Tobacco Documents Online

As can be seen in Table 1, the largest share of Medicare inpatient substance anusee costs--$12.5 billion or 97% of the total--was for medical treatment of illnesses and conditions attributable to the abuse of alcohol; drugs and tobacco. These are condit2ons that do not mention substance abuse in the diagnosis, but are ones that have been repeatedly shown in epidemiologic research to be associated with the use of at least one of these substances. In contrast, treatment for conditions that explicitly mention alcohol or drug abuse account for only 4% of all substance abuse-related discharges, consuming 3'% of those costs. Thi's low percentage of ' alcohol and drug diagnoses is misleading, and is probably more reflective of'& reluctance by physicians to classify the elderly population as alcohol or drug, dependent, than an ~ indication of a low prevalence of alcohol or drug use among the elderly and disabled. Underreporting, of substance abuse as either a primary or secondary condition for this. population is clearly a problern~ as depiicted in Graph 2. While 5% off all Medicare benef ciaries are considered heavy drinkers and 3% report using drugs (see Table 2), only 3% of Medicare patients in the hospital had' a diagnosis that mentioned either alcohol or drug, use or both. Since alcohol and drug users tend to be at' greater risk for medical care, we would expect them! to make up a larger, not smalller, proportion of the hospitalized population. In fact, separate studies measuring, alcoholism alone among the hospitalized elderly indicate that 9-20% actually have a drinking problemL"b This wide range in estimates of alcohol. problems suggest that identification of alcoholism varies considerably across physicians. Thus, relying solely on diagnoses that explicitly mentioni alcohol or drugs on the medical record in order to measure the prevalence and' cost of drugs and/or alcohol problems in hospitals grossly underestimates the full impact of substance abuse on Medicare costs.

TABLE 1: Substance Abuse Costs to Medicare Total Hospital Care, 1991 Discharges Expenditures 1. Direct Treatment for Alcohol 90,659 $319 million and Drug Primary Diagnoses 2. Treatment for Diseases Where 2,076,840 $12.487 billion .~ Substance Abuse Is a Major Risk Factor 3. Additional Days Required for Patients with ~ N/A $112 million w Cr ~ ~ a Seconda -ry Diagnosis of Substance Abuse Substance Abuse Total 2,167,499 $12.9 billion Total Medicare 11.1 million $57 billion Substance Abuse as Percent of Total 20% 23% SOURCES: National Hospital Discharge Survey, 1991; 1992 HCFA Statistics; National Association of Psychiatric Hospitals Annual Survey 1992. SVEV89szo%

Graph 2: Underreporting of Substance Abuse on Hospital Records Alcohol and Drugs Recorded on Medicare Records Compared to Actual Prevalence of Alcohol Alone as Reported in Studies of Heavy Drinking in Hospitalized Elderly Populations Alcohol and Drugs Reported on Record Estimates of Actual Heavy Alcohol Use Alone ., 3% , . ~ i ~ 79% Q Identified Substance Abuse ~ No Identified Substance Abuse National Hospital Discharge Survey, 1991 9% ~E~ES9szo7 . .

Table 2: Consumption Rates for Medicare and General Population Consumption Rate Substance User category Medicar- e General Cigarettes Current Smokers 18.4% 29.6% Former Smokers 36.1 % 23.3 % Alcohol Heavy Drinkers 4.8% V %O Illicit Drugs . Drug Users 3.2% 5% SOURCES: National Afedical Expenditures Survey, 1987; National Health Interview Survey, 1991. iw G. Z. Z.3%4%'07.

Drug use among the elderly also accounted for a very small: percentage of the total dischargEs, even though more than 3% of this population admit to: using drugs in the last year. This lbw number is, in part, indicative of a lack of research co nnecting, illicit drugs wsthh disease, and does not imply that drugs present no probllem for the elderly: Further, since we were also unable to find sufficient data from, the epidem2ologic literature to quantify the health effects of the abuse of prescription drugs, we couldI not estimate the extent of that problem either. Thus, our estimate of the impact of drugs--both legal and illiiciR--on, Medicare is undoubtedly low. There ! is clearly a need for more research to understand and quantify the impact of all' drugs on morbidity and cost. Since: the elderly are: suchl large users of prescription drugs, this research is even more critical wi th respect to, that population. In total, we found more than 60 conditions that are associated~ with substance abuse covering virtually every major disease category (Appendix II)`'. In, the Medicare population, more than half of the substance abuse-related hospital admissions were for cardiovascular diseases, 15% for respiratory diseases, 12% for neoplasms; andi 7°Xa, for burns and trauma (Table 3). These results differ somewhat from the earlier Medicaid study where the adverse impact on birth outcomes represented the major contributor to~ the costs attributable to the substance abuse. Further, in comparing substance abuse problems in the Ivledicaid' and Mediicare populations, the impact oni Medicare was much more a result of the long-term effects of smoking. More than 80% of substance abuse-related Medicare: hospital costs was for treating smoking-related medical conditions -- from lung cancer to chronic pulmonary `This number is lower than the 7?' substance-abuse related conditions identified in the Medicaid becausefurtheranadysis led ustocombine some specific d'iag~nosesialrto broader diagnostic categories. Appendix III provides an even more detailed breakdown of the substancee abuse-related discharges in all the conditions iidentified.

Table 3: Medicare Discharges for Diseases Attributable to Substance Abuse as a Major Risk Factor U.S. General Hospitals, 1991 [)isease/ Condition Attributable Discharges % of Total Attributable Discharges Cardiovascular Diseases 1,156,057 53.3 Respiratory Diseases 328,453 15.2 Neoplasms 249,683 11.5 "I'raunta/I3urns 149,649 6.9 ~ ~ Cerebrovascular Diseases 112,799 5.2 ~ ~ Digestive Diseases 49,798 2.3 Other 30,402 1.4 Direct Alcohol and Drug Diagnoses 90,659 4.2 TOTAL 1991 ATTRIBUTABLE MEDICARE DISCHARGES 2,167,500 1991 ASSOCIATED MEDICARE COSTS ATTRIBUTABLE $12.8 billion TO SUBSTANCE ABUSE AS A RISK FACTOR SOURCES: National Hospital Discharge S«rvey, 1991; CA_ SA Substance Abuse Epidemiologic Database, 1993. GeEes9Sz0z

obstruction, disease (COPD) to coronary artery disease; as opposed! to only 40"/o for the predominantly younger, Medicaid, population. The Medicare populati+oni is at a much higher risk for getting smoking-related diseases because people over 65 who have smoked tend to have done so more heavily and for ]bnger time periods. Nearly 3 out of 5 of current Medicare smokers (58%) and almost one third of former smokers (32'.7%0) smoked an average of more than cigarettes per day for over 35 years (Graph, 3). Substance Abuse as a Complicating Factor in the Treatment of'other Diseases When substance abuse is recordedi as a secondary diagnosis to, an otherwise unrelated condition, it tends to complicate and prolong the treattnent for the underlying, problem. On average„ a secondary diagnosis: of alcohol and/or dhng, abuse increased the length of time patients stay in the hospital. Compared to, Medicare patients with the same primary diagnoses, those with a secondary diagnosis of substance abuse stayed an average of more than a half a day longer, 9:3) days compared wi th 8.6 days. Whnle this is not insignificant, the marginal effect of substance abuse as a secondary diagnosis ini the Medicare population is much smaller than what was found for Med'2caid„ where substance abusers stayed twice as long as non-substance abusers.. This smallll differential betweeni length of stay for Medicare patients with and without substance abuse problems is most likely at signifiicant underestimate of the full effect of aleohol and drugs as a complication. Many cases that actually involved alcohol or drug problems weree noti recorded as having tihis secondary diagnosis. Since, in our analysis. t'hese cases would be counted in the non-substance abuse group, they may be artificiall} inflating the length of stay for that group. If the secondary diagnosis of substance abuse had been correctly noted, the 7

. r Graph 3: Portion of Medicare Smokers Who Have Smoked Heavily (Those Who Ilave Smoked More Than 10 Cigarettes per Day for Over 35 Years) Current Hecz vy 58 Other 42 Former ZVCES9szoz

difference between the two. : groups would have been greater. Nevertheless; despite this problem, complications resulting from a secondary diagnosis of substance abuse accounted for: $108 million in added cost' to Mediicare: These figures represent average lengths of stay in~ Iwledicare but, for some diseases, the difference in length of stay was much higher. For example„ for patients with pancreatitis, those requiring major joint and limb reattachments, and those wirth pathol+ogicall fractures with a, secondary dlagnosis of substance abuse, patients stayed approximately twice as long as theirr non-substanee abusing, counterparts (pancreatitis 17 days vs. 8 days; major joint 24 vs. 10; pathological fractures 17 vs. 10) (Graph 4). However,, for other diseases, the AhAS was, in fact, shorter for some Medicare patients with a secondary diagnosis of substance abuse: For example, patients with respiratory infections and utflflammations„ cellulitis, and G,T obstructiion, who also had a secondary diagnosis of substance abuse, stayed less time in the hospital than non-substance abusers (e:g., respiratory infections 8.4 days vs. 11.7 days, cellulitis 7.4 vs. 816, and GI obstruction 3.7 vs. 7.4) This does nort rnean that patients with, substance abuse needed less care. There : are several possiblie explanations for why these substance abusers had shorter lengths of stay. It may result from an aberration in the data due to the small sample size of patients within these diagnoses, or because of the premature departure caused by some patients signing out against medical advice: However, it may also reflect the f nancial or social undesirability to hospitals of many of these patients which, in turn, may lead to their early discharge or transfer to another faciilit~?: In this context, it is worth noting, that, if this hypothesis is true, as the forces of competition in health~ careintensify., these resulkssugeest that such "undesirable"' pa2ientss maN, . be increasingly pushed out prematurely from some institutions. Further: the nature of the 8

Graph 4: Conditions Where Medicare Patients w/ Substance Abuse Stay Longer Average Length of Stay for Medicare Patients with and without Substance Abuse Pancreatitis Major Joint Procedures Fractures 5 10 15 20 25 Average Length of Stay (Days) SOURCE: National Hospital Discharge Survey, 1991. Ow/o Substance Abuse Ew/ Substance Abuse CVCeN9szoz