The potential of using billing data for emergency department injury surveillance

Development of the BACKIE questionnaire: a measure of children's behaviors, attitudes, cognitions, knowledge, and injury experiences

OBJECTIVE

The objective of this study was to develop a standardized questionnaire (BACKIE) that would assess the Behaviors (B), Attitudes (A), Cognitions (C), Knowledge (K), and Injury Experiences (IE) that elementary-school children possess pertaining to seven types of injuries, including: falls; motor vehicle collisions; burns; drowning; choking/suffocation; poisoning; and bicycle/pedestrian injuries.

METHODS

Over 500 children in grades two through seven completed the questionnaire, with a sub-sample repeating it two months later to assess test-retest reliability of the measure.

RESULTS

Psychometric assessment of the instrument revealed acceptable internal and test-retest reliabilities and results of a Confirmatory Factor Analysis provided support for the hypothesized factor structure.

CONCLUSION

Having a psychometrically sound measure that allows one to assess attitudes, cognitions, and knowledge is an essential first step to exploring the relative influence of these factors on children's risk and safety practices.

Practical introduction to record linkage for injury research

The frequency of early fatality and the transient nature of emergency medical care mean that a single database will rarely suffice for population based injury research. Linking records from multiple data sources is therefore a promising method for injury surveillance or trauma system evaluation. The purpose of this article is to review the historical development of record linkage, provide a basic mathematical foundation, discuss some practical issues, and consider some ethical concerns. Clerical or computer assisted deterministic record linkage methods may suffice for some applications, but probabilistic methods are particularly useful for larger studies. The probabilistic method attempts to simulate human reasoning by comparing each of several elements from the two records. The basic mathematical specifications are derived algebraically from fundamental concepts of probability, although the theory can be extended to include more advanced mathematics. Probabilistic, deterministic, and clerical techniques may be combined in different ways depending upon the goal of the record linkage project. If a population parameter is being estimated for a purely statistical study, a completely probabilistic approach may be most efficient; for other applications, where the purpose is to make inferences about specific individuals based upon their data contained in two or more files, the need for a high positive predictive value would favor a deterministic method or a probabilistic method with careful clerical review. Whatever techniques are used, researchers must realize that the combination of data sources entails additional ethical obligations beyond the use of each source alone.

Evaluating an Inclusive Trauma System Using Linked Population-Based Data

BACKGROUND

Federal and professional programs require "inclusive" trauma systems. We wished to evaluate an inclusive trauma system using administrative data combined from multiple sources.

METHODS

Ambulance reports, outpatient/inpatient discharge data, and/or death certificates were obtained for persons with injury diagnoses who received hospital services and/or died in Maine during 1998 to 2000. Records were unduplicated and joined using probabilistic record-linkage software. Case outcomes, determined from one or more linked records, included place of hospitalization, discharge status, and 30-day mortality.

RESULTS

Per 100,000 population annually, 11,100 injured persons were treated and released, 573 were admitted, and 51.3 died. Trauma centers received 37.0% of major cases directly and another 15.4% in transfer; 51.4% of injury deaths occurred without medical intervention, 21.2% occurred in trauma centers, 20.4% occurred in other hospitals, and 7.0% occurred after discharge from a hospital. Database queries produced comparative hospital statistics and identification of questionable outcomes.

CONCLUSION

Record linkage allows inexpensive description of an inclusive trauma system and may contribute to quality improvement.

Cultural Differences in Young Children's Vulnerability to Injuries: A Risk and Protection Perspective

Interview data from 100 lower income Hispanic and 50 White mothers from a nutritional service clinic extended prior research on cultural differences in the risk for unintentional pediatric injuries. Group differences were expected in reported injury incidence and in the prevalence and impact of contributing factors. As predicted, White mothers reported more injuries for a young child, and among Hispanic mothers, English language preference and use were associated with more reported injuries. Hierarchical regression analysis revealed that risky behaviors, mother's judgment about child compliance, and stressful life events were better predictors of injuries than housing quality, but among Hispanics, the impact of certain factors (e.g. child temperament) was qualified by mother's acculturation level. Stress and child temperament explained injury differences between more- and less-acculturated Hispanic families but only partially accounted for differences between White mothers and less-acculturated Hispanics. Pediatric injury risk and protective factors seemed to operate in concert with cultural processes.

Syndromic analysis of computerized emergency department patients' chief complaints: an opportunity for bioterrorism and influenza surveillance

STUDY OBJECTIVE

Emergency department computerized triage logs might be useful for automated ED surveillance and potentially for early identification of bioterrorism events. We describe a Web-based surveillance program and its feasibility for surveillance.

METHODS

A Web-based surveillance program that receives computerized chief complaint data daily from a large academic urban teaching hospital and performs syndromic analysis on these data was developed. On the basis of preset limits, the Web-based surveillance program sends an alert e-mail message when the syndromic analysis reveals an increase in the number of patients in predefined symptom groups. The feasibility of this system was tested by using historical data during an influenza outbreak (December 1999 to January 2000) and applying the anthrax symptom group.

RESULTS

The Web-based surveillance program identified the influenza outbreak in the first week.

CONCLUSION

Computerized triage logs might be a feasible method for bioterrorism and influenza surveillance. The Web-based nature of the surveillance program creates the opportunity for other hospitals to contribute data, potentially resulting in an automated network of ED computerized triage log surveillance.

Injury surveillance in the ED: design, implementation, and analysis

Comprehensive, population-based surveillance for nonfatal injuries requires uniform methods for data collection from multiple hospitals. To show issues related to design and implementation of multihospital, emergency department (ED), injury surveillance, a city-wide system in the United States is discussed. From October 1, 1995 to September 30, 1996 all injury-related ED visits among District of Columbia residents <3 years of age were ascertained at the 10 hospitals where city children routinely sought care. Information was abstracted from 2,938 injury-related, ED visits (132.7 visits/1,000 person-years). Based on this experience, suggestions to facilitate design of multihospital, injury surveillance in other locations are offered. Importantly, injury-related visits were reliably ascertained from ED logs, and for most variables, a systematic sample of injury-related visits was representative of the total injured population. However, there is a need for more complete documentation of circumstances surrounding injuries and for standardization of data elements on ED logs and treatment records.

Insurance status and admission to hospital for head injuries: are we part of a two-tiered medical system?

Previous studies have shown an association between insurance status and use of resources for inpatient care. We sought to assess whether insurance status influences decisions regarding the evaluation and treatment of head injured patients in the emergency department (ED). Head injured patients were identified from ED data from 4 hospitals reporting to the Kentucky Emergency Medical Services Information System. Multiple regression analysis using admission, ED length of stay, and ED charges as outcome variables was then performed. From 216,137 ED visits there were 8,591 (4%) head injured patients identified from the database. Eliminating those with revisits, transfers to another hospital in the database, and isolated facial lacerations, there were 3,821 cases. Controlling for age, hospital, race, primary diagnosis, and indicators of severity of the injury, insurance status was significantly associated with hospital admission. Those uninsured were the least likely to be admitted (OR 0.41; 95% CI (0.31, 0.50), whereas those with public insurance had an intermediate probability (OR 0.50 95% CI (0.37, 0.68) as compared with those with private insurance. Similarly, ED charges were lower for Medicaid patients than insured patients ($880) and tended to be slightly lower for uninsured patients ($1,043) than insured patients ($1,141) (P =.001). Length of stay in the ED was shorter for publicly insured patients (179 minutes) than uninsured (186 minutes) and privately insured patients (192 minutes) (P =.001). The extent of evaluation and admission for head injured patients is associated with insurance status. This creates a dual standard of care for patients. Practitioners should work to standardize the evaluation of patients independent of paying status.

Injury surveillance in a pediatric emergency department

In this study we have tried to determine physician success in the collection of injury data during the emergency department visit. Prospective data were collected from all children between the ages of 0 to 18 treated for an injury. Data were collected at the time of the visit and by chart review the next day. At an urban, university-affiliated, children's hospital, data were collected on 2,156 injured children. Fifty-one percent of the data forms were completed by the treating physician. Physician completion rate was lower on the weekends (46%) than on weekdays (52%, P = .02). The most common mechanisms of injury were falls (34%), motor-vehicle crashes or pedestrians struck (13%), and nonintentional struck by blunt object (12%). The most common mechanism of injury in all age groups was falls. Our results demonstrate that emergency physicians are not successful data collectors. However, when physician data collection is combined with next-day review of patient records, virtually 100% of patients are captured. Active emergency department data collection is important because in contrast to studies which use hospital discharge and mortality data, we found that overall falls account for more injuries presenting to the ED than transportation-related causes. An active surveillance system in emergency departments that does not require extra work on the part of the treating physician would be ideal and may give a more comprehensive description of the scope of the injury problem.

Hospital trauma registries linked with population-based data

BACKGROUND

We sought to obtain more reliable population-based data for injury epidemiology and trauma system evaluation by linking several sources.

METHODS

In the state of Maine, probabilistic computer methods were used to link data from hospitals contained in a trauma registry for 1995 to 1996 to data from the same years contained in death certificates, ambulance run reports, and hospital discharge abstracts. The most reliable data available from each source were merged to form a standard record for each identifiable case of acute trauma.

RESULTS

A total of 8,924 cases of serious injury were identified that either were in the registry, had a death certificate, or had both an ambulance run report and a hospital discharge abstract. Only 74% of the Trauma Center cases and 33% of the cases overall were contained in the registry. Only 84% of fatal hospitalized cases matched to a death certificate. Incompleteness of the registry and occasional failures to match records from one source to another were attributable to intentional omissions and a variety of human data management problems.

CONCLUSIONS

Combining sources of data for injury victims can produce a resource more descriptive than any single source alone. However, computer-assisted record linkage still requires human review and corrections. Feedback of discrepancies to the individual data sources should further improve the quality of data available for linkage.

Linking large administrative databases: a method for conducting emergency medical services cohort studies using existing data

OBJECTIVE

To evaluate probabilistic matching for linking a cohort of cardiac arrest (CA) patients identified in the Metro Toronto Ambulance (MTA) database in Toronto, Ontario, Canada, to their appropriate record in either the Vital Statistics Information System (VSIS) or the Canadian Institute of Health Information (CIHI) databases and thus establish their clinical outcomes.

METHODS

A linkage of a large administrative database was performed. A cohort of patients who suffered an out-of-hospital CA during the calendar years 1988-1993 was identified. To determine the patients' outcomes, the cohort was probabilistically linked to patient records in the VSIS and CIHI databases. Identifying variables used during the process of linking records included: names (first and last); New York State Identification and Intelligence System (NYSIIS) code; date of event; date of death; city; admitting hospital number; mode of admission to hospital; age; and sex.

RESULTS

A cohort of 7,079 CA patients was identified from the MTA database; 6,448 (91%) patients were accurately linked to records in 1 of the 2 outcome databases (CIHI, VSIS). Missing data for > or = 1 of the linking variables were responsible for unlinked records. Using these longitudinal data, it was possible to determine the number of patients surviving their out-of-hospital CAs to be admitted to hospital (n = 833) (16%). No differences in survival rates (p = 0.06) or median lengths of hospital stay among the survivors (p = 0.15) were observed between admitting hospitals.

CONCLUSIONS

Probabilistic matching is an effective method by which researchers can use existing administrative data to determine outcomes of population cohorts. This is especially valuable in situations where controlled intervention studies are not feasible or may be inappropriate. In this analysis, in-hospital management of admitted CA patients, as determined by hospital-specific survival rates and length of stay, suggests no measurable differences in the care provided to these patients by hospitals in Toronto.