Water use in healthcare facilities: consumption patterns, indicators and demand forecasting models – a literature review

This literature review analysed scientific evidence on water use in healthcare facilities (HCFs). The study synthesizes the available knowledge on water end-uses, water consumption indicators, and water demand forecasting models applicable to different geographic and socioeconomic contexts, discussing their applicability and highlighting opportunities and limitations that may guide future research developments. The research was conducted through a systematic review structured according to the PRISMA protocol, encompassing publications indexed in Scopus, Web of Science, and SciELO. The selected studies were grouped into three thematic axes: water end-uses, water consumption indicators, and predictive modelling. Results reveal strong geographic and methodological asymmetries, with a predominance of studies from developed countries featuring continuous monitoring, historical datasets, and integration with water conservation policies, in contrast to case-based and largely descriptive research in developing countries, often limited by data availability. Number of beds and built area independent variables stand out as the most consistent predictors of water consumption, while staff size exerts a complementary influence, improving the adjustment of multiple models. Findings indicate that progress in water efficiency in healthcare facilities depends on the standardization of indicators, the integration of physical and operational data, and in the development of locally adapted water demand forecast models capable of supporting effective water demand management strategies.HighlightsWater end-uses, consumption indicators and forecasting models were reviewed.Water-use intensity reflects service complexity and operational conditions.Bathrooms are the main sources of water used in healthcare facilities.Efficient water management requires integrated infrastructure, routines and management.Forecasting models remain limited due to scarce data and fragmented monitoring.