An Empirical Study of the Distribution of Episodes of Illness Recorded in the 1970–71 National Morbidity Survey

Episodes of illness reported by 53 general practices participating in the 1970–71 National Morbidity Survey are shown to follow the negative binomial distribution. The morbidity seen in these practices can, therefore, be summarized by m and k, the two parameters of the negative binomial. This distribution is shown to apply in various subdivisions of the data with varying success. Reasons for observed departures from the model are discussed. These results are consistent with a “proneness” hypothesis for episodes of illness as proposed in earlier papers.