Longitudinal patient data has been cited as the largest single gap in healthcare data. This paper shows how a new, patent pending, patient-centric approach to data collection has been designed, using panel recruitment techniques and multi-source data collection. Data are collected on a personal and frequent basis and this is supplemented by full access to the panelists' medical records. The paper describes the complex analytical issues surrounding data of this type, and shows how sophisticated analytical tools and techniques have been designed to overcome these issues. Discussion also includes the implications of this new methodology for ad hoc research, showing that the new data will increase demand for ad hoc research, but that the techniques used are likely to be very different from those currently employed.