1354
Doctor and Patient RELIABILITY OF ASSESSMENT OF ALCOHOL INTAKE BASED ON PERSONAL INTERVIEWS IN A LIVER CLINIC HECTOR ORREGO LAURENCE M. BLENDIS BHUSHAN M. KAPUR JOAN E. BLAKE YEDY ISRAEL Addiction Research Foundation, Toronto; Toronto General
Hospital; and Departments of Pharmacology, Medicine, and Physiology, University of Toronto, Toronto, Ontario, Canada
patients with alcoholic liver disurinary alcohol was measured daily for up to 6 months. Every week the patients were asked about their drinking during the past week. Those who convinced the physicians of their abstinence were recorded as not drinking. Patients with alcohol in their urines convincingly denied alcohol intake 52% of the times that they were questioned. 25% of them denied drinking every time. Only 17% of all patients admitted it at all times. Patients who always admitted to drinking had an average urinary alcohol value of 1420±66 mg/l, compared to 81±5 mg/l in those who denied drinking every time. Those who admitted drinking intermittently had significantly higher urinary alcohol values (1001±57 mg/l) when admitting than when denying (538±30 mg/l). The personal interview should not be used to separate populations of abstainers and nonabstainers in the follow-up of alcoholic patients. On the
Summary
In 37 ease
other hand, deniers appear to consume less alcohol than those who admit their drinking.
On starting the trial the patients were given a week’s supply of stamped mailing containers and were instructed to mail to the clinic daily a sample of urine voided 2 h after ingestion of the treatment capsule at 8 A.M. The sample was used both for monitoring the patients’ compliance in taking the medication (see Introduction) and for testing for the presence of alcohol (and subsequently for the amount). Alcohol was identified in urine injected directly onto a column (Halcomid M-18, 6 ft x t in) at an oven temperature of 108°C, by means of a gas chromatograph with flame ionisation detector.9 Alcohol concentration was measured on an Auto-lab system IV integrator (Spectra-Physics, Mountain View, California) attached to a series 1400 gas chromatograph with autosampler (Varian Instruments, Palo Alto, California). Twenty urine samples from 2 patients were kept at room temperature for 20 days. Alcohol concentration was determined in these samples initially and at 10 and 20 days. The concentration of alcohol in the samples did not change with time, and no alcohol was detected after 20 days among those without alcohol initially. It has been reported’O that urines from non-drinking diabetics can have significant amounts of alcohol, presumably originating from glucose. However, diabetics were excluded from our study. Patients were advised to abstain from alcohol and were seen once a week at the liver clinic. Here they were asked about their symptoms and whether they had been drinking during the previous week. A patient who denied drinking was recorded as not drinking only if the physician was convinced by his denial.
Statistical Analysis The data were expressed as mean +SEM. Comparisons were made with Student’s t-test for unpaired data." Data in percentage were initially subjected to arc-sin transformation. Calculations were done on a pre-programmed Radio Shack (a division of Tandy Corporation, Barrie, Ontario, Canada) TRS-80 microcomputer system. p
Doctor and Patient RELIABILITY OF ASSESSMENT OF ALCOHOL INTAKE BASED ON PERSONAL INTERVIEWS IN A LIVER CLINIC HECTOR ORREGO LAURENCE M. BLENDIS BHUSHAN M. KAPUR JOAN E. BLAKE YEDY ISRAEL Addiction Research Foundation, Toronto; Toronto General
Hospital; and Departments of Pharmacology, Medicine, and Physiology, University of Toronto, Toronto, Ontario, Canada
patients with alcoholic liver disurinary alcohol was measured daily for up to 6 months. Every week the patients were asked about their drinking during the past week. Those who convinced the physicians of their abstinence were recorded as not drinking. Patients with alcohol in their urines convincingly denied alcohol intake 52% of the times that they were questioned. 25% of them denied drinking every time. Only 17% of all patients admitted it at all times. Patients who always admitted to drinking had an average urinary alcohol value of 1420±66 mg/l, compared to 81±5 mg/l in those who denied drinking every time. Those who admitted drinking intermittently had significantly higher urinary alcohol values (1001±57 mg/l) when admitting than when denying (538±30 mg/l). The personal interview should not be used to separate populations of abstainers and nonabstainers in the follow-up of alcoholic patients. On the
Summary
In 37 ease
other hand, deniers appear to consume less alcohol than those who admit their drinking.
On starting the trial the patients were given a week’s supply of stamped mailing containers and were instructed to mail to the clinic daily a sample of urine voided 2 h after ingestion of the treatment capsule at 8 A.M. The sample was used both for monitoring the patients’ compliance in taking the medication (see Introduction) and for testing for the presence of alcohol (and subsequently for the amount). Alcohol was identified in urine injected directly onto a column (Halcomid M-18, 6 ft x t in) at an oven temperature of 108°C, by means of a gas chromatograph with flame ionisation detector.9 Alcohol concentration was measured on an Auto-lab system IV integrator (Spectra-Physics, Mountain View, California) attached to a series 1400 gas chromatograph with autosampler (Varian Instruments, Palo Alto, California). Twenty urine samples from 2 patients were kept at room temperature for 20 days. Alcohol concentration was determined in these samples initially and at 10 and 20 days. The concentration of alcohol in the samples did not change with time, and no alcohol was detected after 20 days among those without alcohol initially. It has been reported’O that urines from non-drinking diabetics can have significant amounts of alcohol, presumably originating from glucose. However, diabetics were excluded from our study. Patients were advised to abstain from alcohol and were seen once a week at the liver clinic. Here they were asked about their symptoms and whether they had been drinking during the previous week. A patient who denied drinking was recorded as not drinking only if the physician was convinced by his denial.
Statistical Analysis The data were expressed as mean +SEM. Comparisons were made with Student’s t-test for unpaired data." Data in percentage were initially subjected to arc-sin transformation. Calculations were done on a pre-programmed Radio Shack (a division of Tandy Corporation, Barrie, Ontario, Canada) TRS-80 microcomputer system. p
