Study of disease associations from linked records.

Comparisons are made between the principal diagnosis on discharge from hospital in successive periods of inpatient care for persons in the Oxford Record Linkage Study area admitted over a period of years. The observed numbers of pairs of diagnoses are compared with expected numbers computed to take account of the discharge rates in the population by age and sex and the number of man years of exposure. Three topics have been selected to display some of the types of analysis possible with the very extensive material: mental disorder and diseases of the central nervous system, hospital discharges preceding those for neoplasm, and discharges following those for tuberculosis. The details of the method of calculation of the observed and expected numbers and the assumptions and approximations involved are given in an Appendix.