Search

Background: Advance care planning documents, including Physician Orders for Life-Sustaining Treatment (POLST), are intended to guide care near end of life, particularly in emergency situations. Yet, research on POLST during emergency care is sparse. Methods: A total of 7,055 injured patients age ≥ 65 years were transported by 8 emergency medical services (EMS) agencies to 23 hospitals in Oregon. We linked multiple data sources to EMS records, including: the Oregon POLST Registry, Medicare claims data, Oregon Trauma Registry, Oregon statewide inpatient data, and Oregon vital statistics records. We describe patient and event characteristics by POLST status at time of 9-1-1 contact, subsequent changes in POLST forms, and mortality to 12 months. Results: Of 7,055 injured older adults, 1,412 (20.0%) had a registered POLST form at the time of 911 contact. Among the 1,412 POLST forms, 390 (27.6%) specified full orders, 585 (41.4%) limited interventions, and 437 (30.9%) comfort measures only. By one year, 2,471 (35%) patients had completed POLST forms. Among the 4 groups (no POLST, POLST-full orders, POLST-limited intervention, POLST-comfort measures), Injury Severity Scores were similar. Mortality differences were present by 30 days (5.0%, 4.6%, 8.0%, and 13.3%, p < 0.01) and were greater by one year (19.5%, 23.9%, 35.4%, and 46.2%, p < 0.01). Conclusions: Among injured older adults transported by ambulance in Oregon, one in 5 had an active POLST form at the time of 9-1-1 contact, the prevalence of which increased over the following year. Mortality differences by POLST status were evident at 30 days and large by one year. This information could help emergency, trauma, surgical, inpatient, and outpatient clinicians understand how to guide patients through acute injury episodes of care and post-injury follow up.

Conducting out-of-hospital research is unique and challenging and requires tracking patients across multiple phases of care, using multiple sources of patient records and multiple hospitals. The logistics and strategies used for out-of-hospital research are distinct from other forms of clinical research. The increasing use of electronic health records (EHRs) by hospitals and emergency medical services (EMS) agencies presents a large opportunity for accelerating out-of-hospital research, as well as particular challenges. In this study, we describe seven key aspects of designing and implementing out-of-hospital research in the era of EHRs: (1) selection of research sites, (2) defining the patient population, (3) patient sampling and sample size calculations, (4) EMS data, (5) hospital selection, (6) handling missing data, and (7) statistical analysis. We use examples from a recent prospective out-of-hospital cohort study to illustrate these topics, including lessons learned.

Objective: To compare the sensitivity of current field triage practices for identifying high-risk trauma patients to strict guideline adherence, including changes in triage specificity, ambulance transport patterns, and trauma center volumes., Methods: This was a pre-planned secondary analysis of an out-of-hospital prospective cohort of injured children and adults transported by 44 EMS agencies to 28 trauma and non-trauma hospitals in 7 Northwest U.S. counties from January 1, 2011 through December 31, 2011. Outcomes included Injury Severity Score (ISS) ≥16 (primary) and early critical resource use. Strict adherence of the triage guidelines was based on evidence in the EMS chart for patients meeting any current field triage criteria, calculated with and without strict interpretation of the age criterion (<15 or >55 years). Due to the probability sampling nature of the cohort, strata and weights were included in all analyses., Results: 17,633 injured patients were transported by EMS (weighted to represent 53,487 transported patients), including 3.1% with ISS ≥16 and 1.7% requiring early critical resources. Field triage sensitivity for identifying patients with ISS ≥16 increased from the current 66.2% (95% CI 60.2-71.7%) to 87.3% (95% CI 81.9-91.2%) for strict adherence without age and to 91.0% (95% CI 86.4-94.2%) for strict adherence with age. Specificity decreased with increasing adherence, from 87.8% (current) to 47.6% (strict adherence without age) and 35.8% (strict adherence with age). Areas under the curve (AUC) were 0.78, 0.73, and 0.72, respectively. Results were similar for patients requiring early critical resources. We estimate the number of triage-positive patients transported each year by EMS to an individual major trauma center (on average) to increase from 1,331 (current) to 5,139 (strict adherence without age) and to 6,256 (strict adherence with age)., Conclusions: The low sensitivity of current triage practices would be expected to improve with strict adherence to current triage guidelines, with a commensurate decrease in triage specificity and an increase in the number of triage-positive patients transported to major trauma centers.

Background: Transport of out-of-hospital cardiac arrest (OHCA) patients expeditiously to appropriately equipped hospitals is of paramount importance., Objective: We sought to test the correlation of the centroids of geographic units with the actual transport distance for OHCA patients in order to determine the most appropriate surrogate marker of location for future planning, protocol development, and research projects., Methods: This was a prospective, observational analysis of OHCA events in Portland, Oregon. Using geographic information systems (GISs), the locations of OHCA events and receiving hospitals were identified and geocoded for visual inspection and analysis. Transport distance was calculated via network transport distance and Euclidean distance from multiple surrogate markers of location (centroids of ZIP code, census tract, census block group, and census block). Actual distance from the location of the event was then compared with these surrogate markers to determine the accuracy of alternative markers of OHCA location., Results: Two hundred seventy patients had location data recorded, 163 of whom were transported to a hospital for further care. The median transport distance was 5.17 miles. The transport distance of OHCA patients from the centroid of the census block had the best correlation (R(2) = 0.99) with actual transport distance, whereas the use of the centroid of ZIP codes as a surrogate location had the lowest correlation (R(2) = 0.21)., Conclusions: The use of centroids of census blocks via network distance is a valid surrogate for actual location of an OHCA event when calculating transport distance.

Background: Statewide emergency medical services (EMS) data linked to outcomes are critical for promoting high-quality emergency care; however, many states do not have such a resource., Objective: To demonstrate the feasibility of creating such a statewide database using a one-month pilot sample., Methods: This was a prospective cohort study of all EMS patient encounters throughout Oregon during May 2008. Eighty-three National EMS Information System (NEMSIS) variables were obtained from EMS agencies via electronic or paper charts. We reformatted raw data, mapped NEMSIS fields, entered hard-copy records, and uploaded data files to a statewide electronic medical records platform. Records from transport and nontransport (first-responder) agencies caring for the same patients were matched using probabilistic linkage, then linked to three statewide outcome databases (Oregon Hospital Discharge Database [OHDD], Oregon Trauma Registry [OTR], and Oregon Department of Transportation [ODOT] Crash File) using similar methodology. We estimated population-adjusted case ascertainment by county and used descriptive statistics to characterize the process., Results: During the one-month period, we collected 27,474 EMS records in 36 (100%) counties from 106 (77%) licensed transport agencies and 10 nontransport agencies, representing 20,673 persons. There were 3,302 admission record matches, 285 trauma registry matches, and 392 crash record matches. Overall, 3,979 hospital outcomes were matched to EMS records for 80 (75%) transport and six (60%) first-responder agencies. Median per-agency match rates were 16.3% for OHDD (interquartile range [IQR] 8.3-22.2%, range 0-56.5%), 0.9% for OTR (IQR 0-2.5%, range 0-60.0%), and 1.6% for ODOT (IQR 0-3.5%, range 0-23.1%)., Conclusion: Developing a statewide EMS database linked to hospital outcomes is feasible. The processes used in this study and match rate estimates may provide a template for other states to follow, enhancing opportunities for outcomes-based EMS research and EMS quality assurance efforts.