Reproducibility of medical information obtained via the telephone vs personal interview

The Mothersick Program, a teratogenic information service, conducts patient interviews over the telephone as well as In a clinic setting. In both inst ances, medical information and exposure history is obtained from the patien t, including such items as pregnancy history, drug exposure, alcohol and sm oking habits. It occurred on several occasions that the clinic interviewer remarked that the same patient had volunteered information to them, that di ffered from the information documented on the intake telephone form. The ob jective of the study was to establish the levels of agreement in the docume ntation, between these two forms of interviews. Two groups of 100 paired te lephone and clinic forms were randomly selected from our data base, 1990-19 91 and 1996-1997. These two groups were chosen to assess if there were any differences over a five year time period. Statistical calculations were per formed using the Kappa statistic, a method that measures agreement. Kappa s cores indicated high reproducibility for both pregnancy and smoking history , good reproducibility for medical history, marginal-good for exposures (al though excellent agreement was found for the primary drug of concern) and m arginal-good for alcohol information. Overall, agreement was superior in th e 1990-1991 group. There were marked differences in consistency, between th e information recorded on the telephone form and the clinic form, most spec ifically relating to secondary exposures and alcohol history. It suggests t hat a person to person interview yields a more complete medical history tha n a telephone interview.