The aim of this study was to analyse the nature and extent of data extracted from case files of deceased individuals in contact with services 6 months prior to drug deaths in Scotland during 2003. A cross-sectional descriptive analysis of 317 case notes of 237 individuals who had drug-related deaths was undertaken, using a data linkage process. All contacts made with services in the 6 months prior to death were identified. Information on clinical and social circumstances obtained from social care, specialist drug treatment, mental health, non-statutory services, the Scottish Prison Service and Criminal Records Office was collated. More than 70% (n = 237) were seen 6 months prior to their drug death. Sociodemographic details were reported much more frequently than medical problems, for example, ethnicity (49%), living accommodation (66%), education and income (52%) and dependent children (73%). Medical and psychiatric history was recorded in only 12%, blood-borne viral status in 17% and life events in 26%. This paucity of information was a feature of treatment plans and progress recorded. The 237 drug deaths were not a population unknown to services. Highly relevant data were missing. Improved training to promote in-depth recording and effective monitoring may result in better understanding and reduction of drug deaths.