Matching Records Methods in the Development of a Child Health Profile

The ability to link records across multiple programs is a critical function for each state and local health department as these agencies develop linked data systems to more rapidly respond to public health events.  These linked data systems are also of benefit for our health care delivery partners particularly as public health organizations develop enhanced bi-directional communications with the health care providers in the community.  In spite of this great need for processes to link public health records for an individual, there has been very little peer reviewed information in the public health field on this topic.  In addition, the process to establish a level of trust in the linked public health data so health professionals in the community have confidence that this information accurately represents the individual patient will be addressed.  This presentation will focus on the efforts from the Indiana State Department of Health (ISDH) in the development of a link public health record system.  The system to be described follows a similar practice to established clinical record matching methodologies by applying both rules-based and probabilistic methods to develop the appropriate linkages.  With over 3 million individuals, this methodology correctly links over 99.8% of the birth/death records, over 99% of the hospital discharge records, and over 98% of the New Born Screening records.  Additionally, the framework for establishing clinical trust with the linked data will be addressed.