Declining trend in cervical cancer incidence in Bombay, India (1964-1985).

Population-based incidence rates for cervical cancer in Bombay showed a decline over the past two decades. The observed rates fitted to a log linear regression model showed a good fit. In an attempt to explain the observed decline, cohort-specific age incidence curves were drawn, cumulative incidence rates over common age ranges estimated, and the log linear model fitted separately for each of the religious groups that showed diverse age-adjusted rates for cervical cancer. It was found that the decline could be explained as a cohort effect, indicating that it was due to an upward shift in age at marriage in the younger cohorts. Furthermore, a significant decline was observed in Hindus in whom the mean age at marriage has increased from 12 years in 1921-1931 to about 17 years in the 1960s, but not in the Christians who even in 1911 had a mean age at marriage of 17 years. It appears that, in India, cervical cancer rates will continue to decline until 1995 after which it will stabilise unless there is a change in other risk factors.