JAMA & ARCHIVES
Arch Fam Med
SEARCH
GO TO ADVANCED SEARCH
HOME  PAST ISSUES  TOPIC COLLECTIONS  CME  PHYSICIAN JOBS  CONTACT US  HELP
Institution: CLOCKSS  | My Account | E-mail Alerts | Access Rights | Sign In
  Vol. 8 No. 2, March 1999 TABLE OF CONTENTS
  Archives
  •  Online Features
  Original Contribution
 This Article
 •Abstract
 •PDF
 •Send to a friend
 • Save in My Folder
 •Save to citation manager
 •Permissions
 Citing Articles
 •Citation map
 •Citing articles on HighWire
 •Citing articles on Web of Science (11)
 •Contact me when this article is cited
 Related Content
 •Similar articles in this journal
 Topic Collections
 •Primary Care/ Family Medicine
 •Alert me on articles by topic

Medication Cost Information in a Computer-Based Patient Record System

Impact on Prescribing in a Family Medicine Clinical Practice

Steven M. Ornstein, MD; Lori L. MacFarlane, PharmD; Ruth G. Jenkins, MS; Qin Pan, MS; Karen A. Wager, MHS

Arch Fam Med. 1999;8:118-121.

ABSTRACT

Background  Medications account for 8% of national health care expenditures, and prescription drugs are a focus of cost containment measures. Physicians have limited knowledge about drug costs, and no method of providing this information has demonstrated sustained cost reductions.

Objective  To determine the impact of cost information in a computer-based patient record system on prescribing by family physicians.

Methods  A yearlong, controlled clinical trial was conducted at the Family Medicine Center, Medical University of South Carolina, Charleston, a group practice staffed by attending physicians and residents. Prescription cost information was included in the computer-based patient record system used at the center. During a 6-month period, cost information was not displayed; during the subsequent 6-month intervention period, costs were displayed at the time of prescribing. An intention-to-treat analysis was used to compare prescription costs between the control and intervention periods for all medications prescribed, and stratified analyses for several medication and physician factors were performed.

Results  A total of 22,883 prescriptions were written during the 1-year study period. The mean ± SD cost per prescription in the control period was $21.83 ± $27.00 (range, $0.01-$510.00), and in the intervention period was $22.03 ± $28.12 (range, $0.01-$435.96) (P = .61, Student t test). Increases in mean prescription cost and proportion of total costs were identified in 4 medication classes: antibiotics, cardiovascular agents, headache therapies, and antithrombotic agents. Decreases in mean prescription cost and proportion of total costs were identified in 5 medication classes: nonsteroidal anti-inflammatory drugs, histamine type 2–receptor antagonists and proton pump inhibitors, ophthalmic preparations, vaginal preparations, and otic preparations.

Conclusions  In this setting, the provision of real-time computerized drug cost information did not affect overall prescription drug costs to patients, although differences in individual medication classes were observed. The negative results of this study may reflect confounding due to the use of historical controls, suboptimal timing of the intervention in the prescribing process, susceptibility bias at the study site, or the insensitivity of prescribing habits to cost information.



INTRODUCTION
 Jump to Section
 •Top
 •Introduction
 •Methods
 •Results
 •Comment
 •Author information
 •References

HEALTH CARE expenditures accounted for 13.9% ($938 billion) of the US gross national product in 1994.1 Projections indicate that by 2005, health care expenditures will reach $2.2 trillion, representing 17.9% of the gross national product.1 Even though physicians and other health care providers are being challenged to assume a greater role in cost containment efforts, they often are not aware of health care costs.

In 1994, $79.0 billion was spent on drugs and nondurable medical supplies in the United States, accounting for approximately 8% of national health care expenditures.1 Appropriate medication use may account for a reduction in costly complications and hospital admissions, impacting overall health care costs beyond the actual medication acquisition cost. On the other hand, the inappropriate use of medications in the ambulatory care setting may cost the US economy more than $76 billion each year.2

Approximately 63% of office-based physician visits result in at least 1 medication being prescribed, making pharmacotherapy the most common form of treatment.3 Decisions governing the use of prescription drugs lie primarily in the hands of practicing physicians. For physicians to make rational, cost-conscious prescribing decisions, they need accurate knowledge about pharmacological principles and the cost of medications.4 Physicians report having limited knowledge of drug costs, but are interested in lowering the cost of drug therapy and in having cost information more readily available.5-7

A 1989 review revealed that traditional approaches to improving physician prescribing, including mailed educational materials, audits, and group education, are largely ineffective,8 whereas "academic detailing" by clinical pharmacists could improve prescribing quality and reduce unnecessary costs.9 More recent studies have shown that audits with weekly reminders10 and academic detailing coupled with computerized drug utilization review11 can lower prescription drug costs.

Computer-based patient record (CPR) systems are ideal tools for providing physicians with needed decision-making information, and there is increasing attention to the wider dissemination of these systems.12 Computer-based patient record systems can provide health care professionals with access to knowledge-based systems and to reminders and alerts at the point of care. Computer-based clinical information systems have improved adherence with preventive services, appropriate drug dosing, and several aspects of patient education.13-15 Computer-based patient record systems have features designed to screen for drug interactions and allergies and to indicate drug costs at the time of prescribing. In 1 study, laboratory test charges were reduced 13% when cost information was provided by the CPR system at the time of test ordering.16 Consequently, it is reasonable to assume that properly designed CPR systems could change physician prescribing habits and decrease prescription drug costs.

This study was designed to determine the impact of cost information in a CPR system on prescribing by family physicians. It was hypothesized that the provision of drug cost information at the time of prescribing would decrease drug costs to patients in various manners. For example, physicians might avoid prescriptions that were unnecessary or considered of marginal benefit, or they might recommend a less costly over-the-counter medication with similar efficacy. Likewise, physicians might choose a generic equivalent rather than the brand name drug, or substitute for another less costly alternative within the same therapeutic class.


METHODS
 Jump to Section
 •Top
 •Introduction
 •Methods
 •Results
 •Comment
 •Author information
 •References

This controlled clinical trial was conducted during a 1-year period from July 1995 through June 1996 at the Family Medicine Center, Department of Family Medicine, Medical University of South Carolina, Charleston. During the control period (July to December 1995), physicians were unaware of drug costs, although this information was collected through the CPR system. During the intervention period (January to June 1996), cost information continued to be collected, and physicians were given the drug cost information while writing prescriptions in the CPR system. Comparisons between the control and intervention periods were conducted to determine the efficacy of the intervention in decreasing drug costs.

The Family Medicine Center has an active patient population of approximately 12,500 patients, with an average of 125 patients seen daily. Sixty-one percent of the patients are women and 39% are men; 69% are black, 29% are white, and 2% are of other or unknown race. The staff includes 10 faculty physicians, 36 residents, nurses, clinical pharmacy faculty and residents, psychologists, and counselors. Staff are divided into 3 practice groups that provide care to a subset of the patient population. Within each group, clinical pharmacy faculty and residents work closely with the faculty physicians and residents in making pharmacotherapy decisions.

Since 1991, the Family Medicine Center has used the microcomputer-based patient record system, Practice Partner Patient Records, from Physician Micro Systems, Inc, Seattle, Wash. The software resides on a local area network, connecting approximately 120 microcomputer workstations throughout the building, including 1 in each examination room, at the nursing stations, and in faculty and staff offices.

The CPR system includes a prescription writing function that allows prescriptions to be written either "free-form" or using a library of medication templates. The template library provides an alphabetized "pop-up" list of various generic and brand name medications from which the physician may choose. The template library was developed and is currently maintained by one of us (L.L.M.), a clinical pharmacist at the Family Medicine Center. Among other prescription data elements, the CPR system records and displays the average wholesale price per unit of medication, the average wholesale price per total amount prescribed, the average generic price per unit, and the average generic price per total amount prescribed. Cost information is updated quarterly based on data provided by the CPR vendor. A pop-up pick list of alternative medications with the same pharmacological classification (eg, angiotensin-converting enzyme inhibitors) and their respective average wholesale price and generic costs are also available in the CPR system.

During the 1-year study period, a complete prescription record was captured in an ASCII text file when each prescription was entered in the CPR system. During the control period, cost data did not appear on the CPR display screen when the physician was writing a prescription. In the intervention period, however, brand name and generic cost data appeared on the screen in real time, both for prescriptions written using the templates and those written in free-form. Less expensive alternatives in the same pharmacological class could be displayed by using 1 keystroke. Cost information was only displayed when an exact match occurred between the medication name and dose in the prescription and CPR data dictionary. When a mismatch occurred, on average for 25% of the prescriptions, cost information was recorded as "no price."

The unit of analysis was the prescription and the mean average wholesale price cost per prescription compared between the control and intervention periods for all analyses. A cost for each prescription was assigned, using the lower of the brand name or generic cost. An intention-to-treat approach was used; therefore, cost information was assigned to all medications, whether or not the information was available to the physician at the time of prescribing. When cost information was not available at the time of prescribing (ie, "no price" was recorded in the database), the average wholesale price during the month of prescribing was assigned. Medications prescribed for less than 30 days were classified as "acute care" medications, and the cost was calculated for the total prescription. Medications prescribed for 30 days or longer were classified as "chronic care" medications, and the cost for a 30-day period was calculated. Medication records excluded from analyses were those for medications prescribed less than 5 times annually, incomplete prescription records (those with no frequency, size, take or amount [ie, the number of pills, milligrams, or milliliters to take in 1 dose], or an amount of zero), and entries for devices, equipment, exercise regimens, and "one-time" injections or treatments.

Differences between the control and intervention periods in the mean cost per prescription were compared using the Student t test and the analyses were confirmed by the Wilcoxon signed rank test. Stratified analyses based on 25 medication classes and physician level of training, gender, and practice groups were performed in a similar manner. The proportion that each medication class contributed to total overall prescription costs in the control and intervention periods was calculated.

Analyses based on cost per prescription may miss changes in prescribing habits characterized by use of fewer, more expensive prescriptions or a greater number of less expensive prescriptions at a patient contact. Therefore, an analysis of the mean prescription cost per patient contact was also performed. All patient contacts during the study period were obtained from the CPR, except those associated with excluded medication records. Differences between the control and intervention periods in the mean cost per patient contact were compared by the Student t test.


RESULTS
 Jump to Section
 •Top
 •Introduction
 •Methods
 •Results
 •Comment
 •Author information
 •References

A total of 22,883 prescriptions were written during the 1-year study period. Of these, 2985 (13.0%) were excluded from the analyses, including 1875 (8.2%) prescribed less than 5 times, 860 (3.8%) with incomplete information or one-time injections or treatments, and 250 (1.1%) for devices, equipment, or exercise regimens. The remaining 19,898 prescriptions represented 600 discrete medications prescribed at 30,461 patient contacts. During the control period, 9013 prescriptions were written at 14,070 patient contacts; 10,885 were written during the intervention period at 16,391 patient contacts. The mean ± SD cost per prescription in the control period was $21.83 ± $27.00 (range, $0.01-$510.00), and in the intervention period was $22.03 ± $28.12 (range, $0.01-$435.96) (P = .61, Student t test). The mean cost per contact in the control period was $12.49 ± $29.35, and in the intervention period was $13.03 ± $30.06 (P = .12, Student t test).

The mean prescription cost and proportion that each medication class contributed to total overall costs in the control and intervention periods are shown in Table 1. Increases in mean prescription cost and proportion of total costs were identified in 4 medication classes: antibiotics, cardiovascular agents, headache therapies, and antithrombotic agents. Decreases in mean prescription cost and proportion of total costs were identified in 5 medication classes: nonsteroidal anti-inflammatory drugs, histamine type 2–receptor antagonists and proton pump inhibitors, ophthalmic preparations, vaginal preparations, and otic preparations. Nonsignificant or inconsistent differences were found in the remaining 16 medication classes.


View this table:
[in this window]
[in a new window]
Table 1. Mean Prescription (Rx) Costs and Percentage of Total Prescription Costs by Medication Class*


The mean prescription costs in the control and intervention periods stratified by physician level of training, physician gender, and practice group are shown in Table 2. No significant differences were observed.


View this table:
[in this window]
[in a new window]
Table 2. Mean Prescription Costs by Physician's Level of Training, Sex, and Practice Group*



COMMENT
 Jump to Section
 •Top
 •Introduction
 •Methods
 •Results
 •Comment
 •Author information
 •References

This study failed to detect an impact of CPR-based prescription drug cost information on overall drug costs to patients among family physicians in an academic family medicine ambulatory clinical practice. Although costs did decrease for certain classes of medication, they increased for others, and we believe it would be inappropriate to overemphasize these subgroup analyses.

There are several possible explanations for the negative findings in this study. First, the use of historical controls may have masked the efficacy of the intervention. Other factors, such as the introduction of newer, more expensive medications, or unexpected cointerventions may have confounded the study findings. The marked increase in the mean costs of cardiovascular agents and antibiotics leads credence to this possibility. An ongoing quality improvement activity designed to increase the use of lipid-lowering therapy was active during the intervention period and resulted in a 50% increase in the frequency of prescribing hepatic hydroxymethyl glutaryl coenzyme A reductase inhibitors and a 2-fold increase in the cost of all prescribed lipid-lowering agents. Although impractical for this particular study, the use of concurrent controls would have mitigated against these types of confounding.

Second, the specific intervention studied may not have been sufficiently potent. Cost information and the cost of alternative class substitutes were only available once the physician initially selected a medication. Since the CPR system does not track how often substitute medications are sought, we were unable to determine how often this option was used. Cost information may have been provided too late in the process of prescribing to affect decision making. An alternative approach, whereby specific cost-effective therapeutic approaches are provided at the initial point of the prescribing process, may be more effective.

Third, the study was done in an academic family practice setting, where clinical pharmacists and pharmacy students work closely with the resident and faculty physicians in making cost-conscious pharmacotherapeutic decisions. Approximately one half of all patient contacts involve some form of clinical pharmacy consultation. Academic detailing has been demonstrated to reduce prescription drug costs9, 11 and it is possible that the practice studied may not have been susceptible to additional cost reduction interventions.

Finally, it may be that physician prescribing habits are relatively insensitive to cost information. Factors such as anticipated efficacy, side effects, patient compliance, and peer recommendations may be more important.5 Particularly for "chronic care" medications that have proven to be effective for an individual patient, cost may be a minor factor. However, it would be premature to adopt this conclusion without further evidence, and better-designed studies, addressing the limitations of this study, are needed.


AUTHOR INFORMATION
 Jump to Section
 •Top
 •Introduction
 •Methods
 •Results
 •Comment
 •Author information
 •References

Accepted for publication November 23, 1997.

This study was supported in part by Physician Micro Systems Inc, Seattle, Wash.

Presented in part at the 24th Annual Meeting of the North American Primary Care Research Group, Vancouver, British Columbia, November 5, 1996.

Reprints: Steven M. Ornstein, MD, Department of Family Medicine, Medical University of South Carolina, 171 Ashley Ave, Charleston, SC 29425 (e-mail: ornstesm{at}musc.edu).

From the Departments of Family Medicine (Drs Ornstein and MacFarlane and Ms Jenkins), Pharmacy Practice (Dr MacFarlane), Biostatistics and Epidemiology (Ms Pan), and Health Administration and Policy (Ms Wager), Medical University of South Carolina, Charleston.


REFERENCES
 Jump to Section
 •Top
 •Introduction
 •Methods
 •Results
 •Comment
 •Author information
 •References

1. Burner S, Waldo DR. National health expenditure projections, 1994-2005. Health Care Financing Rev. 1995;16:221-242. PUBMED
2. Johnson JA, Bootman JL. Drug-related morbidity and mortality. Arch Intern Med. 1995;155:1949-1956. FREE FULL TEXT
3. Schappert SM. National Ambulatory Medical Care Survey, 1991 summary. National Center for Health Statistics. Vital Health Stat 13. 1994;No. 116:2-10.
4. Reidenberg MM. A plea for prices in physician prescribing [editorial]. JAMA. 1991;266:3285.
5. Safavi KT, Hayward RA. Choosing between apples and apples: physicians' choices of prescription drugs that have similar side effects and efficacies. J Gen Intern Med. 1992;7:32-37. ISI | PUBMED
6. Walzak D, Swindells S, Bhardwaj A. Primary care physicians and the cost of drugs: a study of prescribing practices based on recognition and information sources. J Clin Pharmacol. 1994;34:1159-1163. ABSTRACT
7. Barclay LP, Hatton RC, Doering PL, Shands JW. Physicians' perceptions and knowledge of drug costs: results of a survey. Formulary. 1995;30:268-279. ISI | PUBMED
8. Soumerai SB, McLaughlin TJ, Avorn J. Improving drug prescribing in primary care: a critical analysis of the experimental literature. Milbank Q. 1989;67:268-317. FULL TEXT | ISI | PUBMED
9. Avorn J, Soumerai SB. Improving drug-therapy decisions through educational outreach: a randomized controlled trial of academically based "detailing." N Engl J Med. 1983;308:1457-1463. ABSTRACT
10. Frazier LM, Brown JT, Divine GW, et al. Can physician education lower the cost of prescription drugs? a prospective, controlled trial. Ann Intern Med. 1991;115:116-121.
11. Keys PW, Goetz CM, Keys PA, Sterchele JA, Snedded TM, Livengood BH. Computer-guided academic detailing as part of a drug benefit program. Am J Health Syst Pharm. 1995;52:2199-2204.
12. Committee on Improving the Patient Record, Institute of Medicine. The Computer Based Patient Record: An Essential Technology for Health Care. Washington, DC: National Academy Press; 1991.
13. Ornstein SM, Garr DR, Jenkins RG. Computer-generated physician and patient reminders: tools to improve population adherence with selected preventive services. J Fam Pract. 1991;32:82-90. ISI | PUBMED
14. Garr DR, Ornstein SM, Jenkins RG, Zemp LD. The effect of routine use of computer-generated preventive reminders in a clinical practice. Am J Prev Med. 1993;9:55-61. ISI | PUBMED
15. Balas EA, Austin SM, Mitchell JA, Ewigman BG, Bopp KD, Brown GD. The clinical value of computerized information services: a review of 98 randomized clinical trials. Arch Fam Med. 1996;5:271-278. FREE FULL TEXT
16. Tierney WM, Miller ME, McDonald CJ. The effect on test ordering of informing physicians of the charges for outpatient diagnostic tests. N Engl J Med. 1990;322:1499-1504. ABSTRACT


THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES

Evaluation of Outpatient Computerized Physician Medication Order Entry Systems: A Systematic Review
Eslami et al.
J. Am. Med. Inform. Assoc. 2007;14:400-406.
ABSTRACT | FULL TEXT  

The Costs of Denying Scarcity
Alexander et al.
Arch Intern Med 2004;164:593-596.
FULL TEXT  

Engaging Community Mental Health Stakeholders in Pharmacy Cost Management
Baker
Psychiatr. Serv. 2001;52:650-653.
ABSTRACT | FULL TEXT  

Prescription Medication Costs: A Study of Physician Familiarity
Ernst et al.
Arch Fam Med 2000;9:1002-1007.
ABSTRACT | FULL TEXT  




HOME | CURRENT ISSUE | PAST ISSUES | TOPIC COLLECTIONS | CME | PHYSICIAN JOBS | HELP
CONDITIONS OF USE | PRIVACY POLICY | CONTACT US | SITE MAP
 
© 1999 American Medical Association. All Rights Reserved.