Hospitalization for Nontraumatic Disorders of the Eye and Ocular Adnexa Analysis of the Florida Agency for Health Care Administration Data Set
Total Page:16
File Type:pdf, Size:1020Kb
SOCIOECONOMICS AND HEALTH SERVICES SECTION EDITOR: PAUL P. LEE, MD Hospitalization for Nontraumatic Disorders of the Eye and Ocular Adnexa Analysis of the Florida Agency for Health Care Administration Data Set Zuber D. Mulla, MSPH, PhD; Curtis E. Margo, MD, MPH Objective: To study the demographic features and pat- Eighty-three patients (3.9% of eye admissions) were hos- terns of hospital admission in Florida for nontraumatic pitalized for 10 days or longer and accounted for 18.1% disorders of the eye and ocular adnexa. of total hospital-patient days. Prolonged hospital stay was positively associated with hospital transfer (P Ͻ .001) and Methods: The public data set from the Florida Agency facial cellulitis (P=.04). A trend for positive association for Health Care Administration for 2001 was used to iden- with Medicaid coverage was also observed (P=.07). tify persons hospitalized for 24 hours or longer for non- traumatic disorders of the eye and ocular adnexa by us- Conclusions: Nontraumatic eye care composes a small ing International Classification of Diseases, Ninth Revision, proportion of all inpatient care (Ͻ 0.1%) in Florida. Few Clinical Modification codes. of these patients require prolonged hospitalization but use a large proportion of inpatient care on the basis of Results: In 2001, there were 2137 hospital admissions the percentage of gross charges. An opportunity exists for nontraumatic disorders of the eye and ocular ad- to improve hospital efficiency and improve eye care by nexa, most of which were for infections or neuro- targeting the patients at highest risk for prolonged hos- ophthalmologic disorders. The median length of stay was pital stay. 3.0 days (mean±SD, 3.4±3.8 days). On average, 1 pa- tient was admitted per month to 180 Florida hospitals. Arch Ophthalmol. 2004;122:262-266 ODERN HOSPITALS ARE care not related to trauma, we studied the important collective hospital discharge data set of the Florida investments of com- Agency for Health Care Administration munities and the place (AHCA), Tallahassee, Fla, for 2001. most people think of Mgoing when they are seriously ill or in- METHODS jured. Hospitals are typically the care- givers of last resort for persons without The public-use hospital discharge data set medical insurance and the hub of medi- from AHCA for Florida for 2001 was cal education. During the past decade, hos- accessed for clinical and demographic infor- pitals have come under intense financial mation for patients discharged with primary and legislative pressure to reduce waste, diagnoses for eye and ocular disease by using the International Classification of Dis- cut costs, and improve efficiency. The ef- eases, Ninth Revision, Clinical Modification fects of these forces on hospitals are com- (ICD-9-CM). This comprehensive set of plex and often difficult to selectively mea- codes included benign and malignant neo- sure. Data from the Agency for Healthcare plastic diagnoses of the eye, conjunctiva, Research and Quality, for instance, have eyelids, lacrimal gland, ocular adnexa, and From the Discipline of shown declining hospital stays but rising orbit (Table 1). Epidemiology, School of Public hospital charges throughout the United The 2001 AHCA data set includes dis- Health, El Paso Regional States from 1993 through 2000.1 charge summaries from all nonfederal Florida Campus, University of Texas Despite the central role hospitals have hospitals except state tuberculosis hospitals and Health Science Center at in our health care system, there is little state mental health hospitals. After data are en- Houston (Dr Mulla); and the tered into the system, they are subjected to for- Department of Ophthalmology, population-based information on inpa- matting and logic checks. The primary hospi- Watson Clinic, Lakeland, Fla tient eye care. Previous studies of hospi- tal submitting patient information must then (Dr Margo). The authors have tal eye care have dealt with ocular in- certify the data are correct and also verify the no relevant financial interest in jury.2-5 To better understand the nature, accuracy of a summary report before it is re- this article. distribution, and volume of inpatient eye leased by the AHCA. (REPRINTED) ARCH OPHTHALMOL / VOL 122, FEB 2004 WWW.ARCHOPHTHALMOL.COM 262 ©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 We retrieved patient information (age, race, and sex), pri- mary admission diagnosis, principle procedure code, princi- Table 1. Hospital Admissions Identified With ICD-9-CM pal payer, day of week of admission, total gross charges for hos- Diagnostic Codes* pitalization, hospital, length of stay, days before procedure, source of admission (physician referral, clinic referral, health Code Description maintenance organization referral, hospital transfer, emer- 172 Melanoma, eyelid gency room, court or law enforcement, nursing home, other 173 Malignancies, other, eyelid health care facility transfer, and other), and discharge status 360 Disorders of the globe (1 of 9 levels). Admission type was classified by the admitting 361 Retinal detachments and defects physician as emergency, urgent, or elective; when applicable, 362 Other retinal disorders admission type was classified by the coding technician as new- 363 Chorioretinal inflammation and other disorders of choroid born or other. Management systems for medical insurance, such 364 Disorders of iris and ciliary body as health maintenance organization or preferred provider or- 365 Glaucoma ganization, were reported according to the primary insurance 366 Cataract carrier (eg, Medicare, Medicaid, commercial, government, work- 367 Disorders of refraction and accommodation ers’ compensation, CHAMPUS, etc). 368 Visual disturbances For the purpose of this study, diagnostically related ICD- 369 Blindness and low vision 9-CM codes were combined to simplify analysis and minimize 370 Keratitis potential coding variations due to semantic differences in mak- 371 Corneal opacity and other disorder of cornea ing clinical diagnoses. For example, endophthalmitis in- 372 Disorders of conjunctiva cluded 3 ICD-9-CM codes: 36000 endophthalmitis, 36001 acute 373 Inflammation of eyelids endophthalmitis, and 36002 panophthalmitis. 374 Other disorders of eyelids 375 Disorders of lacrimal system Annual rates of admission were calculated by using 2000 376 Disorders of the orbit US census data extrapolated to the study year and were ex- 377 Disorders of optic nerve and visual pathways pressed as number of hospital admissions per 100000 popu- 378 Strabismus and related disorders of motility lation. 190 Malignant neoplasms The SAS System release 8.01 for Windows (SAS Institute 173.1 Malignant neoplasms, eyelid Inc, Cary, NC) was used to analyze the data. Frequencies were 224 Benign neoplasms reported as percentages. Prolonged length of stay was defined 216.1 Benign neoplasms, eyelid as hospitalization for 10 days or more. Frequency distribu- tions were examined before and after stratification according Abbreviation: ICD-9-CM, International Classification of Diseases, Ninth to length of stay. Crude and adjusted odds ratios were calcu- Revision, Clinical Modification. lated by using logistic regression.6 The adjusted odds ratios were *Excludes admissions for ocular trauma identified with specific ICD-9-CM derived from 1 full model that contained the following vari- codes.5 ables: age, male sex, black race (as compared with white), Med- icaid enrollees (as compared with self-pay, commercial insur- tal stay of 24 hours or more. Of these admissions, 2858 were ance, or other), emergency or urgent admission (as compared for primary disorders of the eye and ocular adnexa. Seven with elective admission), and source of admission (hospital trans- fer, as compared with other sources). We did not detect col- hundred twenty-one admissions (0.03%) for injuries to the linearity among these independent variables. eye and ocular adnexa were excluded and are the data set After deleting records that had missing values for the de- of a separate study. The remaining 2137 (0.09%) admis- pendent variable and/or independent variables, records of pa- sions to 180 hospitals for nontraumatic disorders of the eye tients who were not black or white, and records of patients who and ocular adnexa composed the data set of this study. were discharged to another hospital, 1698 records were avail- More than 95% of patients (n=2034) listed a pri- able for logistic regression. The records of patients who were mary residence ZIP code in Florida; 29 (1.4%) resided discharged to another hospital (n=45) were deleted to mini- outside the country. The median number of patients ad- mize the probability of including multiple records for a single mitted per hospital was 6 (mean±SD, 11.9±17.2). 2 patient in the multivariate analysis. The test was used to com- The nontrauma admissions involved 1011 male pa- pare patients deleted from the regression analysis because of an incomplete data field, according to frequency of primary ad- tients (47.3%) and 1126 female patients (52.7%) and were mission diagnosis, with those who remained in the analysis. listed with 204 different ICD-9-CM codes. One thou- An odds ratio greater than 1.00 indicated that the variable in- sand three admissions (46.9%) were classified as emer- creased the odds of prolonged length of stay, while an odds ra- gency, 624 as urgent (29.2%), and 510 as elective (23.9%). tio less than 1.00 indicated that the variable protected against Nearly 48% of the admissions were through hospital emer- it. Ninety-five percent confidence intervals were calculated in gency departments, and 41% were direct admissions by the traditional manner. Results were considered significant (P staff physicians. There were 56 hospital transfers (2.6%). Յ .05) if the confidence interval excluded the null value of 1.00. The mean±SD age at admission was 46.5±29.2 years. One The potential role of secondary diagnoses in hospital length thousand four hundred two patients were white (65.6%), of stay was assessed by examining the frequency distribution 336 were black (15.7%), and 332 were Hispanic (includ- of these diagnoses according to length of stay.