This monthly blog discusses the challenges and emerging trends related to healthcare data and its ever changing life cycle.
By Lorraine Fernandes, RHIA, and Michele O’Connor, MPA, RHIA, FAHIMA
Healthcare is transforming at a rate not previously seen. Many different levers are at work in this transformation, applied either in parallel or jointly. This transformation is occurring with the goal of advancing the IHI Triple Aim—improve care and outcomes, reduce costs, improve health. These levers include:
- Health information exchange to ensure all information is available for care delivery and care coordination, thus freeing data that has historically been isolated and fragmented.
- Population health programs to manage clinical and financial risk. These programs may identify individuals and groups that will benefit from structured care coordination, specialized wellness programs, and much more.
- Analytics to advance a data-driven approach for defining best practices, expected outcomes, and effective innovation models.
- Shift to wellness-focused care programs and value-based reimbursement in contrast to the historical “treat the sick” in a fee-for-service approach
- Cognitive computing (artificial intelligence) to rapidly ingest data that has grown beyond human consumption, advising and guiding care teams and researchers, thus reducing information gaps and time delays that have historically been encountered. Time saving and perhaps life saving.
This list could consume the entire blogpost, but the point is we live in an era where data is finally being used as an asset, and data is being measured in ways not anticipated even a few years ago. Many studies by public health and governmental agencies, healthcare researchers and delivery organizations, and vendors have measured the baseline health data and produced metrics that illustrate improvement. But can you TRUST that the data truly represents a comprehensive view of the patient/consumer or a population, or do the statistics still represented a fragmented view of the patient and therefore a fragmented view of the relevant data?
Creating the complete view of patient data starts with a comprehensive view of the patient across the enterprise—within each system the patient should have a unique medical record number—even before trying to integrate or aggregate data across an enterprise.
Today measuring success in accurately identifying a patient (patient identity integrity) and thus providing a complete, trusted view of patient data is inconsistent at best, and many times not done or erroneous. Specifically, in the 2015 AHIMA survey 44 percent of the respondents did not know the error rate in their EHR system and 58 percent of the respondents reported that they do not know the formula for calculating this important data quality metric. The lack of knowledge around patient identity integrity is partially responsible for the wide range of error rates (1-12 percent) quoted in articles over the past decades.
In 2010 HIMSS published a Patient Identity Toolkit which included suggested metrics (key performance indicators) for measurement and an internet search will reveal many other nuances or opinions in measuring patient identity integrity. Sadly, the AHIMA statistics illustrate the diversity of opinion and no clear best practice, thus errors go undetected (and compromise care delivery or transformation activities) or process improvement is ineffective or non-existent.
We have not resolved the issues surrounding HOW patient identifiers are measured at a local level, nor how to measure “linking” at an enterprise. We have to address the “how” of accurately identifying a person in a market where perfect does not exist—nor will it exist even in the world of biometrics or a national health identifier. As healthcare becomes data-driven and wellness-oriented it is time to call the question and begin consistently measuring the accuracy of patient data in order to build TRUST. Failure to do so, and failure to enact the changes that may be needed, means the billions we are investing in healthcare transformation may produce suboptimal results.