Navigation

Abstract Evaluation is important to evidence-based policy and practice in Health Information Management. Health classifications are important components of information systems and should be evaluated to determine their suitability for the task required. This paper provides a framework for evaluating health classifications that are used for statistical and reporting purposes. The framework revises and updates the fundamental principles that make health classifications effective. It also draws on other frameworks, where relevant, to reflect the influence that informatics has had on nosology. Principles are illustrated with examples, topical issues associated with some principles are discussed, and examples of evaluation in practice are provided.

Abstract Epidemiological research is worthless without verifiable source data. Much of this data is common to the clinical environment. Currently, substantial resources are allocated to data management bureaucracies in attempts to ensure data accuracy. These bureaucracies developed in the era of paper records, but in the present health information climate, the ability to share electronic data presents exciting possibilities, while placing new responsibilities on the gatherers of information and challenging them to develop new work flow paradigms. Radiation oncologists have a pivotal role to play in the processing of oncological data for future epidemiological research because of the substantial overlap in data requirements.

Abstract
Fungal bloodstream infection (BSI) is of increasing concern in the hospital environment. This study compared routine hospital discharge data at two inner Sydney hospitals with a pathology database over a 6-year period. A high level of underreporting was found, with only 42% of the pathology database cases assigned an appropriate code in the hospital discharge data despite evidence of the infection being found in 97% of the medical records identified from the pathology database. The location of the evidence in the medical record had an impact on whether or not the infection was assigned a code. There was a greater likelihood that a code would be assigned if the infection was documented on the front sheet of the medical record. Improvements can be made to the reporting of fungal BSI if clinicians record it on the front sheet and if coders review the whole medical record before coding.

Abstract The Victorian Birth Defects Register (VBDR) is a population-based surveillance system with a primary function of monitoring trends in birth defects. This paper outlines the processes undertaken in Victoria, Australia, to obtain population prevalence rates of birth defects and investigates the effect on the prevalence rates of variations in collection and processing tasks. It includes all birth defects that were notified to the VBDR by 31 December 2004. The overall prevalence rate of birth defects in Victoria for 2003 was 4.0%, with an overall accuracy rate of 88%. However, this proportion varied according to what birth defects were included, the age by which birth defects were diagnosed, changes to sources of ascertainment, inclusion of terminations of pregnancy, or reporting by cases rate (infants affected) or birth defect rate (individual birth defects). Taking all of these factors into consideration, we are confident that 4.0% is an accurate population prevalence rate of birth defects in Victoria for 2003.