The development and use of a new methodology to reconstruct courses of admission and ambulatory care based on the Danish National Patient Registry

IntroductionThe Danish National Patient Registry (DNPR) contains clinical and administrative data on all patients treated in Danish hospitals. The data model used for reporting is based on standardized coding of contacts rather than courses of admissions and ambulatory care.MethodsTo reconstruct a coherent picture of courses of admission and ambulatory care, we designed an algorithm with 28 rules that manages transfers between departments, between hospitals and inconsistencies in the data, e.g., missing time stamps, overlaps and gaps. We used data from patients admitted between 1 January 2010 and 31 December 2014.ResultsAfter application of the DNPR algorithm, we estimated an average of 1,149,616 courses of admission per year or 205 hospitalizations per 1000 inhabitants per year. The median length of stay decreased from 1.58 days in 2010 to 1.29 days in 2014. The number of transfers between departments within a hospital increased from 111,576 to 176,134 while the number of transfers between hospitals decreased from 68,522 to 61,203.ConclusionsWe standardized a 28-rule algorithm to relate registrations in the DNPR to each other in a coherent way. With the algorithm, we estimated 1.15 million courses of admissions per year, which probably reflects a more accurate estimate than the estimates that have been published previously. Courses of admission became shorter between 2010 and 2014 and outpatient contacts longer. These figures are compatible with a cost-conscious secondary healthcare system undertaking specialized treatment within a hospital and limiting referral to advanced services at other hospitals.