Clinical data use is not confined to the medical office and has not been for decades. Clinical data consumers extend from health systems to payers, public health sectors, and the government. Providers may use clinical data values to study overall care delivery or the impact of SDoH on individual patients. Researchers may use clinical data as input to develop machine learning algorithms or build predictive disease progression models. Payers continually gather and synthesize clinical data for mandated quality measures, to make risk adjustments, alternative payment programs, and to determine the value of care; their organizations are also rated according to these metrics.

When discrete clinical data elements are extracted from the electronic health record (EHR) and aggregated with data from different sources for a use case or quality measurement, the potential for error increases thereby making the quality of the foundational data even more crucial. A 2018 paper on clinical data quality in the EHR sums it up nicely: “methodological problems pertaining to data quality may arise when EHR data are used for nonclinical purposes.” If the data is first evaluated in the context of specific dimensions of data quality, then payers, providers, and researchers could avoid such problems. The issue, however, is that “the implementation of data quality appraisal remains elusive and underutilized.” One way to remedy this issue is to consider data quality according to individual data quality dimensions. Pinpointing if data lack in accuracy, consistency, accessibility, timeliness, completeness, or uniqueness reveals how data fall short of being high quality. Knowing this, data users can affect targeted process and system changes that standardize, normalize, and clean data and also help prevent future inaccuracies in areas such as data capture, measure, location, and format.

By evaluating data according to the following six data quality dimensions, payers and providers can determine the overall quality of clinical data and whether the data have the necessary attributes for a specific application or use case. They can also locate where flaws are in data collection and aggregation processes, when possible, and address these flaws, increasing the value and actionability of the data. It is often challenging, however, for organizations to fix upstream errors, as data comes from multiple sources. Data curation solutions that can address these errors after collection can save time and painstaking processes involved with working with multiple siloed groups and departments.

Accurate data are correct, reliable, validated, and reflect real-world values. Examples of inaccurate data include patients in their 50s being coded for pediatric services, or a telephone number transcribed incorrectly during data entry. Accuracy also reflects that individual patients are recognized and that the correct data is associated with the right patient. Enterprise patient management capabilities can be used to manage accurate patient identification. Provider data management governance can be used to tie clinical providers with the services provided to a member patient. One way to verify data accuracy is by returning to the source (i.e., the patient) directly, or by triangulating the data against other sources.

Consistent data are presented in the same format and are compatible with previous data. To ensure consistency, consider the following questions: Are similar data being coded in the same way, and/or in the same format? Do values give conflicting information? Applying data quality validation rules and comparing similar data from other sources can reveal inconsistency.  Standardization of data after receipt can ensure data is consistently reflected with apples-to-apples comparisons.

Data you can’t access are data you can’t use. When data are accessible, it means payers and providers can get what they need quickly and easily–they know where it lives and how to retrieve it. Verinovum’s Data Curation as a Service (DCaaSSM) solution enables payers and providers to share data, streamlining data integration for easier accessibility.

Timeliness depends on the goal for the data. In healthcare sooner is always better than later, and real-time rules supreme. Payers and providers rely on timeliness for varied reasons. For payers, timeliness is essential for reporting requirements such as HEDIS®, which help payers track the success of initiatives, find areas for improvement, and secure reimbursement. For providers, timeliness can have life-or-death stakes when considering patient care. With the current demand for real-time data, timeliness is more important than ever.

Complete data have all the necessary elements to be actionable and can deliver meaningful insights; no essential elements are missing. Defined data rules can find and resolve incomplete data, and data sources can be combined and compared to address incomplete data. Master data management processes can help ensure the completeness and appropriateness of clinical data. Verinovum’s two-phased approach does the heavy lifting, making data more complete and interoperable.

Unique data means that data elements and records are not repeated. Duplicative data is not uncommon, but its presence skews any use case or quality measurement and decreases the credibility of an analysis. Our DCaaS solution finds and reconciles duplicative records within data sets.

At Verinovum, we are focused on solving the problem of poor-quality clinical data. Our DCaaS solution addresses necessary data quality dimensions when assessing and resolving your data challenges. Request a consultation.