Uncertain progress in Swiss perioperative mortality 1998-2014 for 22 operation groups.

BACKGROUND

The perioperative mortality rate (POMR) is used as a quality indicator to monitor health care system performance at regional and national levels. The Swiss Federal Office of Public Health publishes national in-hospital mortality rates for several indicator conditions and indicator operation types (IORs). We investigated long-term time trends of POMRs from 1998-2014. In view of continual advances in perioperative care, we expected to find decreasing trends.

METHODS

Non-cardiosurgical IORs containing aggregated age- and sex-specific data (number of operations and deaths) for all years of the study period were included to calculate age-standardised POMRs using the 2013 European Standard Population. We assessed calendar time trends of POMRs using multivariable Poisson regression. We categorised IORs according to the type of time trend (decreasing, unchanged, or increasing incident rate ratio) and mean risk levels (age-adjusted POMR).

RESULTS

A total of 22 IORs were included, comprising 1,561,012 operations and 22,140 deaths (overall crude POMR 1.42%). POMR trends decreased for 6 IORs representing 26.8% of operations, remained unchanged for 13 IORs (56.9% of operations), and increased for 3 IORs (16.4% of operations). IOR categorisation according to POMR trends and to risk levels yielded four groups. (1) Decreasing POMR trends, low- to intermediate-risk IORs (age-adjusted POMR 0.2-2.2%): cholecystectomy; arterial pelvic/leg aneurysm or dissection operation; femoral neck fracture; trochanteric fractures; gastric, duodenal or jejunal ulcer resection; major pulmonary or bronchial resection. (2) Unchanged POMR trends, low-risk IORs (0.1-0.9%): transurethral resection of the prostate (TUR prostate); hernia repair without intestinal operation; hysterectomy; extracranial vascular operation; nephrectomy; amputation foot, non-traumatic. (3) Unchanged POMR trends, intermediate-risk IORs (1.7-3.8%): hernia repair with intestinal operation; gastric carcinoma resection; non-ruptured abdominal aortic aneurysm (open operation); arterial pelvic/leg thromboembolic operation; colorectal resection, pancreatic resection; complex oesophageal procedure. (4) Increasing POMR trends, low- to high-risk IORs (0.1-5.2%): hip endoprosthesis; cystectomy; amputation lower limb. Impact of sex on POMR: hysterectomy and TUR prostate comprised 19.7% of all operations; among the remaining operations, 68.5% showed significantly lower and 27.1% significantly higher POMRs in females. 4.4% showed no sex difference.

CONCLUSIONS

In Switzerland, in-hospital POMR trends from 1998-2014 were unchanged or even increasing for the majority of IORs (73% of included operations). Our analysis used age-standardisation but cannot account for changes in coding practices and organisation of healthcare delivery. POMR trends should be systematically monitored at the national level and used to guide priorities in national quality improvement strategies.