Accuracy and reliability of the clinical indicators related to hip fractures.

In order to assess the accuracy of the clinical indicators for hip fracture outcomes, we compared the hip fracture outcomes of in-hospital mortality and length of hospital stay as determined by the hospital coded records with that of a prospectively recorded separate hip fracture database over a 5-year period. There was excellent correlation between the two databases with in-hospital mortality figures of 78/1264 versus 79/1299 patients, a difference of 0.1%. For discharge within 28 days the figures for the two databases were 930/1264 versus 987/1299, a difference of 2.4%. The main reason for differences between the databases was caused by ambiguous clinical codes in 66 cases and errors in coding for 57 cases. The next step is to find ways of using these indicators to promote real improvement in the care of these patients.