Association of society of dermatology hospitalist institutions with improved outcomes in Medicare beneficiaries hospitalized for skin disease

To the Editor: In the United States, cutaneous disorders are estimated to be diagnosed in more than 12% of hospitalized adults and account for more than $5 billion in health expenditures annually.1Arnold J.D. Yoon S. Kirkorian A.Y. The national burden of inpatient dermatology in adults.J Am Acad Dermatol. 2019; 80: 425-432https://doi.org/10.1016/j.jaad.2018.06.070Abstract Full Text Full Text PDF PubMed Scopus (24) Google Scholar However, patients with cutaneous disorders are predominantly admitted to inpatient services attended by non-dermatologists.2Madigan L.M. Fox L.P. Where are we now with inpatient consultative dermatology?: Assessing the value and evolution of this subspecialty over the past decade.J Am Acad Dermatol. 2019; 80: 1804-1808https://doi.org/10.1016/j.jaad.2019.01.031Abstract Full Text Full Text PDF PubMed Scopus (13) Google Scholar To this end, the dermatology hospitalist model has emerged over the past decade, and the Society for Dermatology Hospitalists (SDH) was established.2Madigan L.M. Fox L.P. Where are we now with inpatient consultative dermatology?: Assessing the value and evolution of this subspecialty over the past decade.J Am Acad Dermatol. 2019; 80: 1804-1808https://doi.org/10.1016/j.jaad.2019.01.031Abstract Full Text Full Text PDF PubMed Scopus (13) Google Scholar,3Rosenbach M. The logistics of an inpatient dermatology service.Semin Cutan Med Surg. 2017; 36: 3-8https://doi.org/10.12788/j.sder.2017.006Crossref PubMed Scopus (5) Google Scholar Single institution studies have suggested dermatology hospitalist services improve diagnostic accuracy and reduce readmissions.2Madigan L.M. Fox L.P. Where are we now with inpatient consultative dermatology?: Assessing the value and evolution of this subspecialty over the past decade.J Am Acad Dermatol. 2019; 80: 1804-1808https://doi.org/10.1016/j.jaad.2019.01.031Abstract Full Text Full Text PDF PubMed Scopus (13) Google Scholar However, the outcomes of patients treated at SDH hospitals have yet to be described on a national level. In this study, we examined the association of hospitalization at SDH institutions with risk-adjusted outcomes. We utilized a 100% sample of the 2016 to 2018 Center for Medicare and Medicaid Services (CMS) Inpatient Standard Analytic File and Master Beneficiary Summary File. To assign hospital-level characteristics, we used the 2018 American Hospital Association survey data. We included all consecutive dermatology-specific discharges at teaching hospitals from January 2016 through November 30, 2018, with outcomes followed through December 31, 2018. Dermatology-specific discharge were defined as an Medicare Severity Diagnosis Related Groups (MS-DRG) of 595, 596, 606, or 607 (Supplement S1, available via Mendeley at https://doi.org/10.17632/hf7b633xg3.1). The primary outcome was all-cause 30-day mortality, and secondary outcomes were 30-day readmissions, length-of-stay, and cost. The exposure of interest was discharge from a SDH hospital. We used mixed effect logistic regression modeling to adjust for patient-level and hospital-level characteristics. For sensitivity analysis, we compared 30-day mortality and readmissions for non-dermatologic Diagnosis Related Groups (DRGs) using hospital-level data from CMS (Supplement S3, available via Mendeley at https://doi.org/10.17632/hf7b633xg3.1). There were 30,900 encounters across 1912 US teaching hospitals. Table I shows encounter-level demographic and hospital characteristics. Table II shows unadjusted outcomes by DRG. SDH member hospitals experienced significantly lower risk-adjusted 30-day mortality (odds ratio = 0.76; 95% CI: 0.60, 0.97; P = .03) and 30-day readmissions (odds ratio = 0.88; 95% CI: 0.78, 1.00; P = .05) than non-member hospitals. There were no significant differences in length of stay (−0.13 days; 95% CI: −0.37, 0.12; P = .31); however, the average cost per hospitalization was $850 higher (95% CI: $228, $931; P < .001) at SDH hospitals. Sensitivity analyses (Supplement S3, available via Mendeley at https://doi.org/10.17632/hf7b633xg3.1) showed that mortality and readmission rates for non-dermatologic DRGs were not consistently lower at SDH-members.Table IEncounter-level demographic and hospital characteristics comparing society of dermatology hospitalist member hospitals to non-member hospitals (n = 30,900 inpatient encounters from Jan 2016 to Nov 2018 at US teaching hospitals)CharacteristicSDH member n = 1873(6.1%)Non-SDH-member n = 29,027(93.9%)P value∗Chi-squared P value for categorical variables, Wilcoxon rank sum for continuous.Age, median (IQR)69 (62, 78)71 (62, 80)<.0001Sex<.01 Male861 (46.0%)12,234 (42.2) Female1012 (54.0)16,793 (57.8)Race/ethnicity<.0001 White1441 (76.9)21,632 (74.5) Black249 (13.3)5099 (17.6) Asian52 (2.8)581 (2.0) Hispanic of any race56 (3.0)776 (2.7) Other/unknown75 (4.0)939 (3.3)DRG.03 595: Major skin disorders with MCC187 (10.0)2539 (8.8) 596: Major skin disorders without MCC428 (22.9)6931 (23.9) 606: Minor skin disorders with MCC333 (17.8)4618 (15.9) 607: Minor skin disorders without MCC925 (49.4)14,939 (51.5)Elixhauser indices Mortality, mean (SD)8.8 (10.8)7.4 (10.0)<.0001 Readmission, mean (SD)23.2 (16.9)21.6 (16.4)<.0001Total hospital beds<.0001 <500 (0.0)366 (1.3) 50-990 (0.0)703 (2.4) 100-1990 (0.0)3517 (12.1) 200-499230 (12.3)12,761 (44.0) 500-9991044 (55.7)8099 (27.9) 1000+599 (32.0)2745 (9.5) Missing0 (0.0)836 (2.9)Owner<.0001 For-profit0 (0.0)3216 (11.1) Government state353 (18.9)1225 (4.2) Government local75 (4.0)2388 (8.2) Non-profit1445 (77.2)21,346 (73.5) Missing0 (0.0)852 (2.9)Region<.001 Alaska/Hawaii0 (0.0)990 (3.4) Northeast679 (36.3)6762 (23.3) South266 (14.2)10,652 (36.7) West344 (18.4)3673 (12.7) Midwest584 (31.2)6950 (23.9)Dual enrollment.19 Yes589 (31.5)8710 (30.0) No1284 (68.5)20,317 (70.0)MCC, Major complication or comorbidity.∗ Chi-squared P value for categorical variables, Wilcoxon rank sum for continuous. Open table in a new tab Table IIUnadjusted outcomes comparing society of dermatology hospitalist member hospitals to non-member hospitals from Jan 1, 2016 through Nov 30, 2018 at US teaching hospitalsSDH memberNon-SDH-memberP value∗Chi-squared P value for mortality and readmission, Wilcoxon rank sum for length of stay and cost.n = 1873n = 29,02730-day mortality, % DRG 59533/187 (17.7)487/2539 (19.2).61 DRG 59614/428 (3.3)316/6931 (4.6).21 DRG 60623/333 (6.9)413/4618 (8.9).21 DRG 60718/925 (2.0)359/14939 (2.4).3830-day readmission†All-cause readmission, and excluding those who experienced 30-day mortality., % DRG 59546/154 (29.9)667/2052 (32.5).50 DRG 59670/414 (16.9)1448/6615 (21.9).02 DRG 60694/310 (30.3)1332/4205 (31.4).68 DRG 607184/907 (20.3)2921/14580 (20.0).85Length of stay [days, median (IQR)] DRG 5956 (3, 11)5 (3, 9).02 DRG 5963 (2, 5)3 (2, 6).12 DRG 6065 (2, 8)4 (3, 7).14 DRG 6073 (2, 4)3 (2, 4).95Cost per encounter‡Cost = total charges for the claim multiplied by hospital-specific cost-to-charge ratio., median (IQR) DRG 595$12,800 ($6900-$27,600)$8400 ($4900-$15,400)<.001 DRG 596$6200 ($3500-$10,700)$4900 ($3200-$7900)<.001 DRG 606$9800 ($5600-$17,900)$6600 ($4100-$12,100)<.001 DRG 607$5800 ($3400-$10,000)$4200 ($2700-$6700)<.001∗ Chi-squared P value for mortality and readmission, Wilcoxon rank sum for length of stay and cost.† All-cause readmission, and excluding those who experienced 30-day mortality.‡ Cost = total charges for the claim multiplied by hospital-specific cost-to-charge ratio. Open table in a new tab MCC, Major complication or comorbidity. In summary, we found that patients hospitalized at SDH institutions had 24% lower odds of risk-adjusted 30-day mortality and 12% lower odds of risk-adjusted 30-day readmissions. Inpatient dermatology consultations have been associated with increased diagnostic and therapeutic accuracy,2Madigan L.M. Fox L.P. Where are we now with inpatient consultative dermatology?: Assessing the value and evolution of this subspecialty over the past decade.J Am Acad Dermatol. 2019; 80: 1804-1808https://doi.org/10.1016/j.jaad.2019.01.031Abstract Full Text Full Text PDF PubMed Scopus (13) Google Scholar which could explain lower mortality and readmissions rates at SDH hospitals. Similarly, SDH hospitals may achieve improved outcomes through higher rates of outpatient dermatology follow-up.4Hu L. Haynes H. Ferrazza D. Kupper T. Qureshi A. Impact of specialist consultations on inpatient admissions for dermatology-specific and related DRGs.J Gen Intern Med. 2013; 28: 1477-1482https://doi.org/10.1007/s11606-013-2440-2Crossref PubMed Scopus (39) Google Scholar,5Milani-Nejad N. Zhang M. Kaffenberger B.H. Association of dermatology consultations with patient care outcomes in hospitalized patients with inflammatory skin diseases.JAMA Dermatol. 2017; 153: 523https://doi.org/10.1001/jamadermatol.2016.6130Crossref PubMed Scopus (46) Google Scholar However, costs per encounter were $850 higher at SDH hospitals, which could reflect more extensive diagnostic workups.5Milani-Nejad N. Zhang M. Kaffenberger B.H. Association of dermatology consultations with patient care outcomes in hospitalized patients with inflammatory skin diseases.JAMA Dermatol. 2017; 153: 523https://doi.org/10.1001/jamadermatol.2016.6130Crossref PubMed Scopus (46) Google Scholar Nonetheless, the higher costs at SDH hospitals should be weighed against significantly lower 30-day mortality and readmission rates. Limitations of this study include its use of retrospective Medicare claims data and varying levels of access to dermatology consultations at non-SDH hospitals. Taken together, dermatology hospitalist models—as implemented at SDH institutions—may generate added value (outcomes/cost) in the treatment of patients with skin disorders. Further study is needed to define the dermatology-specific care processes and mechanisms that contribute to improved outcomes. Drs Fox, Rosenbach, and Mangold are members of the Society of Dermatology Hospitalists. The Society of Dermatology Hospitalists was not involved in the design of the study, data collection, analysis, or drafting of the manuscript. Author Puri, Dr Pollock, Authors Yousif and Bhullar, and Drs Boudreaux and Pittelkow have no conflicts of interest to declare.