Costs of Care for Hospitalization for Pemphigus in the United States

Importance  Pemphigus is an autoimmune blistering disorder associated with significant morbidity and mortality. However, little is known about the inpatient burden of pemphigus.

Objective  To determine the incidence of and risk factors for hospitalization with pemphigus and cost of care.

Design, Setting, and Participants  The 2002-2012 Nationwide Inpatient Sample provided by the Healthcare Cost and Utilization Project from the Agency for Healthcare Research and Quality was analyzed. A total of 87 039 711 children and adults (mean [SD] age, 57.7 [0.98] years for those with a primary diagnosis of pemphigus; 70.6 [0.32] years for those with a secondary diagnosis of pemphigus; and 47.9 [0.19] years for those without a diagnosis of pemphigus) were studied. Data analysis was performed from June 1 to August 30, 2015.

Main Outcomes and Measures  Hospitalization rates, length of stay, and cost of care.

Results  There were 1185 and 5221 patients admitted with a primary or secondary diagnosis of pemphigus, respectively; when factoring in weights that generalize the sample to the entire hospitalized US cohort, these admissions represented weighted frequencies of 5647 and 24 880, respectively. In multivariable logistic regression models with stepwise selection, increasing age (adjusted odds ratios [95% CIs]: 18-39 years: 5.53 [4.28-7.14], P < .001; 40-59 years: 10.98 [8.46-14.24], P < .001; 60-79 years: 7.54 [5.75-9.89], P < .001; ≥80 years: 7.57 [5.71-10.04], P < .001), female sex (1.10 [1.01-1.20], P = .047), nonwhite race/ethnicity (black: 1.94 [1.75-2.14], P < .001; Hispanic: 4.10 [3.74-4.48], P < .001; Asian: 3.16 [2.68-3.73], P < .001; Native American: 2.11 [1.45-3.08], P < .001), lower household income (quartile 2: 1.19 [1.07-1.32], P < .001), being insured with Medicare (1.56 [1.41-1.74], P < .001) or Medicaid (1.55 [1.39-1.73], P < .001), number of chronic conditions (2-5: 2.36 [2.10-2.65], P < .001; ≥6: 1.47 [1.29-1.69], P < .001), hospital location in a metropolitan area (not metropolitan or micropolitan: 0.60 [0.49-0.72], P < .001), and summer season (1.12 [1.02-1.23], P = .02) were all associated with hospitalization for pemphigus. The total inflation-adjusted cost of care for patients with a primary inpatient diagnosis of pemphigus was $74 466 305, with a mean (SD) annual cost of $14 520.93 ($913.22). The inflation-adjusted cost of care for patients with a primary diagnosis of pemphigus increased significantly from 2002 to 2012 (analysis of variance, P < .001). In particular, length of stay was higher in racial/ethnic minorities compared with whites (survey linear regression, log β [95% CI]: black: 0.076 [0.075-0.076]; Hispanic: 0.021 [0.021-0.022]; Asian: 0.037 [0.036-0.039]; Native American: 0.010 [0.0076-0.013]), lower quartile household income (quartile 1: 0.024 [0.023-0.024]; quartile 2: 0.0029 [0.0022-0.0035]), and those without private insurance (Medicare: 0.12 [0.12-0.12]; Medicaid: 0.082 [0.081-0.083]; no charge: 0.051 [0.047-0.055]).

Conclusions and Relevance  There is a significant inpatient burden for pemphigus in the United States. Moreover, there appear to be racial/ethnic and health care disparities with respect to pemphigus, such that poor, nonwhite, and/or uninsured or underinsured patients have higher odds of hospitalization.

Pemphigus is a debilitating autoimmune disorder characterized by acantholytic blisters of the skin and mucous membranes. The incidence and prevalence of the variants of pemphigus are not completely known.¹ Hospitalization of patients with pemphigus may occur from the intense pain, esophageal involvement leading to difficulty eating,² increased opportunistic infections,³ medication-related complications, and/or comorbid autoimmune disorders associated with pemphigus.⁴ However, to our knowledge, few studies have evaluated the inpatient burden of pemphigus. Previous studies have found racial/ethnic and socioeconomic disparities in hospitalization rates and outcomes for stroke,⁵ cardiovascular disease,⁶ asthma,⁷ and acute respiratory illness.⁸ We hypothesized that pemphigus is also associated with similar racial/ethnic and socioeconomic disparities, possibly related to being underinsured and having less access to dermatologic care. In the present study, we analyzed the incidence of hospitalization, cost of care, and length of stay in US patients with pemphigus.

The 2002-2012 Nationwide Inpatient Sample (NIS) provided by the Healthcare Cost and Utilization Project (HCUP) from the Agency for Healthcare Research and Quality was analyzed. Each year of the NIS contains an approximately 20% stratified representative sample of all inpatient hospitalizations in the United States. Sample weights were created by the NIS that factored the sampling design of hospitals in the United States. These sample weights are needed to provide representative estimates of hospital discharges across the whole country. Data analysis was performed from June 1 to August 30, 2015. All data were deidentified and no attempts were made to identify any of the individuals in the database. Patient consent was not obtained because all patient records were received deidentified. All parties with access to the HCUP were adherent to the HCUP’s formal data use agreement. The study was approved by the institutional review board at Northwestern University.

The databases were searched for a primary and/or secondary diagnosis of pemphigus using International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) codes. The primary diagnosis was defined in the NIS as the condition chiefly responsible for admission to the hospital for care. An ICD-9-CM code of 694.4 corresponds to pemphigus vulgaris, pemphigus vegetans, pemphigus foliaceous, pemphigus erythematosus, paraneoplastic pemphigus, and drug-induced pemphigus. A previous study⁹ validated the use of a single ICD-9-CM code in the inpatient setting for the study of pemphigus. Patients with ICD-9-CM diagnostic codes of both 694.4 and 694.5 (bullous pemphigoid) were excluded.

All data analyses and statistical processes were performed using SAS statistical software, version 9.4 (SAS Institute Inc). Analyses of survey responses were performed using SURVEY procedures. Weighted prevalences of hospitalization with a primary or secondary ICD-9-CM code of pemphigus were determined. The hospital cost for inpatient care was calculated based on the total charge of the hospitalization and the cost to charge ratio estimated by the HCUP. All costs were adjusted for inflation to the year 2014 according to the Consumer Price Index from the US Bureau of Labor Statistics.¹⁰ Summary statistics were generated for length of stay, total charge, and estimated inflation-adjusted cost of care, including sum, mean, SD, minimum, maximum, median, interquartile range, and range.

The control group included all patients without any diagnosis of pemphigus, yielding a representative cohort of hospitalized patients in the United States. Analysis of variance was used to compare hospitalization rates for pemphigus over time. Post hoc comparisons were performed in 2004-2005, 2006-2007, 2008-2009, and 2010-2012 vs 2002-2003 using the Tukey method. To determine the predictors of hospitalization, we constructed survey-weighted binary logistic regression models with hospitalization for pemphigus as the dependent variable. To determine the predictors of cost of care and length of stay, we constructed linear regression models with log-transformed cost of care and length of stay as the dependent variables. Cost of care and length of stay were log transformed because they were not normally distributed. Other independent variables included age (0-17, 18-39, 40-59, 60-79, ≥80 years), sex (male, female), race/ethnicity (white, black, Hispanic, Asian, Native American, multiracial, or other), median annual income of the hospital zip code (quartiles), health insurance coverage (yes vs no; Medicare, Medicaid, private, self-pay, no charge, other), number of chronic conditions (0-1, 2-5, ≥6), season of admission (winter, spring, summer, autumn), hospital location (metropolitan areas [≥1 urban clusters of population ≥50 000] with ≥1 million people, fringe area or metropolitan area with <1 million people, micropolitan [≥1 urban cluster of population of 10 000-49 999], not metropolitan or micropolitan; Northeast, Midwest, South, West), teaching status (yes vs no), and bed size (small, medium, large). Multivariable logistic regression models used stepwise selection from the above-mentioned covariates (P < .10). Multivariable linear regression models were constructed that included all these covariates. Complete case analysis was performed. A 2-sided P < .05 was considered statistically significant.

Overall, there were 87 039 711 discharges captured in the NIS for 2002-2012. There were 1185 and 5221 patients admitted with a primary or secondary diagnosis of pemphigus, respectively; when factoring in weights that generalize the sample to the entire hospitalized US cohort, these admissions represented weighted frequencies of 5647 and 24 880, respectively. The weighted prevalences of primary and secondary hospitalization for pemphigus ranged from 11.7 to 17.1 (mean 13.7) and 50.5 to 72.7 (60.4), respectively, per million patients per year (Figure 1). Hospitalization rates for patients with a primary or secondary diagnosis of pemphigus significantly increased after 2006 compared with 2002-2003 (P < .001; Figure 1).

Patients with a primary (mean [SD] age, 57.7 [0.98] years) and secondary (mean [SD] age, 70.6 [0.32] years) diagnosis of pemphigus were significantly older than those without a diagnosis of pemphigus (mean [SD] age, 47.9 [0.19] years) (Table 1). There were no significant sex differences between patients with and without pemphigus according to survey weighted logistic regression (odds ratios [95% CIs]) (1.03 [0.92-1.16]). Patients who were admitted for pemphigus were significantly more likely to be Hispanic (3.04 [2.61-3.54]), black (1.97 [1.66-2.34]), Asian (2.50 [1.83-3.42]), Native American (2.08 [1.07-4.03]), or multiracial or other (1.82 [1.32-2.50]). Patients with pemphigus had significantly higher odds of being admitted to a hospital in an area with a household income in the lowest quartile (1.42 [1.19-1.68]). Patients who were admitted for pemphigus were less likely to be insured overall (0.50 [0.42-0.62]). Among those with insurance, patients with pemphigus had a higher odds of having Medicare (1.78 [1.53-2.07]) and Medicaid (1.59 [1.33-1.89]) compared with private insurance. Finally, patients who were admitted for pemphigus had a higher odds of multiple chronic conditions (2-5: 2.92 [2.26-3.76]; ≥6: 2.13 [1.62-2.80]).

Admissions for pemphigus were less likely to occur in hospitals in nonmetropolitan areas (fringe area or metropolitan area with <1 million people: 0.66 [0.56-0.78]; micropolitan: 0.63 [0.48-0.84]; not metropolitan or micropolitan: 0.41 [0.27-0.62]) and in midwestern (0.55 [0.46-0.66]) and southern (0.66 [0.57-0.77]) states. However, admissions were more likely to occur in larger (1.31 [1.08-1.59]) and teaching (1.99 [1.76-2.25]) hospitals (Table 1). Bed size categories are defined in the eTable in the Supplement.

In multivariable logistic regression models with stepwise selection, age (18-39 years: 5.53 [4.28-7.14], P < .001; 40-59 years: 10.98 [8.46-14.24], P < .001; 60-79 years: 7.54 [5.75-9.89], P < .001; ≥80 years: 7.57 [5.71-10.04], P < .001), female sex (1.10 [1.01-1.20], P = .047), race/ethnicity (Hispanic: 4.10 [3.74-4.48], P < .001; black: 1.94 [1.75-2.14], P < .001; Asian: 3.16 [2.68-3.73], P < .001; Native American: 2.11 [1.45-3.08], P < .001; multiracial or other (2.21 [1.85-2.65], P < .001), household income (quartile 2: 1.19 [1.07-1.32], P < .001), insurance status (Medicare: 1.56 [1.41-1.74], P < .001; Medicaid: 1.55 [1.39-1.73], P < .001), number of chronic conditions (2-5: 2.36 [2.10-2.65], P < .001; ≥6: 1.48 [1.29-1.69], P = .01), hospital location (fringe area or metropolitan area with <1 million people: 0.90 [0.83-0.96], P = .03; micropolitan: 0.60 [0.51-0.70], P < .001; not metropolitan or micropolitan: 0.60 [0.49-0.72], P = .001), and summer season (1.12 [1.02-1.23], P = .02) were all associated with admission rates for pemphigus (Table 1).

The top 3 primary admission diagnoses for patients with a secondary diagnosis of pemphigus were septicemia, pneumonia, and cellulitis of the leg (Table 2). The primary admission diagnoses that were in the top 20 reasons for admission in patients with pemphigus but not in those without pemphigus were food vomit pneumonitis (reason 8), pulmonary embolism (reason 12), dehydration (reason 15), gastrointestinal hemorrhage (reason 19), and methicillin-sensitive Staphylococcus aureus septicemia (reason 20).

Patients with pemphigus spent a total of 40 757 days and 180 725 days in the hospital for their pemphigus or other reasons, respectively. Mean (SD) length of stay in the hospital was longer for patients with a primary (7.2 [0.2] days) or secondary (7.3 [0.2] days) diagnosis of pemphigus compared with patients who did not have pemphigus (4.6 [0.02]). Overall length of stay decreased by 2.4% from 2002 to 2012, and the mean percentage per year decreased by 0.07% for patients with no diagnosis of pemphigus (Figure 2). In contrast, there was a 13.3% decrease in length of stay in patients with a primary diagnosis of pemphigus from 2002 to 2012 and a mean decrease per year of 2.1%. In patients with a secondary diagnosis of pemphigus, there was a 6.3% decrease in length of stay from 2002 to 2012 and a mean decrease per year of 0.3%.

In weighted multivariable linear regression models, increased length of stay in patients with a primary diagnosis of pemphigus was associated with race/ethnicity (β coefficient [95% CI]: Hispanic: 0.021 [0.021-0.022]; Asian: 0.037 [0.036-0.039]; black: 0.076 [0.075-0.076]; Native American: 0.010 [0.0076-0.013]; multiracial or other: 0.042 [0.041-0.043]), income (quartile 1: 0.024 [0.023-0.024]; quartile 2: 0.0029 [0.0022-0.0035]), insurance status (Medicaid: 0.082 [0.081-0.083]; Medicare: 0.11 [0.12-0.12]), increasing age (60-79 years: 0.011 [0.0098-0.12]; ≥80 years: 0.088 [0.087-0.089]), season (winter: 0.021 [0.020-0.021]; fall: 0.0038 [0.0032-0.0044]), and increasing number of chronic conditions (2-5: 0.31 [0.31-0.31]; ≥6: 0.54 [0.54-0.54]).

The total inflation-adjusted cost of care for patients with a primary inpatient diagnosis of pemphigus was $74 466 305, with a mean (SD) annual cost of $14 520.93 ($913.22). The mean inflation-adjusted cost of care for patients with a primary diagnosis of pemphigus increased significantly from 2002 to 2012 (P < .001) (Figure 2). The actual total cost is likely higher because 486 patients had a missing value for charge and cost.

The mean (SD) cost of care for patients with a secondary diagnosis of pemphigus was $14 817.50 ($365.16). The total cost of care for patients who had a secondary diagnosis of pemphigus from the years 2002 to 2012 was $333 871 514. In contrast, the mean (SD) cost of care for patients without any diagnosis of pemphigus was $9948.74 ($75.75). For patients without a diagnosis of pemphigus, the mean increase per year was 1.3% with an overall increase of 17.2%. In contrast, the mean decrease of cost per year was 0.6% with an overall decrease of 21.8%. For patients with a secondary diagnosis of pemphigus, the mean increase per year was 2.4% with an overall increase of 23.6%.

In weighted multivariable linear regression models, increased cost of care in patients with a primary diagnosis of pemphigus was associated with older age (β coefficient [95% CI]: 18-39 years: 0.82 [0.82-0.82]; 40-59 years: 1.023 [1.023-1.024]; 60-79 years: 1.11 [1.11-1.11]; ≥80 years: 0.98 [0.98-0.98]), season (fall: 0.034 [0.033-0.035]; summer: 0.0097 [0.0090-0.010]; winter: 0.0095 [0.0088-0.010]), race/ethnicity (Asian: 0.067 [0.066-0.068]; black: 0.025 [0.024-0.026]; Native American: 0.040 [0.039-0.041]; Hispanic: 0.040 [0.039-0.040]; multiracial or other: 0.11 [0.11-0.11]), and increasing number of chronic conditions (2-5: 0.49 [0.49-0.49]; ≥6: 0.71 [0.71-0.71]) (Table 3).

In the present study, there were significant racial/ethnic and socioeconomic disparities in hospitalization for pemphigus. In particular, higher rates of hospitalization were found in patients with pemphigus who were nonwhite, had a lower income, or were underinsured or uninsured. All these associations remained significant in multivariable regression models. Moreover, there was wide variation in length of stay and cost of care based on sociodemographic characteristics. Finally, the cost of hospitalization for patients with a primary diagnosis of pemphigus increased significantly from 2002 to 2012.

Rates of hospitalization for pemphigus are likely owing to a combination of higher prevalence of pemphigus in certain racial/ethnic groups and disparities in hospitalization. Pemphigus has been reported in the literature to most commonly affect people of Ashkenazi Jewish or Brazilian descent.¹¹ However, higher rates of hospitalization were found in blacks, Hispanics, Asians, and Native Americans. It may be that the prevalence of pemphigus is higher in these minority groups than previously recognized. Moreover, patients with decreased access to dermatologic care and/or prescription medication coverage may have exacerbation of pemphigus and secondary infections. Previous studies^12,13 in other fields have found the patient characteristics of nonwhite race and uninsured status to be associated with higher rates of potentially preventable amputations and poorer outcomes in terms of morbidity and mortality. Future studies are needed to determine whether decreased access to outpatient specialty care results in increased hospitalization rates for patients with pemphigus.

Patients with pemphigus had a variety of comorbidities that resulted in hospitalization. The 3 most common admitting diagnoses for patients with pemphigus as a secondary diagnosis are septicemia, pneumonia, and lower leg cellulitis. These diagnoses suggest that patients with pemphigus are more prone to infection. Indeed, a prospective study³ found that patients with pemphigus are at higher risk for opportunistic infections, especially after the first year of diagnosis and in older diabetic patients. The increased risk of infection is likely multifactorial, including long-term use of immunosuppressants and corticosteroids as well as disease severity. A study¹⁴ of 155 patients with pemphigus vulgaris found that more severe pemphigus as judged by a severity index of pemphigus (mild, moderate, or severe) is associated with higher rates of infections, especially pulmonary, bacterial skin, and urinary tract infections. Pemphigus has also been associated with higher rates of cardiovascular disease risk and malignant tumor development.^15-17 Thus, aside from dermatologic care, patients with pemphigus require attention to prevent infections and treat their other comorbidities.

The mean cost of care and length of hospitalization were approximately 50% and 60% higher in patients with pemphigus than those without pemphigus, respectively. This was true even for patients with a secondary diagnosis of pemphigus, suggesting that patients with pemphigus tend to be sicker overall and have more comorbidities. Nonwhite race, lower income areas, and Medicaid or Medicare insurance status were all associated with increased length of stay and cost of care. This increase in cost of care may not be unique to pemphigus, however, because a systematic review¹⁸ that analyzed psoriasis costs observed incremental annual costs per patient ranging from $900 to $2000.

Strengths of this study include an analysis of a nationally representative sample of all-payer data for a period of 11 years with more than 85 million records. We previously validated the use of ICD-9-CM codes for identifying pemphigus in the inpatient setting.⁹ Weaknesses of this study include the inability to distinguish among different subsets of pemphigus, such as vulgaris and foliaceous. Moreover, we were unable to determine how many of the hospitalizations were owing to readmissions or transfers among hospitals. However, there were only 25 admissions (2.1%) in the same month for a primary diagnosis of pemphigus in patients who had identical demographic characteristics. Some of these encounters likely represented unique patients who coincidentally had similar demographic characteristics. Thus, hospitalization rates in this study likely represent a cohort of unique patients. Our cost analyses could not examine rituximab or intravenous immunoglobulin infusion as possibly accounting for the increased costs of pemphigus hospitalization because Current Procedural Terminology codes are not captured in the NIS. Lastly, the cost analysis did not include costs of physician services, out-of-pocket expenses, or outpatient costs. Consequently, the total economic burden of pemphigus is likely much higher.

In conclusion, the findings of this study indicate that the inpatient burden of pemphigus is extensive. The cost of care and length of hospitalization were consistently higher for patients with pemphigus than those without. Nonwhite race, lower income, and underinsured or uninsured status were associated with higher rates of hospitalization and increased length of stay. Future research is needed to better understand the reasons for such disparities and identify optimal strategies to reduce these disparities.

Accepted for Publication: December 22, 2015.

Corresponding Author: Jonathan I Silverberg, MD, PhD, MPH, Department of Dermatology, Northwestern University Feinberg School of Medicine, 676 N St Clair St, Ste 1600, Chicago, IL 60611 (jonathanisilverberg@gmail.com).

Published Online: February 10, 2016. doi:10.1001/jamadermatol.2015.5240.

Author Contributions: Dr Silverberg had full access to all the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis.

Study concept and design: Hsu, Silverberg.

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

Drafting of the manuscript: Hsu, Silverberg.

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

Statistical analysis: Hsu, Silverberg.

Obtained funding: Silverberg.

Administrative technical or material support: Silverberg.

Study supervision: Brieva.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by grant K12HS023011 from the Agency for Healthcare Research and Quality and the Dermatology Foundation.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.