OBJECTIVES: Prescription claims databases are commonly used for identifying patients for disease management programs, studying health outcomes and reporting on quality measures.  A shortcoming of claims databases for these purposes is that they include only prescriptions that are adjudicated through insurance plans.  Growth in the use of cash discount generic programs and the frequent use of drug samples suggests that an increasing number of prescriptions dispensed to insured consumers may not be captured on claims databases. We examined the extent to which prescription claims databases do not provide complete records of insured patients’ prescription drug use.  METHODS: We used the 2009 Medical Expenditure Panel Survey (MEPS) dataset. We included participants who purchased at least one prescription medication and who had prescription drug insurance for all of 2009. We quantified the extent to which insured patients used drug samples, drugs paid for by cash only, and/or discount generics. We measured the numbers of prescriptions in each of these categories and the numbers of consumers who had at least one prescription in each category. We reported descriptive statistics. RESULTS: A total of 75.1% of the U.S. non-institutionalized civilian population was insured for prescription drugs. Of the total number of prescriptions dispensed to insured consumers, at least 0.8% were drug samples and 23.3 % were paid for by cash, of which 11.3% were potentially discount generics. Additionally, 11.6 % of insured consumers received at least one sample medication, 68.0% paid for at least one of their prescribed medications by cash, of which 42.5% used at least one potential discount generic product. CONCLUSIONS: Our results indicate that drug samples do not contribute substantially to the problem of missing prescription data on claims databases. On the other hand, substantial number of prescriptions, paid for by cash and discount generics, may be missing from these databases.