Christian Roepstorff

Early postoperative mortality following hysterectomy. A Danish population based study, 1977–1981

Summary. The main objective of this cohort study was to analyse the early postoperative mortality after ‘simple’ hysterectomy for benign indications and to compare it with that of a randomly selected reference group of women matched for age. Registry data covering the entire Danish female population were used. Included in the study were all patients operated in the period 1977–1981. Patients were only included if no cancer was diagnosed and if no major co‐surgery was performed (29 192 patients). Cancer patients were also excluded in the reference group (16182 women). Mortality was studied according to characteristics of patients, their residential area, the surgical approach and operating hospital. Overall 47 patients died within 30 days of admission for hysterectomy (overall mortality 16.1 per 10000). Only seven deaths were expected on the basis of the population sample, and adjusted for age, the relative risk (RR) for hysterectomy patients was 6.38 (95% CI 4.33–9.39). Early postoperative mortality increased with age, and the risk was elevated among emergency patients (RR = 3.22; 1.72–6.04). Patients with more than one diagnosis at discharge (RR = 4.53; 2.12–9.70) were at high risk, but early postoperative mortality was independent of surgical approach. Causes of death are discussed. Compared to the general population, patients who undergo ‘simple’ hysterectomy are faced with a sixfold risk of dying within 30 days, but a complete assessment of the risks and benefits of hysterectomy requires prospective studies of survival and morbidity, including quality of life for longer periods of time following operations.

Complications after hysterectomy : a Danish population based study 1978-1983

We studied complications after hysterectomy among all women in the Danish population who had a simple hysterectomy in the period 1978‐81 based on data obtained from the Danish National Hospital Registry. Among patients, with neither diagnosed cancer nor major co‐surgery (n = 23,386), we identified all the complications which occurred during hospital admission from the time of surgery up to six years from that point. Within 30 days of hysterectomy 2.6% of the patients had been diagnosed in hospitals as having complications according to our definition. The corresponding figures at 90 days and two years after the operation were 3.7% and 9.4%. The most frequently observed complications were post operative wound infections and bleeding, each affecting about 2% of all operated women. Logistic regression and Cox regression were used to identify prognostic indicators of readmission with complications. The probability of readmission with complications within six years after hysterectomy was estimated at 8% among low risk patients. The most pronounced increase in risk of readmission with complication occurred among women who had been admitted to psychiatric or somatic hospitals 0–12 months before they had their uterus removed (OR in the range 1.59 to 1.83). We discuss the prevailing difficulties of comparing observational evidence from different clinical settings reported in the literature, and emphasize the importance of developing a coordinated international strategy for non‐experimental assessment of medical technology.

Prostatectomy in Denmark: Regional Variation and the Diffusion of Medical Technology 1977-1985

In many countries prostatectomy is one of the most common surgical operations in elderly men. We used administrative data for the entire male population of Denmark to study temporal and regional variations in the use of prostatectomy from 1977 to 1985. The total annual number of prostatectomies increased by 43% during the period, when the transurethral procedure (TURP) gradually replaced traditional open surgery. TURP accounted for 56% of all operations in 1977 but increased its share to 92% in 1985. Substantial regional variations occurred with index values for prostatectomy in 72 recruitment areas ranging from 0.56 to 1.62 (SCV x 100 = 5.3). The amount of variation decreased during the process of technology diffusion, but remained at a relatively high level (as in other countries) even after the process had been completed (SCV x 100 = 5.1).