ACES Data Quality and Validation Dashboard
ACES, or Acute Care Enhanced Surveillance (ACES), is an Ontario-wide system that monitors hospital registration records in real time to identify potential disease outbreaks or other public health threats. It uses natural language processing to scan the words or phrases included in the “chief complaint” or "reason for admission" and to sort the records into "syndromes". Monitoring syndromes in real time can identify an outbreak much faster than is possible using laboratory results, but syndromes are not clinical diagnoses.
ACES users (epidemiologists and other healthcare analysts) can access ACES data for their regions directly through the ACES interface. Data products, including the Viral Respiratory Mappers and Ontario Acute Care Surge Monitor, make aggregated data available to a wider array of users. The ACES Data Quality and Validation Dashboard provides an overview of ACES data as compared to the gold standard databases (those maintained by the Canadian Institute for Health Information (CIHI)). The objective of this dashboard is to increase acceptance of the validity of the data, and the natural language processing algorithm. There are two sections to the dashboard: a section on data quality and a section of data validation (of the classifier).
The data quality section summarizes the data holdings of ACES, including the hospitals sending data and the coverage rates for visits and admissions. Two data quality measures have been developed (% capture and % extra) to describe the total number of visits/admissions as compared to CIHI databases. This information is presented over time by PHU, hospital and for gender and age groupings. Summary indicators (good, moderate and poor) for each hospital per year are also provided. More information on the Data Quality elements presented can be found on the “Data Quality Summary” and “Data Quality Notes” pages.
The data validation section aims to assess the validity of the ACES natural language processing classifier for key syndromes of public health significance. It presents the findings of analyses that compare daily visit percentages for ACES syndromes with the percentage of visits for the most probable ICD10 codes in CIHI databases. Currently, it is only available for ED visits. More information on the Data Validation elements presented can be found on the “Data Validation Summary” page.