Elderly persons comprise only 13% of the United States population, yet they account for one third of all prescription drugs used.1 The Kaiser Family Foundation reports that in a 2003 survey of 17,685 Medicare-eligible older persons, respondents with three or more chronic conditions (73%) take five or more medications regularly.2
Before the initiation of Part D, many relied on Medicare as their sole medical insurance. Half of United States residents older than 65 years lack medical insurance for prescription drug coverage, even though 40% receive supplemental insurance from employers or have purchased a Medigap policy and about 10% of Medicare recipients qualify for Medicaid prescription coverage.3 Medicare recipients may spend more than 10% of their income on prescription medications.4 Prescription medication costs can be so prohibitive that some people may be forced to choose between paying for housing, food, or utilities and buying their medication.5
Impact on adhernce to treatment Medication costs can affect patient compliance. On the Kaiser Family Foundation survey, 26% of respondents reported not filling the prescription or not taking the medication as directed because of cost.
2 The Center on an Aging Society reported that many elderly persons, especially those with low incomes, took less medicine than was prescribed.
6 Up to one quarter of people without drug coverage had not filled prescriptions at least once in the past year because of the cost, had skipped doses to make medication last longer, and had spent less in the past year on food and heat in order to purchase medications.
5
Poorer health outcomes Patients who cannot afford to buy their medications or to take their medications as 
prescribed tend to have poorer health outcomes.7-9 Shulman and colleagues found that 36.5% of adult patients with uncontrolled hypertension reported difficulty paying for their medications, compared with 15.5% of those whose BP was controlled.10 The study concluded that drug cost contributed to inadequate control of hypertension in certain population subgroups.10
People are often hospitalized or admitted to nursing homes because of illnesses or sequelae resulting from lack of medicines.7-9,11 The high cost of medication may ultimately lead to a greater number of hospital and nursing home admissions.8,9,11,12 The ability of medical care providers to select cost-effective therapy is important for the patient and for other consumers as well.13
Knowledge of drug costs and clinician attitudes toward the use of generic over branded drugs can have a significant impact on the decision to choose a less expensive regimen over a more expensive one.14,15 In a practical sense, deciding which medication to choose often depends on the provider's knowledge about drug safety and efficacy, beliefs about patient compliance, and attitudes about brand name versus generic preparations.14,15
Physicians' knowledge of drug costs Studies conducted in the 1970s and each decade thereafter have
shown that physicians have a poor knowledge of drug costs.15-23 In the 1970s, physicians mentioned cost as a reason for choosing a drug less than 6% of the time. Many said they did not even consider cost when prescribing.24 Further research suggested that physicians were unaware of the financial impact of the care they provided, leading to the realization that physicians needed education about drug costs.24,25 Unfortunately, later studies indicated that physicians in family medicine,15,19 internal medicine,15 pediatrics,26 neurology,27 and geriatrics13 continued to have poor knowledge of medication costs.21
PAs' knowledge of drug costs Parallel research regarding PAs' knowledge of drug costs or their attitudes about prescribing drugs is limited. A 1984 study assessed PA prescribing behavior and attitudes without assessing drug cost knowledge.28 PAs have prescriptive authority in 48 states,29 and in 2004 alone, they prescribed 250 million medications.30 Since PAs work in more than 90% of the specialties that treat elderly patients, it is important that their knowledge of drug costs also be examined.31
The purpose of the current study was to assess PA attitudes regarding prescribing prescription medications
and their knowledge of drug costs. Goals were to answer the following questions: (1) What knowledge do PAs have regarding actual drug costs? (2) Do PAs actively try to keep drug costs down for their patients? (3) What are the attitudes of PAs regarding prescribing drugs? (4) Do knowledge and prescriptive attitudes of PAs differ by practice specialty? (5) Do knowledge and prescriptive attitudes of PAs differ by practice setting?
Methods
Study design This study employed a mail survey based on one that assessed physicians' prescribing attitudes and knowledge of prescription medication costs.21 The survey was modified for PAs. There were six questions on demographics, followed by eight on prescribing patterns and resources. The final section was designed to measure PA knowledge of the relative costs of 30 different drugs for a 1-month supply. Medications chosen for this part of the survey instrument were taken from a report by the Families USA group in 2003.1
The survey instrument was pretested on six PA faculty from Western Michigan University as well as on six 
committee members from the physician assistant clinical knowledge rating and assessment tool test item writers group from the Association of Physician Assistant Programs (now Physician Assistant Education Association). These participants are PA faculty members from various PA programs across the country.
Subject selection The target population for this study was PAs who reside in the state of Michigan. The author obtained a list of all 1,958 Michigan PAs from the Michigan Department of Consumer and Industry Services. A sample size optimization analysis showed that a survey sample of 323 PAs would sufficiently reflect the target population. Since a realistic response rate for a mail survey is 30% to 60%,32 the survey was sent to 1,079 randomly selected PAs. Hoping to maximize the response rate to more than 30%, the author sent a postcard 3 weeks after the original mailing reminding the selected participants to complete the survey instrument if they had not yet done so or thanking them for their participation if they had completed and returned the survey instrument.33
Data analysis Demographic information was collected regarding gender, professional role, practice setting,
length of practice, professional degree, and practice specialty. Data from the survey instrument were examined
using descriptive statistics utilizing mean and proportions. Inferential statistics explored possible differences
in responses among the different variables of medical specialty and practice setting. The drug cost knowledge portion of the survey instrument was scored by calculating the number of correct responses from the number of those attempted. This method took into consideration that not all specialties utilize all of the drug preparations that were listed on the survey instrument. All statistical analysis was performed utilizing the Statistical Package for the Social Sciences.