readmission rates remains unknown.

Objective  To determine the association of dermatology consultations with patient care in hospitalized patients using objective values.

Design, Setting, and Participants  We retrospectively queried the deidentified database of patients hospitalized between January 1, 2012, and December 31, 2014, at a single university medical center. A total of 413 patients with a primary inflammatory skin condition discharge diagnosis and 647 patients with primary inflammatory skin condition admission diagnosis were selected.

Main Outcomes and Measures  Hospital length of stay and 1-year readmission with inflammatory skin conditions.

Results  The 413 patients with a primary inflammatory skin condition discharge diagnosis were 61.0% female and had a mean (SD) age of 55.1 (16.4) years. The 647 patients with primary inflammatory skin condition admission diagnosis were 50.8% female and had a mean (SD) age of 57.8 (15.9) years. Multivariable modeling showed that dermatology consultations were associated with a reduction of 1-year inflammatory skin condition readmissions among patients who were discharged primarily with an inflammatory skin condition (readmission probability, 0.0025; 95% CI, 0.00020-0.030 with dermatology consult vs 0.026; 95% CI, 0.0065-0.10 without; odds ratio, 0.093; 95% CI, 0.010-0.840; P = .03). No other confounding variable was associated with reduction in readmissions. Multivariable modeling also showed that dermatology consultations were associated with a reduction in the adjusted hospital length of stay by 2.64 days (95% CI, 1.75-3.53 days; P < .001).

Conclusions and Relevance  Dermatology consultations were associated with improvements of outcomes among hospitalized patients. The expansion of the role of dermatology consultation services may improve patient care in a cost-effective manner.

Cutaneous disorders remain an important reason for hospitalization in the United States,^1,2 with an estimated $1.8 billion in Medicare spending in 2009.¹ Over the past couple of decades, more medical centers have relied on nondermatologists to care for patients with cutaneous disorders.³ Many barriers prevent dermatologists from seeing these hospitalized patients, including institutional bureaucracy, time constraints, and poor reimbursement.^4,5

Studies have shown that dermatological conditions are commonly misdiagnosed by nondermatologists, and dermatology consultations often result in changes in both diagnosis and treatment of hospitalized patients.^1,6-9 However, the current literature regarding the role of dermatology consultation services in improving outcomes of hospitalized patients using interchangeable metrics such as hospital length of stay or readmission rate is limited. The goal of this study was to evaluate the association of dermatology consultations with patient care outcomes among hospitalized patients.

The study used deidentified data and was exempt by the Institutional Review Board of the Ohio State University Wexner Medical Center. Deidentified information of patients and all variables of interest were queried from the Ohio State University Medical Center Information Warehouse. This was conducted within a large adult medical center inclusive of 5 specialty hospitals and an ancillary hospital. Primary outcomes included 1-year readmissions and hospital length of stay. Variables that were evaluated include age, sex, race, body mass index, Charlson Comorbidity Index, discharge disposition, hospital (main vs ancillary), history of previous hospitalization, intensive care unit admissions, biopsy procedure, nursing skin care and physician wound care consultations, admission diagnosis, discharge diagnosis, 1-year readmission diagnosis, and dermatology consultations. Inflammatory skin conditions described throughout the article were defined as International Classification of Diseases, Ninth Revision (ICD-9) codes 680 to 710 and 782.

A total of 8246 patients were retrieved from the deidentified database with either an admission or discharge diagnosis from a dermatology-associated ICD-9 code. To minimize diagnosis-related covariates, we focused on patients with primary admission/discharge diagnoses of skin disorders regardless of whether they received a dermatology consult. We identified 413 patients who had a primary discharge diagnosis of an inflammatory skin condition. This population was used for assessing the association of dermatology consultation with 1-year readmission rates. This hospital readmission parameter was defined as any readmission with an inflammatory skin condition. We also identified 647 patients who were admitted primarily for an inflammatory skin condition. These patients were used for evaluating the association of dermatology consultations with the hospital length of stay. The selections process and the top 3 dermatological conditions for the patient populations evaluated in the study are outlined in the Figure.

Statistical analyses were performed in MATLAB (version R2015b, MathWorks). The primary outcomes modeled were adjusted hospital length of stay and 1-year readmission rate. Independent variables were presence of a dermatology consultation and covariates (age, previous hospitalization, body mass index, discharge disposition, biopsy, hospital, nursing skin care consult, physician wound care consult, race, Charlson comorbidity index, and intensive care unit admission). These covariates were chosen for their plausible influence on the primary outcomes. Multivariable linear regression was used to model the effect of dermatology consultation and covariates on adjusted hospital length of stay. Multivariable logistic regression was used to model the association of dermatology consultation and the aforementioned covariates with 1-year readmission rates. Patients who had missing information about the mentioned variables were excluded from the models.

We evaluated factors associated with inflammatory skin condition readmission within a 1-year period among patients who were discharged with inflammatory skin conditions. The demographic characteristics of this patient population are described in Table 1. The multivariable logistic model indicated that dermatology consultation was associated with reduced probability of 1-year readmission (odds ratio, 0.093; 95% CI, 0.010-0.840; P = .03) among this patient population (Table 1). Several other independent variables including age, sex, race, body mass index, Charlson Comorbidity Index, previous hospitalization, hospital site, and physician wound care consultations had no statistically significant association with the readmission rates. Discharge disposition of home health care service was also associated with increased readmissions.

The readmission ICD-9 codes were exactly the same as the initial admission and discharge encounters in approximately 14 of 23 (61%) and 10 of 23 (43%) cases, respectively. The ICD-9 codes of the remainder would generally be included within the same differential diagnosis, for example, lower extremity cellulitis vs lower extremity edema.

There was often a lag between hospitalization and dermatology consultation, even when patients were admitted with a primary diagnosis of inflammatory skin disease (median [SD], 2 [2.08] days). Therefore, we accounted for this lag period by using consult-to-discharge time as the adjusted hospital length of stay. To avoid the influence of hospitalizations lengthened by workup of nondermatologic problems, only patients with a primary admission diagnosis of inflammatory skin disease were included in the evaluation (eFigure in the Supplement). The demographic characteristics of this patient population are outlined in Table 2.

Multivariable modeling showed that dermatology consultations without adjustment for delay in consultation were associated with a mean reduction of 0.41 days (95% CI, −1.36 to 0.53 days) in hospital length of stay; however, it was not statistically significant (P = .39). After adjustment for the lag time, dermatology consultations were associated with a significant mean reduction of 2.64 days (95% CI, 1.75-3.53 days; P < .001) in the adjusted hospital length of stay (Table 2). Body mass index was associated with a nominal but significant decrease in the adjusted hospital length of stay. Performance of biopsy procedure, nursing skin care and physician wound care consults, intensive care unit stay, and discharge dispositions of hospice or death, skilled nursing facility, and others were associated with increase in the adjusted hospital length of stay (Table 2). Other variables studied including age, sex, race, Charlson Comorbidity Index, hospital site, and previous hospitalization had no statistically significant association with the adjusted hospital length of stay.

This study shows that dermatology consultations for skin conditions are associated with (1) a decreased 1-year readmission rate for inflammatory skin conditions and (2) shorter hospital length of stay for patients hospitalized for inflammatory skin conditions. To our knowledge, this is the first study that has shown the positive association of dermatology consultations with these outcomes in hospitalized patients.

Dermatology consultations were associated with a 10-fold reduction in the odds of readmission with an inflammatory skin condition. Importantly, no other variable studied was associated with a reduction in the readmissions, signifying the pivotal role of dermatology consultations in curbing readmissions. One possible explanation for this lower readmission rate is the role of dermatology consults in establishing outpatient follow-up with a dermatologist after discharge. The dermatologist can follow up on disease progress and treatment of these patients and provide early interventions that may reduce the need for rehospitalization. In our study, patients evaluated by inpatient dermatologists had higher rates of clinic follow-up within 1 year after discharge (data not shown). Evaluating this intervention as a possible mechanism for reduced readmissions is beyond the scope of this study’s primary objectives, but it warrants evaluation in future studies. There is also the potential for improved diagnostic accuracy. It is well described in the literature that dermatologists provide more accurate diagnosis and necessary diagnostic tests of cutaneous disorders, including common disorders such as cellulitis, than their nondermatologist colleagues.^6-8,10 Furthermore, involvement of dermatologists in care of patients with dermatological conditions results in change of treatment plan in 50% to 60% of cases.^1,6 Therefore, another possible explanation is that dermatologist consultation results in a more accurate diagnosis and appropriate treatment during the initial hospitalization, which reduces rehospitalization rates. Although education of hospitalists on skin conditions is critical, dermatology-led sessions specific for hospitalists failed to improve the diagnosis concordance rate between dermatologists and nondermatologists after 5 years of implementation of a lecture series.¹¹

The association between dermatology consultations and reduction of 1-year readmission rates has important implications on reducing health care costs and improving quality of health care. The typical readmission cost, depending on patient age and type of insurance, was between $4200 and $14 100 in 2013.¹² This is particularly important because the number of hospitalizations, hospitalization days, and cost for dermatological disorders among Medicare recipients has been gradually increasing.¹ Reducing readmissions among this patient population is 1 mechanism whereby this increasing trend can be curbed. Furthermore, lower readmissions have been shown to be an indicator of the quality of care among hospitalized patients.¹³

The present study also shows that dermatology consultations were associated with a profound reduction in the adjusted hospital length of stay among patients who were admitted primarily for an inflammatory skin condition. We specifically focused on patients who were admitted primarily for an inflammatory skin condition to remove any confounding effects of nondermatology workup or dermatological problems that arose during hospitalization. It was necessary to adjust for the lag period between admission and dermatology consultation because this lag period could artificially increase length of stay prior to involvement of the dermatologist and any change in diagnosis and intervention. Although we hypothesize that dermatologists are consulted for more complex or widespread skin diseases requiring more diagnostic tests and comprehensive interventions that collectively result in longer length of stay, the severity of skin disease is not simply captured in any ICD-9 code, laboratory biomarker, or even by validated mortality indexes such as the Charlson Comorbidity Index. The lag time between the consultation being placed and the attending physician attesting to the consult was not measured; however, the hospital-wide policy requires this period to be less than 24 hours including on weekends. The association between dermatology consultations and shorter adjusted hospital length of stay among patients admitted with inflammatory dermatological conditions suggests that health care teams should be encouraged to consult dermatology services for this patient population at the time of hospitalization. This approach will allow for improved quality of care and reduced cost of care.

Comprehensive wound care management has a positive effect on patient care and improving outcomes.¹⁴ However, in patients with inflammatory skin disease, our data suggest that there is a unique benefit for hospitals to invest in dermatology consultative services. There is a need for dermatologists who attend to and treat hospitalized patients with cutaneous disorders. However, even in the consultative role, the number of such dermatologists is currently limited. A recent survey found that academic dermatologists spend a mean of 2.1 hours per week on hospital consults and nonacademic dermatologists spend even less time.¹⁵ Inconvenience of traveling to the hospital, poor reimbursements, and hospital credentialing requirements are all commonly cited barriers.⁴ Our study indicates that it may be feasible for hospital systems to invest in removing these barriers to improve outcomes by reducing readmissions, hospitalization length, and economic burden of dermatological disorders.

Several other variables were associated with increases in readmission and/or hospital length of stay. Functional impairment, a predictor of use of home health care services,¹⁶ is associated with increased readmission.¹⁷ This likely is a factor in the increased readmission rates of patients discharged with home health care, but it was not evaluated in the present study. The longer hospital length of stay among patients with intensive care unit stays can be attributed to presence of other comorbidities,¹⁸ complex clinical picture, and high incidence of cutaneous manifestation of systemic diseases and life-threatening conditions in this patient population.^18,19 Increased length of stay associated with patients discharged to postacute care settings is likely related to delays in placement in these facilities, and this setting has previously been reported to be associated with longer length of stay.²⁰ Skin biopsies have not previously been reported to increase length of stay, but they are likely associated with delay for diagnosis confirmation by pathologic interpretation of the biopsy and therapy initiation. Future studies are warranted to further evaluate the factors responsible for these observations.

The present study has multiple limitations. This study is a retrospective study of a single medical center and thus may not be generalizable to all medical centers. The adjustment for length of stay is also a limitation. Although it is our best attempt at adjusting for lag, this adjustment may overcompensate and artificially amplify the reduction in length of stay attributed to dermatology consultation. Whereas the odds ratio of readmission for dermatology consultation is statistically significant, the 95% CI is broad due to the small sample size of the present study, which needs to be taken into consideration when the results are interpreted.

In summary, we report that dermatology consultations were associated with reduced readmissions and hospital length of stay of hospitalized patients in a single medical center in the United States. These findings have important implications for providing rationale for expanding the support and role of hospital dermatologists.

Accepted for Publication: December 20, 2016.

Corresponding Author: Benjamin H. Kaffenberger, MD, Ohio State University Dermatology, 915 Olentangy River Rd, Ste 4000, Columbus, OH 43212 (benjamin.kaffenberger@osumc.edu).

Published Online: March 15, 2017. doi:10.1001/jamadermatol.2016.6130

Author Contributions: Drs Milani-Nejad and Kaffenberger had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Milani-Nejad, Kaffenberger.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Milani-Nejad.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Zhang, Kaffenberger.

Obtained funding: Kaffenberger.

Administrative, technical, or material support: Kaffenberger.

Supervision: Kaffenberger.

Conflict of Interest Disclosures: Dr Kaffenberger has served as a consultant to Xbiotech, Xoma, Celgene, and Castle Biosciences and has received research funds from Xbiotech, Celgene, Eli Lilly Co, and Biogen. No other disclosures are reported.