When controversy suddenly erupts about the risk of using a prescription drug, there is an urgent need for fast methods of risk estimation. Some unexpected side-effects of prescription drugs are indications for the prescribing of another kind of drug. If the risk of such a side-effect is high, it should be reflected in clustered prescribing of the side-effect-alleviating drug in sequence with the side-effect-causing drug. The risk of drug-attributable side-effects can be estimated by comparing average incidences of initial prescriptions for the side-effect-alleviating drug before, during, and long after the dispensing of the presumed side-effect-causing drug. We monitored computerized, complete drug dispensing records of anonymous outpatients for use of flunarizine, an anti-vertigo/anti-migraine drug that case reports had suggested causes mental depression and/or Parkinsonism. Among 1284 patients who eventually got flunarizine during a 31 month period, 1 in 7 was started on an anti-depressant before or long after flunarizine; only 1 in 82 might be said to have been started on an anti-depressant because of flunarizine. There was no evidence that anti-Parkinson drugs were started because of flunarizine, though the numbers are small. The analysis takes only a few days, and can help set bounds on risks of the subset of adverse drug reactions that are themselves indications for use of other drugs.