Why You Need Both for Growth Planning and Physician Relations
Every hospital and healthcare system looking to grow could benefit from reliable, clean, and standardized healthcare patient data. It’s critical for strategic planners and physician liaisons, in particular, to have access to data they can trust when looking to expand offerings, build physician relationships, and improve accessibility to care for their service area.
In this post, we’re analyzing state and all-payer claims data. We’ll explore how these datasets differ, why they differ, and how, when armed with insights from both datasets, planners and liaisons can gain a more complete picture of the opportunities within their market.
Let’s start with some foundational definitions of these datasets:
- State Data: A standardized dataset that is usually maintained by state departments of health or hospital associations. Data submission is often required by the state or hospital organization.
- All-Payer Claims Data (APCD): A universal term to describe any data source that contains multiple payers. While typically used to describe medical bill clearinghouses, it can also include EMR data or other insurance claims product. APCD has drastic variation from vendor to vendor: different clearinghouses, different insurers, etc.
There are tremendous differences between these datasets due to reporting requirements, comprehension, quality, documentation, and recency. We’ll discuss each of these differences in more detail below. At a high level, their differences stem from how they are regulated. State data is typically regulated by state legislation or hospital associations, so it has strict data reporting quality control. APCD, however, comes from independent medical clearinghouses so there is no universal standard. Standardization differs across clearinghouses, meaning the formats and layouts used by each will differ as well. Companies pulling from multiple clearinghouses must be able to make sense of APCD raw data—cleaning, categorizing, and standardizing it—before it can be analyzed for insights.
Let’s now dig into more specific differences between state and APCD. We’ll then consider how, when used together, they can be the complimentary one-two punch for successful planning, expansion, and relationship building.
While state data has limited coverage of outpatient claims, typically, all inpatient discharges are reported and there are (usually) very strict, enforced, reporting requirements that must be followed. Emergency and ambulatory surgical claims are commonly reported as well, but reporting requirements differ by state. The strict reporting requirements mean state data is predictable enough that it can be used as a source of truth for the data that is reported. However, with only a fraction of cases required from outpatient and emergency settings, there are definitely deficiencies as to what you can glean from this data. What state data has in confidence, it lacks in completeness.
All-Payer Claims Data
Unlike state data, the clearinghouses do not force data standards or requirements. Providers send their medical and pharmaceutical bills to their chosen claims clearinghouse to bill the patient’s insurer. Because the clearinghouses only process claims that will be sent to payers for reimbursement, the APCD dataset will not capture self-pay or charity care like state data will. There are numerous clearinghouses with no single source. This means ACPD will also not include data for providers not captured within clearinghouses or data which is processed by clearinghouses that they don’t have a relationship with. No two APCD vendors are alike. However, at Stratasan we believe that our vendor offers market-leading coverage across the largest-volume clearinghouses.
Ultimately, knowing what is and is not in any specific dataset is what makes that dataset useful. While APCD can be very powerful, it must first be understood. Specifically, how many covered lives are in the dataset? What payers report? How are duplicated records handled? Here are some considerations regarding volumes reported in the data:
- One row of data equals one claim. Note that this is different than state data where one record typically represents one episode of care. In particularly complex episodes of care, depending on the billing processes in use at the place of care, multiple claims may be filed for the different diagnoses or procedures performed during the episode.
- Patients can be longitudinally tracked based on an anonymized identifier which is assigned before data is received. This identifier uses multiple pieces of identifying information to confirm that it is indeed the same patient across care settings, however sometimes identifying information can be captured so inconsistently between care settings that a patient may incorrectly end up being counted as multiple patients. This nuance should be considered when evaluating any metric around “covered lives”.
Note: at Stratasan we do have a single patient identifier. To preserve patient anonymity, we implement cell-size masking of five so that information can only be reviewed for a group of five patients or larger.
Click to Tweet: "Ultimately, knowing what is and is not in any specific dataset is what makes that dataset useful."
In some states, there is no state data available at all. Check out this post where we reference our State Data Availability map. Most states, however, do have access to data, though strict regulations mean only a narrow slice of this data is available for use. State data is very comprehensive for inpatient care and usually emergency care, but outpatient data is typically limited to ambulatory surgeries within a specific set of procedure codes. Beyond those care settings, state data lacks the robust comprehension strategic planners and physician liaisons need to understand healthcare activity in their markets. This is one compelling driver for the adoption of APCD.
All-Payer Claims Data
APCD is very comprehensive because clearinghouses process medical bills for providers across the healthcare spectrum (hospital, clinics, offices, etc.). It provides data for inpatient, outpatient, and emergency sites of care as well as data for professional services such as office visits, urgent care, imaging, labs, and transportation.
Quality and Documentation
Because state legislations and hospital associations have rules that enforce consistent data reporting, there is usually strong quality control and clear data documentation. Typically, there is a state data or association contact person to field questions. This standardization creates a reliable source for truth for the limited data provided.
All-Payer Claims Data
APCD is the opposite of state data in this category. While encompassing all sites of care, it lacks quality control and clear documentation. Because of the commercial relationship between clearinghouses, payers, and hospitals, there is no single source of truth as there is with state data. Each clearinghouse, payer, and hospital is different. No guarantees or standards are set (or made clear) for APCD. Companies, such as Stratasan, must be ready to run our own analysis to determine quality and usefulness. The following are examples of checks we perform in an attempt to validate the integrity of the data we receive:
- Confirm that the relative proportions for each payer type generally aligns with state data sources
- Confirm that demographic breakdowns of claims do not conflict with published demographics (ex. More covered lives reported than the population which lives in a geographic area)
Such quality checks lead to consistent, standardized data that is reliable for analysis around growth planning and physician relationship management.
State data’s strict regulations, which result in strong quality and clear documentation, also create a lot of process. This process causes state data to be released several months or quarters after it is collected—it is typically six to nine month old when released. Typically, state data releases have a two or three quarter lag. (Q1 data is released during Q3 or Q4 in the calendar year). The most recent state data sets have a 4.5 month lag (Q1 data is released early Q3). The lag in state data is a big driver of APCD’s value.
All-Payer Claims Data
Processing claim bills quickly for payment is paramount for clearinghouses. This means APCD data is extremely current. While not recommended, as the claim needs time to mature and be corrected, APCD can be updated as often as daily. Stratasan receives APCD data updates on a quarterly basis, in two releases. A preliminary release is delivered one-day post close of the quarter with a final release delivered two weeks later. As with all our products and services, there is an approximate ten day data process time to have the data live in our application. With very little lag from the end of the quarter to receipt of the data, we’re able to offer the most current market data available.
As you can see, where state data lacks comprehensiveness, APCD picks up the slack. Conversely, APCD, in its raw form, needs a lot of work before it is ready for analysis and state data is delivered in a way that makes it reliable to use right away. The monumental complexity that comes with APCD, though, means it’s only valuable, when combined with a robust data process or delivered through a partner who is equipped to aggregate, clean, and update it.
It’s fairly easy to conclude how having a combo of both state data (when it is available) and cleaned, standardized APCD, is the ideal situation for planners and physician liaisons. Each data set has its advantages, but together, when optimized, they can create a nearly complete picture of what is occurring in your market.
For more information on state or APCD, or how our Data Processing Service and tools can free up your time for more focus on strategy, relationship building, and growth, contact Sean Conway and schedule a discovery call today.
Article by Haley Attridge, Manager, HR & Administration, Tony Camarata, Product Manager, Rebecca Groner, Product Associate, Lee Ann Lambdin, SVP Healthcare Strategy, and Dave Sellers, Data Specialist for Stratasan