Improving Surgical Ward Care

Objective: To develop a toolkit that covers the clinical, nontechnical, and empathic skills required for effective, safe surgical ward care. Background: Despite the explosion of interest in patient safety, little attention has been placed on the skill set required for safe and effective surgical ward care. Currently, there is a lack of a systematic approach to improving ward care via assessing and improving residents' ward care skills. Methods: A comprehensive evidence-based and expert-derived toolkit was developed, including a novel clinical checklist for ward care (Clinical Skills Assessment for Ward Care: C-SAW-C); a novel team assessment scale for wards rounds (Teamwork Skills Assessment for Ward Care: T-SAW-C); and a revised version of a physician-patient interaction scale (Physician-Patient Interaction Global Rating Scale: PP-GIS). Interassessor reliability (, intraclass correlation), internal consistency (Cronbach ), and convergent validity (Pearson r correlations) were evaluated statistically in 38 simulated scenarios (during which a patient rapidly deteriorated) involving 185 residents. Results: Excellent interassessor reliability was obtained for C-SAW-C [median = 0.89; median intraclass correlation coefficient (ICC) = 0.94], T-SAW-C (median ICC = 0.99), and the revised PP-GIs ( = 1.00; ICC = 0.98 or higher). Internal consistency was also very high for all team skills assessed by T-SAW-C (Cronbach range 0.87-0.94 across 6 skills) and the revised PP-GIS (Cronbach = 0.96)all P's < 0.001. Significant positive correlations were obtained between the 3 assessments (r = 0.73-0.92, P < 0.001) thus showing evidence for convergent validity. Conclusions: We developed a toolkit that captures comprehensively the skills that are required for safe and effective ward care, including the high-risk situation where a patient decompensates. The toolkit offers a systematic evaluation of the quality and safety of surgical ward care and can be used to train and debrief residents' skills and performance.