Analysis of Readmissions Following Hospitalization for Cellulitis in the United States

Key Points

Question

What is the national rate of readmission within 30 days of hospital discharge for cellulitis?

Findings

In this nationally weighted, retrospective cohort analysis of US hospital admissions, 9.8% of 447 080 cellulitis admissions were associated with nonelective readmission within 30 days. All-cause, nonelective readmissions after cellulitis discharge cost more than $500 million; readmissions for skin and subcutaneous infections cost more than $100 million.

Meaning

Readmission after hospitalization for cellulitis is common and costly and may be preventable with improved diagnostics, therapeutics, and discharge care coordination.

This nationally weighted cohort analysis assesses the association of hospitalization for treatment of cellulitis with readmission among patients in the US National Readmissions Database.

Abstract

Importance

Cellulitis commonly results in hospitalization. Limited data on the proportion of cellulitis admissions associated with readmission are available.

Objective

To characterize the US national readmission rate associated with hospitalization for treatment of cellulitis.

Design, Setting, and Participants

This retrospective cohort analysis of cellulitis admissions from the nationally representative 2014 Nationwide Readmissions Database calculated readmission rates for all cellulitis admissions and subsets of admissions. The multicenter population-based cohort included adult patients admitted for conditions other than obstetrical or newborn care. Data were collected from January 1 through November 30, 2014, and analyzed from February 1 through September 18, 2018. Bivariate logistic regression models were used to assess differences in readmission rates by patient characteristics. Costs were calculated for all readmissions after discharge from hospitalization for cellulitis (hereinafter referred to as cellulitis discharge) and by readmission diagnosis.

Exposures

Admission with a primary diagnosis of cellulitis.

Main Outcomes and Measures

Proportion of cellulitis admissions associated with nonelective readmission within 30 days, characteristics of patients readmitted after cellulitis discharge, and costs associated with cellulitis readmission.

Results

A total of 447 080 (95% CI, 429 927-464 233) index admissions with a primary diagnosis of cellulitis (53.8% male [95% CI, 53.5%-54.2%]; mean [SD] age, 56.1 [18.9] years) were included. Overall 30-day all-cause nonelective readmission rate after cellulitis discharge was 9.8% (95% CI, 9.6%-10.0%). Among patients with cellulitis, age (odds ratio for 45-64 years, 0.78; 95% CI, 0.75-0.81; P = .001) and insurance status (odds ratio for Medicare, 2.45; 95% CI, 2.33-2.58; P < .001) were associated with increased readmission rates. The most common diagnosis of readmissions included skin and subcutaneous tissue infections. The total cost associated with nonelective readmissions attributed to skin and subcutaneous infections within 30 days of a cellulitis discharge during the study period was $114.4 million (95% CI, $106.8-$122.0 million).

Conclusions and Relevance

Readmission after hospitalization for cellulitis is common and costly and may be preventable with improved diagnostics, therapeutics, and discharge care coordination.

Introduction

Cellulitis is a common bacterial infection of the skin for which ambulatory care costs alone in the United States were estimated at $3.7 billion in 2006.¹ Of the 2.9 million patients presenting annually to US emergency departments for cellulitis, 12.9% are admitted to the hospital.² However, the proportion of cellulitis admissions that result in subsequent readmission has not been fully characterized, and the cost of these readmissions is unknown. Understanding the national incidence and cost of readmission after cellulitis hospitalization is important for prioritizing targets of readmission reduction programs.

Methods

We performed a retrospective analysis of all admissions from 2014 using the Healthcare Cost and Utilization Project Nationwide Readmissions Database, which allows national estimates via poststratification weighting. Patients admitted for cellulitis were identified according to a list of codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) (681.00, 681.10, 681.9, and 682.0-682.9), that includes all diagnoses containing cellulitis and abscess and excludes several related but distinct conditions such as paronychia and external ear infections. The diagnostic utility of these ICD-9-CM codes was previously validated in the Veterans Affairs database.³ We excluded hospitalizations for obstetrical or newborn care, patients younger than 18 years, and those with missing primary diagnosis and/or length of stay. Insurance status was categorized as Medicare, Medicaid, private insurance, or other, which includes self-pay, worker’s compensation, CHAMPUS (Civilian Health and Medical Program of the Uniformed Services), CHAMPVA (Civilian Health and Medical Program of the Department of Veterans Affairs), Title V, and other non-Medicare/Medicaid government programs. The institutional review board of Partners Healthcare approved this study and waived the need for informed consent for this use of deidentified patient data.

Readmission rates were defined as the number of index admissions from January 1 through November 30, 2014, associated with a nonelective readmission within 30 days of index admission divided by the total number of eligible index admissions during the study period. Admissions during which patients died or that occurred in December 2014 for which 30 days of follow-up were unavailable were excluded as index admissions.

Primary diagnoses assigned at readmission after the index admission for cellulitis were identified using multilevel Healthcare Cost and Utilization Project Clinical Classification Software codes.⁴ Among readmissions with available cost data, the most frequent readmission diagnoses were calculated. Readmission costs were calculated using total charges for each hospital discharge multiplied by hospital-specific cost-to-charge ratios derived from accounting reports collected by the Centers of Medicare & Medicaid Services, as previously described.⁵ For a given readmission period, only the first readmission within 30 days of discharge from a hospitalization for cellulitis (hereinafter referred to as cellulitis discharge) was included in cost calculations. Total cost associated with all first-time readmissions after cellulitis discharge was calculated as well as costs associated with readmissions within each readmission diagnosis category. SURVEYMEANS and SURVEYLOGISTIC from SAS software (version 9.4; SAS Institute, Inc) were used to account for the complex sampling design of the Nationwide Readmissions Database. Two-sided P < .05 calculated using the Wald χ² test indicated significance.

Results

Of 19 882 317 (95% CI, 19 230 335-20 534 298) weighted index admissions, 447 080 (95% CI, 429 927-464 233) or 2.2% (95% CI, 2.2%-2.3%) had an associated primary diagnosis of cellulitis (Table 1) (46.2% women [95% CI, 45.8%-46.5%] and 53.8% men [95% CI, 53.5%-54.2%]; mean [SD] age, 56.1 [18.9] years). Overall 30-day all-cause readmission rate for index admissions with a primary diagnosis of cellulitis was 9.8% (95% CI, 9.6%-10.0%) (Table 2). By bivariate logistic regression analysis, readmission rates were lower for patients aged 18 to 44 years (7.0%; 95% CI, 6.7%-7.3%; P < .001) and for patients aged 45 to 64 years (9.8%; 95% CI, 9.5%-10.1%; P = .001) than for patients 65 years or older (12.3%; 95% CI, 12.0%-12.6%) (Table 2). Rates were higher for patients insured by Medicare (12.8% [95% CI, 12.5%-13.0%]; odds ratio, 2.45 [95% CI, 2.33-2.58]; P < .001) and Medicaid (10.8% [95% CI, 10.4%-11.2%]; odds ratio, 2.03 [95% CI, 1.91-2.15]; P < .001) than for privately insured patients (5.7%; 95% CI, 5.4%-5.9%).

Table 1. Cohort Characteristics of All Index Admissions With Primary Diagnosis of Cellulitis.

Characteristics	Index Admissions, No. (95% CI)	Index Admissions, % (95% CI)
All	447 080 (429 927-464 233)	100
Sex
Female	206 364 (198 202-214 526)	46.2 (45.8-46.5)
Male	240 716 (231 430-250 002)	53.8 (53.5-54.2)
Age, y
18-44	128 379 (122 358-134 401)	28.7 (28.1-29.3)
45-64	166 019 (159 478-172 560)	37.1 (36.8-37.5)
≥65	152 682 (146 337-159 027)	34.2 (33.4-34.9)
Insurance
Medicare	195 063 (187 436-202 690)	43.7 (43.0-44.4)
Medicaid	82 627 (78 245-87 009)	18.5 (17.8-19.2)
Private	104 956 (100 183-109 729)	23.5 (22.9-24.1)
Other	63 817 (59 172-68 462)	14.3 (13.5-15.0)

Table 2. Number of Readmissions, Readmission Rates, and Bivariate Unadjusted Logistic Regression Odds for Nonelective Readmissions Within 30 Days of Index Admission for Cellulitis.

Characteristics	Readmissions, No. (95% CI)	Readmission Rate, % (95% CI)	OR (95% CI)	P Valuea
All	43 916 (41 890-45 943)	9.8 (9.6-10.0)	NA	NA
Sex
Female	21 536 (20 520-22 553)	10.4 (10.2-10.7)	1 [Reference]
Male	22 411 (21 261-23 560)	9.3 (9.1-9.5)	0.88 (0.85-0.91)	<.001
Age, y
18-44	8953 (8321-9585)	7.0 (6.7-7.3)	0.54 (0.51-0.56)	<.001
45-64	16 265 (15 408-17 123)	9.8 (9.5-10.1)	0.78 (0.75-0.81)	.001
≥65	18 729 (17 837-19 621)	12.3 (12.0-12.6)	1 [Reference]	NA
Insurance
Medicare	24 914 (23 758-26 069)	12.8 (12.5-13.0)	2.45 (2.33-2.58)	<.001
Medicaid	8935 (8348-9523)	10.8 (10.4-11.2)	2.03 (1.91-2.15)	<.001
Private	5939 (5547-6332)	5.7 (5.4-5.9)	1 [Reference]	NA
Other	4129 (3689-4568)	6.5 (6.1-6.9)	1.16 (1.07-1.25)	<.001

Abbreviations: NA, not applicable; OR, odds ratio.

^aCalculated using Wald χ² test.

The most common diagnoses at readmission within 30 days of cellulitis index admission were skin and subcutaneous tissue infections and diseases of the heart, respectively (Table 3). The total cost associated with nonelective readmissions within 30 days of a cellulitis discharge from January 1 through November 30, 2014, was $553.3 million (95% CI, $526.5-$580.1 million) (Table 3). The cost from readmissions attributed to skin and subcutaneous tissue infections was $114.4 million (95% CI, $106.8-$122.0 million) (Table 3).

Table 3. 10 Most Frequent CCS Categories of Primary Diagnosis at Readmission Within 30 Days of Cellulitis Discharge.

CCS Code	Primary Diagnosis at Readmission	Frequency, % (95% CI)	Cost, $1 Million (95% CI)a
12.1	Skin and subcutaneous tissue infections	29.7 (28.9-30.4)	114.4 (106.8-122.0)
7.2	Diseases of the heart	8.3 (7.9-8.8)	53.3 (48.1-58.4)
1.1	Bacterial infection	8.0 (7.6-8.5)	67.7 (61.9-73.5)
10.1	Diseases of the urinary system	5.3 (4.9-5.7)	24.6 (22.1-27.1)
3.3	Diabetes with complications	4.3 (4.0-4.6)	25.5 (22.7-28.2)
16.10	Complicationsb	4.0 (3.7-4.3)	30.7 (26.4-35.0)
8.1	Respiratory infections	2.4 (2.2-2.7)	13.1 (11.2-15.0)
17.1	Ill-defined conditionsc	2.4 (2.102.6)	12.4 (10.0-14.7)
7.5	Diseases of veins and lymphatics	2.1 (1.8-2.3)	9.5 (7.9-11.0)
13.1	Infective arthritis and osteomyelitisd	1.5 (1.3-1.7)	10.1 (8.4-11.7)
NA	Other	32.2 (31.5-32.9)	189.4 (178.3-200.5)
NA	All	100.0	553.3 (526.5-580.1)

Abbreviations: CCS, Clinical Classification Software; NA, not applicable.

^aAssociated costs of readmissions are shown for each readmission diagnosis category.

^bIncludes complications of device/implant/graft, surgical procedures, or medical care.

^cIncludes symptoms, signs, and ill-defined conditions such as syncope, fever of unknown origin, shock, gangrene, abdominal pain, allergic reactions, and others.

^dExcept that caused by tuberculosis or sexually transmitted infections.

Discussion

Our estimated 30-day readmission rate after cellulitis discharge of 9.8% in a national cohort approximates a prior estimate of 11.2% within a single hospital system cohort.⁶ The national readmission rate after index admissions for any diagnosis is estimated to be 14.5%.⁷ The slightly lower rate of readmissions after cellulitis discharge relative to those after other illnesses likely reflects the younger mean age and general well-being of patients with cellulitis.⁸

A proportion of readmission is likely due to cellulitis treatment failure; others may be due to comorbid conditions that are distinct from cellulitis. However, misdiagnosis of cellulitis likely contributes to readmission. Multiple common clinical mimics of cellulitis exist, and 30% of patients admitted through the emergency department for cellulitis have been found to have a cellulitis mimic.⁹ Admitted patients who are misdiagnosed with cellulitis commonly receive inappropriate treatment, increasing their likelihood of requiring readmission. We found that the second most common category of readmission diagnoses is diseases of the heart. Because cardiovascular disease is often associated with vascular cellulitis mimics such as stasis dermatitis, this finding suggests that a proportion of readmissions may be related to a misdiagnosed cardiovascular-associated dermatosis.

Several studies have suggested that relatively few readmissions are related to the index admission.^10,11 In contrast, we found that nearly 30% of readmissions after cellulitis discharge are for skin and subcutaneous tissue infections. Other top categories of readmission diagnosis may be closely related to cellulitis index admission as well. Bacterial infections (including sepsis) and osteomyelitis may be sequelae of cellulitis treatment failure; diseases of the urinary system (including urinary tract infections) and respiratory infections may be hospital acquired; and cardiac disease, diabetes, and vein or lymphatic disease may reflect an association with cellulitis mimics as above. Nonetheless, a proportion of readmissions will be unrelated to preceding index admissions, and this limitation is inherent to all readmission analyses.

Our bivariate logistic regression analysis suggests that older patients and those insured by Medicare or Medicaid are at increased risk of readmission after cellulitis discharge. This finding is consistent with previous data showing that these populations are at increased risk of readmission after all-cause admissions.⁷ Although patients insured by Medicare and those 65 years or older overlap, our observed difference in readmissions among these populations is likely accounted for by patients with disabilities and/or end-stage renal disease. Young Medicare patients have been shown to be at uniquely high risk of readmission after all-cause admissions.⁷

Until further research is performed around cellulitis-specific interventions, adaptation of multifaceted interventions shown to be effective in reducing readmissions for other indications, such as bridged care by dedicated transitional care personnel, patient-centered discharge instructions, and telephone follow-up, should be considered.¹² Reducing misdiagnosis of cellulitis through increased use of specialty consultation or point-of-care clinical support with tools or diagnostics may also help reduce readmission by ensuring appropriate care for patients during their index admission.^13,14,15

Limitations

These results must be considered in the context of the study design. One limitation of this study is the lack of adjustment for potential confounders, and prospective, multiple-institution research is required to better understand the root cause and risk factors for readmission. Our analysis is further limited by the grouping of cellulitis with abscess in ICD-9-CM coding as well as limited patient-level data, including patient race/ethnicity and income. In addition, although the list of ICD-9-CM codes used to identify cases of cellulitis was previously validated in the Veterans Affairs database, its diagnostic accuracy is not necessarily translatable to other patient groups. Finally, because we excluded December admissions from analysis owing to incomplete follow-up, our results likely underestimate the true annual prevalence and financial burden of readmission after cellulitis discharge.

Conclusions

Readmissions after cellulitis discharge nationwide are common and costly. Although readmission among patients with cellulitis will never be zero, the more than 40 000 patients affected and $100 million in annual costs demonstrate the scope of opportunity for improvement. Given the limitations of our study, future studies are needed to determine the root cause and risk factors for readmission and to distinguish between readmissions after cellulitis vs abscess. Reducing misdiagnosis of cellulitis through dermatology consultation or diagnostic tools may help reduce readmission.