Portal de Periódicos da CAPES

Substance Abuse-Related Admissions to Adult Intensive Care

1993; Elsevier BV; Volume: 103; Issue: 1 Linguagem: Inglês

10.1378/chest.103.1.21

1931-3543

W Andrew Baldwin, Brian A. Rosenfeld, Michael J. Breslow, Timothy G. Buchman, Clifford S. Deutschman, Richard D. Moore,

Restraint-Related Deaths

The frequency of adult surgical and medical intensive care unit (ICU) admissions related to substance abuse was determined at a large community, trauma, and tertiary referral hospital. Of 435 ICU admissions, 14 percent (95 percent confidence interval [CI], 5 to 23 percent) were tobacco related generating 16 percent of costs, 9 percent (95 percent CI, 0 to 18 percent) were alcohol related generating 13 percent of costs, and 5 percent (95 percent CI, 0 to 14 percent) were illicit drug related generating 10 percent of costs. In all, 28 percent (95 percent CI, 20 to 36 percent) of ICU admissions generating 39 percent of costs were substance abuse related. Substance abuse-related admissions were significantly longer and more costly than admissions not related to substance abuse (4.2 days vs 2.8 days, p = 0.004; $9,610 vs $5,890, p = 0.001). Frequency of substance abuse-related admission was linked with tibe patients insurance status (Medicare, private insurance, uninsured). In the uninsured group, 44 percent of admissions were substance abuse related (95 percent CI, 35 to 52 percent), significandy higher than in the private insurance and Medicare groups, and generating 61 percent of all ICU costs in the uninsured group. Large fractions of adult ICU admissions and costs are substance abuse related, particularly in uninsured patients. The frequency of adult surgical and medical intensive care unit (ICU) admissions related to substance abuse was determined at a large community, trauma, and tertiary referral hospital. Of 435 ICU admissions, 14 percent (95 percent confidence interval [CI], 5 to 23 percent) were tobacco related generating 16 percent of costs, 9 percent (95 percent CI, 0 to 18 percent) were alcohol related generating 13 percent of costs, and 5 percent (95 percent CI, 0 to 14 percent) were illicit drug related generating 10 percent of costs. In all, 28 percent (95 percent CI, 20 to 36 percent) of ICU admissions generating 39 percent of costs were substance abuse related. Substance abuse-related admissions were significantly longer and more costly than admissions not related to substance abuse (4.2 days vs 2.8 days, p = 0.004; $9,610 vs $5,890, p = 0.001). Frequency of substance abuse-related admission was linked with tibe patients insurance status (Medicare, private insurance, uninsured). In the uninsured group, 44 percent of admissions were substance abuse related (95 percent CI, 35 to 52 percent), significandy higher than in the private insurance and Medicare groups, and generating 61 percent of all ICU costs in the uninsured group. Large fractions of adult ICU admissions and costs are substance abuse related, particularly in uninsured patients. confidence interval intravenous drug abuse Johns Hopkins Hospital Diseases related to use of tobacco, alcohol, and illicit drugs are common and can result in admission to intensive care units (ICU). Intensive care is extremely expensive, billing up to $3,000 or more per bed per day. Total annual expenditures for an estimated 63,000 US adult ICU beds1American Hospital AssociationHospital statistics 1989–90 edition. American Hospital Association, Chicago1989Google Scholar may be as much as $50 billion, or approximately 1 percent of the US gross national product.2Raffin T Shurkin J Sinkler W. Intensive care: facing the critical choices. WC Freeman & Co, New York1989Google Scholar To illuminate the relationship between substance abuse and ICU utilization, this study estimates the frequency and costs of adult ICU admissions at The Johns Hopkins Hospital (JHH), a large community hospital, trauma center, and tertiary referral center in Baltimore, which were related to the patient's use of tobacco, alcohol, or illicit drugs. Substance abuse-related ICU admissions were classified as the following: 1.The US Public Health Service (USPHS) Office of Smoking and Health estimates the relative mortality risk of tobacco smokers compared with nonsmokers for tobacco-related diseases.3US Department of Health and Human Services,Office on Smoking and Health. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. DHHS publication No. (CDC) 89–8411, Rockville, Md1989Google Scholar Following USPHS methodology, a smoker was defined as a person who smokes at least one cigarette, pipe, or cigar per day and has done so for at least one years time. A tobacco-related ICU admission was the admission of a smoker due to a tobacco-related disease with a USPHS relative risk estimate in smokers of 2.25 or greater, which includes the following: (a) lung cancer; (b) cancer of mouth, tongue, pharynx, or larynx; (c) cancer of esophagus; (d) chronic obstructive pulmonary disease (COPD); (e) noncoronary arteriosclerosis; and (f) bladder cancer.The USPHS also estimates the relative risk of each of the above diseases in former smokers, which is greater than 2.25 for cancer of lung, mouth, tongue, pharynx, larynx, and esophagus, and COPD (a, b, c, and d, above) in both male and female subjects, and for noncoronary atherosclerosis in male subjects,3US Department of Health and Human Services,Office on Smoking and Health. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. DHHS publication No. (CDC) 89–8411, Rockville, Md1989Google Scholar and these cases occurring in former smokers were also classified as tobacco related. Bronchioloalveolar carcinoma, lung carcinoid, and esophageal adenocarcinoma arising out of Barrett's esophagus were, in all cases, classified as not tobacco related.2.The following were classified as alcohol-related admissions: (a) alcoholic hepatitis or cirrhosis, with bleeding from varices or coagulopathy, peritonitis, encephalopathy, sepsis, hepatorenal syndrome, Le Veen shunt, portal-systemic shunt, or liver transplantation; (b) alcoholic gastritis; (c) alcoholic pancreatitis, pancreatic pseudocyst, or abscess; (d) alcohol overdose; (e) delirium tremens, alcohol withdrawal, or alcohol-related seizures; and (f) trauma with a blood alcohol level greater than 100 mg/dl (the legal threshold of alcohol intoxication for motor vehicle operators in most states, including Maryland).3.Illicit drugs were defined as phencyclidine, cocaine, marijuana, and nonprescribed opiates, barbiturates, and amphetamines. The following admissions were classified as related to the use of illicit drugs: (a) abscess from illicit drug injection; (b) infectious hepatitis, not type A, or acquired immune deficiency syndrome (AIDS) in a patient with intravenous drug abuse (IVDA) history; (c) illicit drug overdose; (d) heroin nephropathy; and (e) trauma with a toxicologic screen positive for illicit drugs, illicit drug metabolites, or cutting agents. Data from all medical ICU and surgical ICU admissions were collected prospectively for 15-week periods between June and October 1989. Determination of the patients disease leading to ICU admission was made in surgical ICU patients as the disease requiring surgery and in medical ICU patients by reference to the patient's physicians and chart notes. With the cooperation of the hospital's billing and accounting departments, the hospital's internal cost of goods and services provided to each patient was determined, including cost of goods sold, associated labor, depreciation, and overhead. The ICU costs included room and board, nursing, supplies, drugs, diagnostic tests, blood products, and special therapies such as mechanical ventilation and dialysis. The billing office also provided the insurance status of each patient, as (1) Medicare (federally paid health program generally available to all US citizens and legal residents age 65 or older and a small number of younger, chronically ill patients), (2) private health insurance, including patients of health maintenance organizations, and (3) uninsured, including those whose bills were paid by Medicaid (federally sponsored, state-administered health care for uninsured patients) and those with no guarantor. This classification was made because of the importance of insurance status to the patient (for example, as insurance often determines the choice of hospitals and availability of substance abuse counseling and treatment programs), to the hospital (as insurance status determines reimbursement rates for many hospital services), and to third party payers, whether private insurers and their policyholders, or state or federal taxpayers. Statistical comparison of rates was performed by χ2 analysis with continuity correction. Significance of differences between groups was assumed only if p<0.01 because multiple comparisons were made between groups. Logistic regression was used to analyze the associations of demographic variables with substance abuse-related admission. Associations of demographic variables with ICU costs and length of stay were analyzed by multiple linear regression; because the distributions of costs and length of stay were skewed by high-cost outliers, log transformation of these variables was done prior to parametric testing. The Wilcoxon rank-sum test was used for nonparametric comparison of ICU costs and length of ICU stay. The 435 study patients generated $3,014,953 in costs, Table 1 shows demographic characteristics of the study patients. Ages were broadly distributed (25 percent younger than 40 years; 40 percent from 40 years through 64 years; and 35 percent 65 and older). Sexes were almost equally represented (52 percent male and 48 percent female). Most patients were white (59 percent). Medicare and private insurance were the most frequent payers (38 percent and 36 percent, respectively) with lesser numbers in the uninsured group (26 percent).Table 1Characteristics of the ICU Admissions (N = 435)No.(%)Age, yr (mean 54.7, SD 18.4) 64153(35)Sex Male225(52) Female210(48)Race White256(59) Black173(40) Other6(1)Insurance status Uninsured (87 Medicaid and 25 no guarantor)112(26) Medicare166(38) Private insurance157(36)ICU Medical136(31) Surgical299(69)ICU stay, days Mean3.2 Median2Costs, dollars Mean6,931 Median2,712Death39(9)SD = standard deviation Open table in a new tab SD = standard deviation Table 2 shows the frequency of admissions related to substance abuse, which was 28 percent of all admissions generating 39 percent of costs ($1,171,940). Tobacco-related admissions were 14 percent of admissions (59/435) and 16 percent of costs ($481,684). All but one (58/59) tobacco-related admissions were cigarette related; one case of otolaryngologic cancer was attributed to cigars. All patients with cigarette-related admissions had more than ten "pack-year" cigarette exposure (one pack per day for ten years or equivalent, or approximately 70,000 lifetime cigarettes). Cigarette exposure in the cigarette-related group averaged 30 to 40 pack-years, with a maximum of 150 pack-years (approximately 1,000,000 lifetime cigarettes).Table 2Frequency and Costs of Substance Abuse-Related Admissions (N = 435)GroupAdmissions% of Total Admissions*95 Percent confidence intervals are as follows: tobacco-related, 5 to 23 percent; alcohol related, 0 to 18 percent; illicit drug related, 0 to 14 percent; all substance abuse related, 20 to 36 percent.ICU Costs, $% of Total ICU CostsTobacco related5914481,68416Alcohol related419398,99313Illicit drug related225291,26310All substance abuse related122281,171,94039Not substance abuse related313721,843,01361Total admissions4351003,014,953100* 95 Percent confidence intervals are as follows: tobacco-related, 5 to 23 percent; alcohol related, 0 to 18 percent; illicit drug related, 0 to 14 percent; all substance abuse related, 20 to 36 percent. Open table in a new tab Alcohol-related admissions were 9 percent of admissions (41/435) and 13 percent of costs ($398,993), and illicit drug-related admissions were 5 percent of admissions (22/435) and 10 percent of costs ($291,263). Two assault victims were admitted with both a blood alcohol level above 100 mg/dl and cocaine metabolite in urine. One was assigned as alcohol related and one as illicit drug related by lot. Substance abuse-related trauma admissions, if culled and aggregated from the alcohol-related and illicit drug-related groups in Table 2, were 5 percent of admissions (22/435) and 7 percent of costs ($197,518). Table 3 compares rates of substance abuse-related admissions in patient subgroups based on age, sex, race, insurance status, and unit (surgical ICU or medical ICU). Statistically significant differences in rates of substance abuse-related admissions between groups was assumed only if p<0.01. Substance abuse-related admissions were more frequent in male subjects than female subjects, in blacks than whites, and in uninsured patients than in Medicare or private insurance patients (column e). Logistic regression analysis demonstrated that male sex and uninsured status were independently associated with higher rates of substance abuse-related admission, but that black race was not. The higher rates of substance abuse-related admissions recorded in blacks appeared to result from a high proportion of black admissions in the uninsured group. In the uninsured group, 44 percent of admissions were substance abuse related, the highest frequency of any patient group.Table 3Association of Substance Abuse-Related Admissions with Demographic Groups (N = 435; Percent is of ICU Admissions in the Group)Group(a) Total Admissions (N = 435)(b) Tobacco Related, No. (%) (n = 59)(c) Alcohol Related, No. (%)(n = 41)(d) Illicit Drug Related, No. (%) (n = 22)(e) Total Substance Abuse Related, No. (%) (n = 122)(f) % of Costs Substance Abuse Related*No p value calculated; statistical testing is not applicable because there is no variation around these percentages.Age, yr <401080 †p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.15 (14)17 (16)†p 6415329 (19)4 (3)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.0 †p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.33 (22)31Sex Male22532 (14)27 (12)20 (9)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.79 (35)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.42 Female21027 (13)14 (7)2 (1)43 (20)36Race‡n = 429; six patients were Oriental or Hispanic. White25639 (15)16 (6)3 (1)58 (23)34 Black17320 (12)25 (14)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.19 (ll)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.64 (37)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.49Payer Uninsured11212 (11)20 (18)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.17 (15)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.49 (44)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.61 Medicare16628 (17)5 (3)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.2 (l)†p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.35 (21)29 Private insurance15719 (12)16 (10)3 (2)38 (24)37ICU Medical13612 (9)20 (15)6 (4)38 (28)33 Surgical29947 (16)21 (7)16 (5)84 (28)41* No p value calculated; statistical testing is not applicable because there is no variation around these percentages.† p<0.01 comparing observed frequency and expected frequency of substance abuse-related admissions in each demographic group by continuity-corrected χ2.‡ n = 429; six patients were Oriental or Hispanic. Open table in a new tab Table 3 also shows the subtypes and costs of substance abuse-related admissions in each patient group. Tobacco abuse-related admissions were less frequent before age 40 years, alcohol abuse-related admissions were more frequent in the uninsured, and less frequent after age 64 years, and illicit drug-related admissions were more frequent in the under 40 years, male, black, and uninsured groups. The frequency of substance abuse-related admissions was less than the percentage of costs they generated in every patient group, suggesting that substance abuse-related admissions generated higher average costs per admission (column f). In the uninsured group, substance abuse-related admissions generated 61 percent of costs, the highest percentage of any patient group. To check for the importance of possible background incidence of smoking-related diseases in the smoking population (ie, although the listed diseases are statistically tobacco related they may occur even if the patient has never smoked), smoking-related admissions in each disease category were multiplied by the attributable risk of smoking ([relative risk – l]/[relative risk]), using the appropriate relative risk for the patient's disease and sex.3US Department of Health and Human Services,Office on Smoking and Health. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. DHHS publication No. (CDC) 89–8411, Rockville, Md1989Google Scholar This adjustment would decrease the estimated number of smoking-related admissions by 22 percent, from 59 to 46, and would make significant the low rates of substance abuse-related admissions seen in the over 64 years and Medicare admissions (column e). Table 4 compares length of ICU stay and costs for substance abuse-related admissions and other admissions. Average length of stay was longer (4.2 days vs 2.8 days) and average costs were higher ($9,610 vs $5,890) for substance abuse-related admissions (p<0.01). Excesses in ICU days and costs of substance abuse-related admissions were adjusted for age, sex, race, insurance status, and unit by multiple linear regression. Longer average ICU stay (by 0.8 days, p<0.001) and higher average costs (by $1,860, p<0.001) were associated with substance abuse-related admissions after this adjustment. Average costs per ICU day were similar in the substance abuse-related admissions and other admissions ($2,280 vs $2,090), suggesting that length of stay and not intensity of therapy primarily explained the excess costs of substance abuse-related admissions.Table 4Association of ICU Length cf Stay and Charges With Substance Abuse (N = 435)ICU Stay, DaysCosts, DollarsMeanSEMedianP*p value by Wilcoxon rank sum test comparing substance abuse category with the not substance abuse-related group.MeanSEMedianP*p value by Wilcoxon rank sum test comparing substance abuse category with the not substance abuse-related group.All substance abuse related(n = 122)4.20.620.0049,6101,6003,7600.001Tobacco related(n = 59)4.20.820.0088,1601,6803,5700.008Alcohol related(n = 41)4.00.720.0139,7302,5604,4900.032Illicit drug related(n = 22)4.72.010.7813,2406,0702,7900.26Not substance abuse related(n = 313)2.80.215,8907702,260* p value by Wilcoxon rank sum test comparing substance abuse category with the not substance abuse-related group. Open table in a new tab Eight percent (39/435) of the patients died, generating 17 percent of total study costs. Substance abuse-related admissions suffered a trend to less favorable outcome, with 13 percent mortality (16/124) vs 7 percent mortality (23/311) for other admissions, but this trend was not statistically significant (p = 0.10). Previous studies have quantified the role of some forms of substance abuse, particularly alcohol, in causing emergency department visits and hospitalizations.4Bush B Shaw S Cleary P Delbanco TL Aronson MD Screening for alcohol abuse using the CAGE questionnaire.Am J Med. 1987; 82: 231-235Abstract Full Text PDF PubMed Scopus (422) Google Scholar, 5Taylor CL Passmore N Kilbane P Davies R. Prospective study of alcohol-related admissions in an inner-city hospital.Lancet. 1986; 2: 265-268Abstract PubMed Scopus (52) Google Scholar, 6Moore RD Bone LR Geller G Mamon JA Stokes EJ Levine DM Prevalence, detection, and treatment of alcoholism in hospitalized patients.JAMA. 1989; 261: 403-407Crossref PubMed Scopus (488) Google Scholar, 7Soderstrom CA Cowley RA A national alcohol and trauma center survey: missed opportunities, failures of responsibility.Arch Surg. 1987; 122: 1067-1071Crossref PubMed Scopus (72) Google Scholar, 8Stephens CJ Alcohol consumption and casualties: drinking in the event.Drug Alcohol Depend. 1987; 20: 115-127Abstract Full Text PDF PubMed Scopus (13) Google Scholar To our knowledge, no study has estimated the total contributions of substance abuse to medical utilization and costs in any medical setting. This study has examined the role of all common forms of substance abuse in the adult ICU (a diverse, costly, and high-risk group of patients) at a large community, trauma, research, and tertiary referral hospital. In this setting, 28 percent of adult ICU admissions were substance abuse related, generating 39 percent of ICU costs. Since these admissions are rooted in patient behavior, they are in theory preventable. Tobacco and alcohol accounted for approximately three quarters of the substance abuse-related ICU admissions and costs. The abundance of morbidity due to these legal activities, compared with illicit drug abuse, is consistent with their larger health burdens in the overall population. The USPHS estimates annual substance abuse-related US deaths as 390,000 tobacco related,3US Department of Health and Human Services,Office on Smoking and Health. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. DHHS publication No. (CDC) 89–8411, Rockville, Md1989Google Scholar 105,000 alcohol related (including trauma),9Alcohol-related mortality and years of potential life lost-United States, 1987.MMWR. 1990; 39: 173-178PubMed Google Scholar and 9,800 illicit drug related (not including trauma or AIDS).10US Department of Health and Human Services,National Center for Health Statistics.Advance report of final mortality statistics. 1987; Monthly Vital Statistics Report. 1989; 38Google Scholar In the uninsured group, 44 percent of ICU admissions and 61 percent of costs were substance abuse related. These findings may reflect the relatively high prevalence of tobacco, alcohol, and illicit drug use among low-income persons3US Department of Health and Human Services,Office on Smoking and Health. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. DHHS publication No. (CDC) 89–8411, Rockville, Md1989Google Scholar, 11US Department of Health and Human Services,National Institute of Alcohol Abuse and Alcoholism.Sixth special report on alcohol and health. NIAAA, Rockville, Md1987Google Scholar, 12US Department of Health and Human Services,National Institute on Drug Abuse, Division of Epidemiology and Practice Research. Drug use patterns and demographics of employed users; data from the 1988 National Household Survey on Drug Abuse. NIDA, Rockville, Md1990Google Scholar who are the primary group in the US population without health insurance.13Davis K Rowland D. Uninsured and underserved: inequities in health care in the United States.Milbank Mem Fund Q. 1983; 61: 149-176Crossref Scopus (104) Google Scholar, 14US Department of Commerce, Bureau of the CensusHealth insurance coverages 1986–88. Current population reports, Series P-70, No. 17. Government Printing Office, Washington, DC1990Google Scholar Substance abuse-related admissions generated significantly more ICU days and higher costs per admission. In an earlier study identifying high-cost hospital patients, tobacco smoking, alcohol abuse, and illicit drug use were associated with high-cost status.15Zook C Moore F. High-cost users of medical care.N Engl J Med. 1980; 302: 996-1102Crossref PubMed Scopus (284) Google Scholar Several hypotheses might explain the increased ICU days and costs of substance abuse-related admissions. Perhaps (1) substance abusers have more advanced disease when admitted to the ICU, or (2) substance abusers have chronic underlying multisystem illness leading to increased costs. In this study, tobacco-related admissions were limited to those tobacco-related diseases listed by the USPHS with a relative risk ratio for smokers greater than 2.25.3US Department of Health and Human Services,Office on Smoking and Health. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. DHHS publication No. (CDC) 89–8411, Rockville, Md1989Google Scholar Adjustment for attributable risk of smoking indicated a possible 20 to 25 percent background incidence of the defined smoking-related diseases in the smoking population. Alcohol-related diseases were those listed as alcohol related by the USPHS-National Institute on Alcohol Abuse and Alcoholism.16US Department of Health and Human Services, National Institute on Alcohol Abuse and AlcoholismHospital discharges with alcohol-related conditions 1979–85. NIAAA, Rockville, Md1989Google Scholar Alcoholic pancreatitis was also included.17Lieber C. Medical disorders of alcoholism. WB Saunders Co, Philadelphia1982Google Scholar Although alcohol drinking is a risk factor for otolaryngologic and esophageal cancer,18De Vita V Heilman S Rosenberg S. Cancer-principles and practice of oncology. JB Lippincott Co, Philadelphia1989Google Scholar these cases were classified as tobacco related in smokers and former smokers,3US Department of Health and Human Services,Office on Smoking and Health. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. DHHS publication No. (CDC) 89–8411, Rockville, Md1989Google Scholar which may have favored tobacco-related against alcohol-related admissions but did not change the total percentages of substance abuse-related admissions and costs. Accidental trauma with a blood alcohol level above 100 mg/dl was assumed to be alcohol related, which is the method used by USPHS.9Alcohol-related mortality and years of potential life lost-United States, 1987.MMWR. 1990; 39: 173-178PubMed Google Scholar, 11US Department of Health and Human Services,National Institute of Alcohol Abuse and Alcoholism.Sixth special report on alcohol and health. NIAAA, Rockville, Md1987Google Scholar Between 40 percent and 50 percent of fatally injured motor vehicle drivers in the United States have blood alcohol levels higher than 100 mg/dl.11US Department of Health and Human Services,National Institute of Alcohol Abuse and Alcoholism.Sixth special report on alcohol and health. NIAAA, Rockville, Md1987Google Scholar Accidental trauma is also strongly associated with use of marijuana19Budd RD Muto JJ Wong JK Drugs of abuse found in fatally injured drivers in Los Angeles County.Drug Alcohol Depend. 1989; 23: 153-158Abstract Full Text PDF PubMed Scopus (52) Google Scholar, 20Fortenberry JC Brown DB Shevlin LT Analysis of drug involvement in traffic fatalities in Alabama.Am J Drug Alcohol Abuse. 1986; 12: 257-267Crossref PubMed Scopus (16) Google Scholar and cocaine.21Marzuk P Tardiff K Leon A Stajic M Morgan E Mann J. Prevalence of recent cocaine use among motor vehicle fatalities in New York City.JAMA. 1990; 263: 250-256Crossref PubMed Scopus (73) Google Scholar, 22Lindenbaum GA Carroll SF Ierachmeil I Kapusnick R. Patterns of alcohol and drug abuse in an urban trauma center: the increasing role of cocaine abuse.J Trauma. 1989; 29: 1654-1658Crossref PubMed Scopus (85) Google Scholar Assault-related trauma is directly linked to alcohol and illicit drug use as well; in one report,23Sloan EP Zalenski RJ Smith RF Sheaff CM Chen EH Keys NI Toxicology screening in urban trauma patients: drug prevalence and its relationship to trauma severity and management.J Trauma. 1989; 29: 1647-1653Crossref PubMed Scopus (81) Google Scholar more than three quarters of violent crime victims had illicit drugs or metabolites in blood or urine, and in another,24Smith S Goodman R Thacker S Barton A Parsons J Hudson P. Alcohol and fatal injuries: temporal patterns.Am J Prevent Med. 1989; 5: 296-302PubMed Google Scholar 53 percent of homicide victims had blood alcohol levels greater than 100 mg/dl. Several other limitations of this study should be considered; (1) Intoxicated motor vehicle drivers often injure nonintoxicated persons, whose injuries might be considered substance abuse related. Also, up to 50 percent of arrested perpetrators of violent crime have toxicologic evidence of alcohol11US Department of Health and Human Services,National Institute of Alcohol Abuse and Alcoholism.Sixth special report on alcohol and health. NIAAA, Rockville, Md1987Google Scholar or illicit drugs25US Department of Justice, Office of Justice ProgramsDrug use forecasting, April to June, 1989. Government Printing Office, Washington, DC1989Google Scholar at the time of their arrest. However, since no information was available about drug and alcohol abuse of nonpatient drivers and nonpatient criminal perpetrators, their possible hospitalized victims could not be counted in the study. (2) Our method also did not allow measurement of comorbidity of abusers of multiple substances. In one report, for example, hospital admissions with a primary alcohol-related diagnosis were accompanied by illicit drug-related comorbidity 20.5 percent of the time.26Measuring comorbidity and overlap in the hospitalization cost for alcohol and drug abuse and mental illness.Inquiry. 1989; 26: 249-260PubMed Google Scholar Alcohol and illicit drug comorbidity is also commonly associated with trauma.19Budd RD Muto JJ Wong JK Drugs of abuse found in fatally injured drivers in Los Angeles County.Drug Alcohol Depend. 1989; 23: 153-158Abstract Full Text PDF PubMed Scopus (52) Google Scholar, 20Fortenberry JC Brown DB Shevlin LT Analysis of drug involvement in traffic fatalities in Alabama.Am J Drug Alcohol Abuse. 1986; 12: 257-267Crossref PubMed Scopus (16) Google Scholar, 21Marzuk P Tardiff K Leon A Stajic M Morgan E Mann J. Prevalence of recent cocaine use among motor vehicle fatalities in New York City.JAMA. 1990; 263: 250-256Crossref PubMed Scopus (73) Google Scholar, 22Lindenbaum GA Carroll SF Ierachmeil I Kapusnick R. Patterns of alcohol and drug abuse in an urban trauma center: the increasing role of cocaine abuse.J Trauma. 1989; 29: 1654-1658Crossref PubMed Scopus (85) Google Scholar, 23Sloan EP Zalenski RJ Smith RF Sheaff CM Chen EH Keys NI Toxicology screening in urban trauma patients: drug prevalence and its relationship to trauma severity and management.J Trauma. 1989; 29: 1647-1653Crossref PubMed Scopus (81) Google Scholar (3) Caution would be appropriate in extending these findings from one institution. Johns Hopkins Hospital, however, was considered a good location to uncover patterns of substance abuse-related admissions between different demographic groups, because it is a large (1,100-bed) university hospital that serves as a tertiary referral center, a clinical research center, an air-evacuation trauma center, core facility of a regional health maintenance organization, and as a community hospital. The JHH patient population, reflecting all these different hospital functions, is very diverse in age, sex, race, disease process, and insurance status, which was confirmed on enrollment as shown in Table 1. (4) The adult ICU was chosen for this investigation because of the diversity of disease, high morbidity, and high cost of this medical setting. Caution would also be advised, however, in extrapolating these findings to other medical settings, for example, adult hospital admissions to non-ICU wards. In summary, a large fraction, almost one third, of adult ICU admissions was related to the patient's substance abuse, substance abuse-related admissions were more lengthy and costly than other admissions, and substance abuse-related admissions were concentrated among patients without health insurance. Because of long latency of many of these diseases and the difficulty of changing substance abuse behavior (or even recognizing it6Moore RD Bone LR Geller G Mamon JA Stokes EJ Levine DM Prevalence, detection, and treatment of alcoholism in hospitalized patients.JAMA. 1989; 261: 403-407Crossref PubMed Scopus (488) Google Scholar), it is hard to predict the medical results of efforts to change the prevalence of substance abuse in the population. However, this report suggests the possibility of significant morbidity and cost savings resulting from such efforts. Future research should test the generalizability of these findings to other institutions and other medical settings. If confirmed, the findings may argue to public health officials for additional emphasis on before-the-fact approaches to substance abuse (counseling, taxes, law enforcement, etc) to potentially decrease the need for expensive after-the-fact interventions such as adult ICU care. We wish to thank Carol Lent and Paulette Webb for expert secretarial assistance, and Beverly Barnes and Betty Holthaus for assistance in collecting billing data.