- Email: [email protected]
Physicians’ low detection rates of alcohol dependence or abuse: a matter of methodological shortcomings?
General Hospital Psychiatry 23 (2001) 133–137
Physicians’ low detection rates of alcohol dependence or abuse: a matter of methodological shortcomings? H.-J. Rumpf, Ph.D.a,*, J. Bohlmann, M.D.b, A. Hill, M.D.c, U. Hapke, Ph.D.d, U. John, Ph.D.d a
Medical University of Lu¨beck, Department of Psychiatry and Psychotherapy, Research Group for Substance Abuse: Treatment, Epidemiology, and Prevention (S:TEP), Lu¨beck, Federal Republic of Germany b Ostholstein-Kliniken, Eutin, Federal Republic of Germany c Universita¨tsklinik Eppendorf, Department of Psychiatry and Psychotherapy, Hamburg, Federal Republic of Germany d University of Greifswald, Institute of Epidemiology and Social Medicine, Greifswald, Federal Republic of Germany
Abstract Previous research may have underestimated physicians’ detection rates of alcohol dependence or abuse because case findings have been based on screening questionnaires instead of using in-depth diagnostic criteria and detection rates have been assessed by analyzing patient records instead of directly interviewing the physician. To test this hypothesis, consecutive patients of a general hospital (N⫽436) and of 12 randomly selected general practices (N⫽929) were examined. A two-step diagnostic procedure included screening questionnaires and a diagnostic interview (SCAN). The analysis compares detection rates based on methods used in previous studies to data using more precise methods. Physicians’ detection rates ranged from 37.0% to 88.9% in the general hospital and from 11.1% to 74.7% in general practices depending on methods used. The physicians’ detection rates could be improved by 10% (general hospital) and 20% (general practice) through the additional use of a screening questionnaire. Of those patients assessed by the physicians as problem drinkers in the general hospital, 13.9% were referred to an addiction consultation-liaison service. Data reveal that physicians’ abilities to detect problem drinkers have been underestimated. Routine screening procedures could play a major role in improving detection rates and reminding the physician to intervene. © 2001 Elsevier Science Inc. All rights reserved. Keywords: Alcohol dependence; General hospital; General practice; Detection; Physician
1. Introduction Physicians are in an outstanding role to detect and manage individuals with alcohol dependence or abuse [1]. In detecting alcohol problems, physicians prefer personal and clinical screening methods to questionnaires [2]. However, only 4% of British general practitioners ask their patients about alcohol consumption all of the time [3]. Therefore, the routine use of screening questionnaires has often been claimed (e.g., [4]). Moreover, several studies found physicians’ detection rates to be unsatisfactory. Between hospital wards, rates ranged from 7% to 65% [5], and general practitioners showed rates below 50% [6,7]. But, previous studies revealed at least one of two shortcomings in determining physicians’ recognition of alcohol problems:
* Corresponding author. Tel.: ⫹451-5002871; fax: ⫹451-5003480. E-mail address: [email protected] (H-J. Rumpf).
First, there is a type of study where case finding is based on screening instruments and not on an in-depth interview providing diagnostic criteria. Although questionnaires have shown high sensitivity (rate of correctly identified individuals having the disorder) and specificity (rate of correctly identified individuals having no disorder) and proved superior to laboratory data [8], screening procedures are restricted by producing a number of false negative and false positive results which leads to a bias in estimating physicians’ ability of detection. The following example shows how this affects the estimation of detection rates: Given a prevalence of 10% in a sample of 100 patients and a screening test with a sensitivity and specificity of .90 each will lead to a misclassification of 10 individuals: Of those 10 subjects with alcohol-related disorders in the given sample, one will not be detected according to the sensitivity of .90, and of those 90 subjects without alcohol-related disorder, 9 will be false positive according to the specificity of .90. Based on the screening results, 18 subjects are deemed to
0163-8343/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S 0 1 6 3 - 8 3 4 3 ( 0 1 ) 0 0 1 3 4 - 7
134
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137
have an alcohol-related disorder (9 are true positive and nine are false positive). Assuming a perfect recognition rate of 100%, the physician will detect 9 of these subjects (those who are true positive) and this is 50% of those scoring positive in the screening test. Therefore, taking the screening diagnosis as a reference test, the detection rate of the physician will be markedly underestimated. In a second type of study, physicians’ detection rates of alcohol problems have been determined by analyzing patient records or case notes instead of specifically interviewing the physician (e.g., [6,9,10]). Diagnoses in patient records do not entirely reflect the physicians’ knowledge and underestimate detection rates [11]. Only one study took account of both methodological requirements and showed a comparatively high rate of 77%, however, data are restricted by a small sample size [1]. The present study has the following objectives: (1) The hypothesis whether physicians’ low detection rates in previous findings are due to the methodological pitfalls described above shall be tested. For this purpose, we use methods of previous studies to determine detection rates and compare this data to rates assessed by using more precise methods in defining cases and assessing the physicians’ detection. (2) The increase of detection rates when using screening questionnaires in addition to the physicians opinion shall be investigated. (3) The percentage of patients referred to a consultation-liaison service among those individuals detected by the general hospital physicians as having alcohol abuse or dependence shall be examined.
2. Method 2.1. General hospital sample Consecutive admissions aged 18 to 64 and staying for more than 24 h on two surgical and two medical wards of a general hospital in Lu¨beck, northern Germany, were examined during a period of 5 months. A two-step diagnostic procedure comprised screening and in-depth diagnostic interview. As screening, two questionnaires were used: the well known CAGE [12,13] comprising 4 items and the Luebeck Alcohol dependence and abuse Screening Test (LAST) [14] consisting of 7 items. The LAST is a combination of items from the CAGE and the Michigan Alcoholism Screening Test (MAST) [15]. This questionnaire showed a higher sensitivity than CAGE and a similar sensitivity as the MAST although comprising less items. The sensitivity in general hospital ranged between 0.82 and 0.87, the specificity between 0.88 and 0.91. As cut-off, 2 points were used for CAGE and LAST. The patient was left alone when filling out the questionnaire which was collected by the research staff the next day. In the first two months of data gathering, all patients were interviewed with section 11 of the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) [16] to provide
DSM-IV [17] and ICD-10 [18] diagnoses of alcohol dependence and abuse/harmful use by research staff. Afterwards, the following two criteria were used to conduct the in-depth diagnostic interview: (1) Patients were positive in CAGE or MAST; (2) The physician who was responsible for the respective patient had the opinion that the patient had alcohol dependence or abuse. Screening and SCAN interview were conducted by research staff. To assess physicians’ detection rates, physicians were specifically asked in a face-to-face interview if the respective patient had alcohol dependence or abuse. The physicians were blind to screening and interview results. In addition, patient records were examined to search for (1) diagnoses of alcohol dependence, alcohol abuse, and synonyms or (2) notes with regard to alcohol problems. Data from an addiction consultation-liaison service were analyzed for the period of data gathering in the general hospital to assess the percentage of patients referred to this service among those individuals detected by the general hospital physicians as having alcohol abuse or dependence. Of 673 admissions, 6.7% refused to take part, 1.8% were too ill or died, 4.8% could not speak German sufficiently, and 22.5% were discharged before finishing the assessment. The data of 436 patients could be entered into analysis. Of this sample, 65.3% were male, 31% were screening positive in one of the questionnaires, 5.3% were only positive in the physicians’ assessment, and 20.9% were interviewed with the SCAN although not positive in screening or the physicians’ judgements; 15.1% had a current alcohol dependence according to either ICD-10 or DSM-IV (fulfilling criteria in the last 12 months) and 7.6% current alcohol abuse/harmful use. Of those having alcohol dependence or abuse, 18.2% were female, 31.3% were married, 41.4% were unemployed, and 22.4% had more than 9 years of schooling; mean age was 44.2 (SD⫽12.0). 2.2. General practice sample Stratified according to town districts, general practices in Lu¨beck were randomly asked to take part in the study. For recruiting 12 practices 18 had to be addressed, a satisfying rate for conditions in Germany where general practitioners work on a private base. In each practice, 80 consecutive patients aged 14 to 75 were screened by using the CAGE and the 13-item Short MAST (SMAST) [19]. As cut-off, 2 points were used for both instruments. Two criteria were used to conduct the in-depth diagnostic interview: (1) Patients were positive in either CAGE or SMAST; (2) The physician had the opinion that the patient had alcohol dependence or abuse. As in the general hospital part of the study, section 11 of the SCAN was used to provide DSMIII-R [20] and ICD-10 diagnoses of alcohol dependence or abuse/harmful use. Screening and SCAN interview were conducted by research staff. To assess physicians’ detection rates, the general practitioners (being blind to the results of the patients’ test)
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137
135
Table 1 General hospital physicians’ detection rates in percent according to different methods (N ⫽ 436)
Table 2 General practice physicians’ detection rates in percent according to different methods (N ⫽ 929)
Assessment of physicians’ detection
Assessment of physicians’ detection
Reference test Screening positive Investigator SCAN by questionnairea interview: alcohol dependence or abuse/ harmful useb
Diagnoses in patient records 37.0 Notes in patient records 57.8 Physician when questioned by 68.1 researcher
48.5 74.7 88.9
a
CAGE or LAST. b DSM IV/ICD-10.
documented their knowledge about the presence or absence of an alcohol problem of the patient on a separate questionnaire. In addition, the physicians’ records were examined according to alcohol-related diagnoses. In this sample, notes with regard to alcohol problems were not assessed in patient records. Of 960 eligible subjects, 2.9% refused to take part and 0.3% were too ill to participate which results in a sample of 929 patients. Of this sample, 38.2% were male; 14.5% were positive in CAGE or SMAST, 8.9% were only positive in the physicians’ assessment. Of the eligible subjects for diagnostics with the SCAN, 6.9% refused to take part in the interview. Of the total sample, 7.2% had a current alcohol dependence according to either DSM-III-R or ICD-10 (fulfilling criteria in the last 12 months) and 3.5% current alcohol abuse/harmful use. Of those having alcohol dependence or abuse, 36.4% were female, 49.0% were married, 12.1% were unemployed, and 36.4% had more than 9 years of schooling; mean age was 43.2 (SD⫽15.9). More details of both samples are described elsewhere [14,21].
Reference test Screening positive Investigator SCAN by questionnairea interview: alcohol dependence or abuse/ harmful useb
Diagnoses in patient records 11.1 Physician when questioned by 50.4 researcher a b
12.1 74.7
CAGE or SMAST. DSM-III-R/ICD-10.
detected by the physicians, groups were compared on grounds of sociodemographic characteristics. No differences were found for gender (Fisher’s Exact Test; P⫽.68), marital status (Fisher’s Exact Test; P⫽.49.), years of schooling (up to 9 vs. more than 9 years; Fisher’s Exact Test; P⫽.23) unemployment rate (Fisher’s Exact Test; P⫽.35), whereas detected patients were significantly younger (t⫽⫺4.07; df⫽97; P⬍.001). In addition, patients on internal wards were more often detected by their physicians compared to surgical wards (Fisher’s Exact Test; P⬍.01). We tested whether using one of the questionnaires could improve physicians’ detection of patients with alcohol dependence or abuse. If the general hospital physicians had also used the CAGE or the LAST as well as giving their diagnosis opinion, they would have detected 98% and 99%, respectively, of patients positive in the SCAN interview. Finally, of those patients assessed by the physicians as problem drinkers in general hospital, 13.9% were referred to an addiction consultation-liaison service. 3.2. General practice
3. Results 3.1. General hospital In the general hospital sample, the physicians showed the lowest detection rate (37%) when screening results were taken as reference test and recognition was assessed by listing diagnoses in patient records (Table 1). If any note with respect to alcohol problems was taken into consideration or physicians were asked specifically, detection rates raise to 57.8% and 68.1%. According to each of the three attempts to assess the physicians’ recognition of alcohol problems, the detection rates increase when DSM-IV and ICD-10 diagnoses are taken as gold standard. Based on interview and diagnostic criteria, physicians detected 88.9% of patients with alcohol dependence or abuse. To examine if these 88.9% of patients differ from those patients not being
Results in the general practice sample show a low detection rate of 11.1% when data are based on physicians’ diagnoses in patient records and screening results (Table 2). Taking diagnostic criteria as gold standard increases the rate slightly to 12.1%. When the general practice physician was asked whether a patient had alcohol dependence or abuse he gave this diagnosis to 50.4% of those who scored positive on either the CAGE or the SMAST questionnaires, and 74.4% of patients who had been diagnosed as a having alcohol dependence or abuse by the SCAN research interview according to DSM-III-R or ICD-10. To examine if these 74.4% of patients differ from those patients not being detected by the physicians, groups were compared on grounds of sociodemographic characteristics. No differences were found for gender (chi2⫽0.28; df⫽1; P⫽.60), marital status (chi2⫽0.12; df⫽1; P⫽.73), years of schooling (up to 9 vs. more than 9 years; chi2⫽1.96; P⫽.16)
136
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137
unemployment rate (Fisher’s Exact Test; P⫽.29), and age (t⫽⫺.77; df⫽97; P⫽.44). If the general practice physicians had also used the CAGE or the SMAST as well as giving their diagnosis opinion, they would have detected 94.9% and 93.9%, respectively, of patients positive in the SCAN interview.
4. Discussion Findings support the hypothesis that results in previous studies underestimate the ability of physicians to detect individuals with alcohol dependence or abuse. In both samples, detection rates increased when the results of the diagnostic interview were taken as gold standard instead of screening diagnoses, and when recognition was assessed by interviewing the physician instead of listing diagnoses in patient records. According to the described methodological approaches, detection rates of physicians showed marked differences ranging from 37% to 88.9% in the general hospital sample and 11.1% to 74.4% in general practices. In the general hospital sample, physicians showed significantly better detection rates in internal medicine (opposed to surgical) and older patients. In general practice, no differences between detected and undetected patients were found. Other factors (e. g. the presence of an alcohol-related disorder) that have not been analyzed in this study might influence physicians’ ability to identify patients with alcohol use disorders. In our samples, three sources of bias to overestimate detection rates have to be considered: Firstly, asking to participate in a research project might stimulate the physicians’ ambition. Secondly, in the general practice study, patient counseling was offered by the research staff, whereas in daily routine physicians might hesitate to diagnose patients because of feeling not sufficiently qualified to offer help. Thirdly, in the general hospital sample, previous research activities and an addiction consultation-liaison service might have had an impact on the physicians’ awareness of alcohol problems among patients. Nevertheless, these assumptions do not affect the discrepancies in detection rates when different methodological approaches are compared. Although the general hospital is an ideal setting to initiate interventions for problem drinkers because 38.2% of alcohol dependent patients have never before received counseling or treatment [22], and motivation to change drinking behavior is enhanced compared to alcohol dependent individuals in the general population [23], our data show that only a minority of patients (13.9%) are referred to an addiction consultation-liaison service. This is in line with other data revealing that only 5% of substance abusing inpatients received adequate treatment [24]. According to our findings, the role of screening for alcohol dependence or abuse in medical settings might be redefined. The main problem seems not to be that physicians show a poor ability
of detecting patients with alcohol problems but that they refrain from intervening. Although data indicates that screening instruments can still improve detection rates by 10% (general hospital) to 20% (general practice) when used in addition to the physicians’ judgement, the major benefit of systematic screening might be to remind the physician to initiate interventions. An approach could be to administer a short screening questionnaire routinely to every new patient by a nurse or receptionist and keep it in the patient records like a medical finding. This could contribute to binding the physician to intervene. Such systematic procedures play an important role in establishing secondary prevention for alcohol problems in health care settings. Acknowledgments This study was supported by German Ministry of Health grant 326-4914-8/38. References [1] Cleary PD, Miller M, Bush BT, Warburg MM, Delbanco TL, Aronson MD. Prevalence and recognition of alcohol abuse in a primary care population. Am J Med 1988;85:466 – 471. [2] Townes PN, Harkley AL. Alcohol screening practices of primary care physicians in Eastern North Carolina. Alcohol 1994;11:489 – 492. [3] Kaner EFS, Heather N, McAvoy BR, Lock CA, Gilvarry E. Intervention for excessive alcohol consumption in primary health care: Attitudes and practices of English general practitioners. Alcohol Alcohol 1999;34:559 –566. [4] Babor TF, Higgins-Biddle JC. Alcohol screening and brief intervention: dissemination strategies for medical practice and public health. Addiction 2000;95:677– 686. [5] Moore RD, Bone LR, Geller G, Mamon JA, Stokes EJ, Levine DM. Prevalence, detection, and treatment of alcoholism in hospitalized patients. JAMA 1989;261:403– 407. [6] Coulehan JL, Zettler-Segal M, Block M, McClelland M, Schulberg HC. Recognition of alcoholism and substance abuse in primary care patients. Arch Intern Med 1987;147:349 –352. [7] Rydon P, Redman S, Sanson-Fisher RW, Reid ALA. Detection of alcohol-related problems in general practice. J Stud Alcohol 1992; 53:197–202. [8] Beresford TP, Blow FC, Hill E, Lucey MR. Comparison of CAGE questionnaire and computer-assisted laboratory profiles in screening for covert alcoholism. Lancet 1990;336:482– 485. [9] Isaacson JH, Butler R, Zacharek M, Tzelepis A. Screening with the Alcohol Use Disorders Identification Test (AUDIT) in an inner-city population. J Gen Intern Med 1994;9:550 –553. [10] Schmidt A, Barry KL, Fleming MF. Detection of problem drinkers: The Alcohol Use Disorders Identification Test (AUDIT). South Med J 1995;88:52–59. [11] Yersin B. Alcoholism in hospitalized patients in Switzerland. JAMA 1989;262:772. [12] Ewing JA. Detecting alcoholism: The CAGE questionnaire. JAMA 1984;252:1905–1907. [13] Mayfield D, McLeod G, Hall P. The CAGE Questionnaire: Validation of a New Alcoholism Screening Instrument. Am J Psychiatry 1974; 131:1121–1123.
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137 [14] Rumpf H-J, Hapke U, Hill A, John U. Development of a screening questionnaire for the general hospital and general practices. Alcohol Clin Exp Res 1997;21:894 – 898. [15] Selzer ML. The Michigan Alcoholism Screening Test: The quest for a new diagnostic instrument. Am J Psychiatry 1971;127:1653–1658. [16] WHO. Schedules for clinical assessment in neuropsychiatry. World Health Organization, Division of Mental Health 1992. [17] American Psychiatric Association, ed. Diagnostic and statistical manual of mental disorders, fourth edition, international version. American Psychiatric Association, 1995. [18] WHO. Tenth revision of the International Classification of Diseases, Chapter V (F), mental and behavioural disorders. World Health Organization 1991. [19] Selzer ML, Vinokur A, Rooijen MA. A self-administered Short Michigan Alcoholism Screening Test (SMAST). J Stud Alcohol 1975;36: 117–126
137
[20] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. American Psychiatric Association 1987. [21] Hill A, Rumpf H-J, Hapke U, Driessen M, John U. Prevalence of alcohol dependence and abuse in general practice in Germany - a representative study. Alcohol Clin Exp Res 1998;22:935–940. [22] Rumpf H-J, Hapke U, John U. Previous help seeking and motivation to change drinking behavior in alcohol dependent general hospital patients. Gen Hosp Psychiatry 1998;20:115–119. [23] Rumpf H-J, Hapke U, Meyer C, John U. Motivation to change drinking behavior: Comparison of alcohol-dependent individuals in a general hospital and a general population sample. Gen Hosp Psychiatry 1999;21:348 –353. [24] Hopkins TB, Zarro VJ, McCarter TG. The adequacy of screening, documenting, and treating the diseases of substance abuse. J Addict Dis 1994;13:81– 87.
Physicians’ low detection rates of alcohol dependence or abuse: a matter of methodological shortcomings? H.-J. Rumpf, Ph.D.a,*, J. Bohlmann, M.D.b, A. Hill, M.D.c, U. Hapke, Ph.D.d, U. John, Ph.D.d a
Medical University of Lu¨beck, Department of Psychiatry and Psychotherapy, Research Group for Substance Abuse: Treatment, Epidemiology, and Prevention (S:TEP), Lu¨beck, Federal Republic of Germany b Ostholstein-Kliniken, Eutin, Federal Republic of Germany c Universita¨tsklinik Eppendorf, Department of Psychiatry and Psychotherapy, Hamburg, Federal Republic of Germany d University of Greifswald, Institute of Epidemiology and Social Medicine, Greifswald, Federal Republic of Germany
Abstract Previous research may have underestimated physicians’ detection rates of alcohol dependence or abuse because case findings have been based on screening questionnaires instead of using in-depth diagnostic criteria and detection rates have been assessed by analyzing patient records instead of directly interviewing the physician. To test this hypothesis, consecutive patients of a general hospital (N⫽436) and of 12 randomly selected general practices (N⫽929) were examined. A two-step diagnostic procedure included screening questionnaires and a diagnostic interview (SCAN). The analysis compares detection rates based on methods used in previous studies to data using more precise methods. Physicians’ detection rates ranged from 37.0% to 88.9% in the general hospital and from 11.1% to 74.7% in general practices depending on methods used. The physicians’ detection rates could be improved by 10% (general hospital) and 20% (general practice) through the additional use of a screening questionnaire. Of those patients assessed by the physicians as problem drinkers in the general hospital, 13.9% were referred to an addiction consultation-liaison service. Data reveal that physicians’ abilities to detect problem drinkers have been underestimated. Routine screening procedures could play a major role in improving detection rates and reminding the physician to intervene. © 2001 Elsevier Science Inc. All rights reserved. Keywords: Alcohol dependence; General hospital; General practice; Detection; Physician
1. Introduction Physicians are in an outstanding role to detect and manage individuals with alcohol dependence or abuse [1]. In detecting alcohol problems, physicians prefer personal and clinical screening methods to questionnaires [2]. However, only 4% of British general practitioners ask their patients about alcohol consumption all of the time [3]. Therefore, the routine use of screening questionnaires has often been claimed (e.g., [4]). Moreover, several studies found physicians’ detection rates to be unsatisfactory. Between hospital wards, rates ranged from 7% to 65% [5], and general practitioners showed rates below 50% [6,7]. But, previous studies revealed at least one of two shortcomings in determining physicians’ recognition of alcohol problems:
* Corresponding author. Tel.: ⫹451-5002871; fax: ⫹451-5003480. E-mail address: [email protected] (H-J. Rumpf).
First, there is a type of study where case finding is based on screening instruments and not on an in-depth interview providing diagnostic criteria. Although questionnaires have shown high sensitivity (rate of correctly identified individuals having the disorder) and specificity (rate of correctly identified individuals having no disorder) and proved superior to laboratory data [8], screening procedures are restricted by producing a number of false negative and false positive results which leads to a bias in estimating physicians’ ability of detection. The following example shows how this affects the estimation of detection rates: Given a prevalence of 10% in a sample of 100 patients and a screening test with a sensitivity and specificity of .90 each will lead to a misclassification of 10 individuals: Of those 10 subjects with alcohol-related disorders in the given sample, one will not be detected according to the sensitivity of .90, and of those 90 subjects without alcohol-related disorder, 9 will be false positive according to the specificity of .90. Based on the screening results, 18 subjects are deemed to
0163-8343/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S 0 1 6 3 - 8 3 4 3 ( 0 1 ) 0 0 1 3 4 - 7
134
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137
have an alcohol-related disorder (9 are true positive and nine are false positive). Assuming a perfect recognition rate of 100%, the physician will detect 9 of these subjects (those who are true positive) and this is 50% of those scoring positive in the screening test. Therefore, taking the screening diagnosis as a reference test, the detection rate of the physician will be markedly underestimated. In a second type of study, physicians’ detection rates of alcohol problems have been determined by analyzing patient records or case notes instead of specifically interviewing the physician (e.g., [6,9,10]). Diagnoses in patient records do not entirely reflect the physicians’ knowledge and underestimate detection rates [11]. Only one study took account of both methodological requirements and showed a comparatively high rate of 77%, however, data are restricted by a small sample size [1]. The present study has the following objectives: (1) The hypothesis whether physicians’ low detection rates in previous findings are due to the methodological pitfalls described above shall be tested. For this purpose, we use methods of previous studies to determine detection rates and compare this data to rates assessed by using more precise methods in defining cases and assessing the physicians’ detection. (2) The increase of detection rates when using screening questionnaires in addition to the physicians opinion shall be investigated. (3) The percentage of patients referred to a consultation-liaison service among those individuals detected by the general hospital physicians as having alcohol abuse or dependence shall be examined.
2. Method 2.1. General hospital sample Consecutive admissions aged 18 to 64 and staying for more than 24 h on two surgical and two medical wards of a general hospital in Lu¨beck, northern Germany, were examined during a period of 5 months. A two-step diagnostic procedure comprised screening and in-depth diagnostic interview. As screening, two questionnaires were used: the well known CAGE [12,13] comprising 4 items and the Luebeck Alcohol dependence and abuse Screening Test (LAST) [14] consisting of 7 items. The LAST is a combination of items from the CAGE and the Michigan Alcoholism Screening Test (MAST) [15]. This questionnaire showed a higher sensitivity than CAGE and a similar sensitivity as the MAST although comprising less items. The sensitivity in general hospital ranged between 0.82 and 0.87, the specificity between 0.88 and 0.91. As cut-off, 2 points were used for CAGE and LAST. The patient was left alone when filling out the questionnaire which was collected by the research staff the next day. In the first two months of data gathering, all patients were interviewed with section 11 of the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) [16] to provide
DSM-IV [17] and ICD-10 [18] diagnoses of alcohol dependence and abuse/harmful use by research staff. Afterwards, the following two criteria were used to conduct the in-depth diagnostic interview: (1) Patients were positive in CAGE or MAST; (2) The physician who was responsible for the respective patient had the opinion that the patient had alcohol dependence or abuse. Screening and SCAN interview were conducted by research staff. To assess physicians’ detection rates, physicians were specifically asked in a face-to-face interview if the respective patient had alcohol dependence or abuse. The physicians were blind to screening and interview results. In addition, patient records were examined to search for (1) diagnoses of alcohol dependence, alcohol abuse, and synonyms or (2) notes with regard to alcohol problems. Data from an addiction consultation-liaison service were analyzed for the period of data gathering in the general hospital to assess the percentage of patients referred to this service among those individuals detected by the general hospital physicians as having alcohol abuse or dependence. Of 673 admissions, 6.7% refused to take part, 1.8% were too ill or died, 4.8% could not speak German sufficiently, and 22.5% were discharged before finishing the assessment. The data of 436 patients could be entered into analysis. Of this sample, 65.3% were male, 31% were screening positive in one of the questionnaires, 5.3% were only positive in the physicians’ assessment, and 20.9% were interviewed with the SCAN although not positive in screening or the physicians’ judgements; 15.1% had a current alcohol dependence according to either ICD-10 or DSM-IV (fulfilling criteria in the last 12 months) and 7.6% current alcohol abuse/harmful use. Of those having alcohol dependence or abuse, 18.2% were female, 31.3% were married, 41.4% were unemployed, and 22.4% had more than 9 years of schooling; mean age was 44.2 (SD⫽12.0). 2.2. General practice sample Stratified according to town districts, general practices in Lu¨beck were randomly asked to take part in the study. For recruiting 12 practices 18 had to be addressed, a satisfying rate for conditions in Germany where general practitioners work on a private base. In each practice, 80 consecutive patients aged 14 to 75 were screened by using the CAGE and the 13-item Short MAST (SMAST) [19]. As cut-off, 2 points were used for both instruments. Two criteria were used to conduct the in-depth diagnostic interview: (1) Patients were positive in either CAGE or SMAST; (2) The physician had the opinion that the patient had alcohol dependence or abuse. As in the general hospital part of the study, section 11 of the SCAN was used to provide DSMIII-R [20] and ICD-10 diagnoses of alcohol dependence or abuse/harmful use. Screening and SCAN interview were conducted by research staff. To assess physicians’ detection rates, the general practitioners (being blind to the results of the patients’ test)
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137
135
Table 1 General hospital physicians’ detection rates in percent according to different methods (N ⫽ 436)
Table 2 General practice physicians’ detection rates in percent according to different methods (N ⫽ 929)
Assessment of physicians’ detection
Assessment of physicians’ detection
Reference test Screening positive Investigator SCAN by questionnairea interview: alcohol dependence or abuse/ harmful useb
Diagnoses in patient records 37.0 Notes in patient records 57.8 Physician when questioned by 68.1 researcher
48.5 74.7 88.9
a
CAGE or LAST. b DSM IV/ICD-10.
documented their knowledge about the presence or absence of an alcohol problem of the patient on a separate questionnaire. In addition, the physicians’ records were examined according to alcohol-related diagnoses. In this sample, notes with regard to alcohol problems were not assessed in patient records. Of 960 eligible subjects, 2.9% refused to take part and 0.3% were too ill to participate which results in a sample of 929 patients. Of this sample, 38.2% were male; 14.5% were positive in CAGE or SMAST, 8.9% were only positive in the physicians’ assessment. Of the eligible subjects for diagnostics with the SCAN, 6.9% refused to take part in the interview. Of the total sample, 7.2% had a current alcohol dependence according to either DSM-III-R or ICD-10 (fulfilling criteria in the last 12 months) and 3.5% current alcohol abuse/harmful use. Of those having alcohol dependence or abuse, 36.4% were female, 49.0% were married, 12.1% were unemployed, and 36.4% had more than 9 years of schooling; mean age was 43.2 (SD⫽15.9). More details of both samples are described elsewhere [14,21].
Reference test Screening positive Investigator SCAN by questionnairea interview: alcohol dependence or abuse/ harmful useb
Diagnoses in patient records 11.1 Physician when questioned by 50.4 researcher a b
12.1 74.7
CAGE or SMAST. DSM-III-R/ICD-10.
detected by the physicians, groups were compared on grounds of sociodemographic characteristics. No differences were found for gender (Fisher’s Exact Test; P⫽.68), marital status (Fisher’s Exact Test; P⫽.49.), years of schooling (up to 9 vs. more than 9 years; Fisher’s Exact Test; P⫽.23) unemployment rate (Fisher’s Exact Test; P⫽.35), whereas detected patients were significantly younger (t⫽⫺4.07; df⫽97; P⬍.001). In addition, patients on internal wards were more often detected by their physicians compared to surgical wards (Fisher’s Exact Test; P⬍.01). We tested whether using one of the questionnaires could improve physicians’ detection of patients with alcohol dependence or abuse. If the general hospital physicians had also used the CAGE or the LAST as well as giving their diagnosis opinion, they would have detected 98% and 99%, respectively, of patients positive in the SCAN interview. Finally, of those patients assessed by the physicians as problem drinkers in general hospital, 13.9% were referred to an addiction consultation-liaison service. 3.2. General practice
3. Results 3.1. General hospital In the general hospital sample, the physicians showed the lowest detection rate (37%) when screening results were taken as reference test and recognition was assessed by listing diagnoses in patient records (Table 1). If any note with respect to alcohol problems was taken into consideration or physicians were asked specifically, detection rates raise to 57.8% and 68.1%. According to each of the three attempts to assess the physicians’ recognition of alcohol problems, the detection rates increase when DSM-IV and ICD-10 diagnoses are taken as gold standard. Based on interview and diagnostic criteria, physicians detected 88.9% of patients with alcohol dependence or abuse. To examine if these 88.9% of patients differ from those patients not being
Results in the general practice sample show a low detection rate of 11.1% when data are based on physicians’ diagnoses in patient records and screening results (Table 2). Taking diagnostic criteria as gold standard increases the rate slightly to 12.1%. When the general practice physician was asked whether a patient had alcohol dependence or abuse he gave this diagnosis to 50.4% of those who scored positive on either the CAGE or the SMAST questionnaires, and 74.4% of patients who had been diagnosed as a having alcohol dependence or abuse by the SCAN research interview according to DSM-III-R or ICD-10. To examine if these 74.4% of patients differ from those patients not being detected by the physicians, groups were compared on grounds of sociodemographic characteristics. No differences were found for gender (chi2⫽0.28; df⫽1; P⫽.60), marital status (chi2⫽0.12; df⫽1; P⫽.73), years of schooling (up to 9 vs. more than 9 years; chi2⫽1.96; P⫽.16)
136
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137
unemployment rate (Fisher’s Exact Test; P⫽.29), and age (t⫽⫺.77; df⫽97; P⫽.44). If the general practice physicians had also used the CAGE or the SMAST as well as giving their diagnosis opinion, they would have detected 94.9% and 93.9%, respectively, of patients positive in the SCAN interview.
4. Discussion Findings support the hypothesis that results in previous studies underestimate the ability of physicians to detect individuals with alcohol dependence or abuse. In both samples, detection rates increased when the results of the diagnostic interview were taken as gold standard instead of screening diagnoses, and when recognition was assessed by interviewing the physician instead of listing diagnoses in patient records. According to the described methodological approaches, detection rates of physicians showed marked differences ranging from 37% to 88.9% in the general hospital sample and 11.1% to 74.4% in general practices. In the general hospital sample, physicians showed significantly better detection rates in internal medicine (opposed to surgical) and older patients. In general practice, no differences between detected and undetected patients were found. Other factors (e. g. the presence of an alcohol-related disorder) that have not been analyzed in this study might influence physicians’ ability to identify patients with alcohol use disorders. In our samples, three sources of bias to overestimate detection rates have to be considered: Firstly, asking to participate in a research project might stimulate the physicians’ ambition. Secondly, in the general practice study, patient counseling was offered by the research staff, whereas in daily routine physicians might hesitate to diagnose patients because of feeling not sufficiently qualified to offer help. Thirdly, in the general hospital sample, previous research activities and an addiction consultation-liaison service might have had an impact on the physicians’ awareness of alcohol problems among patients. Nevertheless, these assumptions do not affect the discrepancies in detection rates when different methodological approaches are compared. Although the general hospital is an ideal setting to initiate interventions for problem drinkers because 38.2% of alcohol dependent patients have never before received counseling or treatment [22], and motivation to change drinking behavior is enhanced compared to alcohol dependent individuals in the general population [23], our data show that only a minority of patients (13.9%) are referred to an addiction consultation-liaison service. This is in line with other data revealing that only 5% of substance abusing inpatients received adequate treatment [24]. According to our findings, the role of screening for alcohol dependence or abuse in medical settings might be redefined. The main problem seems not to be that physicians show a poor ability
of detecting patients with alcohol problems but that they refrain from intervening. Although data indicates that screening instruments can still improve detection rates by 10% (general hospital) to 20% (general practice) when used in addition to the physicians’ judgement, the major benefit of systematic screening might be to remind the physician to initiate interventions. An approach could be to administer a short screening questionnaire routinely to every new patient by a nurse or receptionist and keep it in the patient records like a medical finding. This could contribute to binding the physician to intervene. Such systematic procedures play an important role in establishing secondary prevention for alcohol problems in health care settings. Acknowledgments This study was supported by German Ministry of Health grant 326-4914-8/38. References [1] Cleary PD, Miller M, Bush BT, Warburg MM, Delbanco TL, Aronson MD. Prevalence and recognition of alcohol abuse in a primary care population. Am J Med 1988;85:466 – 471. [2] Townes PN, Harkley AL. Alcohol screening practices of primary care physicians in Eastern North Carolina. Alcohol 1994;11:489 – 492. [3] Kaner EFS, Heather N, McAvoy BR, Lock CA, Gilvarry E. Intervention for excessive alcohol consumption in primary health care: Attitudes and practices of English general practitioners. Alcohol Alcohol 1999;34:559 –566. [4] Babor TF, Higgins-Biddle JC. Alcohol screening and brief intervention: dissemination strategies for medical practice and public health. Addiction 2000;95:677– 686. [5] Moore RD, Bone LR, Geller G, Mamon JA, Stokes EJ, Levine DM. Prevalence, detection, and treatment of alcoholism in hospitalized patients. JAMA 1989;261:403– 407. [6] Coulehan JL, Zettler-Segal M, Block M, McClelland M, Schulberg HC. Recognition of alcoholism and substance abuse in primary care patients. Arch Intern Med 1987;147:349 –352. [7] Rydon P, Redman S, Sanson-Fisher RW, Reid ALA. Detection of alcohol-related problems in general practice. J Stud Alcohol 1992; 53:197–202. [8] Beresford TP, Blow FC, Hill E, Lucey MR. Comparison of CAGE questionnaire and computer-assisted laboratory profiles in screening for covert alcoholism. Lancet 1990;336:482– 485. [9] Isaacson JH, Butler R, Zacharek M, Tzelepis A. Screening with the Alcohol Use Disorders Identification Test (AUDIT) in an inner-city population. J Gen Intern Med 1994;9:550 –553. [10] Schmidt A, Barry KL, Fleming MF. Detection of problem drinkers: The Alcohol Use Disorders Identification Test (AUDIT). South Med J 1995;88:52–59. [11] Yersin B. Alcoholism in hospitalized patients in Switzerland. JAMA 1989;262:772. [12] Ewing JA. Detecting alcoholism: The CAGE questionnaire. JAMA 1984;252:1905–1907. [13] Mayfield D, McLeod G, Hall P. The CAGE Questionnaire: Validation of a New Alcoholism Screening Instrument. Am J Psychiatry 1974; 131:1121–1123.
H.-J. Rumpf et al. / General Hospital Psychiatry 23 (2001) 133–137 [14] Rumpf H-J, Hapke U, Hill A, John U. Development of a screening questionnaire for the general hospital and general practices. Alcohol Clin Exp Res 1997;21:894 – 898. [15] Selzer ML. The Michigan Alcoholism Screening Test: The quest for a new diagnostic instrument. Am J Psychiatry 1971;127:1653–1658. [16] WHO. Schedules for clinical assessment in neuropsychiatry. World Health Organization, Division of Mental Health 1992. [17] American Psychiatric Association, ed. Diagnostic and statistical manual of mental disorders, fourth edition, international version. American Psychiatric Association, 1995. [18] WHO. Tenth revision of the International Classification of Diseases, Chapter V (F), mental and behavioural disorders. World Health Organization 1991. [19] Selzer ML, Vinokur A, Rooijen MA. A self-administered Short Michigan Alcoholism Screening Test (SMAST). J Stud Alcohol 1975;36: 117–126
137
[20] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. American Psychiatric Association 1987. [21] Hill A, Rumpf H-J, Hapke U, Driessen M, John U. Prevalence of alcohol dependence and abuse in general practice in Germany - a representative study. Alcohol Clin Exp Res 1998;22:935–940. [22] Rumpf H-J, Hapke U, John U. Previous help seeking and motivation to change drinking behavior in alcohol dependent general hospital patients. Gen Hosp Psychiatry 1998;20:115–119. [23] Rumpf H-J, Hapke U, Meyer C, John U. Motivation to change drinking behavior: Comparison of alcohol-dependent individuals in a general hospital and a general population sample. Gen Hosp Psychiatry 1999;21:348 –353. [24] Hopkins TB, Zarro VJ, McCarter TG. The adequacy of screening, documenting, and treating the diseases of substance abuse. J Addict Dis 1994;13:81– 87.