Data-based modeling of the failure rate of repairable equipment.

A database of failures of many types of medical equipment was analysed, to study the dependence of failure rate on equipment age and on time since repair. The intention was to use this large dataset to assess the validity of some widely-used models of failure rate, such as the power-law and loglinear Poisson processes, and so to recommend simple and adequate models to those practitioners having little data to discriminate between rival models. The aim is also to illustrate a methodology for computing policy costs from failure databases. The power-law process model was found to fit slightly better overall than did the loglinear and linear processes. Some related models were created to fit an observed peaking of failure rate. The data showed a decreasing hazard of (first) failure after repair for some equipment types. This can be due to imperfect or hazardous repair, and also to differing failure rates among a population of machines. Two simple models of imperfect repair were used to fit the data, and an Empirical Bayes method was used to fit a model of variable failure rate between machines. Neglect of such variation can lead to an over-estimate of the hazardousness of repair.