Brief communication and research noteThe feasibility of linking hospital and police road crash casualty records without names
References (16)
- et al.
The Construction of a road injury database
Accid. Anal. Prev.
(1993) - et al.
Application of Exact ODDS for partial agreements of names in record linkage
Comput. Biomed. Res.
(1991) Matching fatal accident reporting system cases with National Center for Health Statistics motor vehicle deaths
Accid. Anal. Prev.
(1989)- et al.
The relationship of the retail availability of alcohol and alcohol sales to alcoholrelated traffic crashes
Accid. Anal. Prev.
(1995) - et al.
A generalized iterative record linkage computer system for use in medical follow-up studies
Comput. Biomed. Res.
(1981) - et al.
A time comparison study of the New York State Safety Belt Use Law utilizing hospital admission and police accident report information
Accid. Anal. Prev.
(1990) - et al.
The California Automated Mortality Linkage System
Am J. Public Health
(1984) Record linkage
Am J. Public Health
(1946)
Cited by (22)
A systematic review and meta-analysis of data linkage between motor vehicle crash and hospital-based datasets
2024, Accident Analysis and PreventionFifty Years of Accident Analysis & Prevention: A Bibliometric and Scientometric Overview
2020, Accident Analysis and PreventionCitation Excerpt :In AA&P, based on the construction of an RTI database (Ferrante et al., 1993), considerable related work has been conducted, and produced valuable results. It has included comparisons of hospital and police RTI data (Rosman and Knuiman, 1994; Aptel et al., 1999), linkages between hospital and police records (Rosman, 1996; Rosman, 2001), and assessments of under-reporting of RTI data (Alsop and Langley, 2001; Amoros et al., 2006; Watson et al., 2015; Huang et al., 2017; Ahmed et al., 2019). It should be noted that issues of under-reporting (especially for less-severe injuries) affect both crash-frequency and crash-injury severity analyses (clusters 2 and 5), since omitting minor crashes, and having a bias toward severity in the model parameters, could lead to erroneous inferences (Yamamoto et al., 2008; Yasmin and Eluru, 2013; Mannering and Bhat, 2014).
Classifying, measuring and improving the quality of data in trauma registries: A review of the literature
2016, InjuryCitation Excerpt :Specific statistical methods to improve data capture in trauma registries operated concurrently with the methods used to measure data capture. Linkage methods were either “deterministic” [87], where cases were merged if there was an exact match (unlikely), or “probabilistic” [39,43,83,87] (preferred), where various weights were applied according to the level of matching. Working in tandem with dataset linkage, “capture-recapture” methods were used to estimate the eligible population and the proportion of these eligible cases actually included in the registry [32,43,80,83,86,87].
Pedestrian injuries in eight European countries: An analysis of hospital discharge data
2010, Accident Analysis and PreventionCitation Excerpt :We are missing fatalities on-the-spot and pedestrians that have been admitted to hospital but were never recorded as pedestrians in the hospital discharge data. The need of the linkage of different trauma and crash databases have been pointed out repeatedly in the literature (Loo and Tsui, 2009; Cryer et al., 2001; Rosman, 1996). Also, we do not claim that our participating countries are representative from Europe at large.
Using data linkage to generate 30-day crash-fatality adjustment factors for Taiwan
2006, Accident Analysis and PreventionSan Francisco pedestrian injury surveillance: Mapping, under-reporting, and injury severity in police and hospital records
2005, Accident Analysis and PreventionCitation Excerpt :The absence of specific pedestrian identifiers limited our ability to match pedestrians across datasets. Rosman (1996) reported that 50% of the correctly linked hospital and road casualty records were identified in absence of names or phonetic name codes when compared with a 90% linkage when these personal identifiers were used. We suspect that we obtained a greater match rate than 50% because our hospital data was from a single main source (SFGH, the trauma center for the city), and was very specific—patients admitted within a limited time frame who had an ICD9 E-code indicating a traffic-related pedestrian injury.