MarketScan Data

---

MarketScan is a robust secondary data source of national claims data. Persons in the dataset are largely those with employer-sponsored insurance (ESI), as well as some persons enrolled in Medicare and some enrolled in Medicaid (see below).  Data are submitted by employers, most of whom are self-insured, and from payers (insurers).  There are approximately 25 million enrolled lives per year in the MarketScan data, with the ability to track patients longitudinally as long as they remain within the same data submitter (employer or insurer).

The data provided are claims data – largely covering outpatient care, inpatient care (including emergency department visits), labs, imaging, and drugs.  There is also an enrollment file.   

• MarketScan provides only Medicare supplemental claims (i.e., payments reimbursed by commercial plans after Medicare payments; no Centers for Medicaid and Medicare (CMS) claims included). Medicare advantage members were added in the 2020 data year. Medicare data are available only for persons who remain in ESI and whose employers pay for their Medicare supplemental insurance (aka “Medigap”).
• For the subsets of commercial and Medicare enrollees, lab results and dental claims are also available. These data are not missing at random.
• For the subset of Medicaid enrollees, long term care data is included, but lab results and dental claims are not.  The data vendor denotes that Medicaid data come from “numerous geographically dispersed states” but due to data privacy concerns will not disclose which states these are.  
• The data include individuals of all ages in the commercial data and 65+ in the Medicare Supplemental data (supplemental and advantage plans).

This MarketScan description also provides a general overview of the data.

Once the Secondary Data Resource is fully operational, access to the MarketScan data will be available to all faculty of the University of Colorado Anschutz Medical Campus within the terms and conditions of the licensing agreement with Merative, the company that curates and provides MarketScan data.

No. However, for the subsets of commercial and Medicare enrollees, Merative has created weights to project this population to the national population of individuals with ESI. The MarketScan Commercial Insurance Weights were constructed using the Public Use Microdata Sample (PUMS) of the American Community Survey (ACS) conducted by the U.S. Census Bureau. These estimates are designed to weight individuals in the MarketScan databases to reflect the national ESI population by demographic stratum; health care use varies appreciably across demographic strata.

1. The three main databases available for use are each composed of a convenience sample for one of the following patient populations:
   Beneficiaries/Enrollees with employer-based health insurance from contributing employer/payers
2. Medicare beneficiaries who are enrolled in a Medicare supplemental insurance plan paid by their current/former employers
3. Medicaid beneficiaries/members in one of eleven participating states. Per MarketScan, they cannot release the names of the states contributing.
   
   While the exact figure varies year over year, there are approximately 350 different data submitters to MarketScan.
   

Inpatient and outpatient claim data elements include beneficiary demographics, limited geographic information, service start and end dates, admission type, principal and secondary diagnosis codes, discharge status, major diagnostic category, principal and secondary procedure codes, DRGs, length of stay, place of service, and quantity of services. The financial data elements include total and net payments, payments to physician, hospital, and total admission payments.  Financial data include both the allowed amount paid by the insurer as well as the patient cost-sharing (deductible, co-insurance/co-payment).  One the outpatient side, there are up to 4 ICD diagnosis codes per claim. 

Outpatient pharmacy data elements include generic product name, average wholesale price, prescription drug payment, therapeutic class, days supply, national drug code, number of refills, and therapeutic group.  
Lab results include LOINC code, reference high and low ranges, lab result categories (e.g. abnormal, equivocal, etc.) and service date.

Dental data include provider specialty, procedures, procedure group and service type.

 

CU Anschutz has licensed commercial Sets A and B (see FAQ answer below), Medicare Sets A and B, and Medicaid data. Individual researchers and analysts will only have access to either Set A or Set B at a time. Please contact the Secondary Data Resource for additional information on the features of Set A and Set B to determine which set best meets your needs.

There are two datasets (Set A and Set B) based on the financial data that are available.

Starting in data year 2019, actual cost data became unavailable for a small subset (approximately 15 percent) of the population in the databases. For this part of the population, Merative (the MarketScan data vendor) offers imputed cost data. While use of imputed data is a common industry practice, we understand that, depending on a client’s intended use for the data, including their objectives and specific study concepts and researcher preferences, the addition of some imputed data may not be a preferred solution. Hence, Merative offers clients requesting 2019 datasets and beyond a choice between one of the two datasets: 

Set A: Contains 100 percent of the population with actual cost data for ~85% and imputed cost data for the additional 15%. This dataset does not contain an imputation flag to protect the privacy of patients as well as the privacy of our data contributors and suppliers. The methodology used for the imputed cost data is a combination of hotdecking and stochastic regression.  No further information on imputed data is available from the data vendor.  

Set B: Contains approximately 85 percent of the population and contains actual cost data for all observations.

The Secondary Data Resource currently has 10 years of MarketScan data. Available data are for calendar years 2014-2023. The duration of the current Merative license agreement to the Anschutz Medical Campus is 3 years, and the available data will be updated annually to provide 10 years of data. As additional years of data are added, earlier years of data will be removed. We have the option to retain analysis files that include earlier years of data; however, they can only include 10 years of data

Strengths:

• Cost data for covered, billable services (including both the care received and the costs associated with that care).  For example, an investigator could use these data to construct costs of episodes of care.  
  
• The large starting sample size which can provide researchers the opportunity to examine rare illnesses and medical conditions. 
  
• Includes private payer insurance plans for extended periods of time, Marketscan has the ability to conduct longitudinal analyses, especially of commercially-insured populations.
  
• While MarketScan is not nationally representative, there are weights that can be utilized to allow the data to reflect a nationally representative population. 
  
  

Limitations:

• Demographic information is limited to some geographic and employment information to maintain de-identification, and there is no race/ethnicity data. Certain demographic data like race/ethnicity, and other social determinants of health, are not available 
• Because MarketScan is based on medical claims data, information on over-the-counter medications or clinical aspects of health (e.g., diastolic and systolic blood pressure) is unavailable. 
• While prescription medication fills are captured, this doesn't necessarily reflect actual medication adherence.
• Demographic information is limited to geographic and employment information. 
• While MarketScan is a large private payer database, it represents a convenience sample of enrollees across the country representing private payers with varying market penetration in the insurance market within their region. 
  

Note: Once the Secondary Data Resource is established, to sustain its operations there will be a fee for accessing the data resources and the information technology platform for analyzing these data. We will provide advanced notice before beginning to charge a fee. 

Currently there are fees for using the MarketScan data that are specified in our licensing agreement. Use of the MarketScan data for grant-funded research (foundation, state, or federal grant) requires payment of an additional licensing fee that depends on the total amount of the grant award. If the award is less than or equal to $250,000, the incremental fee is $10,000 or 10% of the total award per study, whichever is greater. If the award exceeds $250,000, the incremental fee is fixed at $25,000 per study. The incremental fee for commercially funded research will be negotiated on a case-by-case basis with Merative, the provider of MarketScan data.  This fee covers one specific project.  

Preliminary and exploratory analyses to determine if the MarketScan data can be used to address your research questions would generally not be considered research and would not need COMIRB approval. Once you determine the MarketScan data will be used to address your research questions, you should submit a secondary data analysis protocol to COMIRB for review and approval or a determination that your research is not human subjects research. COMIRB review will facilitate publication of your analysis. 

To start the process, please complete and submit a consult request form here and select the Pragmatic Research and Trials Core.

We will reach out within 5 business days to schedule a one-hour consultation to discuss your request. If needed, a second consultation with our MarketScan Subject Matter Expert will help you decide if the MarketScan data is appropriate for your project and can answer your research question(s).  

The MarketScan data are only available on the Secondary Data Resource server and may not be copied or exported. Please contact the Secondary Data Resource by submitting a consult request form here to discuss access to the Secondary Data Resource server and the MarketScan data with the ACCORDS Pragmatic Research and Trials Core.
While the Secondary Data Resources Data Steward will help you obtain access to the MarketScan data, you will need to do your own analysis or arrange for a data manager/analyst to work with our Data Steward. If you need analyst support, you can also select the Biostatistics and Analysis Core on the consult request form or there is a campus resource, called CUBE (insert hyperlink here) that can help connect you with biostatistical and analytic support.
We recommend that your data manager/analyst participate in the consultation(s). The discussions during the consultation will include a review of the Secondary Data Resource Project Description and Data Use Application Form and the MarketScan Data Element Selection Form. 
Once approved, you and your data manager/analyst will be allocated user accounts on the Secondary Data Resource server and the Data Steward will support the extract of the data elements from the Data Element Selection Form to your project folder on the server.