Data Normalization & Validation
HCRM collects raw data consisting of medical and pharmacy claims, eligibility, lab data and HRA results from a wide variety of insurance carriers, third party administrators, pharmacy benefit managers, reference labs and wellness vendors. To simplify and accelerate the implementation process, HCRM accepts data in standard layouts already in use with each data source. Since each data source has their own file layouts, HCRM receives literally hundreds of different file types and layouts each month to process. Therefore, HCRM has developed a proprietary system that normalizes all the different data files into a single, common data structure.
Once the data are in a common structure, its accuracy and completeness is validated. HCRM uses over 300 edits to assess errors, missing data, erroneous or home-grown codes that health care data often contain. The system also converts or transforms systemic errors or home-grown codes into industry standard codes. Data problems that cannot be fixed in-house are discussed with the relevant data source for correction and re-submission.
Once all the data has been normalized, validated and reviewed by HCRM personnel, it is assembled into a large data set that is placed into a database containing data for health plan groups going back as far as data has been processed. In some cases, the normalized data goes back eight years. The minimum amount of normalized data for each group is 24 months.
HCRM can generate three data files containing member level detail of the results from the predictive model including cost summaries by diagnosis group, forecast costs by clinical risk driver, forecast days inpatient and forecast ER visits, gaps in care and pharmacy therapeutic class. These files are loaded into another database at HCRM. The normalized database and the predictive modeling results database are related across several dimensions.
Extracts from these two databases are available to clients. Many HCRM clients have data provided by a number of data sources and need the normalized claims and eligibility data to populate their own data warehouse for claims history reports. Other clients utilize the predictive modeling results files to populate a data warehouse for integration into their own reporting system for further analysis. These extracts are provided in tab delimited text files which can be loaded into any database system. HCRM will also provide the data in an Access database, if requested.
Using these underlying databases gives HCRM the ability to create custom and ad-hoc reports for its clients. These reports can be one-time to satisfy a particular project or they can be created monthly. If there are other vendors involved we can provide connectivity of data to those vendors at your request.