The feasibility of linking hospital and police road crash casualty records without names

• Published in: Accident analysis and prevention Vol. 28; no. 2; pp. 271 - 274
• Main Author: Rosman, Diana L.
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.03.1996
• Subjects: Accidents, Traffic - legislation & jurisprudence; Accidents, Traffic - statistics & numerical data; Cross-Sectional Studies; Data Collection; Data Interpretation, Statistical; Feasibility Studies; Hospital Records - statistics & numerical data; Humans; Incidence; Mathematical Computing; Medical Record Linkage; Patient Admission - statistics & numerical data; Record linkage; Road crash; Social Control, Formal; Software; Statistics; Western Australia - epidemiology; Wounds and Injuries - epidemiology; Western Australia; Road crash; Statistics; Record linkage
• DOI: 10.1016/0001-4575(95)00058-5
• ISSN: 0001-4575

Government and other agencies routinely collect complementary information on road crash casualties and there may be opportunities to widen the scope of data available for research and policy evaluation if mechanisms can be established to combine these data sources. Where unit record data are available, direct links within and between individual records can often be achieved using record linkage techniques. Without the benefit of unique identifiers, sufficient other identifying information such as the full name, date of birth and sex of the casualty, their role in the crash and the date of the crash are required to enable correct links between individual records to be made. However, all these data items may not be available and it is therefore important to investigate the effect this may have on the quality and quantity of links between records. The research reported here used the hospital admission and police records from the Western Australian Road Injury Database, which had previously been linked using all the identifying characteristics mentioned above, special purpose software and comprehensive manual checking. This set of linked records were considered to be a “gold standard” which could be used to measure the quantity and quality of links produced using fewer identifying characteristics. Results from this process showed that about 90% of the original links could be identified when a phonetic code of the family name of the casualty was used with age, sex, road user type and crash date. However, only about 50% of the original links were found if linking was performed without using names or phonetic name codes.