To understand how the price estimates were created, please read the documentation below or contact our team by emailing: alexnielson@utah.gov
If you are interested in the Cannabis Search Methodology, please click this button: Cannabis Methodology
On May 14, 2019, Utah Code 67-3-11 went into effect, which, among other things, requires the state auditor to create and maintain a health care price transparency tool (Tool) that is accessible by the public. As permitted by the new law, the Tool incorporates existing health care data from Utah's All Payer Claims Database (APCD). The Utah Department of Health and Human Services's Office of Health Care Statistics (OHCS) is responsible for managing the APCD. Utah Code (see § 26-33a Utah Health Data Authority Act) and Utah Administrative Rule (R428-15, Health Data Authority Health Insurance Claims Reporting) requires health insurance carriers covering 2,500 or more Utahns to submit medical claims data to OHCS “to facilitate the promotion and accessibility of quality and cost-effective health care (see Utah Code § 26-33a-104).” According to the Department of Health, the APCD "represents approximately 90 percent of Utah's non-Medicare population."
The Tool provides median cost information for various healthcare procedures. We define procedure as a healthcare encounter used to diagnose, measure, or treat a medical problem. A single patient may have multiple procedures performed during a single visit. Each patient and their bundle of procedures during a visit was given a unique visit id. Visit information like procedure performer (provider), facility, procedure cost, and duration were recorded.
Visits that contained the exact same facility, provider, and procedures were aggregated and averaged to get a median price and frequency. These aggregated visits with a list of common bundled procedures performed at a common provider-facility are called a "Bundled Procedure".
Since some Bundled Procedures at a given facility with the same doctor are similar in price, and differences are neglible, we then aggregated Bundled Procedures into "Grouped Bundled Procedures" and gave plain language names to each group.
An example of a neglible difference might be slightly different Colonoscopy bundled procedures at a specific doctor-facility pair. Lets say Bundled Procedure A cost $1000 and contained the procedure cpt codes (45378, 99214), and Bundled Procedure B cost $1010 and had the procedure cpt codes (45378, 99214, 90772). The two bundled procedures are exactly the same except for the code 90772 which corresponds to an antibiotic injection. Since these two procedures are highly similar and are close in price except for a negligible $10 antibiotic cost, our method would warrant grouping these two bundled procedures together to calculate a new median estimated price. The nuance of each Grouped Bundled Procedure is documented in the Bundled Procedure Methodology website found below.
For each Grouped Bundled Procedure, a primary facility and primary doctor was determined algorithically. Primary facility was ranked on the following order: Hospital > Center > Clinic > Everything Else. The Primary Doctor was determined by a more complex heuristic which included analysis of NPI taxonomy codes and this hueristic changed depending on the grouped bundled procedure. For example, when looking at a Knee Surgery, providers with an NPI taxonomy class of 'Orthopaedic Surgery' would be considered the primary doctor over a provider in the same visit bundle with the taxonomy class of 'Nurse Practitioner'.
Provider addresses and names came from the National Provider Identifier (NPI) Registry, which is hosted by the Centers for Medicare & Medicaid Services (CMS) (our data was retreived on 11/16/2020). We matched NPI data with APCD data based on common provider and facility NPIs.
While NPIs are useful in identifying providers, some claims submitted to the APCD do not contain NPIs and multiple NPIs can be associated with one provider. In an effort to group claims belonging to the same provider, we manually cleaned provider information based on matching service addresses and/or provider names.
To see reproducible R scripts used to create the bundled procedures. Please visit the https://hccdocs.web.app
Procedures in this tool which are marked as "CPT Code Only" are from a legacy version of this tool and have a different methodology than the bundled procedures.
We define procedure as a healthcare encounter used to diagnose, measure, or treat a medical problem. A procedure may involve multiple medical claims and billing codes. We chose which procedures to analyze based on a judgmental sample of high frequency and/or high cost procedures. Because multiple billing codes exist for a particular procedure, we chose the code(s) most frequently used for each respective procedure.
The first public iteration of the Tool uses data from April 1, 2018 to September 29, 2019. OHCS pulled data for our requested procedures on December 5, 2019.
We excluded the top and bottom five percent of each provider's claims based on overall dollar amount. We also excluded provider information for a particular procedure if the provider did not perform a procedure, within the 18-month period, more than ten times. In cases where the median cost was less than 25 percent of the average median cost across all providers, the provider's information was flagged for review.
We included only data from providers located in Utah and for primary liability and commercial claims. We also excluded Medicaid and Medicare data.
Provider addresses and names came from the National Provider Identifier (NPI) Registry, which is hosted by the Centers for Medicare & Medicaid Services (CMS) (file last updated on 11/10/2019). We matched NPI data with APCD data based on the billing provider NPI. While NPIs are useful in identifying providers, some claims submitted to the APCD do not contain NPIs and multiple NPIs can be associated with one provider. In an effort to group claims belonging to the same provider, we manually cleaned provider information based on matching service addresses and/or provider names.
Many provider groups perform healthcare procedures at multiple locations throughout Utah. Typically, each location of a provider group has its own billing NPI. However, some provider groups have told us they do not consistently use the billing NPI associated with the location where a procedure was performed. Generally, the Tool assumes the procedure occurred at the location associated with the billing NPI. Under this assumption, each location has a unique name and has a separately calculated median cost for a given procedure. In a few cases, we have combined the claims of a provider group across all locations. We have made this exception to the general rule for provider groups that (a) have less than five locations and (b) have requested that their claims be combined due to certain billing practices. Under this exception, all locations are combined under one name and a single median cost is calculated for the provider group.
We calculated total costs by using the amount allowed of a claim or encounter. The amount allowed includes plan paid, prepaid, copay, coinsurance, and deductible amounts. It is important to note that we used actual amounts paid in our Tool, not charged amounts.
For more complex, surgical procedures, we discovered that a single claim and billing code were inadequate in accounting for the full cost of the procedure. For example, the cost of a knee replacement normally includes a claim for the facility, anesthesiologist, and physician. Though these costs are likely associated with the overall cost of the knee replacement, or other procedure, they are often not captured in a single claim, within a single procedure code, or on the same date. To better capture associated costs, we combined all claims for a patient within the service start and end date of the procedure code of interest. We combined the claim amounts within each respective window to form an overall cost per patient for each encounter.
Imaging procedures (e.g. x-rays, MRIs, etc.) often contain a professional and technical component that may be billed separately. To capture the total cost of MRIs and x-rays, we included claims on the same day as the procedure code of interest. This method would also include the cost of an office visit. For CT scans, we looked at only the single claim associated with the procedure code of interest. Trying to combine all claims on the same day as the CT-scan procedure code resulted in including expensive claims not associated with the CT scan. As a result, to provide a more accurate representation of the cost of just the CT scan, we do not include such claims.