Life sciences companies operate in an increasingly dynamic environment shaped by evolving patient needs and behaviors, payer requirements, provider behaviors, and new or changing policies and regulations. As a result, it is more important than ever for life sciences companies to rely on sound methods and comprehensive data when making strategic and tactical decisions. One type of data that can support this need is administrative claims data, which provide insights into healthcare trends and market dynamics for the services a patient population receives that are covered by insurance companies.
What are administrative claims data?
Administrative claims data capture real-world healthcare utilization across sites of care and over time for large insured populations. Records included in these datasets are generated when healthcare providers submit claims to payers (e.g., commercial insurers, Medicare, and Medicaid) for reimbursement of medical services and prescription drugs. These data include information such as patient enrollment and demographics, diagnoses and procedures, prescriptions, dates of service, and costs paid by the payer and patient (as well as other stakeholders, when applicable). Administrative claims data do not typically include laboratory or imaging results or detailed clinical notes.
Administrative claims data are especially useful for assessing healthcare utilization and diagnosed patient populations for three key reasons:
- The paid data collected are complete. Administrative claims databases are “closed,” meaning that every claim in the database was submitted to a patient’s insurance provider and paid (i.e., not denied) regardless of when or where the service or treatment was incurred. This differs from “open” claims databases, which may include data from only a subset of providers and can contain denied claims or prescriptions that were not filled.
- The data include comprehensive care information. Administrative claims data include both medical and prescription drug claims, with diagnosis codes (ICD-10-CM), procedure codes (CPT/HCPCS, revenue codes), inpatient claim codes (DRGs, ICD-10-PCS), and prescription drug details (NDCs, days’ supply, and units).1
- Enrollment files accompanying administrative claims data include all enrollees, not just those who utilize healthcare services. Having visibility into the full insured population enables accurate calculation of prevalence rates, incidence rates, utilization rates, and per-member-per-month (PMPM) costs. Enrollment files also include patient characteristics (e.g., age, sex, and geographic location), supporting demographic and regional comparisons. Because enrollment data are available longitudinally, they also enable multiyear analyses of patient behavior, utilization, and/or outcomes (e.g., changes in services or clinical events, such as readmissions or acute exacerbations).
5 business questions administrative claims data can help life sciences companies answer
#1 – What is the anticipated payer mix of a patient population?
- Value of answering this question: Understanding the anticipated payer mix of a potentially eligible population is important for forecasting financial implications, including the expected impact of rebates and mandated discounts on net revenue for the life sciences company. These insights can also inform engagement and contracting strategies by identifying insurance channels or payer types with a large share of potential patients, enabling brand and market access teams to better tailor messaging and focus resources.
- How administrative claims can be used to answer this question: Robust administrative claims datasets include multiple payer lines of business (e.g., commercial, Medicare, and Medicaid) and, in some cases, can be extrapolated to national enrollment estimates. Populations of interest are identified using algorithms informed by clinical expertise that account for diagnostic patterns and disease-specific nuances. Identified populations are divided by total enrollment to estimate disease prevalence rates by line of business, which can then be applied to national enrollment figures. In databases that capture 100% of claims within a given line of business, extrapolation is not required, as experience and enrollees within the channel can be directly observed.
#2 – What is the diagnosis or treatment journey for a patient population?
- Value of answering this question: Analyses of diagnosis and treatment codes reveal how patients interact with the healthcare system and are treated in the real world over time, highlighting delays, drop-offs, and variation across provider types, sites of care, and geographies. These analyses can also provide insight into patient experiences prior to diagnosis or treatment initiation and how costs, utilization, and/or measurable outcomes change following diagnosis or treatment initiation (e.g., a change in rate of events/complications post-treatment initiation). Published findings from these studies can provide real-world evidence to support effectiveness claims, inform value propositions, and aid contracting and negotiation discussions with insurers, pharmacy benefit managers (PBMs), and plan sponsors.
- How claims can be used to answer this question: Medical and pharmacy claims are longitudinal and include consistent member identifiers, enabling patients to be followed over time. This allows for identification of index events (e.g., incident diagnosis or treatment initiation) and assessment of ongoing experience, such as outcomes, duration on therapy, and evolving treatment patterns. Claims data can also be used to compare healthcare utilization, costs, and clinical events before and after treatment initiation or between treated and comparison cohorts, supporting quantification and statistical assessment of differences attributable to an intervention. However, because claims data capture the services patients receive that are paid for by payers, they do not capture potential services that may have been ordered but not rendered (e.g., prescriptions that were not filled, lab tests or diagnostics that were not completed).
#3 – What is the healthcare resource utilization of a diagnosed or treated population?
- Value of answering this question: Understanding healthcare resource utilization and associated costs for a specific population can help quantify disease burden. These insights can inform budget impact modeling, effectiveness studies, and broader characterization of patient populations associated with a given condition or treatment.
- How administrative claims can be used to answer this question: Administrative claims data enable granular assessment of healthcare utilization and costs across inpatient, outpatient, professional, and pharmacy settings. This includes evaluation of emergency department visits, inpatient admissions and readmissions, and use of specific procedures or therapies. Claims can also be grouped into resource categories to better understand cost drivers and summarized as annual or PMPM metrics. Enrollment data support geographic analyses, while facility and provider identifiers (e.g., National Provider Identifiers (NPIs)) allow identification of commonly used treatment providers and locations.
#4 – What are the out-of-pocket costs for patients treated with a drug of interest?
- Value of answering this question: Understanding the out of pocket costs patients incur for a particular drug can help identify situations in which patient assistance programs—such as free drug programs or copay card programs—may be valuable. Out-of-pocket cost analysis can also support pricing strategy discussions for new products by benchmarking against analogs currently on the market and helping quantify changes in drug uptake associated with policy changes (e.g., the introduction of an out of pocket cap under the Inflation Reduction Act).
- How claims can be used to answer this question: Pharmacy claims include claim level detail for each prescription filled, including patient cost sharing components, such as deductibles, copays, and coinsurance. Enrollment data include information on plan type, which can be used to segment patients and model potential changes in drug uptake driven by differences in cost sharing or formulary tiering. Medicare prescription drug data also include coverage phase indicators, which influence patient cost sharing amounts and enable more detailed assessment of out of pocket exposure across benefit designs. Note that copay card utilization is not directly observable in the data but may be inferred in some cases.
#5 – What is the financial impact of the introduction or change of a policy or regulatory action on patients, payers, and manufacturers?
- Value of answering this question: New policies and regulatory changes can meaningfully affect costs for patients, payers, and manufacturers. Administrative claims data can be used to quantify these impacts and inform adjustments to rebate structures, contracting strategies, or access approaches in response to policy and/or regulatory changes.
- How claims can be used to answer this question: Historical claims can be re-adjudicated under actuarial models to simulate future policy or regulatory scenarios, enabling estimation of financial impacts across stakeholders. Claims data can also be used to measure real-world effects following implementation by tracking changes over time using datasets updated on a monthly basis, such as the Centers for Medicare and Medicaid Services’ (CMS’s) 100% Parts A, B, and D data.
How advanced analytics enhance administrative claims data
Advanced analytics, such as predictive modeling and machine learning, can further extend the value of administrative claims data beyond descriptive insights. These approaches can identify patterns within diagnosed populations and estimate the potential size of undiagnosed populations, helping life sciences companies better understand disease scope and inform strategic planning and decision making.
Conclusion: Administrative claims data can inform decision making for life sciences companies
Administrative claims data offer life sciences companies a robust, real-world view of healthcare utilization, costs, and treatment patterns across large insured populations. Applied thoughtfully, it can answer key strategic questions—from payer mix and patient journeys to resource use, out-of-pocket burden, and policy impacts—while advanced analytics can extend these insights even further. Leveraging these datasets supports more informed planning, evidence generation, and decision making in a rapidly changing healthcare landscape.
Examples of administrative claims data across U.S. healthcare payer channels
At Milliman, we have access to robust administrative claims data across all major payer channels in the United States, which can be used to answer questions like those highlighted previously. Our data types include the following.
| Dataset | Data included |
|---|---|
| CMS 100% fee-for-service (FFS) Medicare data | Includes 100% of all claims and enrollment for FFS beneficiaries nationally and is updated monthly, with only a two-to-four-week delay (not accounting for claims runout). This dataset includes eligibility, patient characteristics (including race/ethnicity), and identifiable provider information. |
| CMS 100% fee-for-service (FFS) Medicare data | Includes 100% of all claims and enrollment for FFS beneficiaries nationally and is updated monthly, with only a two-to-four-week delay (not accounting for claims runout). This dataset includes eligibility, patient characteristics (including race/ethnicity), and identifiable provider information. |
| CMS 100% Medicare Advantage (MA) encounter claims data | Includes 100% of all Medicare managed care encounters nationally. This dataset does not contain cost information but does include service utilization and enrollment. Eligibility, patient characteristics (including race/ethnicity), and identifiable provider information are available as well. |
| CMS 100% Part D Prescription Drug Event (PDE) files | Includes 100% of Part D Prescription Drug Plan (PDP) and MA Part D (MAPD) prescription drug claims. This dataset is updated monthly, with only a two-to-four-week delay (not accounting for claims runout), providing timely access to prescription drug claims. Information about the prescription drug plan, prescriber, and pharmacies is included, and this dataset can be linked to the CMS 100% FFS and MA beneficiary information. |
| CMS 100% Transformed Medicaid Statistical Information System (T-MSIS) Medicaid claims data | Includes 100% of state FFS claims and managed care encounters (service utilization only) for Medicaid and Children’s Health Insurance Program (CHIP) beneficiaries. This dataset includes eligibility, patient characteristics (including race/ethnicity), and identifiable provider information. This dataset can also be linked with the CMS 100% FFS and MA beneficiary information, which ensures that individuals with both Medicare and Medicaid coverage are not double counted when conducting analyses that include both lines of business. |
| Milliman Consolidated Health Cost Guidelines™ Sources Database (CHSD) | Includes claims and enrollment for a large sample of commercial group, individual, managed Medicaid, and MA insured lives. Enrollment, patient characteristics, cost and patient pay metrics, and blinded provider information are included. |
1 ICD-10-CM: International Classification of Diseases-10th Revision-Clinical Modification; CPT: Current Procedural Terminology; HCPCS: Healthcare Common Procedure Coding System; DRGs: Diagnosis-Related Groups; ICD-10-PCS: International Classification of Diseases-10th Revision-Procedure Coding System; NDCs: National Drug Codes.