Role of key informants and direct patient interviews in epidemiological studies of substance abuse.

PURPOSE: The present study was undertaken to assess the degree to which key informants' perceptions of relative abuse liability of opioids in their communities correspond to actual drug use in their patients.
METHODS: Key informants (n = 49) and their patients (n = 738) were asked to complete anonymous, independent paper surveys about perceived or actual abuse of nine targeted opioid drugs: buprenorphine, fentanyl, heroin, hydrocodone, hydromorphone, methadone, morphine, oxycodone and tramadol.
RESULTS: There was a strong correspondence (R = 0.98) between KI rankings of relative abuse of opioids in their patient population and the actual rank order calculated from their patients' selection of a primary euphorigenic opioid. The patient surveys not only validated the KI's conclusions, but also provided a great deal of important information about the nature of the abuse patterns. Notably, significant co-morbidity was found in the patient population: 55% had a history of psychiatric problems; chronic pain of moderate intensity was common; 67% were nicotine dependent; and 43% were diagnosed alcoholics. DISCUSSION: Our results demonstrate that KI's perceived rank order of drugs of choice correlated almost perfectly with their patients' choice of a primary drug. These data suggest that key informants, as they have for decades in epidemiological research, provide a reliable, high-level view of substance abuse in their catchment area. As such, KIs provide valid and reliable information on scope and patterns of abuse, but if more detail is required about the characteristics of abuse, such as that required for FDA-mandated Risk Evaluation and Mitigation Strategies, then direct interviews with the abusers themselves seem more appropriate.