Experimental study on Fuzzy Trace Theory and paramedic clinical reasoning

collection

Experimental study on Fuzzy Trace Theory and paramedic clinical reasoning

Fuzzy Trace Theory and Paramedic Reasoning

Dataset

Research School of Psychology

10.25911/kt49-4n92

English

English

This is the first study attempting to experimentally alter clinical reasoning in paramedics, using a theory-based approach.

Background: Observational studies have found paramedics form rapid intuitive impressions on first meeting a patient and these impressions subsequently affected their clinical reasoning with the most experienced paramedics more likely to rely upon their intuition, than on deliberate reasoning. We report an experiment where theory-based interventions are used to alter clinical reasoning by paramedics and paramedic students in simulated patients. Method: Australian paramedics (n = 213; 49% female) and paramedicine students (n = 83; 55% female) attending paramedic conferences completed a 2x2 fully between-participants experiment. They saw a written clinical vignette in which key clinical information was precise or degraded (stimulus), they then either chose the single most likely diagnosis from a list, or ranked competing diagnoses (response). Outcome variables were diagnostic rate and response time. Results: There were no differences in diagnosis rate across the four stimulus-response conditions (0.75 [0.65, 0.84] vs 0.79 [0.68, 0.87] vs 0.78 [0.65, 0.87] vs 0.72 [0.59, 0.82], p = 0.42), and no effect of experience. Conclusion: This is the first study attempting to experimentally alter clinical reasoning in paramedics. Neither of the interventions tested succeeded in altering measures of clinical reasoning. Similar to previous research on physicians, paramedic reasoning is robust to manipulation.

Design We conducted a 2x2 between-subjects experiment with two independent variables: stimulus and response. The primary dependent variable was the proportion of participants diagnosing Acute Coronary Syndrome; the secondary outcome was participant response time. The study was pre-registered on AsPredicted.org (URL to be made public once accepted; private PDF to be provided to reviewers). Ethics The study protocol was considered and approved by the Australian National University Human Research Ethics Committee (2017/141). Participants Participants were Australian registered paramedics and paramedic students studying an accredited undergraduate paramedicine degree at an Australian university. A total of 296 participants were recruited at the Australian and New Zealand College of Paramedicine annual conference in Melbourne, Australia in August 2019, and the Dare to Know Student Paramedic Conference in Bathurst, Australia in September 2019. Table 1 outlines the details of participants. Table 1. Characteristics of participants. Group n (%) Age (median; IQR, years) Gender (n;%) Years of experience (median; IQR) Paramedics 213 (72%) 32 (28-41) Female: 106 (49%) Male: 105 (49%) Other: 2 (1%) 6 (3-14) Year of study Students 83 (28%) 23 (20-30) Female: 46 (55%) Male: 37 (45%) 1st year: 20 (24%) 2nd year: 26 (31%) 3rd year (single degree): 29 (35%) 4th year (combined degree): 1 (1%) Not answered: 7 (8%) Materials Acute Coronary Syndrome (ACS) is an umbrella term for several life-threatening conditions including acute angina pectoris, ST-elevation myocardial infarction, and non-ST-elevation myocardial infarction. We employed a clinical vignette previously designed and used as highly representative of ACS in an out-of-hospital setting (Keene et al., 2022). In that study, the 80% of participants exposed to the vignette recorded a final diagnosis of ACS, providing a useful baseline for this study. The vignette consisted of five diagnostic cues, along with several non-diagnostic cues. The latter were non-diagnostic by virtue of being innocuous or within normal ranges. Non-diagnostic cues were added because they represented information routinely sought as part of a paramedic patient assessment. Two variations of the vignette were created for the stimulus manipulation (Table 2). In the verbatim version, all non-diagnostic cues were precisely defined, to provide a verbatim account of the patient (Verbatim Stimulus; VS). In the gist version, the non-diagnostic cues were presented as a range or categorised as normal versus abnormal (Gist Stimulus; GS). In both cases, the diagnostic cues were presented identically. Following presentation of the stimulus, participants were randomly assigned to one of two response conditions. In the Gist Response (GR) manipulation, participants were asked: “What do you think is wrong with the patient? (Choose one option)”. In the Verbatim Response (VR) manipulation, participants were asked: “What do you think is wrong with the patient? (Drag and drop the options so the most likely is at the top and least likely at the bottom)”. In both manipulations, participants were presented with four response options in random order: Acute Coronary Syndrome; Respiratory Tract Infection; Musculo-skeletal injury; Pulmonary embolism. The survey materials and raw data can be found at (URL). Table 2. Stimulus manipulation. Verbatim Stimulus (VS) Gist Stimulus (GS) You have a 53 year old male patient. He is sitting partially reclined, is alert, oriented, and speaking clearly in complete sentences. You observe he is sweating profusely and he is moving his limbs normally with no loss of power. He reports pain in his chest radiating to the right shoulder and arm. When you examine him, you find him hypotensive. He has a history of peripheral artery disease and a previous abnormal stress test. You note his vital signs as: Pulse 94 regular Respiratory Rate 18 Blood Pressure 90/70 Pulse oximetry 98% (room air) Temperature 37.5C Blood Sugar Level 5.2 Glasgow Coma Score 15 Pain 7/10 ECG: Normal sinus rhythm You have a male patient in his 50s. He is sitting, is alert, oriented, and speaking normally. You observe he is sweating and he is moving all his limbs. He reports pain in his chest radiating to the right shoulder and arm. When you examine him, you find him hypotensive. He has a history of peripheral artery disease and a previous abnormal stress test. You note his vital signs as: Pulse normal Respiratory Rate normal Blood Pressure low Pulse oximetry normal Temperature normal Blood Sugar Level normal Glasgow Coma Score normal Pain severe ECG: Normal Procedure The experimental materials were presented on Qualtrics survey software (Qualtrics, Provo, UT). Participants were recruited during a plenary conference presentation; data collection occurred at the start of the presentation to limit any effect of the presentation itself. Conference attendees were invited to participate and shown an internet link to the experimental materials. No inducements or incentives were offered for participation. Participants accessed the materials on phones or tablets over the internet. On accessing the internet link, participants viewed a participant information sheet and provided consent to participate. They were then randomly assigned to one of four conditions: gist stimulus-gist response (GS-GR); gist stimulus-verbatim response (GS-VR); verbatim stimulus-gist response (VS-GR); or verbatim stimulus-verbatim response (VS-VR). Participants were instructed to take as much time as they wanted to read the vignette. After reading the vignette, participants pressed a “next button” to load the response manipulation. The software recorded the time in seconds from the loading of the response options to the participant’s choice; this is referred to as response time. Following the vignette and response, participants completed a distractor task prior to a conjoint recognition activity. Details of the conjoint recognition activity are described in the supplementary materials and are not discussed further here. Statistical Analysis For an 80% chance of detecting a change of ten percentage points in diagnosis at a baseline of 80% with 0.05 level of significance, 62 participants were required for each group (Cohen, 1988). Statistical analysis was conducted using ‘R’ version 3.4.3 (R Core Team, 2017). Frequency data are reported as count, proportion (95%CI). Continuous variables are reported as median (IQR). Comparisons of categorical data were conducted using Fisher’s Exact Test and continuous variables were compared using the relevant non-parametric test. P-values were set at 0.05 two-tailed, with Bonferroni adjustments for multiple comparisons. In order to analyse the effect of the stimulus-response manipulations on conjoint recognition, the proportions of correct categorisations were calculated for each participant, along with the confidence intervals. These proportions were compared across conditions in order to determine the effect of the manipulation.

toby.keene@anu.edu.au

Research School of Psychology Australian National University Acton, ACT, 2600

+61 402357545

Toby Keene

Kristen Pammer

Eryn Newman; Bill Lord

460102 - Applications in health; 520402 - Decision making

280112 - Expanding knowledge in the health sciences

paramedic; decision-making; dual process theory; fuzzy trace theory; emergency medical services

Applied Research

2019

2022

The Australian National University Data Commons

Open Access allowed (see rights held in and over resource below)

Open

This dataset is made available under the Public Domain Dedication and Licence v1.0. Full text can be found at http://www.opendatacommons.org/licences/pddl/1.0/

CC-BY-NC-SA - Attribution-NonCommercial-SharedAlike (Version 4.0)

Attribution-NonCommercial-ShareAlike - This licence lets others remix, tweak, and build upon your work non-commercially, as long as they credit you and licence their new creations under the identical terms.

7 years

2

902KB

No