Predictive Value of Emergency Physician-coded Cincinnati Prehospital Stroke Scale Scores Based on Witness Reports and Their Association With Action Decision Time: A Cross-sectional Emergency Department Validation Study
Predictive Value of Emergency Physician-coded Cincinnati Prehospital Stroke Scale Scores Based on Witness Reports and Their Association With Action Decision Time: A Cross-sectional Emergency Department Validation Study
The Cincinnati Prehospital Stroke Scale (CPSS) is traditionally scored directly by a healthcare professional or a layperson who independently observes facial droop, arm weakness, and speech disturbance, constituting the primary CPSS score. In contrast, this study focuses on witness-derived secondary CPSS scoring, which is performed indirectly by an emergency physician who is blinded to the patient's physical examination findings and relies exclusively on symptom descriptions provided by witnesses of the patient's initial presentation.
This prospective analytical cross-sectional study aims to evaluate the predictive accuracy of secondary CPSS scores obtained by emergency physicians through structured interviews with witnesses of suspected stroke cases. Interrater reliability will be assessed between secondary CPSS scores independently assigned by a senior emergency medicine resident and an attending emergency physician based on witness-reported symptom descriptions. Secondary CPSS scores will then be compared with final neuroimaging diagnoses (CT/MRI), which will serve as the diagnostic gold standard. The study will also investigate the association between physician-coded secondary CPSS scores, witness demographic characteristics, and delays in seeking emergency medical care.
Research Questions
Research Hypotheses
A total of 235 stroke witnesses, including family members, friends, neighbors, or bystanders who observed the patient's initial symptoms and accompanied the patient to the emergency department with acute stroke or stroke-like symptoms, will be included. All eligible participants will provide written informed consent prior to enrollment. Witnesses will undergo a structured, physician-administered interview in which they will be asked to recall and describe the patient's first symptoms. Based solely on these reports, an emergency physician blinded to the patient's clinical examination will assign a secondary CPSS score focusing on facial droop, arm weakness, and speech impairment. Two investigators will independently score the secondary CPSS to enable interrater agreement analysis.
Participants will also provide demographic data and report three key time points: symptom recognition, decision to seek care, and arrival at the emergency department. These data will be used to calculate onset-to-door time (prehospital delay), recognition-to-action time (decision delay), and call-to-door time (transfer delay). Data collection and emergency care provision will be conducted by investigators independent of those performing secondary CPSS scoring to minimize bias.
Study Design
Prospective, analytical, cross-sectional validation study
Conducted in the Emergency Department of a tertiary academic hospital in Türkiye Two phases
Phase 1:
Interrater reliability assessment: Between a senior emergency medicine (EM) resident and an emergency medicine (EM) attending physician
Phase 2:
Predictive validity assessment The association between secondary CPSS scores, witness demographic characteristics, and decision delays
Phase 1:
Inter-rater agreement between a senior EM resident and an EM attending was assessed using CPSS scoring based on the same witness reports. All witness interviews were conducted face-to-face by trained emergency physicians during the initial ED evaluation. Independent, blinded scoring was performed using CPSS items (0-3 points). Agreement was quantified through percent agreement and Cohen's κ statistics.
Phase 2:
Secondary CPSS scores were compared to final radiological diagnoses. This study investigates the usability and clinical value of witness-derived CPSS scoring through structured interviews conducted by emergency physicians. Given that witnesses are often the first observers of neurological deficits, their descriptions may play a substantial role in early stroke recognition. Recognition-to-action time (decision delay) was obtained from structured witness questionnaires. Witness demographics were recorded.
Data collection used two research forms (RF: RF-1 for reliability assessment; RF-2 for BT/MRI result, witness characteristics and decision delay). All CT/MRI interpretations were performed by independent radiologists. Data were anonymized, coded (HY1, HY2…), and entered daily into SPSS by a research assistant.
Missing or ambiguous time data were clarified during interviews; incomplete CPSS items were excluded. Bias was minimized through consecutive sampling, standardized training, blinding procedures, and a controlled interview environment.
The study follows STROBE guidelines.
Study definitions This section outlines the key terms and operational definitions used throughout the study.
Witness or bystander: Defined as the individual who was with the patient at symptom onset, discovered the patient, or was contacted by the patient.
Onset-to-door time: Defined as the interval from the onset of the patient's symptoms to arrival at the emergency department. Onset-to-door time (ODT) consists of the recognition-to-action time (RAT), determined by the patient experiencing the event or by the stroke witness, and the call-to-door time (CDT), determined by emergency medical services teams. [ODT = RAT + CDT].
Recognition-to-action time: The secondary objectives of this study focus on the recognition-to-action time as determined by the stroke witness. During this interval, the witness undergoes several sequential decision-making processes. Recognizing the suspected stroke symptom(s) constitutes the first decision mechanism; assessing the seriousness of the situation represents the second; inferring whether the condition may be a stroke is the third; deciding whether to call an ambulance or transport the patient to the emergency department is the fourth; and finally, initiating the act of calling EMS-transitioning from decision to behavior-represents the combined decision and action mechanism.
Because, in practice, it is difficult to draw sharp boundaries among these sequential, decision-driven intervals, the time between recognition and taking action was evaluated as a single combined measure in this study. Witnesses were first asked to report the time at which they initially noticed the symptoms, and then the time at which they took action to call an EMS ambulance or transport the patient to the emergency department. The difference between these two time points was defined as the recognition-to-action time.
Delay: In this study, the term prehospital delay (PD) is used to describe delays occurring within the onset-to-door period. Delays within this recognition-to-action interval were referred to as decision delay (DD). The term transfer delay (TD) is used to describe delays occurring within the call-to-door interval. [PD = DD + TD].
Primary CPSS score: Defined as the score obtained directly from the healthcare professional or the layperson (patient or witness), who independently observes the patient's facial droop, arm weakness, and speech using the scale.
Secondary CPSS score: Defined as the score indirectly derived from witness reports by an emergency physician who was blinded to the patient's physical examination and relied on the observations of individuals who witnessed the patient's initial symptoms.
Inclusion Criteria:
Exclusion Criteria: