- Email: [email protected]
Evaluation of Consumer Drug Information Databases
•
RESEARCH
Evaluation of Consumer Drug Information Databases Jean-Ah Choi, Johanna Sullivan, Marvin Pankaskie, and Jill Brufsky
Objectives: To evaluate prescription drug information contained in six consumer drug information databases available on CD-ROM, and to make health care professionals aware of the information provided, so that they may appropriately recommend these databases for use by their patients. Design: Observational study of six consumer drug information databases: The Corner Drug Store, Home Medical Advisor, Mayo Clinic Family Pharmacist, Medical Drug Reference, Mosby's Medical Encyclopedia, and PharmAssist. Setting: Not
applicable. Patients or Other Participants: Not applicable. Interventions: Information on 20 frequently prescribed drugs was evaluated in each database. The databases were ranked using a point-scale system based on primary and secondary assessment criteria.
Main Outcome Measures: For the primary assessment, 20 categories of information based on those included in the 1998 edition of the USP 01 Volume I/, Advice for the Patient: Drug Information in Lay Language were evaluated for each of the 20 drugs, and each database could earn up to 400 points (for example, 1 point was awarded if the database mentioned a drug's mechanism of action). For the secondary assessment, the inclusion of 8 additional features that could enhance the utility of the databases was evaluated (for example, 1 point was awarded if the database contained a picture of the drug), and each database could earn up to 8 points. Results: The results of the primary and secondary assessments, listed in order of highest to lowest number of points earned, are as follows: Primary assessment-Mayo Clinic Family Pharmacist (379), Medical Drug Reference (251), PharmAssist (176), Home Medical Advisor (113.5), The Corner Drug Store (98), and Mosby's Medical Encyclopedia (18.5); secondary assessment- The Mayo Clinic Family Pharmacist (8), The Corner Drug Store (5), Mosby's Medical Encyclopedia (5), Home Medical Advisor (4), Medical Drug Reference (4), and PharmAssist (3). Conclusion: The Mayo Clinic Family Pharmacist was the most accurate and complete source of prescription drug informa-
tion based on the USP 01 Volume /I and would be an appropriate database for health care professionals to recommend to patients. JAm Pharm Assoc. 1999;39:683-7.
Previous studies have evaluated both written and verbal communication as effective means of disseminating drug information
Received October 6, 1998, and in revised form February 19, 1999. Accepted for publication March 3, 1999. Jean-Ah Choi, PharmD, is assistant professor, Division of Information Systems, Bernard J. Dunn School of Pharmacy, Shenandoah University, Winchester, Va.; at the time of this study she was specialized resident in drug information, University of Pittsburgh. Johanna (Jodi) Sullivan, PharmD, is research manager in applied health outcomes, Pharmacia & Upjohn, Tampa, Fla.; at the time of this study she was assistant professor, School of Pharmacy, University of Pittsburgh. Marvin Pankaskie, PhD, is software developer, clinical pharmacology, Gold Standard Multimedia, Inc., Tampa, Fla.; at the time of this study he was with the Center for Biomedical Informatics, Pittsburgh, Pa. Jill Brufsky, PharmD, is chief, Drug Information and Pharmacoepidemiology Center, University of Pittsburgh Medical Center Health System, and assistant professor, Department of Pharmacy and Therapeutics, School of Pharmacy, University of Pittsburgh. Correspondence: Jean-Ah Choi, PharmD, Division of Information Systems, Bernard J. Dunn School of Pharmacy, Shenandoah University, 1775 N. Sector Court, Winchester VA 22601. Fax: 540-665-1283. E-mail: [email protected].
Vol. 39, No.5
September/October 1999
to patients. Morris l ascertained that written resources as a complement to verbal consultation may provide an effective means of conveying prescription drug information. Morris and Halperin2 conducted a literature review on the effects of written drug information on patient knowledge and compliance; several of the studies3- 5 they looked at suggested that written drug information can improve patient compliance and knowledge of drug therapy. Culbertson and colleagues6 surveyed patients at six pharmacy health care sites to determine consumer preferences for verbal or written drug information. They randomly selected patients who received new or refill prescriptions during a period of 4 months. Each site had to contribute at least 10 patients per week. In their analysis, they observed that 62% of the 317 respondents preferred a combination of written and verbal communication. In addition to written and verbal drug information, electronic resources also provide a useful means for educating consumers. Belgado and coinvestigators7 evaluated the utility of electronic drug information resources for answering questions received by decentralized pharmacists. They assessed seven databases that
Journal of tbe American Pharmaceutical Association
683
RESEARCH
Drug Information Databases
were designed specifically for health care professionals, not consumers. The investigators assessed the utility of each database in answering 100 questions, randomly selected from 500, that had been collected from decentralized pharmacists. They also evaluated the cost and user-friendliness of the databases. The present study was designed to assess the types of drug information provided to patients on computerized databases. We evaluated six consumer drug information databases based on 20 categories of information provided in the 1998 edition of the USP DI Volume II, Advice for the Patient: Drug Information in Lay Language. 8 To date, no studies evaluating consumer databases as a source of prescription drug information have been published. Text resources for consumer drug information, such as The Pill Book, The PDR Family Guide, and The USP Guide to Medicines, are available for purchase at bookstores; however, increasingly, many consumers may rely on computerized databases as a source of prescription drug information. CD-ROMs containing information on prescription drugs are written specifically for consumers and are available with a multitude of features at varying costs.
Objectives Our objectives were to evaluate the prescription drug information in six consumer databases by analyzing the accuracy and completeness of the information provided, and to ascertain the inclusion of additional features that could enhance the utility of the databases. Our overall intent was to make health care professionals aware of the information provided, so that they may make better-informed recommendations to their patients.
Methods The six consumer drug information databases we evaluated were the only currently available CD-ROMs found to contain prescription drug information at the initiation of the study in October 1997. They were obtained through various software suppliers. Table 1 lists their prices and availability. We selected 20 prescription drugs (see Table 2) from the list of the top 200 prescription drugs contained in the 1997 edition of the Drug Topics Redbook. 9 Every tenth drug was selected, beginning with the most commonly dispensed drug (as measured by the National Prescription Audit of IMS America,1O for the lO-month period ending October 1996, based on dispensed volume of drugs through independent, chain, and food store retail pharmacies). The primary end points of this study were the accuracy and completeness of the prescription drug information provided in each of database, as compared with the USP DI Volume II, Advice for the Patient: Drug Information in Lay Language, 1998 edition. This annually updated and reputable source of prescription drug information is written in lay language and is intended for use by consumers. The preface encourages health care professionals to
684
Journal of the American Phannaceutical Association
reproduce pages for distribution to their patients or clients. ~ son and colleagues6 also used the USP DI Volume II as the sarce of written drug information in their study of consumer prefereoces for written or verbal drug information. Information on those drugs that were included in each database was assessed using a point. scale system based on 20 categories included in the USP Dl. The point-scale system used in the primary assessment allowed for objective comparison of the accuracy and completeness of drug information provided in the databases. Allocating 0, ~, or I point for each category enabled an assessment of the complete· ness of drug information provided. Assigning 0 points for any mistakes in the information allowed the accuracy of drug infor· mation to be assessed. If more than 50% of the information contained in the USP DI for each category was provided in the database for the drug, 1 point was awarded. If 50% or less of the information was includ· ed, 7.! point was given. If there was no information on the specific category, or if the information provided was inaccurate, 0 points were awarded. Inaccuracy was defined as any information thatdif· fered from that in the USP DI, including misspelled names. For example, if 10 drug interactions were listed in the USP Dl and at least 6 of those 10 drug interactions were correctly listed in the database, 1 point was awarded in this category. If one to five of the drug interactions in the USP DI were correctly listed in the database, 7.! point was awarded in this category. If no drug interacI tions were listed, or if any of the drug interactions listed were inac· curate, 0 points were awarded. Thus, each drug in the database could eam at most 20 points (I point for each category), and each database could eam up to 400 points for the primary assessment , The secondary end point of this study was the inclusion of eight additional features that could possibly enhance the utility of the database to consumers. Use of lay language was detennined by employing the Readability Graph,1O developed by Edward Fry at the Rutgers University Reading Center. According to the direc· tions for using the Readability Graph, three 100-word passages from each database and the USP DI were chosen from near the beginning, middle, and end of each database. The average number of syllables and average number of words per sentence were plot· ted on the graph to determine the approximate reading grade level. The USP DI was determined to be written at a ninth-grade reading level. Therefore, any database that was at the ninth-grade reading level or below was awarded I point. Any database that was above the ninth-grade reading level was given 0 points. Each additional feature that was included in the database was assigned I point; therefore, each database could eam up to 8 points in the secondary assessment.
Results Tables 3 and 4 show the points awarded in each category for the primary and secondary assessments, respectively. Mayo Clinic Family Pharmacist earned the highest score
SeptemberlOctober 1999 Vol. 39,No.5
I
Drug Information Databases
RESEARCH
Table 1. Cost and Availability of Databases& Cost
Database The Corner Drug Store
$29.95
Manufacturer
Availability Software supplier
Pixel Perfect Software, 1995
Home Medical Advisor
$29.95
Computer store, software supplier
The Learning Company, 1997
Mayo Clinic Family Pharmacist
$29.99
Computer store, software supplier
IVI Publishing Inc., 1997
Medica/ Drug Reference
$29.00
Software supplier
Parsons Technology, 1997
Mosby's Medical Encyclopedia
$49.99
Computer store, software supplier
The Learning Company, 1997
PharmAssist
$49.95
Software supplier
N8 Productions and Softkey Multimedia, Inc., 1993
'Purchased in November 1997.
(379/400) for the primary assessment. The information contained in this database was arranged by categories as listed in the USP D1. The database lacked information in the areas of foreign names and non-FDA-approved indications/additional information. Medical Drug Reference (251/400) listed all 20 drugs and contained accurate information on the majority of categories assessed. It lacked information in the areas of non-FDA-approved indicationsladditional information and uses in children and the elderly. PharmAssist scored less than half of the total possible points (176/400) and did not give any information about non-FDAapproved indications, dosing instructions, and proper storage directions. Home Medical Advisor and The Comer Drug Store received similar scores, with 113.5 and 98 total points, respectively. Although most of the drugs were listed in these two databases, the information included in the majority of the categories was incomplete according to our study criteria. Mosby's Medical Encyclopedia (18.5) did not list 12 of the 20 drugs assessed, and minimal amounts of information were provided for the few drugs that were included. In general, only the name and indications were listed in this database. Mayo Clinic Family Pharmacist earned the highest possible Score of 8 for the secondary assessment. The other databases included at least three but not more than five of the enhanced features.
Discussion It is important for health care professionals to be aware of the infonnation provided in consumer drug information databases, so that they can make infonned recommendations to patients. Consumer drug infonnation databases are easily accessible to patients. We were able to purchase these databases at local computer stores and through software suppliers advertised on the Internet. CD-ROMs containing prescription drug information for consumers cost approximately $30 to $50; drug information text resources for consumers are available for approximately $10 to $ I 00. There were obvious differences among the six databases in the amount of information included for each category in the primary
Vol. 39, No.5
September/October 1999
assessment. Perhaps one explanation for the lack of prescription drug infonnation included in some of these databases is that they do not concentrate solely on providing infonnation about medications. Home Medical Advisor and Mosby's Medical Encyclopedia contain infonnation on diseases, symptoms, human anatomy, and other medical topics in addition to drug infonnation. Therefore, these two sources may be more useful for general medical information than for specific drug information. The Comer Drug Store, Mayo Clinic Family Pharmacist, Medical Drug Reference, and PharmAssist contain information only on medications. The content and accuracy of prescription drug information in the databases as compared with the USP DI Volume II were measured as the primary end points in this study, and recommendaTable 2. Prescription Drugs Chosen for Study Brand Name
Generic Name Conjugated estrogen
Premarin
Warfarin sodium
Coumadin Sodium
Furosemide
NL
Propoxyphene-N/acetaminophen
NL
Metformin
Glucophage
Acetaminophen/codeine
NL
Quinapril
Accupril
Atenolol
NL
Sumatriptan
Imitrex
Medroxyprogesterone
NL
Ibuprofen
NL
Norgestrel!ethinyl estradiol
Lo/Ovral-28
Loracarbef
Lorabid
Propoxyphene-N 100 mgt acetaminophen 650 mg
Darvocet-N 100
Glipizide
NL
Fosinopril
Monopril
Losartan
Cozaar
Norgestimate/ethinyl estradiol
Ortho-Cyclen 28
Metoprolol
Lopressor
Indapamide
Lozol
NL = not listed.
Journal of the American Pharmaceutical Association
685
RESEARCH
Drug Information Databases
-
Table 3. Points Awarded in Each Drug Information Category for Primary Assessment" CDS
Category from USP DI
HMA
MCFP
MDR
MME
PA
20.0
8.0
13.0
17.0
4.0
12.0
Generic names
16.0
18.0
20.0
Brand names
15.5
18.0
20.0
Foreign names
9.5
9.0
12.0
6.5
Mechanism of action
4.0
8.0
20.0
14.5
15.5
15.0
20.0
19.5
0
10.0
0.5
FDA-approved indications Non-FDA-approved indications (" Additional Info")
0
5.0
0 1.0
13.0
2.5
12.5
0
0
Various dosage forms
3.5
6.5
20.0
19.0
0
12.0
Allergy warnings
0.5
0
20.0
12.0
0.5
12.0
Use in pregnancy
3.0
5.0
20.0
20.0
0
13.0
Use in breast-feeding
1.0
3.5
20.0
19.0
0
13.0
0
12.0
0
12.0 8.5
Use in children
3.0
2.5
19.0
2.0
Use in elderly
0
0
20.0
2.0
Drug interactions
8.0
9.0
20.0
12.0
0
Use in concomitant medical conditions
1.0
0.5
20.0
13.0
0.5
7.0
20.0
12.5
0
6.0
Proper use of medication
0
0
Dosing instructions
7.5
9.5
20.0
9.0
0
0
Instructions for missed dose
0
0
19.0
19.0
0
12.0
Proper storage directions
0
0
19.0
11.0
0
0
Precautions
2.0
0
20.0
11.5
1.0
6.5
9.0
20.0
11.0
1.0
6.5
18.5
176.0
8.0
Side effects
98.0
Total
251.0
379.0
113.5
aMaximum points for each category = 20; maximum total points = 400. CDS = The Corner Drug Store,' HMA = Home Medical Advisor; MCFP = Mayo Clinic Family Pharmacist; MDR = Medical Drug Reference; MME = Mosby's Medical Encyclopedia; PA = PharmAssist.
tions cannot be made about other types of information included in these databases, All of the databases were available for less than $50. The USP DI Volume 1/ is considerably more expensive, approximately $100. The differences in pricing among the databases could be due to the inclusion of information other than medications. Both Mosby's Medical Encyclopedia and PharmAssist had the highest total cost, but PharmAssist contains only prescription drug infor-
mati on. All of the databases were copyrighted in 1997, except for The Comer Drug Store (1995) and PharmAssist (1993), Howev· er, this had no bearing on the information provided, as many of the drugs evaluated were marketed before 1993. The secondary end points were chosen to objectively detennine whether or not eight features thought to enhance the database were included. These enhancements were thought to increase the utility of the databases to consumers, thereby increasing their
Table 4. Points Awarded in Each Category for Secondary Assessmenta Feature Use of lay language
CDS
HMA
0
0
MCFP
MDR
MME
PA
0
0
0
0
0
Mouse capabilities Capability to search by drug name Pictures of medications
0
Nonprescription drug information Sound features
0
0
0 0
0
Video capabilities
0
Internet links
0
Total
5
4
8
0
0
0
0
4
5
3
aMaximum total points = 8, CDS = The Corner Drug Store; HMA = Home Medical Advisor; MCFP = Mayo Clinic Family Pharmacist; MDR = Medical Drug Reference; MME = Mosby's Medical Encyclopedia; PA = PharmAssist,
686
Journal of the American Pharmaceutical Association
SeptemberlOctober 1999
Vol, 39,No.~
Drug Information Databases
RESEARCH
understanding of the drug information. Other secondary end points which were not evaluated but which may add to the utility of the databases include information about alternative medical therapies or general health and well-being; computer specifications needed to run the databases (RAM, Pentium, etc.); a cost analysis (what is included for the price); whetherlhow often the databases are updated and the prices for updating the database (if applicable); quality of the Internet links provided; accuracy of the nonprescription drug information (our study evaluated only whether the information was present); and the usefulness of sound and video capabilities in enhancing patient education about medications. Since the time of this study, only Medical Drug Reference has been updated (May 1999).
Consumers may purchase these CD-ROM databases at local computer stores or via software suppliers on the Internet. Such easy access makes it even more important for health care professionals to be aware of the accuracy and completeness of the drug information provided. Several of the databases were found to be incomplete and outdated. Mayo Clinic Family Pharmacist was determined to be the most accurate and complete source of prescription drug information out of the six databases evaluated when using the 1998 edition of the USP DI Volume II, Advice for the Patient: Drug Information in Lay Language as the standard reference for comparison. In addition, it contained more enhanced features than the other databases, which may increase its utility to consumers.
Limitations
The authors declare no conflicts of interest or financial interests in any product, or service mentioned in this manuscript, including grants, employment, gifts, or honoraria.
One potential bias in this study is the fact that Mayo Clinic Family Pharmacist references the 1997 edition of the USP DI, Advice for the Patient; however, this detail was not known at the initiation of this study. Another limitation involves the measurement of consumer usability as part of the secondary assessment. Features such as mouse capabilities and type-in search areas should enhance consumer usability. However, what a patient would actually consider useful was not determined. The secondary end points were assessed only by whether or not the enhanced feature was included in the database. No comparisons were made as to the quality of these enhanced features or whether these features added value to the database. In addition, the quality of nonprescription drug information was not evaluated. Future studies should actually have consumers use each of the databases. This would allow the investigators to assess the value of the enhanced features in increasing the utility of the databases.
Conclusion It is important for health care professionals to be aware of the prescription drug information provided in consumer databases, so that they may appropriately recommend the most accurate and complete resources to patients. With the recent trend toward the use of computer technology as a source of knowledge, motivated consumers may increasingly rely on these databases for prescription drug information.
Vol. 39, No.5
September/October 1999
Supported in part by a grant from the University of Pittsburgh Pharmacy Associates. Presented at the Eastern States Conference for Pharmacy Residents and Preceptors, Baltimore, Md., April 24, 1998.
References 1. Morris LA. Rationale for patient package inserts. Am J Hosp Pharm. 1978:35: 179-84. 2. Morris LA, Halperin JA. Effects of written drug information on patient knowledge and compliance: a literature review. Am J Public Health. 1979:69:47-52.
3. Sharpe TR. Mikeal RL. Patient compliance with antibiotic regimens. Am J Hosp Pharm. 1974;31 :479--84. 4. Lima J, Nazarian L, Charney E, et al. Compliance with short-term antimicrobial therapy. Pediatrics. 1976;57:383--6. 5. Linkewich JA, Catalano RB, Flack HL. The effect of packaging and instruction on outpatient compliance with medication regimens. Drug Intell Clin Pharm. 1974;8: 10--15. 6. Culbertson VL, Arthur TG, Rhodes PJ, et al. Consumer preferences for verbal and written medication information. Drug Intell Clin Pharm. 1988;22:390--6. 7. Belgado BS, Hatton RC, Doering PL. Evaluation of electronic drug information resources for answering questions received by decentralized pharmacists. Am J Health-Syst Pharm. 1997;54:2592--6. 8. Halperin JA, ed. USP 01 Volume II: Advice for the Patient: Drug Infor· mation in Lay Language. Englewood, Colo: Micromedex; 1997; Rockville, Md: United States Pharmacopeial Convention, Inc; 198096:115-1802. 9. Cardinale V, ed. 1997 Drug Topics Redbook. Montvale, NJ: Medical Economics Company, Inc; 1997:124. 10. IMS America, Ltd. National Prescription Audit (NPA). Westport, Conn: IMS America, Ltd; 1996. 11. Fry E. The readability formula that saves time. J Reading. 1968;11(7):513--78.
Journal of the American Pharmaceutical Association
687
RESEARCH
Evaluation of Consumer Drug Information Databases Jean-Ah Choi, Johanna Sullivan, Marvin Pankaskie, and Jill Brufsky
Objectives: To evaluate prescription drug information contained in six consumer drug information databases available on CD-ROM, and to make health care professionals aware of the information provided, so that they may appropriately recommend these databases for use by their patients. Design: Observational study of six consumer drug information databases: The Corner Drug Store, Home Medical Advisor, Mayo Clinic Family Pharmacist, Medical Drug Reference, Mosby's Medical Encyclopedia, and PharmAssist. Setting: Not
applicable. Patients or Other Participants: Not applicable. Interventions: Information on 20 frequently prescribed drugs was evaluated in each database. The databases were ranked using a point-scale system based on primary and secondary assessment criteria.
Main Outcome Measures: For the primary assessment, 20 categories of information based on those included in the 1998 edition of the USP 01 Volume I/, Advice for the Patient: Drug Information in Lay Language were evaluated for each of the 20 drugs, and each database could earn up to 400 points (for example, 1 point was awarded if the database mentioned a drug's mechanism of action). For the secondary assessment, the inclusion of 8 additional features that could enhance the utility of the databases was evaluated (for example, 1 point was awarded if the database contained a picture of the drug), and each database could earn up to 8 points. Results: The results of the primary and secondary assessments, listed in order of highest to lowest number of points earned, are as follows: Primary assessment-Mayo Clinic Family Pharmacist (379), Medical Drug Reference (251), PharmAssist (176), Home Medical Advisor (113.5), The Corner Drug Store (98), and Mosby's Medical Encyclopedia (18.5); secondary assessment- The Mayo Clinic Family Pharmacist (8), The Corner Drug Store (5), Mosby's Medical Encyclopedia (5), Home Medical Advisor (4), Medical Drug Reference (4), and PharmAssist (3). Conclusion: The Mayo Clinic Family Pharmacist was the most accurate and complete source of prescription drug informa-
tion based on the USP 01 Volume /I and would be an appropriate database for health care professionals to recommend to patients. JAm Pharm Assoc. 1999;39:683-7.
Previous studies have evaluated both written and verbal communication as effective means of disseminating drug information
Received October 6, 1998, and in revised form February 19, 1999. Accepted for publication March 3, 1999. Jean-Ah Choi, PharmD, is assistant professor, Division of Information Systems, Bernard J. Dunn School of Pharmacy, Shenandoah University, Winchester, Va.; at the time of this study she was specialized resident in drug information, University of Pittsburgh. Johanna (Jodi) Sullivan, PharmD, is research manager in applied health outcomes, Pharmacia & Upjohn, Tampa, Fla.; at the time of this study she was assistant professor, School of Pharmacy, University of Pittsburgh. Marvin Pankaskie, PhD, is software developer, clinical pharmacology, Gold Standard Multimedia, Inc., Tampa, Fla.; at the time of this study he was with the Center for Biomedical Informatics, Pittsburgh, Pa. Jill Brufsky, PharmD, is chief, Drug Information and Pharmacoepidemiology Center, University of Pittsburgh Medical Center Health System, and assistant professor, Department of Pharmacy and Therapeutics, School of Pharmacy, University of Pittsburgh. Correspondence: Jean-Ah Choi, PharmD, Division of Information Systems, Bernard J. Dunn School of Pharmacy, Shenandoah University, 1775 N. Sector Court, Winchester VA 22601. Fax: 540-665-1283. E-mail: [email protected].
Vol. 39, No.5
September/October 1999
to patients. Morris l ascertained that written resources as a complement to verbal consultation may provide an effective means of conveying prescription drug information. Morris and Halperin2 conducted a literature review on the effects of written drug information on patient knowledge and compliance; several of the studies3- 5 they looked at suggested that written drug information can improve patient compliance and knowledge of drug therapy. Culbertson and colleagues6 surveyed patients at six pharmacy health care sites to determine consumer preferences for verbal or written drug information. They randomly selected patients who received new or refill prescriptions during a period of 4 months. Each site had to contribute at least 10 patients per week. In their analysis, they observed that 62% of the 317 respondents preferred a combination of written and verbal communication. In addition to written and verbal drug information, electronic resources also provide a useful means for educating consumers. Belgado and coinvestigators7 evaluated the utility of electronic drug information resources for answering questions received by decentralized pharmacists. They assessed seven databases that
Journal of tbe American Pharmaceutical Association
683
RESEARCH
Drug Information Databases
were designed specifically for health care professionals, not consumers. The investigators assessed the utility of each database in answering 100 questions, randomly selected from 500, that had been collected from decentralized pharmacists. They also evaluated the cost and user-friendliness of the databases. The present study was designed to assess the types of drug information provided to patients on computerized databases. We evaluated six consumer drug information databases based on 20 categories of information provided in the 1998 edition of the USP DI Volume II, Advice for the Patient: Drug Information in Lay Language. 8 To date, no studies evaluating consumer databases as a source of prescription drug information have been published. Text resources for consumer drug information, such as The Pill Book, The PDR Family Guide, and The USP Guide to Medicines, are available for purchase at bookstores; however, increasingly, many consumers may rely on computerized databases as a source of prescription drug information. CD-ROMs containing information on prescription drugs are written specifically for consumers and are available with a multitude of features at varying costs.
Objectives Our objectives were to evaluate the prescription drug information in six consumer databases by analyzing the accuracy and completeness of the information provided, and to ascertain the inclusion of additional features that could enhance the utility of the databases. Our overall intent was to make health care professionals aware of the information provided, so that they may make better-informed recommendations to their patients.
Methods The six consumer drug information databases we evaluated were the only currently available CD-ROMs found to contain prescription drug information at the initiation of the study in October 1997. They were obtained through various software suppliers. Table 1 lists their prices and availability. We selected 20 prescription drugs (see Table 2) from the list of the top 200 prescription drugs contained in the 1997 edition of the Drug Topics Redbook. 9 Every tenth drug was selected, beginning with the most commonly dispensed drug (as measured by the National Prescription Audit of IMS America,1O for the lO-month period ending October 1996, based on dispensed volume of drugs through independent, chain, and food store retail pharmacies). The primary end points of this study were the accuracy and completeness of the prescription drug information provided in each of database, as compared with the USP DI Volume II, Advice for the Patient: Drug Information in Lay Language, 1998 edition. This annually updated and reputable source of prescription drug information is written in lay language and is intended for use by consumers. The preface encourages health care professionals to
684
Journal of the American Phannaceutical Association
reproduce pages for distribution to their patients or clients. ~ son and colleagues6 also used the USP DI Volume II as the sarce of written drug information in their study of consumer prefereoces for written or verbal drug information. Information on those drugs that were included in each database was assessed using a point. scale system based on 20 categories included in the USP Dl. The point-scale system used in the primary assessment allowed for objective comparison of the accuracy and completeness of drug information provided in the databases. Allocating 0, ~, or I point for each category enabled an assessment of the complete· ness of drug information provided. Assigning 0 points for any mistakes in the information allowed the accuracy of drug infor· mation to be assessed. If more than 50% of the information contained in the USP DI for each category was provided in the database for the drug, 1 point was awarded. If 50% or less of the information was includ· ed, 7.! point was given. If there was no information on the specific category, or if the information provided was inaccurate, 0 points were awarded. Inaccuracy was defined as any information thatdif· fered from that in the USP DI, including misspelled names. For example, if 10 drug interactions were listed in the USP Dl and at least 6 of those 10 drug interactions were correctly listed in the database, 1 point was awarded in this category. If one to five of the drug interactions in the USP DI were correctly listed in the database, 7.! point was awarded in this category. If no drug interacI tions were listed, or if any of the drug interactions listed were inac· curate, 0 points were awarded. Thus, each drug in the database could eam at most 20 points (I point for each category), and each database could eam up to 400 points for the primary assessment , The secondary end point of this study was the inclusion of eight additional features that could possibly enhance the utility of the database to consumers. Use of lay language was detennined by employing the Readability Graph,1O developed by Edward Fry at the Rutgers University Reading Center. According to the direc· tions for using the Readability Graph, three 100-word passages from each database and the USP DI were chosen from near the beginning, middle, and end of each database. The average number of syllables and average number of words per sentence were plot· ted on the graph to determine the approximate reading grade level. The USP DI was determined to be written at a ninth-grade reading level. Therefore, any database that was at the ninth-grade reading level or below was awarded I point. Any database that was above the ninth-grade reading level was given 0 points. Each additional feature that was included in the database was assigned I point; therefore, each database could eam up to 8 points in the secondary assessment.
Results Tables 3 and 4 show the points awarded in each category for the primary and secondary assessments, respectively. Mayo Clinic Family Pharmacist earned the highest score
SeptemberlOctober 1999 Vol. 39,No.5
I
Drug Information Databases
RESEARCH
Table 1. Cost and Availability of Databases& Cost
Database The Corner Drug Store
$29.95
Manufacturer
Availability Software supplier
Pixel Perfect Software, 1995
Home Medical Advisor
$29.95
Computer store, software supplier
The Learning Company, 1997
Mayo Clinic Family Pharmacist
$29.99
Computer store, software supplier
IVI Publishing Inc., 1997
Medica/ Drug Reference
$29.00
Software supplier
Parsons Technology, 1997
Mosby's Medical Encyclopedia
$49.99
Computer store, software supplier
The Learning Company, 1997
PharmAssist
$49.95
Software supplier
N8 Productions and Softkey Multimedia, Inc., 1993
'Purchased in November 1997.
(379/400) for the primary assessment. The information contained in this database was arranged by categories as listed in the USP D1. The database lacked information in the areas of foreign names and non-FDA-approved indications/additional information. Medical Drug Reference (251/400) listed all 20 drugs and contained accurate information on the majority of categories assessed. It lacked information in the areas of non-FDA-approved indicationsladditional information and uses in children and the elderly. PharmAssist scored less than half of the total possible points (176/400) and did not give any information about non-FDAapproved indications, dosing instructions, and proper storage directions. Home Medical Advisor and The Comer Drug Store received similar scores, with 113.5 and 98 total points, respectively. Although most of the drugs were listed in these two databases, the information included in the majority of the categories was incomplete according to our study criteria. Mosby's Medical Encyclopedia (18.5) did not list 12 of the 20 drugs assessed, and minimal amounts of information were provided for the few drugs that were included. In general, only the name and indications were listed in this database. Mayo Clinic Family Pharmacist earned the highest possible Score of 8 for the secondary assessment. The other databases included at least three but not more than five of the enhanced features.
Discussion It is important for health care professionals to be aware of the infonnation provided in consumer drug information databases, so that they can make infonned recommendations to patients. Consumer drug infonnation databases are easily accessible to patients. We were able to purchase these databases at local computer stores and through software suppliers advertised on the Internet. CD-ROMs containing prescription drug information for consumers cost approximately $30 to $50; drug information text resources for consumers are available for approximately $10 to $ I 00. There were obvious differences among the six databases in the amount of information included for each category in the primary
Vol. 39, No.5
September/October 1999
assessment. Perhaps one explanation for the lack of prescription drug infonnation included in some of these databases is that they do not concentrate solely on providing infonnation about medications. Home Medical Advisor and Mosby's Medical Encyclopedia contain infonnation on diseases, symptoms, human anatomy, and other medical topics in addition to drug infonnation. Therefore, these two sources may be more useful for general medical information than for specific drug information. The Comer Drug Store, Mayo Clinic Family Pharmacist, Medical Drug Reference, and PharmAssist contain information only on medications. The content and accuracy of prescription drug information in the databases as compared with the USP DI Volume II were measured as the primary end points in this study, and recommendaTable 2. Prescription Drugs Chosen for Study Brand Name
Generic Name Conjugated estrogen
Premarin
Warfarin sodium
Coumadin Sodium
Furosemide
NL
Propoxyphene-N/acetaminophen
NL
Metformin
Glucophage
Acetaminophen/codeine
NL
Quinapril
Accupril
Atenolol
NL
Sumatriptan
Imitrex
Medroxyprogesterone
NL
Ibuprofen
NL
Norgestrel!ethinyl estradiol
Lo/Ovral-28
Loracarbef
Lorabid
Propoxyphene-N 100 mgt acetaminophen 650 mg
Darvocet-N 100
Glipizide
NL
Fosinopril
Monopril
Losartan
Cozaar
Norgestimate/ethinyl estradiol
Ortho-Cyclen 28
Metoprolol
Lopressor
Indapamide
Lozol
NL = not listed.
Journal of the American Pharmaceutical Association
685
RESEARCH
Drug Information Databases
-
Table 3. Points Awarded in Each Drug Information Category for Primary Assessment" CDS
Category from USP DI
HMA
MCFP
MDR
MME
PA
20.0
8.0
13.0
17.0
4.0
12.0
Generic names
16.0
18.0
20.0
Brand names
15.5
18.0
20.0
Foreign names
9.5
9.0
12.0
6.5
Mechanism of action
4.0
8.0
20.0
14.5
15.5
15.0
20.0
19.5
0
10.0
0.5
FDA-approved indications Non-FDA-approved indications (" Additional Info")
0
5.0
0 1.0
13.0
2.5
12.5
0
0
Various dosage forms
3.5
6.5
20.0
19.0
0
12.0
Allergy warnings
0.5
0
20.0
12.0
0.5
12.0
Use in pregnancy
3.0
5.0
20.0
20.0
0
13.0
Use in breast-feeding
1.0
3.5
20.0
19.0
0
13.0
0
12.0
0
12.0 8.5
Use in children
3.0
2.5
19.0
2.0
Use in elderly
0
0
20.0
2.0
Drug interactions
8.0
9.0
20.0
12.0
0
Use in concomitant medical conditions
1.0
0.5
20.0
13.0
0.5
7.0
20.0
12.5
0
6.0
Proper use of medication
0
0
Dosing instructions
7.5
9.5
20.0
9.0
0
0
Instructions for missed dose
0
0
19.0
19.0
0
12.0
Proper storage directions
0
0
19.0
11.0
0
0
Precautions
2.0
0
20.0
11.5
1.0
6.5
9.0
20.0
11.0
1.0
6.5
18.5
176.0
8.0
Side effects
98.0
Total
251.0
379.0
113.5
aMaximum points for each category = 20; maximum total points = 400. CDS = The Corner Drug Store,' HMA = Home Medical Advisor; MCFP = Mayo Clinic Family Pharmacist; MDR = Medical Drug Reference; MME = Mosby's Medical Encyclopedia; PA = PharmAssist.
tions cannot be made about other types of information included in these databases, All of the databases were available for less than $50. The USP DI Volume 1/ is considerably more expensive, approximately $100. The differences in pricing among the databases could be due to the inclusion of information other than medications. Both Mosby's Medical Encyclopedia and PharmAssist had the highest total cost, but PharmAssist contains only prescription drug infor-
mati on. All of the databases were copyrighted in 1997, except for The Comer Drug Store (1995) and PharmAssist (1993), Howev· er, this had no bearing on the information provided, as many of the drugs evaluated were marketed before 1993. The secondary end points were chosen to objectively detennine whether or not eight features thought to enhance the database were included. These enhancements were thought to increase the utility of the databases to consumers, thereby increasing their
Table 4. Points Awarded in Each Category for Secondary Assessmenta Feature Use of lay language
CDS
HMA
0
0
MCFP
MDR
MME
PA
0
0
0
0
0
Mouse capabilities Capability to search by drug name Pictures of medications
0
Nonprescription drug information Sound features
0
0
0 0
0
Video capabilities
0
Internet links
0
Total
5
4
8
0
0
0
0
4
5
3
aMaximum total points = 8, CDS = The Corner Drug Store; HMA = Home Medical Advisor; MCFP = Mayo Clinic Family Pharmacist; MDR = Medical Drug Reference; MME = Mosby's Medical Encyclopedia; PA = PharmAssist,
686
Journal of the American Pharmaceutical Association
SeptemberlOctober 1999
Vol, 39,No.~
Drug Information Databases
RESEARCH
understanding of the drug information. Other secondary end points which were not evaluated but which may add to the utility of the databases include information about alternative medical therapies or general health and well-being; computer specifications needed to run the databases (RAM, Pentium, etc.); a cost analysis (what is included for the price); whetherlhow often the databases are updated and the prices for updating the database (if applicable); quality of the Internet links provided; accuracy of the nonprescription drug information (our study evaluated only whether the information was present); and the usefulness of sound and video capabilities in enhancing patient education about medications. Since the time of this study, only Medical Drug Reference has been updated (May 1999).
Consumers may purchase these CD-ROM databases at local computer stores or via software suppliers on the Internet. Such easy access makes it even more important for health care professionals to be aware of the accuracy and completeness of the drug information provided. Several of the databases were found to be incomplete and outdated. Mayo Clinic Family Pharmacist was determined to be the most accurate and complete source of prescription drug information out of the six databases evaluated when using the 1998 edition of the USP DI Volume II, Advice for the Patient: Drug Information in Lay Language as the standard reference for comparison. In addition, it contained more enhanced features than the other databases, which may increase its utility to consumers.
Limitations
The authors declare no conflicts of interest or financial interests in any product, or service mentioned in this manuscript, including grants, employment, gifts, or honoraria.
One potential bias in this study is the fact that Mayo Clinic Family Pharmacist references the 1997 edition of the USP DI, Advice for the Patient; however, this detail was not known at the initiation of this study. Another limitation involves the measurement of consumer usability as part of the secondary assessment. Features such as mouse capabilities and type-in search areas should enhance consumer usability. However, what a patient would actually consider useful was not determined. The secondary end points were assessed only by whether or not the enhanced feature was included in the database. No comparisons were made as to the quality of these enhanced features or whether these features added value to the database. In addition, the quality of nonprescription drug information was not evaluated. Future studies should actually have consumers use each of the databases. This would allow the investigators to assess the value of the enhanced features in increasing the utility of the databases.
Conclusion It is important for health care professionals to be aware of the prescription drug information provided in consumer databases, so that they may appropriately recommend the most accurate and complete resources to patients. With the recent trend toward the use of computer technology as a source of knowledge, motivated consumers may increasingly rely on these databases for prescription drug information.
Vol. 39, No.5
September/October 1999
Supported in part by a grant from the University of Pittsburgh Pharmacy Associates. Presented at the Eastern States Conference for Pharmacy Residents and Preceptors, Baltimore, Md., April 24, 1998.
References 1. Morris LA. Rationale for patient package inserts. Am J Hosp Pharm. 1978:35: 179-84. 2. Morris LA, Halperin JA. Effects of written drug information on patient knowledge and compliance: a literature review. Am J Public Health. 1979:69:47-52.
3. Sharpe TR. Mikeal RL. Patient compliance with antibiotic regimens. Am J Hosp Pharm. 1974;31 :479--84. 4. Lima J, Nazarian L, Charney E, et al. Compliance with short-term antimicrobial therapy. Pediatrics. 1976;57:383--6. 5. Linkewich JA, Catalano RB, Flack HL. The effect of packaging and instruction on outpatient compliance with medication regimens. Drug Intell Clin Pharm. 1974;8: 10--15. 6. Culbertson VL, Arthur TG, Rhodes PJ, et al. Consumer preferences for verbal and written medication information. Drug Intell Clin Pharm. 1988;22:390--6. 7. Belgado BS, Hatton RC, Doering PL. Evaluation of electronic drug information resources for answering questions received by decentralized pharmacists. Am J Health-Syst Pharm. 1997;54:2592--6. 8. Halperin JA, ed. USP 01 Volume II: Advice for the Patient: Drug Infor· mation in Lay Language. Englewood, Colo: Micromedex; 1997; Rockville, Md: United States Pharmacopeial Convention, Inc; 198096:115-1802. 9. Cardinale V, ed. 1997 Drug Topics Redbook. Montvale, NJ: Medical Economics Company, Inc; 1997:124. 10. IMS America, Ltd. National Prescription Audit (NPA). Westport, Conn: IMS America, Ltd; 1996. 11. Fry E. The readability formula that saves time. J Reading. 1968;11(7):513--78.
Journal of the American Pharmaceutical Association
687