Leveraging Narrative Feedback in Programmatic Assessment: The Potential of Automated Text Analysis to Support Coaching and Decision-Making in Programmatic Assessment.

Introduction: Current assessment approaches increasingly use narratives to support learning, coaching and high-stakes decision-making. Interpretation of narratives, however, can be challenging and time-consuming, potentially resulting in suboptimal or inadequate use of assessment data. Support for learners, coaches as well as decision-makers in the use and interpretation of these narratives therefore seems essential. Methods: We explored the utility of automated text analysis techniques to support interpretation of narrative assessment data, collected across 926 clinical assessments of 80 trainees, in an International Medical Graduates' licensing program in Australia. We employed topic modelling and sentiment analysis techniques to automatically identify predominant feedback themes as well as the sentiment polarity of feedback messages. We furthermore sought to examine the associations between feedback polarity, numerical performance scores, and overall judgments about task performance. Results: Topic modelling yielded three distinctive feedback themes: Medical Skills, Knowledge, and Communication & Professionalism. The volume of feedback varied across topics and clinical settings, but assessors used more words when providing feedback to trainees who did not meet competence standards. Although sentiment polarity and performance scores did not seem to correlate at the level of single assessments, findings showed a strong positive correlation between the average performance scores and the average algorithmically assigned sentiment polarity. Discussion: This study shows that use of automated text analysis techniques can pave the way for a more efficient, structured, and meaningful learning, coaching, and assessment experience for learners, coaches and decision-makers alike. When used appropriately, these techniques may facilitate more meaningful and in-depth conversations about assessment data, by supporting stakeholders in interpretation of large amounts of feedback. Future research is vital to fully unlock the potential of automated text analysis, to support meaningful integration into educational practices.

Composite Reliability of Workplace Based Assessment of International Medical Graduates.

This article was migrated. The article was marked as recommended. Introduction All developed countries depend on International Medical Graduates (IMGs) to complement their workforce. However, the assessment of their fitness to practice and acculturation into the new system can be challenging. To improve this, we introduced Workplace Based Assessment (WBA), using a programmatic philosophy. This paper reports the reliability of this new approach. Method Over the past 10 years, we have assessed over 250 IMGs, each cohort assessed over a 6-month period. We used Mini-Cex, Case Based Discussions (CBD) and Multi-Source Feedback (MSF) to assess them. We analysed the reliability of each tool and the composite reliability of 12 Mini-Cex, 5 CBDs and 12 MSF assessments in the tool kit. Results A reliability coefficient of 0.78 with a SEM of 0.19 was obtained for the sample of 236 IMGs. We found the MSF to be the most reliable tool. By adding one more MSF to the assessment on two occasions, we can reach a reliability of 0.8 and SEM of 0.18. Conclusions The current assessment methodology has acceptable reliability. By increasing the MSF, we can improve the reliability. The lessons from this study are generalisable to IMG assessment and other medical education programs.

The reliability of a portfolio of workplace-based assessments in anesthesia training.

PURPOSE: Competency-based anesthesia training programs require robust assessment of trainee performance and commonly combine different types of workplace-based assessment (WBA) covering multiple facets of practice. This study measured the reliability of WBAs in a large existing database and explored how they could be combined to optimize reliability for assessment decisions. METHODS: We used generalizability theory to measure the composite reliability of four different types of WBAs used by the Australian and New Zealand College of Anaesthetists: mini-Clinical Evaluation Exercise (mini-CEX), direct observation of procedural skills (DOPS), case-based discussion (CbD), and multi-source feedback (MSF). We then modified the number and weighting of WBA combinations to optimize reliability with fewer assessments. RESULTS: We analyzed 67,405 assessments from 1,837 trainees and 4,145 assessors. We assumed acceptable reliability for interim (intermediate stakes) and final (high stakes) decisions of 0.7 and 0.8, respectively. Depending on the combination of WBA types, 12 assessments allowed the 0.7 threshold to be reached where one assessment of any type has the same weighting, while 20 were required for reliability to reach 0.8. If the weighting of the assessments is optimized, acceptable reliability for interim and final decisions is possible with nine (e.g., two DOPS, three CbD, two mini-CEX, two MSF) and 15 (e.g., two DOPS, eight CbD, three mini-CEX, two MSF) assessments respectively. CONCLUSIONS: Reliability is an important factor to consider when designing assessments, and measuring composite reliability can allow the selection of a WBA portfolio with adequate reliability to provide evidence for defensible decisions on trainee progression.

RéSUMé: OBJECTIF: Les programmes de formation en anesthésie basés sur les compétences nécessitent de solides évaluations des performances des stagiaires et combinent habituellement des évaluations sur le lieu de travail (ÉLT) couvrant de nombreux aspects de la pratique. Cette étude a mesuré la fiabilité des ÉLT dans une grande base de données existante et a exploré comment elles pourraient être combinées pour accroître leur fiabilité pour des décisions sur les évaluations. MéTHODES: Nous avons utilisé la théorie de la généralisation pour mesurer un critère composite de fiabilité de quatre types d'ÉLT utilisés par les collèges d'anesthésiologistes d'Australie et de Nouvelle-Zélande : un exercice de mini-évaluation clinique (mini-CEX), l'observation directe des habiletés procédurales (DOPS), une discussion de cas (CbD) et une rétroaction de multiples sources (MSF). Nous avons alors modifié le nombre et la pondération des combinaisons d'ÉLT pour optimiser la fiabilité avec moins d'évaluations. RéSULTATS: Nous avons analysé 67 405 évaluations de 1 837 stagiaires et 4 145 assesseurs. Nous avons supposé une fiabilité acceptable pour les décisions intérimaires (enjeux intermédiaires) et définitives (enjeux élevés) à, respectivement, 0,7 et 0,8. Selon la combinaison des types d'ÉLT, 12 évaluations ont permis d'atteindre le seuil de 0,7 lorsqu'une évaluation de chaque type a le même poids, alors qu'il en a fallu 20 pour que la fiabilité atteigne 0,8. Si la pondération des évaluations est optimisée, la fiabilité acceptable pour les décisions intérimaires et finales est possible avec, respectivement, neuf évaluations (p. ex., deux DOPS, trois CbD, deux mini-CEX, deux MSF) et quinze évaluations (p. ex. deux DOPS, huit CbD, trois mini-CEX, deux MSF). CONCLUSIONS: La fiabilité est un facteur important dont il faut tenir compte quand on conçoit les évaluations et la mesure d'une fiabilité composite permet la sélection d'un éventail d'ÉLT avec une fiabilité adéquate pour l'obtention de données probantes et la défense de décisions sur les progrès des stagiaires.

The reliability of multisource feedback in competency-based assessment programs: the effects of multiple occasions and assessor groups.

PURPOSE: Residency programs around the world use multisource feedback (MSF) to evaluate learners' performance. Studies of the reliability of MSF show mixed results. This study aimed to identify the reliability of MSF as practiced across occasions with varying numbers of assessors from different professional groups (physicians and nonphysicians) and the effect on the reliability of the assessment for different competencies when completed by both groups. METHOD: The authors collected data from 2008 to 2012 from electronically completed MSF questionnaires. In total, 428 residents completed 586 MSF occasions, and 5,020 assessors provided feedback. The authors used generalizability theory to analyze the reliability of MSF for multiple occasions, different competencies, and varying numbers of assessors and assessor groups across multiple occasions. RESULTS: A reliability coefficient of 0.800 can be achieved with two MSF occasions completed by at least 10 assessors per group or with three MSF occasions completed by 5 assessors per group. Nonphysicians' scores for the "Scholar" and "Health advocate" competencies and physicians' scores for the "Health advocate" competency had a negative effect on the composite reliability. CONCLUSIONS: A feasible number of assessors per MSF occasion can reliably assess residents' performance. Scores from a single occasion should be interpreted cautiously. However, every occasion can provide valuable feedback for learning. This research confirms that the (unique) characteristics of different assessor groups should be considered when interpreting MSF results. Reliability seems to be influenced by the included assessor groups and competencies. These findings will enhance the utility of MSF during residency training.