Psychosocial and Biomedical Aspects of Deaths Associated with Heroin and Other Narcotics


Reliable and relevant data are scarce concerning the etiology of deaths due to psycho-active drugs. As a result, nationwide efforts to combat an apparently growing use and abuse of dangerous drugs have been seriously hampered.

To begin to obtain the kind of information needed to appreciate some epidemiological aspects of drug-associated deaths, the National Institute on Drug Abuse (NIDA), in collaboration with the Special Action Office of Drug Abuse Prevention (SAODAP), contracted with a research team from the University of California at Irvine. The goals of the resulting project were, broadly, threefold: (a) to develop and test a comprehensive form for recording information on psychoactive drug-associated deaths (); (b) to use this form to collect data on 2,000 cases from the medical examiners or coroners in nine major urban cities (); and (c) to get an estimate of the quality of toxicological investigations carried out in the laboratories of these nine urban reporting centers, with the long-term goal of exploring means of improving the uniformity and accuracy of such analytical determinations so that nationwide surveys in this area might rest on a more valid and consistent foundation (). The resulting UCI Reporting Form for Drug-Involved Deaths consists of about 135 items of inquiry in such data areas as biography, demography, on-site investigation, toxicology, post-mortem, treatment prior to death, and suicide.

For this conference, presentation is limited to: (a) a brief report of some psycho-social and biomedical characteristics of 551 heroin-associated and 302 other narcotic-associated deaths from the sample of 2,000; and (b) intercity differences in the quality of toxicological examinations.

Psychosocial and Biomedical Characteristics of Narcotic-Associated Deaths

Study Methodology

Cases included for study were those in which psychoactive drugs were involved as a primary, contributing, or indirect cause of death. Cases in which alcohol was combined with a psychoactive drug, and in which the combinations were considered by the medical examiners or coroners to play a significant role in death, were accepted.

Limitations of funds and other constraints led to a decision by the NIDA to limit the study to 2,000 cases from the jurisdiction of the medical examiners or coroners of nine specific cities. The cities included such heavily populated ones as New York, Chicago, Los Angeles, and Philadelphia, as well as the less populated areas of Miami, San Francisco, Cleveland, Washington, D.C., and Dallas.

Based on 1970 census figures, quotas were set for each jurisdiction on the basis of the ratio between its population and that of the others. Thus, if City A has twice the population of City B, City A would contribute approximately twice as many cases. Accordingly, cities with the larger populations were targeted to provide more cases for this study (for example, New York, 400, Chicago, 300; and Los Angeles, 300). Smaller cities were assigned smaller quotas, but slightly more than their percentage share in order to achieve a minimal number for purposes of statistical analysis (for example, Dallas, 150; Cleveland, 150; and Washington, D.C., 150). During data collection, certain cities had difficulties meeting these quotas, and some minor readjustments were made. Because of local problems the rate of data collection varied. Therefore, though the cases were chosen consecutively over weekly or monthly time periods, none were selected during some time periods for some cities. Hence, an extended period of time (1972-1974) was required to fill all quotas. However, these samples were considered by the medical examiners or coroners from each of the nine cities to be representative of all psychoactive drug-involved deaths over this time period.

The forms were filled out by personnel who were selected by each office but paid directly by the investigators and monitored by telephone and personal visits of research team members. Each person was paid a very adequate fee per form, and the motivation and talent applied to the task can be considered to have been very high.

When a report form was received from the office of the medical examiner or coroner it was scanned for obvious omissions, and when units of measurement different from those that were recommended were used, they were corrected or converted when possible. A computer program designed specifically to detect a variety of potential errors in filling out the forms was developed and utilized by the research team, and after processing each case was added to a data bank..

The nine cities are not representative of the entire nation, but probably constitute a fair picture of urban America. In combination, these data allow a meaningful analysis of within-city variables such as age and sex. In addition, percentages of such characteristics associated with drug abuse across cities are important and illustrate the regional variability of the drug abuse problem in the United States.

Study Findings

Heroin-Associated Deaths. Of the 2,000 cases, 551 or 28%, were listed as deaths involving heroin or morphine. Most medical examiners did not differentiate heroin from morphine deaths since morphine is the form in which heroin is found in the body. The medical examiners reported heroin as instrumental in the death when direct evidence was present, such as the presence of a syringe and cooker at the site of death, especially if the syringe was left in the body and contained traces of heroin. However, these nine offices were unanimous in the opinion that “practically all” of the cases listed as morphine were heroin-associated.

Role of Heroin in Death. Heroin may contribute to a death in a variety of ways. It may have a fatal outcome owing to accidental misuse, an inadvertent toxic combination with other chemical substances, or may be a deliberate instrument of suicide or homicide. Therefore, a schema for defining and categorizing the role of a drug in each of these types of death was devised. Its application to the study’s heroin deaths, presented in Table Role of Heroin in 2,000 Drug-Involved Deaths.

As noted, the role of heroin in these deaths was considered accidental or “unexpected” in about 60% of the cases. In 36% of the cases, heroin was specifically the cause of death with no other agent playing a significant role, while in approximately another 40%, heroin in combination with some other potentiating agent, such as alcohol or a barbiturate was the cause of death. It was definitely used as the instrument of suicide in four cases and homicide in an additional four. In 82, or 15%, of the cases it coexisted with homicide by other means (gunshot, stabbing, etc.); the latter finding presumably reflects the criminal environment with which heroin is associated.

Intercity Differences in Heroin-Associated Deaths. As shown in Table Percentage of Drug-Involved Deaths Listed as Heroin-Associated, by City, among the 2,000 deaths the larger cities had a higher percentage, 20-40%, of heroin-associated deaths in comparison with the 7-13% rate of the smaller cities of Cleveland, Dallas, and Miami. This is consistent with the fact that these smaller cities report a less serious heroin problem.

Table Percentage of Drug-Involved Deaths Listed as Heroin-Associated, by City


Total in sample

Heroin-associated deaths

% of sample total













Los Angeles








New York








San Francisco




Washington, D.C.








Demographic Differences in Heroin-Associated Deaths. The majority, 56%, of the heroin victims were in their twenties with an additional 20% in their thirties. Nearly 80% of them were male, and the majority, 57%, were black. Thirty percent of the victims were reported as white. Of the remaining ethnic groups, the Mexican-Americans were most represented, at 7%.

More than 50% of the victims were never married, probably reflecting the youth of the population. Somewhat surprisingly, 47% of the group were listed as employed at the time of their demise, and a relatively small percentage seems to have been on welfare. However, since about 32% of them were reported as unskilled in their main occupational pursuit, it appears that their level of employment was generally not very high.

Heroin, Methadone, and Other Narcotics Involved in Drug-Associated Deaths

Among the 2,000 psychoactive-drug involved deaths, there was a total of 853 narcotic deaths. Deaths due to heroin (including those due to morphine) totaled 64%, and another 32% of the deaths were associated with methadone. Codeine was involved in 12 (1.4%) of these deaths and Demerol in another 11 (1.3%). Dilaudid added only three cases.

Besides heroin, thus, the only narcotic of significant fatal involvement appeared to be methadone, which was associated with 276 cases or 32% of the 853. In these cases methadone was the direct and sole cause of death 40% of the time, and about 40% of the time it was combined with other drugs, such as alcohol or the barbiturates. Therefore, it may be said that while heroin is the Number One agent in narcotic-associated death, methadone is a strong second.

Most of these methadone-associated deaths were from New York and Washington, D.C. — 177 from New York and 62 from Washington. For New York this represented 70% of narcotic deaths in the data and for Washington, 60%. Miami, Philadelphia, and Cleveland were in the 20 to 30% range for deaths associated with methadone, while San Francisco, Los Angeles, and Chicago reported that only 0-3% of their narcotic-category deaths were methadonerelated. It is therefore quite apparent that large regional differences exist in the abuse of methadone.

Some of these differences are undoubtedly reflective of real geographic differences in drug availability and drug abuse habits. Another factor that may contribute to this variation may be the failure in some locales to look for or detect methadone in the body’s biological fluids. This issue is further discussed below. Our data suggest, in summary, that methadone deaths are, indeed, a significant problem, especially among Eastern cities, and that to some unknown extent the problem may be underestimated because of variations in detection and reporting procedures.

Suicide and Narcotics

An estimate was made of the probability that any of these 853 narcotic-involved deaths was a consequence of suicide. Only 17, or 2% were listed as “definitely suicide,” but 25% were thought “possible suicide,” and another 19% were considered “suspicious.” Thus, in nearly half of these narcotic-associated deaths, suicide was not entirely ruled out. Even though those people who commit suicide by use of barbiturates and similar drugs, as a group, tend not to be addicts or to have the socio-economic characteristics of the narcotic user, it is apparent that suicide is not to be dismissed as a consideration among narcotic users. There is ample evidence elsewhere that the suicide correlates of despair and depression exist among the narcotic-using population.

Narcotic-Associated Deaths by Race in Selected Cities

Whites exceeded blacks — 41% versus 30% — in narcotic deaths in Los Angeles, with Mexican-Americans at the level of 24%.

In New York this ratio was the opposite, for blacks exceeded whites in narcotic-involved deaths (55% versus 30%); Puerto Ricans added about 8% and Latin Americans another 6%.

In San Francisco, like Los Angeles, whites exceeded blacks (62% versus 30%), with Orientals producing 4% of the narcotic-associated deaths. In Chicago, narcotic deaths were confined almost entirely to whites and blacks — 29% versus 67% — with Puerto Ricans contributing an additional 3% to make up the entire sample.

Racial differences in psychoactive drug-involved deaths in different American cities are sometimes, but not always, correlated both with differences in local population and with the kinds of drugs locally used or abused (Gottschalk et al., 1976). Between the two major drugs, heroin and methadone, the differences between drug-involved death rates for whites and blacks were not very great, each race showing a split of about 60% heroin deaths and 313% methadone deaths. The 26 Puerto Ricans and Latin-Americans showed about a 50-50 split, while all of the 39 Mexican-American deaths were associated with heroin and no methadone or any other narcotic.

Intercity Differences in the Quality of Toxicological Examinations

A proficiency testing program was performed in collaboration with the nine major cities (). Since the appraisal of drug-associated deaths reported from coroners and medical examiners offices depends, along with other relevant data, on test results obtained from their toxicological laboratories, toxicological examinations are crucial. Differences in laboratory procedures, thoroughness of screening, and limits of detection could result in sizable differences in some details within mortality statistics from various reporting areas. Brief proficiency studies were, therefore, conducted to determine the differences in the quality of toxicological examinations performed by these toxicological laboratories.

Five standard “unknown” samples were sent to each city. Three samples consisted of drugs added to clean urine and two samples were drugs added to a 3% solution of human albumin. Some drugs were repeated at the same concentration in different samples. All five unknown-specimens were designed to contain 6.5 µg/ml of secobarbital to measure the variations in the secobarbital assay for a given laboratory over a period of time. The concentrations chosen for these unknown samples were at low toxic levels to provide a moderate challenge to the toxicological methods. Some samples were sent as complete unknowns, some with partial information (such as “Drugs in this vial are neutrals and volatiles” or “This vial contains morphine and methadone”), and some samples were sent with all drugs contained identified. This program was set up to simulate the actual situation prevailing in offices of medical examiners, where variable amounts of information are available for each case.

Since an agreement was made with each toxicological laboratory that the results for this proficiency testing would be strictly confidential with respect to what laboratory had obtained what results, all reports and findings were coded numerically, and these findings were not identified by city.

All participating laboratories were found to have adequate instrumentation and methodologies to quantitate the psychoactive drugs when they were known to be in the sample. Errors in quantitation, including both false negatives and false positives, were made much more frequently in the process of screening for the drugs than in quantitating them. Overall, a wide variance was found among the nine cities in detection, accuracy, and precision of toxicological analysis. This is likely to be a factor in the contribution of variations in the certification of the cause of psychoactive-drug-involved deaths in these and, by inference, other cities in the United States.

For example, one of the completely unknown samples contained morphine (3.5 µg/ml) and methadone (2.5 µg/ml). Only three of the nine participating toxicological laboratories found and quantitated the morphine, and an additional four laboratories detected morphine in this urine sample but did not quantitate it. Two laboratories did not detect the morphine present in this urine sample. With respect to methadone, the same three toxicological laboratories located and quantitated methadone in the urine sample that quantitated morphine in it, but only an additional two toxicological laboratories detected the methadone. Four laboratories did not even detect the methadone present in the sample. It was of interest to note, also, that although methadone was present in this urine sample in the concentration of 2.5 µg/ml, the range of concentration measured by the three toxicological laboratories that quantitated this chemical was from 2.8 to 10 µg/ml.

Clearly, the lack of uniformity of proficiency or quality control indicates the possibility of, at least, a modest error variance in national death statistics as reported by some cities and introduces some question into the relative accuracy of toxicological evaluations. One must not interpret these results as being overly critical. In a separate study by the UCI group, the same standard samples were sent for analysis to 19 laboratory members of a state association of toxicologists, a mixed group of forensic, commercial, and clinical laboratories. This group produced 33% false negatives, exactly that of the nine target cities ().


Our findings point to sizable intercity differences in the United States among certain psychosocial and biomedical aspects of deaths associated with narcotics. Secondly, narcotic-involved deaths are not purely accidental, but many are motivated by suicidal goals and a smaller percentage by homicidal intentions. And finally, in addition to the errors that have been surmised to occur in estimates of psychoactive drug deaths from heroin and other narcotics owing to inadequate reporting or other shortcomings in data collection, there are biomedical errors due to variations in the quality control of toxicological laboratories.


Selections from the book: “The Epidemiology of Heroin and Other Narcotics”. Joan Dunne Rittenhouse, Ph.D., editor. Task Force report on measurement of heroin-narcotic use, gaps in knowledge and how to address them, improved research technologies, and research implication. National Institute on Drug Abuse Research Monograph 16. November 1977.