Overcoming 5 challenges in using healthcare data
While healthcare data is becoming more available by the day, the challenges inherent in working with these data are increasing in number and complexity.
Since most available healthcare data is derived from financial reporting and accounting systems (EHRs, by design, are billing systems, and most third-party healthcare data comes from the ‘switches’ that connect providers with payers to facilitate payment), the data structure, field availability and ease of analysis are all skewed toward financial reporting purposes not so much toward extracting other business value.
Healthcare data structure, field availability, and ease of analysis are skewed towards financial reporting purposes.
That said, it’s not all bleak. As you overcome these 5 challenges to effectively using healthcare claims data, you’ll start realizing the inherent value in the data. With the right team and approach, backed by quality processes and good tools, this data can be marshaled to inform marketing campaigns, align teams to sales opportunities and reach your commercialization objectives.
Here are the top 5 challenges in using healthcare data for business growth:
- Data formats vary from source to source, and even from a single data vendor
In late 2017, CMS (Centers for Medicare & Medicaid Services) announced that their Limited Dataset files will change data format. New fields will be added, and some fields will be changed. CMS is now on version K of their data dictionary, meaning this is the 11th iteration of the data format! These data changes are small individually, but compound quickly, once you take into account the number of affected data sources and downstream processes.
At Carevoyance, we keep up with data availability changes, new datasets and updates to existing data sources, and have automated systems to intake, process and normalize this data, delivering standard data views across all available sources. It’s literally our job to make sure that you don’t have to think about how versions K & J of the Inpatient LDS data consolidate!
2. There are many ontologies that are inherent in the data, and keeping them up to date is painful
An ontology is a description of the concepts and relationships that can exist in the data. In healthcare there are at least a dozen different ontologies that describe diagnoses (ICD9 and ICD10), procedures performed (CPT, ICD9 and ICD10, APC, DRG), provider specialties (NUCC Taxonomies) and many others. These ontologies change at least annually, and sometimes are replaced by altogether different sets of codes (in October 2015, ICD9 was replaced by ICD10, an entirely new and completely incompatible set of codes). Keeping up with ontology changes, and mapping the various ontologies in each data source is time consuming and error prone.
We built a data engine with healthcare in mind. Our analytics & search engines, import pipelines and all other data and application tools are ontology aware, and we keep up on the changing landscape of ontologies, mapping specific versions of ontologies to when the data was created. Our built-in support for ontologies means that you don’t have to continually update your processes as these ontologies change.
Upward of 75k physicians changed employment from private practice to employed between 2014 and 2016. Keep track of these physicians
3. Physicians change employment regularly, facilities change ownership and that’s not in the data
Billing and claims data, by definition, is taken at a moment in time. A physician who practiced at The Toledo Hospital in 2016 may have gone to Henry Ford Hospital in 2017. If you rely on your claims data to tell you where physicians are, your sales efforts will suffer, as your field sales reps will be knocking on wrong doors. According to a recent AMA study, physician practice ownership dipped to 47.1%, down from 53.2%. in 2014. That means that upward of 75k physicians changed employment status between 2014 and 2016.
In order to help our customers maintain their own provider directories, we built what we refer to as our SingleSource DB, our own provider directory, that contains data on over 1.2 million physicians in the US. For SingleSource DB we source physician and facility data from state, federal and non-government sources. Once that data is sourced, the hard work of consolidating that data starts. As you can imagine, each state and federal agency has their own identifiers for each physician and facility, so keeping them in sync is a major function of our system. Once the data is co-mingled, now comes the work of standardizing schema and data elements between more than 50 disparate data sources. Not to bore you with details, but keeping a consistent provider directory is at least 2 FTEs a year.
We help our customers clean up their provider directories by connecting their systems of record (Salesforce, SAP, other CRMs and ERP systems) to the SingleSource DB, so that whenever a change appears for any current or target physician in Carevoyance, that change is propagated to our customers’ systems of record, automatically.
Maintaining an up-to-date provider directory for a medical technology company takes is at least 2 FTEs a year.
4. Keeping up with location data, and attributing location to physicians and facilities is a huge effort
Once you have your provider data all configured, you still have to find affiliations between providers and their facilities, have to find locations of their practices, assign these locations to cities, counties, states, Hospital Referral Regions, MSAs, and ensure that as physicians move, that location data stays consistent.
Did you know that USPS does not maintain an up-to-date list of ZIP codes, and where those ZIP codes actually are in the world? What about matching Census data to ZIP code data, so you could use the fantastic wealth of information that Census collects to understand the healthcare market. What about using publicly available research, based on Hospital Referral Regions and Hospital Service Areas to align your salesforce to the opportunity? You must have a solid provider directory in place before you even begin to think about location data.
We maintain our own in-house data system that’s expressly designed to keep up-to-date location and other geospatial data. This system serves as our system of record for geospatial zip code data, census and healthcare-specific locations. Using this system, we can precisely geotag physicians, practices and facilities, and run all manner of location-based aggregations on this data. We call it geo-spatial intelligence.
Our hope is by using Carevoyance, our customers can focus on the business value of data, instead of the operational and capital cost of maintaining it.
5. Automating processes to keep up with data updates, changes and new dataOnce you have an up-to-date provider directory, location data, claims data warehouse ontology mappings from data to real-world concepts, everything suddenly changes. Maybe a new dataset comes out, or maybe an ontology changes versions, or something mundane, like physician data updates or ZIP code updates, and all your data janitorial work now is obsolete, and needs to be updated, or worse, re-done. This is one of the leading causes of data frustration in medical technology businesses. The data work is never done, and it only gets more difficult, more detail-oriented and has to be done at a faster rate. In order to get these challenges under control, automated systems must be in place, and no system can be bottlenecked by an individual.
We’ve been building data systems for decades. Our founders come from technical backgrounds, building data-driven healthcare products and high-availability, low latency networking hardware. Performance, as applied to healthcare, is in our blood. That’s why we built Carevoyance; we saw a clear need to maintain these systems of record for all the medical technology companies that are downing in data, spend hundreds of thousands annually on efforts that are not scalable, not repeatable and that don’t have longevity. Our hope is by using Carevoyance, our customers can focus on the business value of data, instead of the operational and capital cost of maintaining it.
If you think you have a business challenge that we can solve with our industry-leading data platform, please reach out and say hello!
Originally published at www.carevoyance.com.
