Kelly Jones1,2. 1. Indigenous Eye Health Unit, Centre for Health Equity, School of Population and Global Health, The University of Melbourne, Melbourne, VIC, Australia. 2. The Australian Research Centre for Population Oral Health, The University of Adelaide, SA, 5005, Australia.
Abstract
OBJECTIVE: To determine the clinical validity and accuracy of an empirically derived triage model when compared with traditional priority assessment by receptionists. BACKGROUND: The use of predictive models to support evidence-based delivery of health care is increasing. Predictive models assist in predicting priority of need or treatment planning for patients and facilitate transparent and consistent decision-making. A predictive model for determining priority of need for dental care was evaluated against a reference standard dentist assessment and traditional receptionist assessment. METHODS: We sampled 310 patients seeking dental care. Participants were selected from people requesting care at two community dental clinics and who agreed to answer eight questions. Receptionists recorded their judgements of participant priority into three categories: care needed <48 h, 2-7 days or 8+ days. The 'reference standard' priority was determined by dental officers using the same categories. Model coefficients generated a predicted probability of requiring care. Sensitivity (Se), specificity (Sp), and predictive (PPV and NPV) and area under the curve (AUC) values were computed for two thresholds, <48 h versus 2+ days and <2-7 days versus 8+ days. RESULTS: At <48-h threshold, the model PPV was higher and NPV lower than receptionists in predicting patients not needing care. At the 2- to 7-day threshold, the model also performed better than receptionists in predicting those needing care for 2-7 days. AUC statistics show the model performed better than the traditional receptionist method. CONCLUSIONS: The predictive model outperformed traditional receptionist screening in predicting priority of care.
OBJECTIVE: To determine the clinical validity and accuracy of an empirically derived triage model when compared with traditional priority assessment by receptionists. BACKGROUND: The use of predictive models to support evidence-based delivery of health care is increasing. Predictive models assist in predicting priority of need or treatment planning for patients and facilitate transparent and consistent decision-making. A predictive model for determining priority of need for dental care was evaluated against a reference standard dentist assessment and traditional receptionist assessment. METHODS: We sampled 310 patients seeking dental care. Participants were selected from people requesting care at two community dental clinics and who agreed to answer eight questions. Receptionists recorded their judgements of participant priority into three categories: care needed <48 h, 2-7 days or 8+ days. The 'reference standard' priority was determined by dental officers using the same categories. Model coefficients generated a predicted probability of requiring care. Sensitivity (Se), specificity (Sp), and predictive (PPV and NPV) and area under the curve (AUC) values were computed for two thresholds, <48 h versus 2+ days and <2-7 days versus 8+ days. RESULTS: At <48-h threshold, the model PPV was higher and NPV lower than receptionists in predicting patients not needing care. At the 2- to 7-day threshold, the model also performed better than receptionists in predicting those needing care for 2-7 days. AUC statistics show the model performed better than the traditional receptionist method. CONCLUSIONS: The predictive model outperformed traditional receptionist screening in predicting priority of care.