A new algorithm for hip fracture surgery

Abstract

Background and purpose Treatment of hip fracture patients is controversial. We implemented a new operative and supervision algorithm (the Hvidovre algorithm) for surgical treatment of all hip fractures, primarily based on own previously published results.

Methods 2,000 consecutive patients over 50 years of age who were admitted and operated on because of a hip fracture were prospectively included. 1,000 of these patients were included after implementation of the algorithm. Demographic parameters, hospital treatment, and reoperations within the first postoperative year were assessed from patient records.

Results 931 of 1,000 operative procedures were performed according to the algorithm, as compared to only 726 of 1,000 prior to its introduction (p < 0.001). After implementation of the algorithm, junior registrars still performed half of the operations, but unsupervised procedures declined from 192 of 1,000 to 105 of 1,000 (p < 0.001). The rate of reoperations declined from 18% to 12% (p < 0.001 in a multiple Cox regression analysis), with a decline of 24% to 18% for intracapsular fractures and a decline of 13% to 7% for extracapsular fractures. The proportion of bed-days caused by reoperations was reduced from 24% of total hospitalization before the algorithm was introduced to 18% after it was introduced.

Interpretation It is possible to implement an algorithm for treatment of all hip fracture patients in a large teaching hospital. In our case, the Hvidovre algorithm both raised the rate of supervision and reduced the rate of reoperations. The reduced reoperation rate saved many hospital bed-days.

All authors took part in the planning and design of the study and all revised and approved the final manuscript. Hypothesis (HP, HK). Suggestion of algorithm (HP) and development of algorithm (HP, MK, NF). Fracture classification, surgery, supervision, and daily rounds (HP, MK, KH, TL). Statistical analysis and writing of the manuscript (HP).

We thank bio-statistician Janne Petersen, Clinical Research Center, Hvidovre University Hospital for statistical support.