Improving Inferences Regarding Patient Events Using Emergency Medical Services Response-Based Data.

Area of Science:

• Emergency Medicine
• Health Services Research
• Data Science

Background:

• The National Emergency Medical Services Information System (NEMSIS) public release data is valuable for EMS research.
• Current data organization by EMS response poses challenges for patient-level analysis due to multiple EMS units attending single patients.
• Accurate patient-level event inference is crucial for effective emergency medical services research.

Purpose of the Study:

• To evaluate if 911 call time and patient characteristics can effectively screen for multiple EMS records representing a single patient encounter.
• To improve the accuracy of patient-level data derived from the NEMSIS public release dataset.
• To address challenges in EMS research caused by duplicate response records.

Main Methods:

• A cross-sectional study utilizing 2024 New York City EMS response data for assaults.
• Identified EMS responses with identical 911 call time, patient age, sex, race/ethnicity, and location (longitude-latitude).
• Assessed the validity of matching on call time, demographics, and location to screen for single patient encounters.

Main Results:

• Out of 32,202 EMS responses, 5,143 (approximately 16%) matched other responses across all variables, indicating 26,451 unique patient encounters.
• Matching on 911 call time, patient age, sex, race, and ethnicity achieved 100% sensitivity.
• The specificity ranged from 91.3% to 98.6% for identifying responses that also matched on location.

Conclusions:

• The NEMSIS public release data can be utilized to screen for duplicate EMS responses.
• This screening method has the potential to significantly improve patient-level event inferences in EMS research.
• Utilizing readily available NEMSIS data offers a practical approach to enhance data quality for research purposes.