This article by Crawford and colleagues examines trauma registries across many different applications and countries of origin to determine what barriers to application still exist and what other opportunities they made still hold. The authors conducted a comprehensive literature search and ultimately identified 30 studies examining various registries across applications, subpopulations, and nationalities. Each national registry offers examples of the kinds of changes studying large national databases can bring. In the Netherlands, use of a broad inclusion criteria and standard reporting template (Utstein template) allowed for identification of high-risk groups for further study. Other Scandinavian countries like Sweden also use the Utstein template, demonstrating the effect it can have on data completeness, accuracy, and timely inclusion in the database. The UK uses a best practice tariff, incentivizing even more timely entry of clinical data into the registry. Finally, Australia’s system also includes New Zealand, making it a rare example of a multinational registry; it has been used to inform road safety strategy and targeted injury interventions (e.g. ladder falls).
The authors also use the literature to demonstrate the drawbacks of registries and their associated studies. Most notable is the shift in focus away from mortality and towards long-term outcomes for both patients and families. This transition will better reflect the experienced needs of trauma survivors but highlights a major challenge with registry data – namely that registries are only as good as the data contained within. These data might be limited by funding, difficulties with patient follow up, or other logistical challenges and failure to surveil this aspect of registries can lead to erroneous conclusions. Some registries exclude certain injury patterns associated with frail elderly patients or even elderly patients directly, which may lead to underrepresentation of the societal injury burden. Standardization of inclusion criteria can improve this to some degree, but even registries with standardized data reporting usually don’t follow the recommendation to collect all 35 core variables, suggesting further opportunity to maximize the capabilities of registries. Future registry utilization can be improved by establishing more robust infrastructure and durable funding sources. Registries should also aspire to complete inclusion of all data from all traumatically-injured patients to best reflect the nature of reality; effective registries require more than 95% completeness. Although AI and other emerging technologies might enhance the ability to achieve this, registry stewards will need to view it with a skeptical eye to ensure accurate data processing. As these improvements in standardization proceed, registries available for international collaboration may contribute to stronger data validity and generalizability of findings. Registries will also need to examine better ways to collect long-term outcomes to represent the full consequence of traumatic injury.
This national survey completed in 2024 set out to better understand the evolution of US state-level trauma registries over time. The authors delivered a national survey of trauma registries like one completed 20 years prior and compared their findings. The intent of this study was to illustrate state-level trauma registry maturation over time. Trauma registry managers or emergency medical personnel from all 50 states and the District of Columbia were surveyed regarding state-level infrastructure, data collection, reporting, quality assurance, and sharing. Over the 20-year comparison period, 15 states started new registries for a total of 47 state registries (92%). While many states have shifted to maintain online registries, limitations in data collection and standardization abound. Among the surveyed state registries, 38% mandate data reporting; 72% require manual data abstraction; 60% are not integrated with electronic medical records (EMR); and 43% contribute to national data collection efforts. The authors conclude that while significant progress has been made since 2004, significant gaps around standardization, comprehensive data collection, mandatory reporting, and integration with EMR national databases remain.
This national survey presents compelling evidence. Despite improvement over time, the results highlight persistent limitations of state-level trauma registries. The authors eloquently portray the frustrating limitations of incomplete registry data collection and reporting. They suggest that state trauma registries have opportunity to improve by standardization, modernization, and electronic integration. Reliable, readily available data is often the primary limitation for investigators and is paramount to quality monitoring and process improvement. Without comprehensive, accurate, and actionable data, trauma systems are limited in their capability to understand trends and identify opportunities. The authors suggest that enhanced state-level registries offer unique advantage to the national trauma registry system by enhancing completeness of available data. While registry overhaul is an enormous undertaking, it portends a huge advantage to trauma quality improvement and registry utility. The authors should be lauded for illustrating frustrating gaps in trauma registries and providing a jumping-off-point to support trauma registry evolution.
American College of Surgeons NSQIP benchmarking reports provide quarterly benchmarking data to the participating hospitals for 14 outcomes, including mortality, composite morbidity and 12 other individual complications. Data entered by the participating hospitals includes a primary CPT code. The reporting system allows to enter additional 20 CPT codes for procedures performed. However, benchmarking report only uses the primary CPT code to calculate all 14 performance data points. This does not account for some of the complexity of the procedures performed and therefore causes a certain bias for estimation of operative risks and outcomes. The authors of this paper used a Catboost (CATB) machine learning (ML) algorithm allowing for multiple CPT codes to calculate benchmarking assessments and compared them to the logit score derived from a single principal CPT code.
Standard logit score derivation of NSQIP benchmarking data was applied to the 2023 fiscal year semiannual reporting data, comprising 994,332 patients from 676 hospitals. This data was then compared with CATB method of benchmarking analysis. They found that 26.8°/o of hospitals reported more than 1 CPT code and only 1.3°/o had more than 6 CPT codes. They found some advantage with the CATB method for hospitals submitting more than 1 CPT code. This advantage was more pronounced with higher decile ranking. Hospitals that reported more CPT codes per patient did better in 13 of 14 variables, but the improvement was only significant for 5°/o of the hospitals. For Clostridium difficile outcome, there was no difference. This demonstrates that including multiple CPT codes when reporting data can improve hospital's benchmarking data and ranking, which can result in improved quality metrics and compensation.
Have you renewed your EAST 2026 Membership? Renew Today
This Literature Review is being brought to you by the EAST Manuscript and Literature Review Committee. Have a suggestion for a review or an additional comment on articles reviewed? Please email [email protected]. Previous issues available on the EAST website.
