7103 Completeness and accuracy of voluntarily reported data to a national case registry of laparoscopic cholecystectomy.

Background: During the first eight years of laparoscopic cholecystectomy, a voluntary registration with participation of all 58 departments in Denmark offering this procedure has been carried out ( Gastrointest Endoscopy ; 48: 332-4), with main objectives being quality assurance on local and national level, and outcomes research including safety issues. Registration is ongoing. The value of clinical databases like this depends on several factors, of which accuracy and completeness are of critical importance. Aims: To validate completeness and accuracy of registry data reported from three randomly chosen departments to The Danish National Registry of Laparoscopic Cholecystectomy. Methods: The contents of all consecutive case reports comprising all laparoscopic cholecystectomies performed over two years in three randomly chosen surgical departments were compared with data reported to the registry. Discrepancies were identified, and complication rates in non-reported (missing) cases were compared with rates in reported cases. Results: 431 of 509 eligible patients had been reported to the registry. Completeness of reporting from the three departments were 69%, 80% and 95%, respectively, significantly higher in the single department with a formalised registration and reporting procedure. Eight filled forms were found in the case notes (all from the same department), and another 8 patients with inaccessible case notes were excluded from the validation. Inaccuracies were noted in 28%, 41% and 51% of the cases, but none regarding serious complications such as bile duct injury or perioperative death. Complication rates or length of stay did not differ between reported and missing cases. The most frequently occurring inaccuracies regarded the indications for surgery (in 6%, 11% and 17% in the respective departments) and the date of discharge (in 1%, 7% and 6%) Conclusion: The information on serious complications in the registry may be accurate provided that the present findings can be extrapolated to all participating departments, but the overall accuracy should be improved. A formalised procedure of handling and forwarding information to the registry is necessary to ensure completeness. Continuous validation, support and motivation to the contributing departments by means of, e.g., visits by registry staff for direct validation with case notes may increase costs, but would result in a higher accuracy and complete registration, and should be accounted for in funding.