ESTIMATION OF TOBACCO- OR ALCOHOL-ATTRIBUTABLE DISEASE RATES IN NATIONAL HOSPITAL CARE: AN APPROACH BASED ON ROUTINE IN-PATIENT DISEASE REGISTER DATA AND SYSTEMATIC DIAGNOSIS OF ALCOHOL USE DISORDERS

Ulrich John*, Hans-Jürgen Rumpf1, Monika Hanke, Peter Gerke2 and Ulfert Hapke

University of Greifswald, Institute of Epidemiology and Social Medicine, D-17487 Greifswald,
1 University of Lübeck, Department of Psychiatry and Psychotherapy, Lübeck and
2 University of Freiburg, Medical Department, Freiburg, Germany

Received 21 October 2002; in revised form 7 February 2003; accepted 12 March 2003


    ABSTRACT
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Aims: The goal of this paper was to estimate and compare the numbers and rates of tobacco- or alcohol-attributable disease (TAAD) in in-patient-treated cases in a high tobacco smoking and alcohol consumption country, based on different estimates. Methods: Two samples, three TAAD estimates, and tobacco- or alcohol-attributable fractions were used. Sample 1 included all disease cases aged 25–64 years and treated more than 24 h as in-patients during the year 1997 (n = 7 344 079) in the hospitals in Germany. Sample 2 included all in-patients aged 25–64 years (n = 1136) consecutively admitted to one general hospital. The first estimate of the TAAD was the routine main diagnosis based on the treating physician’s report to the in-patient disease register (IDR) in sample 1. The second estimate included up to three routine treatment diagnoses in sample 2, and the third estimate a diagnosis of alcohol dependence or misuse according to DSM-III-R or ICD-10, as well as harmful or hazardous alcohol consumption, in sample 2. The tobacco- and alcohol-attributable fractions were calculated based on the method for the estimation of tobacco- and alcohol-attributable mortality, originally provided for the Centers for Disease Control in the USA. Results: When the three estimates were combined, a total of 37.8% of all in-patient treatment cases had at least one diagnosis that was attributable in part or fully to tobacco smoking, alcohol dependence, alcohol misuse, or harmful or hazardous alcohol drinking. When the tobacco- and the alcohol-attributable fractions were considered, of all treatment cases, 10.5% could be revealed as attributable to smoking or alcohol consumption according to the one main diagnosis based on the IDR. When all three estimates were combined, the rate was 30.2%. This corresponded to 32.2% of the national in-patient hospital care costs. Conclusions: The TAAD rate is underestimated when using one routine diagnosis alone. Additional alcohol misuse or dependence diagnoses are needed, which may be obtained with a reasonable level of resources in a sample of hospitals. TAAD rates may be used for the planning and practice of brief intervention and as an outcome measure for population-based intervention.


    INTRODUCTION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Providing estimates of tobacco- or alcohol-attributable disease (TAAD) for nations is a goal of high priority, especially for the planning of brief intervention and for the estimation of treatment costs (cf. Moyer et al., 2002Go). In particular, we lack nationally representative data about the prevalence of TAAD. Little evidence is available so far, and this may be due in part to the considerable effort needed to diagnose alcohol misuse and dependence in routine medical care.

More than 90 causes of death have been attributed to tobacco smoking (Shultz et al., 1991aGo) and more than 30 to alcohol consumption (Shultz et al., 1991bGo; Corrao et al., 1999Go). The tobacco-attributable causes of death have been derived from the US Cancer Prevention Study II, in which over 1.2 million persons from the adult population were enrolled. The individuals gave information about their smoking status at baseline in 1988, and up to 6 years later the fact of death was ascertained (Garfinkel and Heath, 1992Go). Relative risks of death found in smokers for single causes of death based on this limited time period have been corroborated by another prospective study, which covered 40 years of life of the individuals enrolled (Doll, 1998Go). In both investigations, the risk of death was analysed in relation to smoking status, and relative risk estimates were calculated for smokers and for single causes of death (US Department of Health and Human Services, 1989Go). Several studies have reported the co-morbidity between diseases attributable to the two substance use behaviours (for examples cf. Gulliver et al., 1995Go; Burton and Tiffany, 1997Go; Castellsagué et al., 1999Go; Daeppen et al., 2000Go).

There have been several population-related studies about the number of tobacco-attributable (Peto et al., 1994Go; McGinnis and Foege, 1999Go; Single et al., 1999Go; Thun et al., 2000aGo,bGo; Ridolfo and Stevenson, 2001Go) and about alcohol-attributable (Holman et al., 1996Go; Her and Rehm, 1998Go; Corrao et al., 1999Go; Chikritzhs et al., 2001Go; Rehm et al., 2001Go) death cases; however, only a few population-based studies focused on the synergy between tobacco smoking and hazardous alcohol drinking (cf. John and Hanke, 2003Go). For Germany for example, 19.5% (167 845 death cases) of the total mortality (age: –1 year, 35 years or above) has been reported to be tobacco or alcohol attributable (John and Hanke, 2002aGo). This kind of evidence is based on comparisons between non-smokers or non-risky drinkers and smokers or hazardous drinkers in the population regarding the relative risk of death from a defined disease in a given time frame. This information is used for calculating the tobacco-attributable fraction (TAF) and alcohol-attributable fraction (AAF) of a disease, which are expressed as a rate per cause of death, in which 0 stands for no attributable fraction and 1 for completely attributable to tobacco or alcohol (Shultz et al., 1991aGo,bGo; Ridolfo and Stevenson, 2001Go).

To our knowledge, no studies on representative estimates of in-patient TAAD cases are available as yet. It has been reported by one general hospital that 68.5% of current alcohol dependents detected were treated as in-patient for alcohol-attributable disease (Gerke et al., 1997Go); this was the case for only 3.2% of patients who did not show any alcohol use disorders (AUD). AUD include alcohol dependence or misuse, as defined by the Diagnostic and Statistical Manual, 3rd edition, revised (DSM-III-R), of the American Psychiatric Association (1987)Go, or the International Classification of Diseases (ICD-10) (Gerke et al., 1997Go). Among 436 general hospital in-patients who were identified as alcohol dependent, misusers or harmful alcohol consumers by a standardized diagnostic interview (the alcohol section of the Schedules for Clinical Assessment in Neuropsychiatry, SCAN; WHO, 1992Go), only 48.5% were detected as such according to routine diagnoses in the patient records (Rumpf et al., 2001Go). In accordance with the scarcity of evidence, there are only few data available on calculating hospital costs due to TAAD. For Germany, an estimate based on TAF and hospital days for the year 1993 showed costs of ¤2163 million for hospital treatment, which represented 45.7% of the total estimated smoking-attributable direct costs (¤4737 million for in-patient and out-patient treatment, prescribed drugs, and rehabilitation) (Welte et al., 2000Go).

In principle, there are five sources of data that may contribute to the estimate of the burden of disease caused by tobacco smoking or risky alcohol drinking on in-patient care in a nation: population survey, sales, mortality and morbidity statistics, and single studies focusing on TAAD. All these approaches have their own bias. Population survey data may suffer from under-reporting caused by denial or recall error, as well as sample selection bias, particularly by non-participation of subjects with hazardous or harmful consumption or AUD. Sales data provide the opportunity for a crude approximation of the burden on health, but cannot provide information about the time lag between the selling of tobacco or alcohol and the manifestation of a disease. Mortality and morbidity statistics allow indirect estimates that come closer to the true prevalence; however, AUD are underestimated (Gerke et al., 1997Go). Single investigations consume a lot of resources when examining a wide range of TAAD. This approach seems to be adequate only for a subsample of, and not all, medical care facilities of a nation. It may be assumed that each single data source will be able to cover only part of the information needed for the estimation of the burden of TAAD in hospital care. Bias may particularly be expected when a single approach to estimation is used. A combination of approaches might help to minimize this bias. The goal of the present paper was therefore to estimate and compare TAAD rates among hospital patients for one country based on four steps, two samples, three estimates, and TAF and AAF (Fig. 1Go): (1) the estimation of the TAAD rates based on routine national in-patient disease register (IDR) data; (2) the estimation of the TAAD rates on grounds of up to three routine diagnoses in one hospital; (3) the diagnosing of AUD and hazardous as well as harmful alcohol use among in-patients in one hospital; and (4) combining the data for the provision of a TAAD estimate for the nation. The TAAD cases were calculated without considering the TAF and the AAF. These results can be used for planning intervention resources, since brief intervention is needed for each TAAD case, irrespective of how large the TAF or AAF is. In an additional analysis, the TAF and the AAF were considered. These results are needed in order to calculate the burden of TAAD in in-patient care in terms of treatment days and costs.



View larger version (18K):
[in this window]
[in a new window]
 
Fig. 1. Rates of tobacco- or alcohol-attributable disease (TAAD) cases based on three different estimates (attributable fraction considered). aIDR, in-patient disease register; bAUD, alcohol use disorders.

 

    SUBJECTS AND METHODS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Samples
Sample 1 included all hospital in-patient treatment cases with a stay of more than 24 h in Germany in 1997. Rehabilitation was not included. Germany has high levels of tobacco smoking and alcohol consumption (Corrao et al., 2000Go). There were 2244 hospitals in 1997 (Federal Statistics Office, 1999Go), of which 2010 were general, 196 psychiatric and 38 other specialized hospitals. Considering only general hospitals for the data analysis was not possible. A total of 16 425 233 cases were eligible, and 16 388 614 (99.8%) had been included in the data file. Of these, 930 481 were discharged or died during the first 24 h in hospital, and in 632 cases data were missing. Of the remaining 15 457 501 cases (94.3% of all cases in the data file), 7 762 819 were aged 25–64 years; of these, 418 740 were no disease cases (uncomplicated delivery), leaving 7 344 079 cases that constituted the final sample 1. There were 3 754 795 females (51.1%) and 3 589 284 males (48.9%).

Sample 2 included all consecutive admissions to a municipal hospital over a time period of 6 months. The local ethics commission gave consent for the study (John et al., 1999Go; Rumpf et al., 1999Go). The hospital consisted of two departments (internal medicine and surgery). Patients aged 18–64 years and staying for at least 24 h were asked to take part in the study. Those readmitted during the data collection period and those from a specialized oncology ward, on which alcohol misuse or dependence was one cause of rejecting patients, were excluded. Thus, a total of 1733 patients were eligible for the study. Of these, 62 were too ill to participate, 61 were foreigners unable to speak German sufficiently, 128 were discharged before data about alcohol consumption were complete, 122 refused to participate, and 72 could not be investigated due to different reasons of treatment or illness. There were 1288 in-patients with TAAD data. Of these, 1136 were aged 25–64 years: these constituted the final sample 2. Sample 2 contained 484 (42.6%) females and 652 (57.4%) males.

Estimates
The TAAD rates were estimated for both samples on the grounds of routine treatment diagnoses and, in part, based on the diagnosis of AUD in sample 2.

Estimate 1: national IDR data. The IDR provided the main treatment diagnosis according to ICD-9 (three digits), number of treatment days, age and gender for the year 1997 for each in-patient (Federal Statistics Office, 1999Go). The provision of the main diagnosis at discharge, the dates of admission and discharge, age and gender to the Federal Statistical Office is mandatory for each hospital in Germany. The IDR included the possibility for the treating physician to enter a four-digit diagnosis voluntarily. Among the eight three-digit diagnosis groups in which the four-digit codes were necessary for the calculation of the TAF or AAF, these four-digit diagnoses were provided in 59.1–72.4% of the cases concerned. The total numbers of the four-digit diagnoses were estimated based on a 10% random sample drawn, for reasons of data security, from all three-digit diagnosis cases within which the four-digit diagnoses occurred. Only data about treatment cases aggregated per diagnosis could be obtained from the Federal Statistical Office. Therefore, the data may include cases that have been admitted more than once with the same diagnosis during the year.

Estimate 2: three routine treatment diagnoses. In sample 2, up to three routine treatment diagnoses according to ICD-9 could be provided by the treating physician on the third day after admission. Of all 1136 patients, 657 (57.8%) received one, 397 (26.1%) received two and 182 (16.0%) received three diagnoses.

Estimate 3: diagnosis of AUD. In sample 2, AUD were diagnosed by screening and a standardized diagnosis (Rumpf et al., 1999Go). Every patient answered the CAGE (Ewing, 1984Go) and the Michigan Alcoholism Screening Test (MAST; Selzer, 1971Go). If a patient screened positive in one of the questionnaires, had a withdrawal medication, or if there was information about alcohol problems in the charts or from the treating physician, an interview was conducted face to face by research staff with the alcohol section of the SCAN (WHO, 1992Go), which provided a diagnosis according to DSM-III-R. The SCAN was chosen on the grounds of its feasibility for experienced clinicians in clinical settings. As part of this investigation, alcohol consumption was assessed within the interview using quantity–frequency questions. Hazardous alcohol consumption was defined as drinking the amount of 20 to 40 g of pure alcohol per day in females, 30 to 60 g in males, harmful alcohol consumption as more than 40 g in females and more than 60 g in males (British Medical Association, 1995Go).

Attributable fractions
For the calculation of the TAF, the available data on the disease-specific relative mortality risk-rates of current and former tobacco smokers, each compared with never smokers (John and Hanke, 2002aGo), were calculated according to the formula provided by Shultz et al. (1991a)Go for the Centers for Disease Control in the USA. The mortality rates were taken on the grounds of data of the cancer prevention studies carried out in the USA (US Department of Health and Human Services, 1993Go). This database was used because a sufficiently large sample and a large number of death cases are needed for mortality risk estimates. No such data exist for Germany. The smoking rates were estimated on the grounds of data from the largest population survey in Germany (micro-census 1995) carried out on a regular basis. The micro-census 1995 provided the most recent data available for scientific use. The micro-census is a 1% household and individual person survey representative for Germany (Schimpl-Neimanns, 1998Go). It is mandatory for every German citizen to answer the questions. Half of the individuals selected at random were asked about their smoking behaviour, on a voluntary basis, face to face (for details see John and Hanke, 2002aGo). A 70% random sample data set of all respondents was available. The sample of the voluntary inquiry included 257 470 individuals, of which 146 029 were aged 25–64 years. Of these, 131 604 (90.1%) answered the questions about smoking. Never smokers in our data are individuals who denied smoking at that time or ever. Former smokers are individuals who said that they did not smoke currently but had smoked in the past.

The AAF was estimated in a similar way. It includes previously published alcohol-attributable mortality rates (Single et al., 1999Go; WHO, 2000Go; Ridolfo and Stevenson, 2001Go). From Corrao et al. (1999)Go the relative risks for 25/50/100 g of pure alcohol per day were used. Data about hazardous and harmful alcohol drinking were taken from the German National Health Survey carried out in 1991 (n = 6949 aged 25–64 years; Stolzenberg, 1995Go; calculations of grams of pure alcohol according to Meyer et al., 1998Go). If the cause of death was attributable to tobacco smoking as well as to alcohol consumption, we used the higher attributable fraction, even when results showed lives saved by alcohol consumption (John and Hanke, 2002aGo).

The methods developed for tobacco- or alcohol-attributable mortality rates include the attributable fractions multiplied by the TAAD. The result was the number of TAAD cases, attributable fractions considered. There were 61 three-digit diagnoses plus three four-digit diagnoses which constituted the tobacco-attributable diagnoses (mainly cancer, and heart, cardio-vascular and respiratory diseases), 29 three-digit diagnoses plus one four-digit diagnosis constituted the tobacco- and alcohol-attributable diagnoses (predominantly cancers, hypertension, ischaemic heart disease and cerebrovascular disease), and six three-digit diagnoses plus 14 four-digit diagnoses were alcohol attributable (predominantly liver cancer, alcohol-related psychiatric diseases, diseases of the lower digestive tract). Traffic accidents were not included in the IDR.

We distinguished four diagnosis groups: (1) tobacco-attributable diagnoses are all those for which evidence shows a TAF, but not an AAF, of >0; (2) tobacco- and alcohol-attributable diagnoses are all those for which evidence shows a TAF > 0 and an AAF deviant from 0; (3) alcohol-attributable diagnoses are all those for which evidence shows an AAF, but not a TAF, of >0; and (4) neither tobacco- nor alcohol-attributable are all those diagnoses which show a TAF and an AAF of 0. In-patient hospital treatment costs were calculated on the grounds of the mean costs for one in-patient treatment day. In 1997, these were ¤292.23 (Federal Statistics Office, 2000Go).

Data analysis
The analysis was carried out in four steps, using the two samples, the three estimates and the attributable fractions (Table 1Go). The first step was to calculate the TAAD for the nation based on sample 1 and estimate 1 (one routine diagnosis according to the IDR); attributable fractions were left unconsidered in one, but were considered in another, analysis. The second step included the calculation of the TAAD based on sample 2 and estimate 2 (three routine treatment diagnoses); the attributable fractions were not considered. The third step was to include the AUD in the estimate of the TAAD, for which we used sample 2. We made a cross-tabulation of AUD by the number of tobacco- or alcohol-attributable diagnoses and, after that, displayed the TAAD cases detected by the first routine diagnosis, by the three routine diagnoses and by the three routine diagnoses plus the AUD. The fourth step was the calculation of the TAAD based on the first, the first plus the second, and all three estimates; the attributable fractions were unconsidered in one analysis and considered in another.


View this table:
[in this window]
[in a new window]
 
Table 1. Steps of data analysis, samples, estimates and attributable fraction (AF)
 
For estimating the TAAD based on all three measures and both samples, we first calculated the TAAD rate based on the IDR. Secondly, on grounds of the diagnoses in sample 2, we divided the cases detected additionally on the grounds of estimate 2 by the cases based on the IDR alone, and added the additional cases as percentage to those identified by the IDR alone. Thirdly, we divided the AUD cases detected additionally to those on the grounds of estimate 2 and added them as percentage to the ones found by estimate 2. The data were analysed by SPSS 11.0 and Microsoft Excel 2000.


    RESULTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
In the first step of data analysis, using sample 1, of all TAAD cases, attributable fractions unconsidered, the TAAD rate was 22.6%. When the attributable fraction was considered, the IDR data revealed a TAAD rate of 10.5% of all, 4.9% of the female and 16.4% of the male cases aged 25–64 years (Table 2Go). Of all cases, 4.0% were attributable to tobacco smoking as well as alcohol drinking. In the second step of data analysis, in sample 2, among the 1136 patients, 27.0% were detected as TAAD cases by one, a further 8.0% by two and a further 2.7% by three diagnoses, amounting to 38.8% of all cases (Table 3Go). The tobacco-attributable, the tobacco- and alcohol-attributable and the alcohol-attributable diagnoses each contributed between 12.4 and 15.5% to the TAAD.


View this table:
[in this window]
[in a new window]
 
Table 2. Tobacco- or alcohol-attributable disease (TAAD) based on sample 1 and estimate 1
 

View this table:
[in this window]
[in a new window]
 
Table 3. Number of tobacco- or alcohol-attributable diagnoses (TAAD) based on sample 2 and estimate 2
 
The third step of data analysis shows that there were 12.9% current and 2.6% remitted alcohol dependents, 4.2% alcohol misusers in sample 2, and 80.3% had no AUD; however, among these, there were 27 harmful and another 72 hazardous alcohol consumers (Table 4Go). The rate of AUD cases increased from 13.6% among those who had no TAAD diagnosis to 38.5% (plus 2.5% harmful and 4.9% hazardous consumers) among those with two or three TAAD diagnoses.


View this table:
[in this window]
[in a new window]
 
Table 4. Alcohol use disorders and number of tobacco- or alcohol-attributable diagnoses (TAAD) based on sample 2 and estimate 3
 
Among the 1136 patients in sample 2, attributable fractions unconsidered, there were 27.8% detected by the first treatment diagnosis, an additional 10.0% detected by giving the physicians the opportunity to declare three treatment diagnoses instead of one, and an additional 8.4% of all cases were detected by the AUD diagnosis, amounting to 46.2% of sample 2 (Table 5Go). When the attributable fractions of the single diseases were considered and when all three approaches of detecting TAAD were used, the TAAD rate was 33.3%. It increased by the estimate from 11.6% (first TAAD diagnosis) to 24.9% (three diagnoses), and to 33.3% (three diagnoses + AUD).


View this table:
[in this window]
[in a new window]
 
Table 5. Cases detected by three estimates based on sample 2 and estimate 3
 
In the fourth step of data analysis, the IDR data, the three diagnoses and the AUD diagnoses in the one general hospital were used for the calculation of the number of TAAD cases. Among male cases 53.0%, and of all in-patient cases 37.8%, were expected as TAAD cases, attributable fractions unconsidered (Table 6Go). When considering the attributable fractions, the two additional approaches of detecting TAAD cases even led to a larger gain. While 10.5% of all treatment cases in the nation were attributable to smoking or drinking based on IDR alone, this was true for 22.6% when adding the detection by three diagnoses, and 30.2% of all cases were tobacco- or alcohol-attributable when using the three approaches together.


View this table:
[in this window]
[in a new window]
 
Table 6. Number of tobacco- or alcohol-attributable disease (TAAD) diagnoses based on samples 1 and 2, and estimate 4
 
The female TAAD cases showed a mean of 11.1 treatment days, compared with 10.2 days in female non-TAAD cases. Attributable fractions considered and based on estimate 3, which revealed 480 020 TAAD cases, 11.1 days per case and ¤292.23 per treatment day, the costs for the female TAAD in-patient treatment in 1997 were ¤1 557 066 315. Male TAAD cases were treated 11.9 days in the mean compared with 11.4 days in male non-TAAD cases. Based on estimate 3 (1 734 980 TAAD cases, 11.9 days per case and ¤292.23 per treatment day) the costs for the male in-patient TAAD treatment cases amounted to ¤6 033 457 144. In total, the in-patient TAAD treatment costs were ¤7 590 523 459 for cases in the age range 25–64 years. This represents 32.2% of the total in-patient treatment costs of ¤23 537 539 691 in 1997 for cases aged 25–64 years old.


    DISCUSSION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The routine main hospital diagnosis underestimates the TAAD rate. The inclusion of three treatment diagnoses more than doubles the TAAD rate, and the further systematic diagnosis of AUD leads to nearly a 3-fold increase in the rate of detection of TAAD cases, compared with the IDR alone. The three routine treatment diagnoses plus the detection of AUD provide a base for the estimate of TAAD among hospital in-patients. The strength of such an estimate is that it combines the representative routine diagnostic data with the non-representative, but highly valid, measure of AUD, and is thus cost-saving. For the purposes of monitoring, a promising approach could be to draw representative samples of hospitals, diagnose AUD and tobacco smoking with greater scrutiny and estimate the TAAD on those grounds, coupled with national routine diagnostic data.

The results show that 15.0% of sample 2 had TAAD attributable to tobacco smoking plus alcohol drinking. This rate corresponds to the rates of disease cases that are tobacco- or alcohol-attributable alone. Thus, the data corroborate the approach to analyse both substance use behaviours in conjunction. It helps to minimize, if not prevent, overestimations that may occur when tobacco- and alcohol-attributable disease rates are presented separately. The combined analysis of both substance use behaviours corresponds to the co-morbidity and the interactions of several factors inherent in the use of both substances (Gulliver et al., 1995Go; Burton and Tiffany, 1997Go; Castellsagué et al., 1999Go; Daeppen et al., 2000Go).

The female as well as the male TAAD rates, attributable fractions considered and based on the first diagnosis in sample 2, are consistent with the TAAD rates in sample 1. Thus, it seems justified to use the measures from sample 2 for a provisional estimate at the national level. The TAAD rates in females, compared with males, seem plausible in the light of knowledge of differences in prevalence rates of AUD between females and males. Based on all three estimates, males revealed roughly the 3-fold TAAD rate compared with females, when attributable fractions were considered.

The estimate of 2.2 million TAAD cases or 30.2% of all in-patient treatment cases in the age range 25–64 years seems plausible in the light of the national tobacco- or alcohol-attributable mortality of 167 845 tobacco- or alcohol-attributable mortality cases or 19.5% of all mortality cases at age 35 years or above (John and Hanke, 2002aGo). A large proportion of the in-patients with attributable disease probably die later of tobacco- or alcohol-attributable causes. The in-patient treatment costs caused by TAAD came to 32.2% of the national in-patient treatment costs for cases aged 25–64 years. This finding demonstrates the huge impact of tobacco smoking and hazardous or harmful alcohol use or AUD on in-patient care. The results provide a base for the calculation of resources for reducing TAAD. One practical approach for this, among others, is brief intervention in medical settings (Moyer et al., 2002Go). The TAAD rate may be used to work on denial or precontemplation in patients who are not ready to admit the detrimental effects of smoking, risky alcohol drinking or AUD, and the reference to the TAAD might help to further abstain from or not to take up smoking or hazardous alcohol drinking.

Limitations of our study are that: (1) the national IDR included cases only, not patients. We could not identify patients with more than one hospital stay during the year of the investigation. Thus, the TAAD rate may decrease when looking at patients instead of cases. However, the planning of brief intervention as well as the burden of TAAD for health care are determined by the treatment case, not the patient. (2) In sample 1, there were four-digit diagnoses available only in hospitals that provided these voluntarily. (3) We only analysed one hospital according to AUD. This is not representative, even for one area. (4) If the opportunity is given at the national level to provide three diagnoses, there may be variation by hospital in preferences for making more or less use of it. However, the variation should be random due to the large number of hospitals. (5) Of all hospitals, 8.7% were psychiatric in-patient facilities, which usually have a particularly high AUD patient load (Wienberg et al., 1993Go). Thus, our results are likely to be inadequate when only specialized hospitals are in focus. (6) The age range 25–64 years in both samples is small. This may have caused an underestimation of tobacco-attributable diseases, since a considerable part may become apparent and lead to death at the age of 65 years or above. This seems to be particularly true for tobacco-attributable diseases, many of which are cancers, while on the other hand, it has been found that among alcohol-attributable death cases, 74.6% occurred before the age of 65 years (John and Hanke, 2002bGo). (7) Smoking status and nicotine dependence were not included in the systematic diagnosis in sample 2. However, we did not expect this to change the results substantially, because there is no TAF of 1, as is the case in AUD. (8) For the AUD diagnoses in sample 2, we decided to rely on ICD-10 or DSM-III-R, because the diagnoses are more precise than in ICD-9. (9) For the calculation of the TAF for the TAAD in sample 2, we had to use smoker rates from the general population, although evidence shows that smoker rates among general hospital patients must be expected to be higher than in general population samples (John et al., 2003Go). (10) The rates of alcohol consumption estimated by survey data are probably an underestimate because of under-reporting. This, however, is not the case for tobacco smoking (Vartiainen et al., 2002Go).

Altogether, the results allow us to conclude that: (1) a routine treatment diagnosis does not suffice to estimate the TAAD in routine in-patient hospital care; and (2) monitoring of TAAD with, for example, annual reports about its prevalence, including combined estimates on grounds of routine diagnoses plus systematic AUD diagnoses from hospital samples stratified by region, should be established. We propose the planning of brief intervention and the provision of an outcome measure for population-based interventions, with the goal of reducing the burden of tobacco smoking and alcohol consumption on health. Such a monitoring system may be set up at manageable costs, particularly when the potential of saving treatment costs by preventive efforts is taken into consideration.


    ACKNOWLEDGEMENTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The work reported in this paper was funded by the German Federal Ministry of Health (grant no. 326-4914-8/38) and the Social Ministry of the Federal State of Mecklenburg-West Pomerania (grant: gesundheitsziele).


    FOOTNOTES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
* Author to whom correspondence should be addressed. Back


    REFERENCES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
American Psychiatric Association (1987) Diagnostic and Statistical Manual of Mental Disorders, 3rd edn, revised. American Psychiatric Association, Washington, DC.

British Medical Association (1995) Guidelines on Sensible Drinking. British Medical Association, London.

Burton, S. M. and Tiffany, S. T. (1997) The effect of alcohol consumption on craving to smoke. Addiction 92, 15–26.[CrossRef][ISI][Medline]

Castellsagué, X., Munoz, N., De Stefani, E., Victora, C. G., Castelletto, R., Rolon, P. A. et al. (1999) Independent and joint effects of tobacco smoking and alcohol drinking on the risk of esophageal cancer in men and women. International Journal of Cancer 82, 657–664.[CrossRef][ISI]

Chikritzhs, T. N., Jonas, H. A., Stockwell, T. R., Heale, P. F. and Dietze, P. M. (2001) Mortality and life-years lost due to alcohol: a comparison of acute and chronic causes. Medical Journal of Australia 174, 281–284.[ISI][Medline]

Corrao, G., Bagnardi, V., Zambon, A. and Arico, S. (1999) Exploring the dose–response relationship between alcohol consumption and the risk of several alcohol-related conditions: a meta-analysis. Addiction 94, 1551–1573.[CrossRef][ISI][Medline]

Corrao, M. A., Guindon, G. E., Sharma, N. and Shokoohi, D. F. eds (2000) Tobacco Control Country Profiles. American Cancer Society, Atlanta, GA.

Daeppen, J. B., Smith, T. L., Danko, G. P., Gordon, L., Landi, N. A., Nurnberger, J. I., Jr et al. (2000) Clinical correlates of cigarette smoking and nicotine dependence in alcohol-dependent men and women. The Collaborative Study Group on the Genetics of Alcoholism. Alcohol and Alcoholism 35, 171–175.[Abstract/Free Full Text]

Doll, R. (1998) Uncovering the effects of smoking: historical perspective. Statistical Methods in Medical Research 7, 87–117.[Medline]

Ewing, J. A. (1984) Detecting alcoholism. The CAGE questionnaire. Journal of the American Medical Association 252, 1905–1907.[Abstract]

Federal Statistics Office (1999) Gesundheitswesen, Fachserie 12, Reihe 6.2, Diagnosedaten der Krankenhauspatienten [Diagnoses of Hospital Patients]. Metzler-Poeschel, Stuttgart.

Federal Statistics Office (2000) Statistisches Jahrbuch 2000 für die Bundesrepublik Deutschland [Statistical Yearbook for the Federal Republic of Germany]. Metzler-Poeschel, Stuttgart.

Garfinkel, L. and Heath, C. W. (1992) Cancer Prevention Study II. The American Cancer Society Prospective Study. Statistical Bulletin 73, 21–29.[Medline]

Gerke, P., Hapke, U., Rumpf, H. J. and John, U. (1997) Alcohol-related diseases in general hospital patients. Alcohol and Alcoholism 32, 179–184.[Abstract]

Gulliver, S. B., Rohsenow, D. J., Colby, S. M., Dey, A. N., Abrams, D. B., Niaura, R. S. et al. (1995) Interrelationship of smoking and alcohol dependence, use and urges to use. Journal of Studies on Alcohol 56, 202–206.[ISI][Medline]

Her, M. and Rehm, J. (1998) Alcohol and all-cause mortality in Europe 1982–1990: a pooled cross-section time-series analysis. Addiction 93, 1335–1340.[CrossRef][ISI][Medline]

Holman, C. D., English, D. R., Milne, E. and Winter, M. G. (1996) Meta-analysis of alcohol and all-cause mortality: a validation of NHMRC recommendations. Medical Journal of Australia 164, 141–145.[ISI][Medline]

John, U. and Hanke, M. (2002a) Tobacco smoking- and alcohol drinking-attributable cancer mortality in Germany. European Journal of Cancer Prevention 11, 11–17.[CrossRef][ISI][Medline]

John, U. and Hanke, M. (2002b) Alcohol-attributable mortality in a high per capita consumption country — Germany. Alcohol and Alcoholism 37, 581–585.[Abstract/Free Full Text]

John, U. and Hanke, M. (2003) Tobacco- and alcohol-attributable mortality and potential life years lost in Germany. European Journal of Public Health 13, in press.

John, U., Rumpf, H.-J. and Hapke, U. (1999) Estimating prevalence of alcohol abuse and dependence in one general hospital — an approach to reduce sample selection bias. Alcohol and Alcoholism 34, 786–794.[Abstract/Free Full Text]

John, U., Hill, A., Rumpf, H. J., Hapke, U. and Meyer, C. (2003) Alcohol high risk drinking, abuse and dependence among tobacco smoking medical care patients and the population. Drug and Alcohol Dependence 69, 189–195.[CrossRef][ISI][Medline]

McGinnis, J. M. and Foege, W. H. (1999) Mortality and morbidity attributable to use of addictive substances in the United States. Proceedings of the Association of American Physicians 111, 109–118.[ISI][Medline]

Meyer, C., Rumpf, H.-J., Hapke, U. and John, U. (1998) Regionale Unterschiede in der Prävalenz riskanten Alkoholkonsums: Sekundäranalyse des Gesundheitssurveys Ost-West [Differences by area in prevalence of alcohol risk drinking: Analysis of the health survey east–west]. Das Gesundheitswesen 60, 486–492.[Medline]

Moyer, A., Finney, J. W., Swearingen, C. E. and Vergun, P. (2002) Brief interventions for alcohol problems: a meta-analytic review of controlled investigations in treatment-seeking and non-treatment-seeking populations. Addiction 97, 279–292.[CrossRef][ISI][Medline]

Peto, R., Lopez, A. D., Boreham, J., Thun, M. and Heath, C. (1994) Mortality from Smokers in Developed Countries 1950–2000. Oxford University Press, New York, NY.

Rehm, J., Greenfield, T. K. and Rogers, J. D. (2001) Average volume of alcohol consumption, patterns of drinking, and all-cause mortality: results from the US National Alcohol Survey. American Journal of Epidemiology 153, 64–71.[Abstract/Free Full Text]

Ridolfo, B. and Stevenson, C. (2001) The Quantification of Drug-Caused Mortality and Morbidity in Australia, 1998. Australian Institute of Health and Welfare, Canberra.

Rumpf, H.-J., Hapke, U., Meyer, C. and John, U. (1999) Motivation to change drinking behavior: Comparison of alcohol dependent individuals in a general hospital and a general population sample. General Hospital Psychiatry 21, 348–353.[CrossRef][ISI][Medline]

Rumpf, H.-J., Bohlmann, J., Hill, A., Hapke, U. and John, U. (2001) Physicians’ low detection rates of alcohol dependence or abuse: A matter of methodological shortcomings? General Hospital Psychiatry 23, 133–137.[CrossRef][ISI][Medline]

Schimpl-Neimanns, B. (1998) Analysemöglichkeiten des Mikrozensus [Potential analyses of the microcensus]. ZUMA-Nachrichten 42, 91–122.

Selzer, M. L. (1971) The Michigan Alcoholism Screening Test: the quest for a new diagnostic instrument. American Journal of Psychiatry 127, 89–94.[ISI]

Shultz, J. M., Novotny, T. E. and Rice, D. P. (1991a) Quantifying the disease impact of cigarette smoking with SAMMEC II Software. Public Health Reports 106, 326–333.[ISI][Medline]

Shultz, J. M., Rice, D. P., Parker, D. L., Goodman, R. A., Stroh, G. and Chalmers, N. (1991b) Quantifying the disease impact of alcohol with ARDI software. Public Health Reports 106, 443–450.[ISI][Medline]

Single, E., Robson, L., Rehm, J. and Xi, X. (1999) Morbidity and mortality attributable to alcohol, tobacco, and illicit drug use in Canada. American Journal of Public Health 89, 385–390.[Abstract]

Stolzenberg, H. (1995) Gesundheitssurvey Ost-West. Befragungsund Untersuchungssurvey in den neuen und alten Bundesländern. Dokumentation des Public Use File OW91 [Health survey East–West Germany. Manual of the public use file OW91]. Robert-Koch-Institut, Berlin.

Thun, M. J., Apicella, L. F. and Henley, S. J. (2000a) Estimating the numbers of smoking-related deaths. Journal of the American Medical Association 284, 2319–2320.[Free Full Text]

Thun, M. J., Apicella, L. F. and Henley, S. J. (2000b) Smoking vs other risk factors as the cause of smoking-attributable deaths: confounding in the courtroom. Journal of the American Medical Association 284, 706–712.[Abstract/Free Full Text]

US Department of Health and Human Services (1989) Reducing the Health Consequences of Smoking — 25 Years of Progress. A Report of the Surgeon General. Centers for Disease Control, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, DHHS Publication no. (CDC) 89-8411, Atlanta, GA.

US Department of Health and Human Services (1993) Cigarette smoking-attributable mortality and years of potential life lost — United States, 1990. Morbidity and Mortality Weekly Report 42, 645–649.[Medline]

Vartiainen, E., Seppala, T., Lillsunde, P. and Puska, P. (2002) Validation of self reported smoking by serum cotinine measurement in a community-based study. Journal of Epidemiology and Community Health 56, 167–170.[Abstract/Free Full Text]

Welte, R., König, H.-H. and Leidl, R. (2000) The costs of health damage and productivity losses attributable to cigarette smoking in Germany. European Journal of Public Health 10, 31–38.[Abstract/Free Full Text]

WHO (1992) Schedules for Clinical Assessment in Neuropsychiatry. World Health Organization, Geneva.

WHO (2000) International Guide for Monitoring Alcohol Consumption and related Harm. World Health Organization, Geneva.

Wienberg, G., Andritsch, F., Bertram, W., Drees, E., Fleischmann, H., Gerber, H.-G. et al. (1993) Abhängigkeitskranke in psychiatrischer Krankenhausbehandlung. Ergebnisse einer Erhebung in 14 Kliniken der Bundesrepublik [Dependent patients in psychiatric hospital treatment. Results of a data collection in 14 hospitals of the Federal Republic of Germany]. Sucht 39, 264–275.





This Article
Abstract
FREE Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (2)
Request Permissions
Google Scholar
Articles by John, U.
Articles by Hapke, U.
PubMed
PubMed Citation
Articles by John, U.
Articles by Hapke, U.