Methodological issues in health human resource planning: cataloguing assumptions and controlling for variables in needs-based modelling.

Health Human Resource Planning (HHRP) models approximate future nursing requirements based on a variety of factors specific to the model being employed. There is an urgent need to develop a better understanding of the sources of bias in statistical modelling in order to ensure that we are guided by accurate and robust formulae. This paper addresses these issues as they apply in the context of needs-based HHRP research for nursing by presenting a review and discussion of the relevant literature as it relates to: (1) the testing of assumptions, (2) avoiding ecological and atomistic fallacies, (3) how need is directly or indirectly related to health care, and (4) alternatives to aggregate analysis for assessing the relationship between health needs and utilization of nursing services. The paper concludes that multilevel modelling is useful for the simulation analysis of individuals and their ecologies, and that small area variation modelling holds promise for assessing the relationship between health needs and utilization of nursing services.