Secondary Logo

Journal Logo

RFS – Education and Training

Evaluating Physician Knowledge of Commonly Prescribed Inpatient Rehabilitation Unit Discharge Medication's Costs

An Observational Study

Capizzi, Allison Nuovo MD; Kirk, Melissa Nuntapreda MD; Lee, Venessa Ann MD; Cushman, Daniel M. MD

Author Information
American Journal of Physical Medicine & Rehabilitation: February 2020 - Volume 99 - Issue 2 - p e15-e18
doi: 10.1097/PHM.0000000000001219


Prior studies demonstrate physicians have a poor understanding of the cost of medications that they commonly prescribe.1 In addition to the possible knowledge-deficit physicians may have, simply determining current medication cost is a complicated undertaking. A medication's cash price may differ between two pharmacies in the same city, and larger price variations are documented by region.1 The actual cost to the patient for a medication also varies by insurance coverage and whether a generic version of the drug exists. These factors limit current integration of accurate cost tools into the electronic medical record. The literature in this area also suggests that physicians are more likely to underestimate the cost of medications they prescribe.2,3 Although studies document physician interest in learning the cost of medications, research evaluating knowledge of cost in the field of Physical Medicine and Rehabilitation (PM&R) is lacking. The aim of the study was to evaluate PM&R physician knowledge of medication costs, with the hypothesis that PM&R residents, fellows, and attending physicians equally underestimate the cost of medications that they commonly prescribe to patients.


Participants were recruited from the Accreditation Council for Graduate Medical Education (ACGME) PM&R residency programs across the United States. Surveys were sent to program coordinators from all institutions listed on the Association of Academic Physiatrists listserv. Participating program representatives sent the survey to all resident, fellow, and attending physicians within their department or division. This study was approved by the institutional review board at the University of Utah. Informed consent was obtained within the survey from all subjects, and this method of consent was approved by the institution's institutional review board.

The anonymous survey tool (see Appendix 2, Supplemental Digital Content 2, was developed using REDCap software, a secure, Health Insurance Portability and Accountability Act-compliant, research-grade survey system.4 A demographic assessment included geographic region, level of training (resident, fellow, or attending physician), postgraduate year (PGY) of training for those in residency, sex, age, and prior tools used by participants to identify cost of medications (see Appendix 1, Supplemental Digital Content 1, The next section included 17 medications commonly prescribed by physiatrists at discharge from inpatient rehabilitation units. Both brand name and generic names were included. Each question was framed as follows: “What is the cost of [medication] per pill?” The survey contained an embedded monthly cost calculator for each question, such that if a participant chooses US $10 per pill for a twice-daily pill, the calculator would automatically show the cost of 30 days (60 pills) at US $600.

The medications used in this survey were chosen by co-investigators based on a pilot study evaluating the most common medications prescribed at discharge from the Inpatient Rehabilitation Facility at the investigators' home institution for a 3-mo period.

Medication cost was determined as an average of the cash price of each medication gathered from 92 pharmacies, distributed in four regions (Northeast, South, Midwest, West) of the United States, to account for regional variations (see Appendix 3, Supplemental Digital Content 3, Investigators contacted individual pharmacies and requested the cash price of each medication. Medications on the Walmart $4 list were also identified. The Walmart $4 list is an evolving formulary of prescription medications, which are subsidized at a set low price (typically US $4–$10 for a 1-mo supply), regardless of insurance coverage.

For each participant, a “differential cost” was calculated for each medication by subtracting the actual cost by the participant's “guessed” (estimated) cost. For example, if a participant estimated the cost of a medication to be US $15, but the actual cost was US $10, then the differential cost was US −$5. Then, a total differential cost was created by adding the 17 differential costs for each medication. Both the percentage differential and the differential cost were calculated; however, only differential cost was chosen as the outcome variable because patients would be more concerned with the dollar discrepancy than a percentage discrepancy. The differential cost and percentage inaccuracies are included in Table 2. Nonparametric testing (Kruskal-Wallis and Mann-Whitney tests, respectively) was performed to identify differences in total differential costs between groups (PGY of training and regular usage of a medication cost tool). Because of small sample sizes, the only groups included in PGY of training analysis were PGY-2, PGY-3, and PGY-4 residents. This study conforms to all STROBE guidelines and reports the required information accordingly (see Supplemental Checklist, Supplemental Digital Content 4,

Survey respondents' accuracy for medication cost


Forty-three participants responded to the survey, as seen in Table 1. Most respondents were residents (88%), with most in their third PGY (n = 27 or 59% of residents), from the Midwest (43%) and female (55%). Fifty-seven percent of respondents used a tool to determine medication cost. Tools commonly reported by respondents included GoodRx (21.6%), UpToDate (18.3%), eMedicine (3.3%), and “other” (13.3%).

Demographic information of survey respondents (N = 43)

There was a large range in participants' answer choices. The lowest estimated cost was US $3/month for warfarin and the highest estimated cost was enoxaparin US $1704. A similar amount of individual responses overestimated or underestimated the cost of the medications chosen for the survey (53% of all responses were underestimates). When those responses were averaged, however, as Figure 1 and Table 2 display, the average value was more commonly an overestimation. The largest variability (both percentage and dollar) was seen in the expensive medications (enoxaparin range US $4–$1704, levetiracetam range US $7.71–$800, ondansetron range US $10–$935.89), where most participants underestimated their cost. When evaluating participants' estimated cost-per-pill and cost-per-month, only enoxaparin, atorvastatin, and ondansetron were underestimated. There was no significant difference between medication cost knowledge and use of a medication cost tool (P = 0.497) or PGY of training (P = 0.593).

Estimated and actual costs of the study medications. The black bars represent the actual costs of the medications, whereas the grey bars represent the physicians' median estimated costs, with interquartile ranges displayed as error bars.


Our study showed that physiatrists have difficulty identifying the monthly cost of commonly prescribed medications on discharge from inpatient rehabilitation. These data are mixed on the hypothesis that most physiatrists underestimate medication cost because only enoxaparin, atorvastatin, and ondansetron were underestimated in this study. This suggests that physiatrists are unaware of the true cost of expensive medications. This suggests that physiatrists are unaware of the true cost of expensive medications. One explanation for this is that physicians feel that medicines, such as enoxaparin, atorvastatin, and ondansetron, are medically necessary and as such may not take the time to research alternatives. For atorvastatin, providers may think that because this medication is on the Walmart Web site (US $9 for 30-day supply), the cash price should reflect this. Alternatively, these medications may commonly be started by other hospital providers (such as internists, surgeons, or neurologists) and then continued on inpatient rehabilitation. In this case, physiatrists may feel less obliged to know their cost because they are temporarily taking over the prescribing responsibility until the patient is discharged and can then follow up with their specialist in clinic.

Two prior studies, based in the United States and in France, respectively, determined that physicians are more likely to underestimate the cost of medications they prescribe.2,3 Similar to this study, a systematic review by Allan et al.5 noted physicians tend to underestimate expensive medications and overestimate inexpensive ones. All studies demonstrate a significant discrepancy between true cost and healthcare provider perception.2,3,5–7 To our knowledge, no prior published study used the current cash price to determine medication cost. It should be noted that although relatively large variability existed, as evidenced by large standard deviations of the differentials, one could argue that accuracy was obtained for some medications, because some had differential values that were smaller than their standard deviations (eg, acetaminophen). Although physicians who use pricing tools did not have better estimates, this does not invalidate the utility and cost-effectiveness of pricing tools. It is important to note that cost is often a barrier to patient medication compliance.8 If physiatrists are not aware that they are prescribing an expensive medication that may not be covered by insurance, this may affect quality of patient care and patient satisfaction.

It is challenging to know the current price of medications here in the United States. This may be partially due to the free market system, which allows a variety of insurance carriers to exist. Although many follow Medicare guidelines, specific medication coverage can differ significantly between health insurance plans, even within the same company. In addition, politics between pharmaceutical companies, hospitals, and insurance companies may drive rapid cost changes, which are difficult to keep up with. Finally, tools, such as GoodRx and eMedicine, are helpful but do not give specifics on cost within each health insurance plan. Future studies might further evaluate the relationship between a cost tool and insurance coverage.

This study was limited by a low response rate and small sample size. Resident physicians and the Midwest region constituted most of the sample population, leading to selection bias. In addition, there was a low response rate from attending physicians. Among the residents who participated in the study, most were at the PGY-3 level, which limits comparison among year in training.

This survey does not take insurance coverage into consideration. All medications in the survey, which are also on the Walmart $4 list, were underestimated (fluoxetine, lisinopril, warfarin, clonidine, levothyroxine, metformin), suggesting that participants may know that these medications are on the $4 list and therefore do not know their cash price. Although efforts were made to provide a representative sample of medications physiatrists commonly prescribe, the final medication list may not represent the prescribing practices of each PM&R resident, fellow, or practicing physician who participated in this survey. The cash prices listed for medications in this study represented a national average and may not represent the respondent's region. Finally, the medications within this survey are most representative of those typically used on an inpatient rehabilitation unit and may not be generalizable to all of PM&R, which includes outpatient and subspecialty practices.

In conclusion, physiatrists seem to underestimate the cost of expensive medications. This highlights the discrepancy between the presumed cost of medications and their cash price and serves as a resource for PM&R residents and practicing physicians.


1. Alghanem N, Abokwidir M, Fleischer AB, et al.: Variation in cash price of the generic medications most prescribed by dermatologists in pharmacies across the United States. J Dermatolog Treat 2017;28:119–28
2. Hammond DA, Chiu T, Painter JT, et al.: Nonpharmacist health care providers' knowledge of and opinions regarding medication costs in critically ill patients. Hosp Pharm 2018;53:188–93
3. Reichert S, Simon T, Halm EA: Physicians' attitudes about prescribing and knowledge of the costs of common medications. Arch Intern Med 2000;160:2799
4. Harris PA, Taylor R, Thielke R, et al.: Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inf 2009;42:377–81
5. Allan GM, Lexchin J, Wiebe N: Physician awareness of drug cost: a systematic review. PLoS Med 2007;4:e283
6. Hernu R, Cour M, de la Salle S, et al.:, Costs in French ICU Study Group: Cost awareness of physicians in intensive care units: a multicentric national study. Intensive Care Med 2015;41:1402–10
7. Schutte T, Tichelaar J, Nanayakkara P, et al.: Students and doctors are unaware of the cost of drugs they frequently prescribe. Basic Clin Pharmacol Toxicol 2017;120:278–83
8. Brown MT, Bussell JK: Medication adherence: WHO cares? Mayo Clin Proc 2011;86:304–14

Medication Cost; Drug Cost; Medical Education; Physical Medicine & Rehabilitation

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.