Validation of ICDPIC software injury severity scores using a large regional trauma registry Nathaniel H Greene, 1 Mary A Kernic, 2 Monica S Vavilala, 3,4,5 Frederick P Rivara 2,4,5 ▸ Additional material is published online only. To view please visit the journal online (http://dx.doi.org/10.1136/ injuryprev-2014-041524). 1 Department of Anesthesiology and Pediatrics, School of Medicine, Duke University, Durham, North Carolina, USA 2 Department of Epidemiology, School of Public Health, University of Washington, Seattle, Washington, USA 3 Department of Anesthesiology and Pain Medicine, School of Medicine, University of Washington, Seattle, Washington, USA 4 Department of Pediatrics, School of Medicine, University of Washington, Seattle, Washington, USA 5 Harborview Injury Prevention and Research Center, Harborview Medical Center, Seattle, Washington, USA Correspondence to Dr Nathaniel H Greene, Department of Anesthesiology and Pediatrics, School of Medicine, Duke University, MD, DUMC Box 3094, Durham, NC 27710, USA; nathaniel.greene@dm.duke.edu Received 16 December 2014 Revised 20 April 2015 Accepted 21 April 2015 Published Online First 18 May 2015 To cite: Greene NH, Kernic MA, Vavilala MS, et al. Inj Prev 2015;21: 325–330. ABSTRACT Background Administrative or quality improvement registries may or may not contain the elements needed for investigations by trauma researchers. International Classification of Diseases Program for Injury Categorisation (ICDPIC), a statistical program available through Stata, is a powerful tool that can extract injury severity scores from ICD-9-CM codes. We conducted a validation study for use of the ICDPIC in trauma research. Methods We conducted a retrospective cohort validation study of 40 418 patients with injury using a large regional trauma registry. ICDPIC-generated AIS scores for each body region were compared with trauma registry AIS scores (gold standard) in adult and paediatric populations. A separate analysis was conducted among patients with traumatic brain injury (TBI) comparing the ICDPIC tool with ICD-9-CM embedded severity codes. Performance in characterising overall injury severity, by the ISS, was also assessed. Results The ICDPIC tool generated substantial correlations in thoracic and abdominal trauma (weighted κ 0.87–0.92), and in head and neck trauma (weighted κ 0.76–0.83). The ICDPIC tool captured TBI severity better than ICD-9-CM code embedded severity and offered the advantage of generating a severity value for every patient (rather than having missing data). Its ability to produce an accurate severity score was consistent within each body region as well as overall. Conclusions The ICDPIC tool performs well in classifying injury severity and is superior to ICD-9-CM embedded severity for TBI. Use of ICDPIC demonstrates substantial efficiency and may be a preferred tool in determining injury severity for large trauma datasets, provided researchers understand its limitations and take caution when examining smaller trauma datasets. BACKGROUND The Centers for Medicare & Medicaid Services (CMS) emphasises and incentivises the adoption of electronic medical records in healthcare allowing researchers to use the power of large medical data- sets. 1 However, data can be stored in a variety of ways and in an inconsistent manner, which can make harvesting and interpreting administrative data for research purposes problematic. Researchers can face a difficult choice of working around potential limitations when analysing administrative datasets or using their limited resources to create datasets for research purposes at the expense of smaller sample sizes. International Classification of Diseases (ICD)-9 codes are usually embedded in large-scale administrative healthcare data giving researchers the ability to identify patients by clinical diagnoses. 2 These issues are especially important when assessing outcomes and quality of care. Historically, the gold standards of injury severity have been the Abbreviated Injury Scale (AIS) 3 and the injury severity score (ISS) (derived from the AIS). The AIS classifies injury into six distinct body regions (head and neck, face, thorax, abdomen, extremities and other) and is an ordinal value between 1 and 6 with an increase in score repre- senting an exponential increase in severity (an abdominal injury severity of 4 is much more than twice as worse as a 2). The ISS uses the worst injury score in three separate body regions (the New ISS uses the three worst injuries, regardless of body region) to calculate a final score. AIS scores are only directly available from trauma registries which employ skilled personnel to review medical records, which can be both time-intensive and expensive. 4 Historically, the first software to produce AIS scores from ICD-9 codes was the ICDMAP, 5 first published in 1989 and made avail- able at a cost; it has not been updated to be current with the latest versions of ICD-9 and AIS coding. More recently, a freeware statistical program imple- mented in Stata statistical software called ICD Program for Injury Categorisation (ICDPIC) has become available. While ICDPIC is free and readily accessible, it has been validated in only one study in the peer-reviewed literature. 6 However, this one study has a small sample size and is limited in its generalisability. Given the potential utility of this freeware tool and its frequent use in the recent medical literature, 7–16 we conducted a validation study in a large trauma population. METHODS Overview of study design This retrospective medical record review was con- ducted to assess the validity of the ISSs generated by the publically available user-generated ICDPIC program 17 implemented in Stata statistical software (Statacorp LP, College Station, Texas, USA). AIS spe- cific to body region and ISS derived by trained trauma nurses based on medical record reviews were considered the gold standards and were compared with AIS and ISS derived from the ICDPIC program. ICDPIC is designed to calculate AIS and ISS from ICD-9 diagnosis codes alone. For traumatic brain injury (TBI) specifically, two methods were compared with the gold standard method of gener- ating head AIS: ICDPIC generated scores and those derived from the 5th digit of ICD-9 diagnosis codes. Study samples and data source The study sample included all trauma-registry- eligible admissions to Harborview Medical Center from April 2005 to August 2012. Paediatric Greene NH, et al. Inj Prev 2015;21:325–330. doi:10.1136/injuryprev-2014-041524 325