New Zealand hospital records insufficient for vitamin K study.

Aims: To assess the availability and completeness of existing hospital records, required for a proposed case-control study of vitamin K administration to neonates and childhood cancer.
Methods: We surveyed 44 hospitals, to see whether they had kept records for children born during 1962-87. Additionally, we sought the neonatal records of 44 Dunedin-born people who would be in the proposed study if it went ahead. We abstracted details of vitamin K administration from the records found.
Results: There were 36 responses to our survey (82%). For two-thirds of the hospitals, neonatal records were not available for all of the period 1962-87. The commonest reason for unavailability of records was that they had been destroyed. Neonatal records could be found for 34 (77%) of those born in Dunedin. Although 30 records indicated that vitamin K had been given, only 17 gave the route of administration.
Conclusions: Our proposed study is not feasible because of the poor availability of records and the lack of details recorded. We support the introduction of regulations requiring hospitals to retain health information for a specified time, and we argue that the minimum period for hospitals should be longer than has been proposed.