Hysterectomy prevalence and adjusted cervical and uterine cancer rates in England and Wales

Objective To present recent trends in cervical and uterine cancer adjusted for true population at risk, using accurate estimates of the prevalence of hysterectomy where the cervix has been removed or not. To describe trends a nd projections of hysterectomy incidence and prevalence with and without ce rvix removal. Design Collation of available NHS and private sector information. Setting England and Wales. Sample NHS operations from Hospital Inpatient Enquiry, Hospital Episode Sta tistics and Hospital Activity Analysis for England and Wales. Private secto r data from surveys with up to 97% coverage. Methods and Main Outcome Measures NHS data by 5-year age group, year and op eration type were collated for 1961-1995. non-NHS operations for 1981, 1986 , and 1992/3 were back-projected. Hysterectomy incidence rates, 1961-95, we re back-projected to estimate prevalence rates by accumulation. True popula tions at risk of disease and hysterectomy were calculated by applying one m inus the relevant hysterectomy prevalence rates to the population by age gr oup and year. Results When based on the true population at risk, the age standardised cer vical cancer incidence rate in 1992 was 14.4 per 100,000, compared with 12. 6 when based on the all women population estimate. Incidence rates for earl ier years were also affected, but there was no important effect on the rate of change over time. Absolute changes for uterine cancer are greater becau se the true population at risk is proportionally smaller particularly at th e older ages, but there are again no major effects on the rate of change. B y 1995 2.3 million women in England and Wales were without a uterus, with a peak prevalence of 21.3% in the age group 55-59. Projections based on 1995 incidence rates show hysterectomy prevalence for the screened age groups, 25-64, will now fall. Subtotal hysterectomy is 3.5% of operations and incre asing. Conclusions True populations at risk must be used to assess the impact of s creening if further reductions in cervix cancer incidence rates are not to be masked. It is essential to monitor hysterectomy by type, as subtotal hys terectomy is becoming more common. Hysterectomy incidence may have peaked. Hysterectomy prevalence in England and Wales may not be as high as would be estimated from some regional studies.