Automating Quality and Safety Benchmarking for Children: Meeting the Needs of Health Systems and Patients

Our aim is to implement and test electronic measures of outpatient pediatric healthcare quality and safety using the PCORI Common Data Model.The PCORnet Common Data Model is a way of organizing data, including data mined directly from the electronic health record, into a standard structure. This research will take place in two PCORI funded networks which use the PCORI Common Data Model: PEDSnet and OneFlorida. This will be the first study to use the PEDSnet and OneFlorida data, and will include analysis of the data pulled directly from the electronic health record for over 5 million children, and validation of this data using a manual review of a sample of approximately 500 charts. We will test the creation of a benchmarking report so health systems can compare their performance to others. Without this information, hospitals do not know if they have the best quality care or the worst, and cannot learn from each other, which limits improvement. We will focus on 3 well-established measures of quality: (1) transcranial Doppler screening for children with sickle cell disease; (2) appropriate antibiotics for ear infections; and (3) cholesterol and glucose testing for children on antipsychotic medications. Only 30% of children with sickle cell disease receive transcranial Doppler screening, which identifies high risk for stroke and initiates treatments that reduce risk by 92%. Only 70% of children with ear infections receive the appropriate antibiotics. Over 600,000 children take antipsychotic medications which greatly increase the risk for diabetes and high cholesterol; only 19% of these receive glucose and cholesterol tests.

We aim to: 1.Using the PCORnet Common Data Model, implement and evaluate electronic measures of pediatric safety and quality for three high-priority conditions. 2.Determine the value of electronic measurement benchmarking to health system leaders and parents. We will implement the three measures using the PCORnet Common Data Model and then evaluate the validity of these electronic measures using manual chart review on a sample of charts. We will also evaluate the ability of the PCORnet Common Data Model to correctly identify diagnoses, medications, and radiographic and laboratory testing. Our study staff includes a Health System Leader Advisory Panel and a Parent Advisory Panel who will advise the scientific and data teams for the development of a mock-up benchmarking report, and will assist in interpretation of findings from the interviews and focus groups. In addition, we will interview health system leader study participants and perform skype/phone focus group meetings with parents to understand who will use this data and how to make it as useful as possible.