![[ Visit Client Website ]](images/banner.png)
BACKGROUND: In Florida, staff in 67 county health departments (CHDs) enter data on reportable disease cases into Florida’s secure web-based reportable disease surveillance system, Merlin. State-level case review occurs for most diseases, but not the highest burden diseases such as chronic hepatitis and salmonellosis. With over 1,000 Merlin users, over 20 state-level case reviewers and thousands of cases that are not individually reviewed, data quality is challenging. Merlin features automated logic checks that prevents impossible data from being entered (e.g., birth date after onset date), but does not address improbable scenarios (e.g., patient age over 100 years), which depend on astute state-level case-reviewer to identify and question. In 2013, an additional series of data quality checks was added to Merlin to more systematically address this issue.
METHODS: A systematic review of 2012 case data identified 12 improbable data scenarios that required additional follow-up with counties. In March 2013, a series of data quality checks based on these 12 scenarios was implemented in Merlin. Quality checks occur at the time a CHD indicates a case is ready to be reported to CDC. When a quality check scenario is violated, the user must check an override box acknowledging the data are correct and provide a text explanation.
RESULTS: At the end of 2012, 8,241 cases were identified that violated quality check scenarios; many of which had already undergone state-level case review. These cases were returned en mass to CHD users for clarification, creating significant burden. At the end of 2013, only 634 cases were identified that violated quality check scenarios, representing a 92% decrease.
CONCLUSIONS: Data quality is a key attribute of surveillance systems, and can be challenging for systems with hundreds of users. The addition of quality checks in Merlin significantly reduced the number of data errors in submitted cases, improved data quality, decreased time spent on state-level case review and reduced CHD time spent updating cases. Performing data quality checks at the time of report ensures that accurate data is available for immediate decision-making, rather than waiting until the case review process is completed. Waiting until later in the process to identify data issues reflects poorly on the state and is problematic, as CHD users may not remember or be inclined to research details from cases reported in the past. Harnessing the power of system automation to improve accuracy and consistency of reportable disease surveillance data is efficient and effective and should be considered a best practice.