Recent Australian data from the Programme for International Student Assessment (PISA) indicate that around 14% of children are performing poorly on reading. Comparisons with earlier data show that the relative numbers of Australian children with low reading status have not declined in this century. This is despite a surge of robust research documenting evidence-based principles that should guide reading intervention for children who make a poor start in learning to read. In this article, we outline an integrated model of best practice based on empirical research, and contrast this with current practices in schools in the Australian state of Victoria to support poor readers. Practices appear to comply with a model of early identification of reading difficulty and an intensive intervention program in early school years. However, significant shortcomings are apparent in the theoretical orientation of some interventions, the resourcing of reading intervention programs and intensity and continuity of intervention for children with persisting difficulty. We suggest that major changes are required to support vulnerable readers at school.