As many as 20 percent of hospital outpatients may be alcoholics, a new study reports, and most of them go undiagnosed.
One reason: Doctors have been asking the wrong questions. Straightforward questions -- such as "Have you ever had a drinking problem?" -- the study found, work better than might be expected in pointing up a patient's alcoholism.
The study, by researchers at Brown University and Rhode Island Hospital in Providence, is based on interviews of 232 patients at the hospital's ambulatory care clinic. Results were reported in the Journal of the American Medical Association.
In trying to find out whether a patient is an alcoholic, doctors traditionally have used questions focusing on how much and how often a person drinks. But those are "unsuitable screening techniques for the diagnosis of alcoholism," the Rhode Island researchers said.
A much more sensitive question -- better at eliciting valid evidence of alcoholism -- is, "Have you ever had a drinking problem?" The sensitivity is even higher when that question is asked along with another simple one, "When was your last drink?"
The researchers advise doctors to incorporate such simple questions into the routine medical history "as a first step toward diagnosis."
Though a series of sophisticated screening questionnaires has been developed since 1970, the Rhode Island researchers said that "it is likely that most physicians in practice screen for alcoholism by simply asking the standard questions taught in medical school: 'How much and how often do you drink?' " The validity of that approach, they added, "has never been established."
Some studies have shown that doctors fail to make the alcoholism diagnosis in as many as 90 percent of their alcoholic patients.
The new survey covered first-time patients in the medical primary care unit of Rhode Island Hospital. The patients were interviewed while they waited for their scheduled appointments.
Researchers used a questionnaire that included the Michigan Alcoholism Screening Test (MAST), a proven alcoholism screening tool. Of the 232 patients interviewed, 47 were identified by MAST as alcoholics. For those patients, researchers measured how good the other questions were in pinpointing a patient's alcoholism.
For example, the question "How much do you drink?" had a relatively low sensitivity of 34 percent. "How often do you drink?" was only slightly better at 47 percent. But "Have you ever had a drinking problem?" yielded a 70 percent sensitivity that rose to 92 percent when asked in combination with the question "When was your last drink?"
The high sensitivity of such straightforward questions was surprising in light of the widespread belief that most alcoholics deny or minimize their drinking problems. In the Rhode Island study, "it was actually the minority of alcoholic patients who denied having a drinking problem."