We describe the development of a method for estimating and modelling future demand for suband non-acute inpatient activity across New South Wales, Australia to 2016. A time series linear regression equation was used, which is consistent with projection models found in the literature.Results of the modelling indicated an increase in rehabilitation, palliative care and maintenance episodes and bed-days. Projections for other categories of care are problematic due to smaller levels of activity and data quality issues. This project indicated a need for ongoing monitoring of type-changing by facilities and management of data quality. Local planners will need to consider a range of factors when considering the applicability activity projections at a local level, particularly THE ACUTE INPATIENT projection tool used by the NSW Health Department, aIM2005, incorporates subacute activity in its modelling, applying the same methodology for acute and subacute activity. 1 This methodology had not been validated for subacute care, which was seen as a key missing element within the planning process. Health service planners, particularly in rural regions, had requested assistance to develop a more consistent and systematic approach to sub-and non-acute care given the disparate nature of this activity.In late 2005, the New South Wales Health Department initiated a project requiring development of a methodology for estimating and modelling future demand for sub-and non-acute inpatient activity across NSW to 2016. 2 The objective of the project was to provide NSW with a well researched and credible projection methodology that can be used to model future demand (and potentially supply) of sub-and non-acute inpatient services. The methodology included consideration of factors that influence overall demand for these services and also the variation in demand across NSW, including: ■ Population demographic effects such as growth and ageing to 2016;