Spanish version of emergency teamwork scale validated across four rater perspectives

A European Spanish version of the Team Emergency Assessment Measure, one of the most widely used tools for evaluating how emergency teams perform, has been translated, culturally adapted and psychometrically validated in a large simulation-based study that also broke new ground by collecting ratings from four different perspectives at once. The work, published in Advances in Simulation, involved 216 high-fidelity cardiology emergency scenarios run within a national training programme for second-year cardiology residents organised by the Spanish Society of Cardiology, and produced 1,414 valid ratings — described by the authors as the largest analytical sample reported in a TEAM validation study to date.

The instrument, originally developed by Cooper and colleagues, assesses non-technical skills in emergency teams, covering leadership, communication, cooperation, coordination, team climate, adaptability, situation awareness, prioritisation and adherence to clinical standards. It comprises 11 items originally grouped into three domains — leadership (2 items), teamwork (7 items) and task management (2 items) — each rated on a five-point scale from 0 (Never) to 4 (Always), giving a total score from 0 to 44, plus a Global Rating Scale from 1 to 10 that captures an overall holistic judgement of team performance. Although the TEAM has robust psychometric evidence across multiple languages and settings, no European Spanish version had been validated before, and no previous TEAM study had simultaneously assessed teamwork from instructors of the learners’ own specialty, instructors from other specialties, self-assessment by performing participants, and peer-assessment by observing peers.

The translation followed internationally recommended guidelines for cross-cultural adaptation. Seven translators produced independent forward translations, including physicians, nurses with critical care and simulation expertise, a professional scientific translator and a professional medical translator. A consensus version was then synthesised by three bilingual group members with cultural adaptation to the European Spanish context, followed by independent back-translations by two professional translators blind to the original instrument. Harmonisation with the original TEAM author resolved a single terminological discrepancy — whether the English title should be rendered as a reference to a tool or to a measurement — and confirmed semantic and conceptual equivalence. The final version was designated s-TEAM, with the 12-item structure fully retained.

The validation setting was a two-day national simulation programme delivered at IAVANTE, a regional simulation centre in Andalusia, over six months and nine course editions. Each course included eight high-fidelity cardiology emergency scenarios, and each resident took part in four scenarios as an active performer and observed four as a peer assessor. Every scenario session lasted about an hour — a 10-minute briefing, a 10-minute simulation and a 40-minute debriefing structured according to the PEARLS framework — and was facilitated by two instructors, one cardiologist and one non-cardiologist. Crucially, all ratings were completed immediately after each scenario and before the debriefing, to avoid bias from the group discussion, and each rater had just two minutes to complete a printed form.

In total, 145 residents and 30 instructors — 15 cardiologists and 15 non-cardiologist specialists from anaesthesiology, angiology and vascular surgery, emergency medicine, family and community medicine, and intensive care medicine — took part. Of 1,541 collected forms, 43 were excluded for illegibility or blank responses and 84 for missing or out-of-range item data, leaving the analytical sample of 1,414 ratings: 192 from instructor-cardiologists, 158 from instructor non-cardiologists, 522 from residents as peer-assessors and 542 from residents as self-assessors. Data collection was fully anonymous, with no demographic data recorded, and no imputation was performed.

The psychometric results were strong. Internal consistency was excellent for the total sample, with Cronbach’s alpha of 0.921 for the 11-item scale, and ranged from 0.901 to 0.938 across all four rater groups. The total score correlated strongly with the Global Rating Scale (r = .816), supporting concurrent validity, and all corrected item-total correlations exceeded the 0.30 retention threshold, with no item deletion improving the overall alpha. Floor effects were negligible, though ceiling effects exceeded 15% on all items, reaching 55.9% for the morale item — a pattern the authors note may reflect the high overall performance of trained teams in simulation settings and could attenuate variance.

Factor analysis, using polychoric correlations appropriate for the ordinal response scale, was conducted on two independent subsamples created by splitting the total sample. Optimal parallel analysis recommended retaining a single factor, and the one-factor exploratory model showed excellent fit, with all 11 items loading substantially (0.633–0.858) and 64.8% of variance explained — a figure virtually identical to the 64% reported in the original development study. However, inspection of standardised residuals revealed a consistent pattern of local dependence among the two leadership items, motivating a two-factor model that was confirmed on the independent subsample. This solution identified a distinguishable but highly correlated Leadership/Communication facet, with interfactor correlations of 0.904–0.908. The authors conclude the scale is best characterised as predominantly unidimensional, supporting the total score as the primary metric while suggesting leadership-related items may warrant specific attention during debriefing.

Two findings stood out. First, peer-assessment — evaluation of a team by residents observing from outside the scenario — produced internally consistent scores (alpha = 0.901) and converged closely with self-assessment, with no significant differences between the two resident groups in total score, domain scores or Global Rating Scale. This is, to the authors’ knowledge, the first time peer-assessment has been examined within the TEAM framework, and it means multi-source feedback can be gathered during debriefing without additional faculty. Second, instructor-cardiologists scored team performance significantly lower than all three other groups (d = −0.42 for total score), with differences concentrated in Teamwork and Task Management, while Leadership ratings did not differ significantly across perspectives. The authors offer two interpretations: specialty-specific expertise may increase scoring stringency, or raters without cardiology-specific expertise may share a degree of leniency. Without an external criterion measure, the design cannot distinguish between the two, but the practical implication is that averaging across perspectives may provide a more balanced assessment.

Notably, reliable measurement was achieved without formal rater training, calibration exercises or practice ratings — only a brief familiarisation of roughly 10 minutes for instructors and 5 minutes for residents. Internal consistency was already excellent from the first scenario (alpha = 0.900) and remained stable throughout. The authors are careful not to claim that rater training is unnecessary: internal consistency does not guarantee calibration, and the specialty-related stringency differences are precisely the kind of effect formal training might mitigate. Total scores also increased progressively across the eight scenarios, from 31.68 in scenario 1 to 37.41 in scenario 8 (r = .266), a trend consistent across all four rater roles and compatible with genuine gains in team performance — though the authors stress this interpretation remains tentative, since no objective clinical performance data were available and shared expectations of improvement over a course may have played a role.

The study has acknowledged limitations: it was conducted at a single simulation centre using cardiology scenarios, the anonymous design precluded formal inter-rater reliability coefficients and multilevel modelling of the nested data, test–retest reliability was not assessed, and the absence of an external criterion measure limited criterion validity to the internal Global Rating Scale. Even so, the authors conclude that the s-TEAM is a valid, reliable and feasible European Spanish adaptation whose consistent scores across four rater perspectives support its integration into simulation curricula as both an assessment and a formative tool — potentially extending structured teamwork assessment to every participant in a simulation activity if the findings are confirmed in other settings.

Subject of Research: Psychometric validation of a European Spanish adaptation of the Team Emergency Assessment Measure in simulated cardiology emergencies

Article Title: Translation, cross-cultural adaptation, and psychometric validation of the Team Emergency Assessment Measure (TEAM) into European Spanish: a multi-perspective study including peer-assessment

Article References: Leon-Castelao, E., Medina-Moya, J. L., Durante-Lopez, A., Castro-Rebollo, P., Cooper, S., Ruiz-Bueno, A., & Bañeras, J. (2026). Translation, cross-cultural adaptation, and psychometric validation of the Team Emergency Assessment Measure (TEAM) into European Spanish: a multi-perspective study including peer-assessment. Advances in Simulation, 11(1), Article 61. https://doi.org/10.1186/s41077-026-00484-1

Image Credits: AI Generated

DOI: 10.1186/s41077-026-00484-1

Keywords: TEAM instrument, psychometric validation, cross-cultural adaptation, simulation-based education, teamwork assessment, peer-assessment, cardiology residents, emergency teams, non-technical skills, internal consistency, factor analysis, Spanish translation