Inter-rater agreement on the protocol for care and risk classification in obstetrics

Abstract Objective To determine the degree of agreement, sensitivity and specificity of the priority of care determined by inter-rater nurses, based on the use of the protocol for care and risk classification in obstetrics, in an obstetric emergency unit. Method Cross-sectional study with a methodological approach, carried out in a maternity school in Belo Horizonte-MG-Brazil, from September to November 2020. It was carried out in two stages: 1) Documental with an evaluation of the records of nurse classifiers in the medical records of pregnant women, parturients or puerperal women; 2) Interviews with trained and not trained nurses in risk classification. Sensitivity and specificity were analyzed and the Kappa coefficient (k) was used to assess agreement. Results The degree of inter-rater agreement (trained and not trained nurses) was found to be moderate to strong (k= 0.47 and 0.77). There was a tendency to underestimate the red (sensitivity of 85%; specificity of 99%) and yellow priorities (sensitivity of 54%; specificity of 85%), as well as overestimate the green (sensitivity of 62%; specificity of 84%) and blue priorities (sensitivity of 89%, specificity of 98%), although there were no significant differences. Despite satisfactory agreement and specificity, sensitivity was low, due to the rates of underestimation and overestimation in risk classification. Conclusion The protocol is reliable for determining priority of care in obstetrics, but its sensitivity was low when applied to determining priority of care by trained and not trained nurses.