Why Validate?
Relying on data in healthcare has increased exponentially in recent years, but trust in data doesn’t match that level reliance. Data validation can explain why the data are wrong, if they’re wrong, and uncover what is required to correct inaccuracies and inconsistencies. Data validation can also prove – surprise! – that the data is right and the expectation or perception of what the data should look like is skewed.
Trust in the data increases when errors are promptly addressed and resolved. Those involved in validation not only have a better understanding of the data, the measures, the documentation workflow, etc., they see the fruits of their labor and are motivated to find more inaccuracies to improve the data further.
Data accuracy can improve due to siloed IT fixes in a limited way, but clinical expertise and knowledge of documentation workflows are required to comprehensively review and correct the inevitable issues that exist with all healthcare data reports. Data validation activities require interchange and communication between clinical teams and the IT staff or vendors responsible for developing and maintaining critical reports.
Not only do data validators uncover issues with reports and workflow documentation, the process of validation inevitably leads to a deeper understanding of measures and evidence-based guidelines. This benefits the entire organization; physicians, quality and performance improvement teams, and all staff involved in documentation of patient data, from the front desk to the back office, and exam rooms to billing.
Sometimes the data in the reports is accurate enough, but it’s not usable. Data validation efforts should include a review of the report accessibility and format, with feedback to report developers about how the data could be more actionable or could be presented in a way that is easier for users to understand.
A classic example is ‘drill-down’ functionality. When a report user views a report and sees a performance percentage value, the user should be able to click on that number and generate a numerator, denominator, and patient list.
As report development matures, consider which reports are relevant and most actionable by role. Usability of specific reports, prioritization of focused dashboards, and drill-down capability may vary significantly by the needs of executive or practice leaders, physicians, and high-risk care managers.
Almost all healthcare providers are now involved in some level of value-based payment, in which compensation is based not only on fee-for-service but also on the quality of care delivered (i.e., pay for performance, QPP, etc.). A smaller percentage of practices are involved in shared risk models, in which demonstrating clinical quality of care has even more of an impact on payment. Assessing, improving, and demonstrating quality of care requires reporting on clinical quality measures extracted from an EHR or provided by a third party. Data validation helps ensure that the good work healthcare providers do is captured in submitted reports that dictate payment for quality performance.
Reports are increasingly used to help guide patient care at the time of the visit, to facilitate outreach to patients between visits, to risk stratify and coordinate care for patient populations, drive decision support, and assist with a host of other patient care activities. All of this depends on accurate data. Missing or inaccurate data can lead to increased gaps in care and potential patient safety issues.
What Does Data Validation Uncover?
In general, measure-related data elements need to be documented in specific ways to be picked up by reports. Free-texting information or relying on scanned imaging results are insufficient for population management reporting. Data validation uncovers the need to reduce variation in documentation habits and implement workflows that ensure entry of data in standardized, discrete, searchable fields, avoiding the ‘junk in, junk out’ phenomenon. This is an essential first step to increase accuracy and reliability. Performance improvement efforts will fall flat, and buy-in will stall if data is not pulling correctly into reports.
Providers often question why a specific patient is listed or not listed in a report, or why an alert displays for a patient indicating a gap in care. These questions are frequently cleared up by understanding the measure parameters and current guidelines and present a great opportunity for discussion and provider education.
Measure definitions sometimes differ from current or accepted guidelines because it takes some time for measures to ‘catch-up’ with clinical best practices. Alternatively, there may be ambiguity or controversy about the guidelines behind the measure. This can be tricky because certified reports required for program participation are very strict and can’t be changed, but might conflict with providers’ best judgment. Providers and care teams should avoid getting caught up in a ‘teaching to the test’ approach and temptation to allow care to be driven by reporting requirements.
It is important to understand that while measure definitions have vastly improved and continue to be updated and aligned by the entities who create them, they’re rarely perfect and exceptions for individual patients are bound to occur. Talking through these issues and understanding when care can and should deviate from the measure or reporting requirements can help alleviate the frustration providers feel when a given measure doesn’t make sense for an individual patient. Goals for measure performance should almost never be 100%, as exceptions that make sense should be expected, even outside of the expanded exclusions built into many current measure definitions. It is important to focus on improvement trends over time rather than achieving a “perfect score.”
Sometimes reports are just wrong! The key thing here is to identify inaccuracies and have a process in place to remediate errors quickly. Errors in reports that are fixed quickly actually result in increased trust in data. Providers/staff feel they have some control over the data that reflects their performance. Report errors are commonly the result of issues with data mapping, documentation workflow, patient attribution, calculation, deviation from measure definition, report logic, or interface problems (access to the report). Provide staff or vendor(s) developing your reports clear expectations for timely responses when validating findings and initiating corrective fixes.
Data validation never ends! There are continual changes related to new data sources, new methods of data capture, EHR upgrades, measure definition updates, clinical guidelines changes, quality program requirements, etc. As such, healthcare reporting and data extraction require on-going review and remediation. Data validation needs are certainly elevated when implementing a new analytics system or a new set of reports but don’t underestimate or overlook the efforts required to continuously validate reports for maintenance purposes.
Data validation sometimes uncovers technical issues with the EHR. For example, a field for data entry that should reflect in the report does not, or information entered in one place doesn’t carry over to a place where the reports are able to pick it up. Or, the data standards built into the EHR aren’t appropriately mapped or configured to count in the reports. Data validation can help identify configuration or setup issues, and guide the development of fixes or customizations necessary for the data to be picked up appropriately.
How to Validate Data:
Attribution of patients to practices/providers is a common sticking point. Review and understand the attribution rules and logic to know when patients should show up in reports and when they should not.
- Date and number of last qualifying encounter(s) (qualifying encounters are usually billed face-to-face visits)
- Age/sex
- Practice/provider (see attribution above)
- Chronic condition diagnosis
- Lab results; medications, etc.
The data elements that determine inclusion differ according to the measure focus and definition.
- The date range of the data report is vital to be aware of when undertaking validation, i.e., if data is as of July, you’ll want to avoid referencing patient encounters past July
- Be sure to look at both ‘met’ and ‘not met,’ as well as patients who are ‘excluded,’ to comprehensively validate a measure report or gaps in care report. It is sometimes useful to start with patients who meet the requirements of the measure or report being validated, to understand what does count towards a measure.
- It is sometimes useful to validate one report against another. Comparing a report generated by the EHR to a custom developed report can help uncover issues in both reports and accelerate the process of validation.
- Try to involve care team individuals who touch all parts of the documentation required for accurate reporting.
- Spread the wealth and assign many team members to validate a small number of records, summarize their findings, then report back to the team.
- Dedicate team time to work through the more complicated or mysterious findings – it often takes the power of a team to uncover the root cause of the issue found.
Key Takeaway
Data validation is the critical process of comparing data from one source to another—and while straightforward in concept, it can reveal complex discrepancies and highlight opportunities for improving data accuracy, reporting, and overall care quality.
