Medical Costs in Community Subjects With Irritable Bowel Syndrome

Transcription

1 GASTROENTEROLOGY 1995;109: ALIMENTARY TRACT Medical Costs in Community Subjects With Irritable Bowel Syndrome NICHOLAS J. TALLEY, SHERINE E. GABRIEL, W. SCOTT HARMSEN, ALAN R. ZINSMEISTER, and ROGER W. EVANS Division of Gastroenterology and Internal Medicine and Department of Health Sciences Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota See editorial on page Background & Aims: Costs of management of irritable bowel syndrome (IBS) are unknown. The direct medical charges in community subjects with IBS were estimated. Methods: An age- and sex-stratified random sample of residents of Olmsted County, Minnesota, ranging in age from 20 to 95 years, was mailed a valid self-report questionnaire. Subjects were categorized as having IBS, having some symptoms but inadequate criteria for IBS, and controls. All charges (in 1992 U.S. dollars) for health services rendered in the year before completing the survey were obtained (except outpatient medications). Results: A total of 88% of subjects with IBS, 86% of subjects with some symptoms of IBS, and 83% of controls incurred direct medical charges during the study year. The odds of incurring charges were 1.6 times greater in subjects with IBS relative to those without symptoms (P < 0.01) adjusting for age, sex, education, marital status, and employment. Overall median charges incurred by subjects with IBS were S742 compared with $429 for controls and 5614 for subjects with some symptoms. Among those subjects with nonzero charges, there were significant positive associations with age, higher education, and symptom groups (all P < 0.01) but not sex. Conclusions: The economic impact of IBS is significant. A better understanding of the determinants of these costs is needed so that cost-saving strategies can be implemented. p ber capita health care expenditures in the United States increased from $302 in 1970 to $3160 in Health care expenditures now constitute 14% of the gross domestic product compared with 7.4% in Much of this increase is attributed to expensive technological innovations. 4 However, common disorders of modest expense can contribute substantially to the total cost of health care, as is the case with selected gastrointestinal disorders. 5'6 Irritable bowel syndrome (IBS) represents one of the most common conditions encountered by gastroenterolo- gists and general internists alike. It has been estimated that there are between 2.4 and 3.5 million visits to physicians yearly in the United States by patients with IBS. 7 In the general population, symptoms consistent with IBS are reported by 15% of persons 8-11 and, although only a minority seek treatment, IBS still accounts for 20%- 50% of referrals to gastroenterology clinics Despite this level of use, data on the economic impact of IBS are sparse. In an effort to assess the cost of managing IBS, we performed a community survey in Olmsted County, Minnesota, and obtained data on all charges incurred for direct medical services (except medications) during a period of 1 year. Materials and Methods Subjects The population of Olmsted County comprises more than 100,000 persons, of whom 96% are white; sociodemographically, the community is similar to the U.S. white population, 8 and the Mayo Clinic is the major provider of medical care for this population. It has been determined that about 15 % of all Mayo Clinic registrations are from the local population. Each year, more than half of the population of Olmsted County is seen at one of the Clinic facilities. During any given 4-year period, more than 95% of local residents will have had at least one medical contact with a local care provider (e.g,, for dental x-rays, sports physicals, pre-employment examinations, and minor illness as well as routine medical care). All health care services and associated charges used by residents of Olmsted County are available in a computerized database. 15'16 An important feature of the Rochester environment is that each of the medical care providers uses a dossier (or unit record) system whereby all medical information for each individual is accumulated in a single record. The pertinent clinical data are Abbreviations used in this paper: IBS, irritable bowel syndrome by the American Gastroenterological Association /95/$3.00

2 December 1995 EXPENDITURES FOR IBS 1737 accessible because the Mayo Clinic has maintained extensive indices based on clinical and histological diagnoses and surgical procedures since The system was further developed by the Rochester Epidemiology Project, which created similar indices for the records of the other providers of medical care to residents of Rochester and Olmsted County.17 The Rochester Epidemiology Project records linkage system therefore provides what is essentially an enumeration of the population from which samples can be drawn. Using this system, we randomly selected persons living in Olmsted County ranging in age from 20 to 64 years and, in a separate random sample, from 65 to 95 years, stratified by age (in 5-year intervals) and sex (equal numbers of men and women). This approach provided a representative sample of residents of Olmsted County who are 20 years of age and older. Data Collection Medical record review. The medical records (inpatient and outpatient) of candidate subjects were reviewed. Subjects were excluded if they were not white (n = 50), had been diagnosed as having a major psychotic episode or dementia (n = 195) precluding them from answering a postal survey, lived in a nursing home (n = 252), or had undergone major abdominal surgery or currently had a major organic medical disease (n = 236). A letter was sent to all remaining eligible subjects (n = 4108) outlining the study and requesting their participation. Included with the letter was either the Bowel Disease Questionnaire I8 or, in subjects 65 years of age or older, the Elderly Bowel Symptom. Questionnaire) 9 Both questionnaires have been shown to be understandable, easily completed, and reliable; they have also been shown to have adequate validity. ~8'19 The Bowel Disease Questionnaire has 46 gastrointestinal symptom items, whereas the Elderly Bowel Symptom Questionnaire has 33 symptom items; 26 of these items are identical on both questionnaires, and these alone were considered in the analyses here. Reminder letters were sent as needed after 2 weeks, 4 weeks, and 7 weeks to nonresponders. The remaining persons were then telephoned at 10 weeks. Subjects who indicated at any point that they did not wish to participate were not contacted further. Questionnaire. Subjects were classified a priori into those with and without IBS as follows. Symptoms compatible with IBS. This category included persons who experienced abdominal pain more than six times in the prior year in combination with two or more of the following symptoms (referred to as the Manning symptom criteria): (1) pain that was relieved by defecation often (more than 25% of the time), (2) looser stools when pain began often, (3) more frequent stools when pain began often, (4) visible abdominal distention often, (5) a feeling of incomplete evacuation often, and (6) mucus per rectum. It has been shown that the more of these criteria that are present, the higher the probability of IBS Based on the available literature, we used a cutoff score of two or more criteria to identify symptoms compatible with IBS because this has been considered optimal for epidemiological studies. *-1 Controls. This category included all subjects who did not report gastrointestinal symptoms in the year before the survey. Other gastrointestinal symptoms. This category included subjects with abdominal pain or disturbed defecation in the year before the survey who failed to meet the criteria for IBS. Charge data. Charge data were derived from the institutional planning database. Olmsted County is a valuable resource for the analysis of use of medical care services. Virtually all primary, secondary, and tertiary medical care for residents is provided by the Mayo Clinic or the Olmsted Medical Group. Both facilities are nonprofit, private group practices with salaried physicians. Of the three hospitals staffed by these physicians, two (St. Mary's Hospital and Rochester Methodist Hospital) are affiliated with the Mayo Clinic and the third (Olmsted Community Hospital) is affiliated with the Olmsted Medical Group. The medical record, utilization, and billing data systems for the two practice groups and the three affiliated hospitals are linked to allow "capture" of inpatient and outpatient activity for community residents regardless of the health plans in which they participate. The resulting data is known as the Olmsted County Health Care Utilization and Expenditures Database. Information on all hospital episodes (inpatient and outpatient) is obtained from the Hospital Utilization Review Database from Mayo hospitals and a similar database for the Olmsted Community Hospital. Information on charges for all ambulatory surgical procedures and inpatient physician services is obtained from the billing data systems for Olmsted County providers. All other inpatient charges for the Olmsted Medical Group are obtained from billing files. For Mayo, charge data are compiled from both the Hospital Utilization Review Database (inpatient) and the Institutional Planning Database (outpatient). Together with the Rochester Epidemiology Project, the Olmsted County Health Care Utilization and Expenditures Database provides a unique opportunity to perform population-based analyses of health services used and costs incurred. Billing data for all health services used among residents of Olmsted County have been archived and are available electronically through the Olmsted County Health Care Utilization and Expenditures Database beginning on January 1, Although the majority of health services in this database are identified through the physicians' current procedural terminology, a number of other coding systems were also in use, including physician specialty codes (for both the Mayo Clinic and Olmsted Medical Group), inpatient universal billing codes (UB 82), other hospital codes contained in the Hospital Utilization Review Database, outpatient department/service codes, and the Health Care Financing Administration's common procedure codes. These multiple coding systems make the assembly of use and expenditures data for research purposes difficult and time consuming. We have circumvented this problem by creating a single financial and procedural coding system that categorizes each health service used by every resident of

3 1738 TALLEY ET AL. GASTROENTEROLOGY Vol. 109, No. 6 Table 1. Demographic Characteristics of the Study Population No Some gastrointestinal gastrointestinal symptoms symptoms ibs Total no, Age (yr)" 64 (48, 73) 61 (42, 73) 53 (37, 70) Sex (% male) Employed (%) Married (%) Education (,% with more than high school education) % who completed high school "Median (25th percentile, 75th pementile). Olmsted County, whether it is inpatient or outpatient, and its associated cost and charge to one of 17 body systems (according to the original current procedural terminology categories and seven major service categories). This mechanism provides a method for summarizing all financial health service data for the population of Olmsted County in a standardized form. This system, which is known as the ReSTRUCT coding system, is available both as an INGRES and an SAS file and also has the capability of providing line-item detail for any category. Data are thus accumulated for all visits, examinations, tests, and procedures. All charges for health services used by the subjects (except outpatient medication use) from Mayo and non-mayo sources were obtained. Using the Consumer Price Index (medical), all charges were inflation-adjusted to 1992 dollars. 23'2a Statistical Analysis The distributions of overall charges and charges broken down by outpatient vs. inpatient/emergency room and also by physician, laboratory, and radiology categories were summarized as median (and interquartile range) due to the positive skewness in these distributions (i.e., some subjects incurred very high charges, whereas others incurred no charges during the study year). In view of these distributions, the data analysis first examined the association between symptom groups (no gastrointestinal symptoms, some gastrointestinal symptoms, and IBS) and charges (0 vs. >0) using a logistic regression analysis. The logistic model included age, sex, year of survey mailing, education level, marital status, employment status, and gastrointestinal symptom group (coded as two dummy variables with no symptoms as the reference). Second, the data in subjects with only nonzero charges were analyzed after a rank transformation of the charge values using a multiple regression analysis of charges (rank scale) on age, sex, education level, year of mailing, and dummy regression variable coding of symptoms (no symptoms, some symptoms, and IBS) to assess the association between charges and these variables. Results A completed questionnaire was returned by 3022 subjects, giving a response rate of 74%. Responders and nonresponders had a similar age and sex distribution. A total of 536 subjects (18%) had symptoms indicating IBS, whereas 775 (25%) had no gastrointestinal symptoms. An additional 1711 (57 %) had other gastrointestinal symptoms but did not meet diagnostic criteria for IBS. The IBS subjects tended (P < 0.01) to be younger and had a greater proportion of women than the group with some gastrointestinal symptoms as well as compared with those without symptoms (Table 1). A summary of the charges for all subjects is presented in Table 2. A total of 88% of subjects with IBS and 83% of controls incurred some direct medical charges during the study year. The odds for incurring some charges were 1.6 times greater in subjects with IBS relative to those without symptoms (P < 0.01) adjusting for age, sex, education, marital status, and employment. The overall median charges incurred by subjects with IBS was $742 compared with $429 for controls and $614 for subjects with some gastrointestinal symptoms. Charges increased with age (P < 0.001) for all three symptom groups, and the median values were greatest for subjects with IBS in all but one age group (Figure 1). A similar pattern was observed for just outpatient charges with subjects with IBS again having the highest overall median charges. The distributions of charge data among those subjects with nonzero charges are summarized in Figure 2 and Table 3. Median charges for subjects with IBS who used services was $893 compared with $659 for controls (Figure 2). Among those subjects with nonzero charges, the multiple linear regression analysis indicated significant positive associations with age, higher education, and symptom groups (all P < 0.01) but did not detect an association with sex. Similar results were obtained in an analysis of outpatient charges only and separately for each category of charges (physician, laboratory, and radiology). Interestingly, the inpatient/emergency room charges among subjects with some gastrointestinal symptoms Table 2. Charges for All 3022 Subjects in the Study Population Some No gastrointestinal gastrointestinal symptoms symptoms (n = 775) (n = 1711) OveraIP 429 (75, 1255) 614 (124, 1532) % with charges Outpatient only a 444 (95, 1240) 606 (138, 1364) Physician a 141 (23, 349) 177 (50, 409) Laboratory a 68 (0, 277) 110 (0, 344) Radiology ~ 46 (0, 163) 69 (0, 214) NOTE. Values are expressed as 1992 inflation-adjusted dollars. amedian (25th percentile, 75th Percentile). IBS (n = 536) 742 (177, 1654) (193, 1508) 228 (6O, 47O) 117 (0, 345) 79 (0, 279)

4 December 1995 EXPENDITURES FOR IBS No Symptoms Some Symptoms ae.""'" "''..... IBS / "" e ~f I I I I t I Figure 1. Medical charges (1992 inflation-adjusted dollars) in a random sample of residents of Olmsted County, Minnesota (n = 3022), by age. were the highest, whereas IBS had the highest average for all other categories (Table 3). Discussion We report the first population-based study estimating the direct medical care costs of a well-defined cohort of individuals with IBS compared with controls from the same community. The use of a reliable and valid postal survey helped to ensure the accuracy of the diagnostic classification and provided other clinical and demographic data. Our results show that individuals with IBS incur substantial excess charges in all service categories examined compared with control subjects of similar age and sex from the same community. Our data suggest that, on average, a subject with IBS incurs approximately $300 more direct medical charges per year compared with a control subject of the same age and sex. Given that the population of people older than 20 years of age in Olmsted County in 1992 was roughly 75,000 and the prevalence of IBS was 18%, we would estimate the total excess charges for IBS alone to be approximately $4 million for 1992 in Olmsted County (although it is not necessarily the case that the excess occurs every year for patients with IBS vs. asymptomatic patients). Extrapolating to the population of U.S. white persons would result in excess charges of $8 billion for IBS yearly. These results exceed previous estimates. 25'26 Sonnenberg and Everhart calculated that the total costs of IBS (including constipation and abdominal pain of unknown cause) approached approximately $1.3 billion annually in We believe that our results provide a reasonable estimate because we studied a random sample of the population, we had an excellent response rate to the questionnaire, and our population-based data re- Age sources ensured virtually complete ascertainment of all health services rendered and charges incurred among the study subjects. Moreover, comparison of previous population-based studies of a variety of chronic diseases in Rochester with those from other communities in the United States 2v suggest that the results for Olmsted County can be generalized to white persons in the United States. Indeed, with the exception of the proportion of the working population employed in the health care industry (approximately 10%) and the higher level of education (84% had > 12 years of education compared with 71% nationally), the sociodemographic characteristics of the population of Olmsted County are similar to those of U.S. white persons. Our study, however, did have some limitations. First, because our study was limited to direct medical costs, data on prescription drug use, indirect costs (such as lost wages), and nonmedical costs (such as home care) were not considered. Thus, our results represent only a proportion of the total socioeconomic impact of IBS in the community. Second, charges were used as a surrogate for costs. Although the terms "costs" and "charges" are often used interchangeably, 2s-3~ the distinction between the two concepts is critical. True costs that reflect the consumption of resources are rarely available Although billed charges are not an accurate estimate of costs, we believe that incremental charges (the primary outcome measure for this study) are a reasonable estimate for incremental costs, particularly in Olmsted County given the relatively small number of providers in the community who were all included (both Mayo and non-mayo). One limitation may have been the variations in charges among the providers. Using publicly available data resources for Medicare beneficiaries treated at the three Olmsted County hospitals, we compared the severity-adjusted = i' 1600 ' _= 1000 "o Controls Other GI IBS Figure 2, Overall medical charges (1992 inflation-adjusted dollars) by symptom group in only those with nonzero charges.

5 1740 TALLEY ET AL. GASTROENTEROLOGY Vol. 109, No. 6 Table 3. Subjects With Nonzero Charges No IBS symptoms (n = 775) Some GI symptoms (n = 1711) IBS symptoms (n = 536) Outpatient n = 617 (80%) a n = 1377 (80%) n = 455 (85%) 646 (215, 1413) o 806 (301, 1527) 875 (322, 1607) Inpatient/emergency n = 73 (9%) n = 272 (16%) n = 93 (17%) 553 (76, 2454) 1349 (140, 4159) 1024 (101, 3380) Physician n = 594 (77%) n = 1390 (81%) n = 447 (83%) 220 (101, 422) 251 (112, 480) 300 (135, 522) Laboratory n = 470 (61%) n = 1129 (66%) n = 374 (70%) 241 (96, 396) 251 (117, 493) 254 (106, 444) Radiology n = 445 (57%) n = 1059 (62%) n = 344 (64%) 132 (69, 283) 167 (85, 340) 207 (91, 430) NOTE. Values are expressed as 1992 inflation-adjusted dollars. anumber of subjects with type of charge (% of total in respective group). ~Median (25th percentile, 75th percentile). GI, gastrointestinal. charges by diagnosis-related group for several digestive diseases) 6-38 After severity adjustment, the length of stay for each diagnosis-related group was similar across hospitals, as was the average charge per case. Based on these results, we conclude that total expenditures for medical care do not vary significantly across Olmsted County providers. Another potentially important issue is medical insurance and whether this influenced the results. Insurance coverage in Olmsted County is high, but we have no data available on the insurance status of individuals in the study. However, for this factor to have confounded our results, there would have to be substantial differences in insurance coverage among each of the subject groups, which seems very unlikely. As we hypothesized, this study found that people with functional bowel complaints accrued more medical charges than other people without symptoms. Our major aim was to estimate the magnitude of any differences in an effort to calculate the individual dollars that society is paying for the care of IBS compared with normal. We did not aim in this study to compare IBS with organic gastrointestinal disease; indeed, those with organic disease were excluded from the population studied. Although organic disease is less common in the community, the next logical step is to put the costs of IBS in perspective by comparing these two groups, and this should be the goal of a future study. It is of interest to speculate why charges for IBS continued to increase more than charges for other gastrointestinal symptoms in older subjects. We did not identify the specific disorders in those with other gastrointestinal symptoms, but they presumably also represent functional complaints. We speculate that subjects with colonic symptoms (and their physicians) become increasingly concerned about the possibility of colon cancer, even if the symptoms have been chronic and previously investigated in older age groups. Possibly those with IBS are also more demanding than subjects with other symptoms. More work is required to explore these issues. We conclude that IBS is a costly disorder in the community. Research is now needed to determine the appropriateness of these services and their outcomes for patients with IBS. References 1. Schieber G J, Poullier J-P, Greenwald LM. U.S. health expenditure performance: an international comparison and data update. Health Care Financing Rev 1992; 13: Burner ST, Waldo DR, McKusick DR. National health expenditures projections through Health Care Financing Rev 1992; 14: Rublee DA, Schneider M. International health spending: comparisons with the OECD. Health Affairs 1991; 10: Congressional Budget Office. Trends in health care spending: an update. Washington, DC: U.S. Government Printing Office, June Jacobs P. The economics of health and medical care. 3rd ed. Rockville, MD: Aspen Publishers, Schramm CJ. Health care and its costs. New York: Norton, Sandier RS. Epidemiology of irritable bowel syndrome in the United States. Gastroenterology 1990;99: Talley N J, Zinsmeister AR, Van Dyke C, Melton LJ III. Epidemiology of colonic symptoms and the irritable bowel syndrome. Gastroenterology 1991; 101: Jones R, Lydeard S. Irritable bowel syndrome in the general population. BMJ 1992;304: Heaton KW, O'Donnell LJD, Braddon FEM, Mountford RA, Hughes AO, Cripps PJ. Symptoms of irritable bowel syndrome in a British urban community: consulters and nonconsulters. Gastroenterology 1992; 102: Drossman DA, Li Z, Andruzzi E, Temple RD, Tailey N J, Thompson WG, Whitehead WE, Janssens J, Funch-Jensen P, Corazziari E, Richter JE, Koch GG. U.S. householder survey of functional GI disorders: prevalence, sociodemography and health impact. Dig Dis Sci 1993;38: Switz DM. What the gastroenterologist does all day. Gastroenterology 1976; 70: Mitchell CM, Dressman DA. Survey of the AGA membership relat-

6 December 1995 EXPENDITURES FOR IBS 1741 ing to patients with functional gastrointestinal disorders. Gastroenterology 1987; 92: Harvey RF, Mauad EC, Brown AM. Prognosis in the irritable bowel syndrome: a five-year prospective study. Lancet 1987;1: Campion ME, Naessens JM, Leibson CL, Shaller D, Ballard DJ. The OImsted County benchmark project: primary study findings and potential implications for corporate America. Mayo Clin Proc 1992;67: Gabriel SE, Campion ME, O'Falion WM, Melton LJ II1. The health care costs of nonsteroidal anti-inflammatory drug-related gastric ulcer in Olmsted County, MN, Post Marketing Surveillance 1992;6: , 17. Kurland LT, Molgaard CA. The patient record in epidemiology. Sci Am 1981; 245: Talley N J, Phillips SF, Melton L J, Wiltgen C, Zinsmeister AR. A patient questionnaire to identify bowel disease. Ann Intern Med 1989; 111: O'Keefe EA, Talley N J, Tangalos EG, Zinsmeister AR. A bowel symptom questionnaire for the elderly. J Gerontol Med Sci 1992;47:Ml16- M Manning AP, Thompson WG, Heaton KW, Morris AF. Towards positive diagnosis of irritable bowel. BMJ 1978;2: Talley N J, Phillips SF, Melton L J, Wiltgen C, Zinsmeister AR. Diagnostic value of the Manning criteria in irritable bowel syndrome. Gut 1990; 31: Talley NJ. Diagnosing an irritable bowel: does sex matter? Gastroenterology 1991; 100: U.S. Bureau of the Census. Statistical abstract of the United States: th ed. Hyattsville, MD: Public Health Service, National Center for Health Statistics. Health, United States, DHHS Pub No. (PHS) Hyattsville, MD: Public Health Service, Brown DM, Everhart JE. Total cost of digestive diseases in the United States, In: Everhart JE, ed. Digestive diseases in the United States: epidemiology and impact. Washington, DC: U.S. Government Printing Office, Sonnenberg A, Everhart JE. Socio-economic determinants of digestive diseases. Gastroenterol int 1993;6: Kurland LT, Molgaard CA. The patient record in epidemiology. Sci Am 1981;245: Finkler SA. The distinction between cost and charges. Ann Intern Med 1982; 96: Lundberg GD. Cost, charges, conscience, and control (lost). JAMA 1984;250: Newhouse JP, Cretin S, Witsberger CJ. Predicting hospital accounting costs. Health Care Financing Rev 1989; 11: Price KF. Pricing Medicare's diagnosis-related groups: charges versus estimated costs. Health Care Financing Rev 1989;11: Evans RW. The socioeconomics of organ transplantation. Transplant Proc 1985;17(Suppl 4): Evans RW. Cost effectiveness analysis of transplantation. Surg Clin North Am 1986;66: Evans RW. Organ transplantation costs, insurance coverage, and reimbursement, in: Terasaki PI, ed. Clinical transplants. Los Angeles, CA: UCLA Tissue Typing Laboratory, 1990: Evans RW, Manninen DL, Dong FB. An economic analysis of liver transplantation: costs, insurance coverage, and reimbursement. Gastroenterol Clin North Am 1993; 22: Love JR, Pashos CL, Anderson GF, et al. Costing medical care: using Medicare administrative data. Medical Care 1994; 32:JS1- JS lezzoni Ll. Risk adjustment for measuring health care outcomes. Ann Arbor, MI: Health Administration Press, U.S. General Accounting Office. Medicare: changes to HMO rate setting method are needed to reduce program costs. Report GAO/HEHS Washington, DC: U.S. General Accounting Office, September Received February 14, Accepted July 31, Address requests for reprints to: Nicholas J. Talley, M.D., Ph.D., University of Sydney, Clinical Sciences Building, Nepean Hospital, P.O. Box 63, Penrith, New South Wales 2751, Australia. Fax: (61) Supported in part by grants AG09440 and AR30582 from the National Institutes of Health.