- Email: [email protected]
A Survey of Selected Internet Pharmacies in the United States
RESEARCH
A Survey of Selected Internet Pharmacies in the United States Andrew M. Peterson
Objective: To determine whether differences in the provision of pharmacy services exist among different types of Internet pharmacies. Design: Survey of selected pharmacies with a presence on the Internet. Data were abstracted onto a data collection form for further analysis. Data collection was limited to 3 weeks. Setting: U.S.-based Internet pharmacies that allow patients to purchase prescription medications online. Pharmacies were identified using a metasearch engine with the search terms "Internet pharmacy" and "Internet pharmacist." Intervention: Survey. Main Outcome Measures: Comparisons of availability of 10 commonly used products representing a variety of product categories, prescription verification methods, and privacy issues; and determinations of site navigability, drug information and provider access, and payment methods. Sites were categorized as "chain pharmacy extensions," "mail order pharmacies," "independent pharmacy extensions," and "online pharmacies." Results: Thirty-three sites were reviewed. There was significant variation among the four types of pharmacies selling prescriptions over the Internet. Most pharmacies provided all of the drugs in the survey. Patients were required to provide their own prescription at 88% of the sites, and 75% of sites used mail or fax to verify prescription integrity. More than 50% of sites had privacy policies posted, and 64% used cookies. Chain pharmacy extensions required completion of an average of 10.2 pages to order drugs versus 2.4 to 4 pages for all other site types. Drug information was written at an eighth-grade reading level at 36% of the sites. More than two-thirds of the sites provided a toll-free telephone for a health care professional. Nearly 80% of the sites accepted health insurance, and 95% accepted credit cards; however, only 40% used a secure transmission mechanism for patient or payment information. Conclusion: Internet pharmacies provide varying levels of service. Policies regarding the use of the Internet for obtaining medications should focus on improving the privacy of consumer information and ensuring the secure transmission of financial information. JAm Pharm Assoc. 2001;41:205-12.
The provision of health-related services via the Internet has grown tremendously over the past several years.' Health sites on the World Wide Web provide a broad range of products, as well as clinical and diagnostic services, and connect patients directly with health care providers. Providing rapid, 24-hour access to information is a primary function of the Web. In fact, this convenience has often been considered the driving force behind the growth of the Web in the health care sector. 2 However, the occasional display of inconsistent or inaccurate infonnation may be detrimental to patients. 3 Pharmacies are a growing presence on the Web. Many chain pharmacies have developed an extension of their brick-and-rnorReceived August 4, 2000, and in revised form December 1, 2000. Accepted for publication December 12, 2000. Andrew M. Peterson, PharmD, BCPS, is assistant professor of clinical pharmacy, Philadelphia College of Pharmacy, University of the Sciences in Philadelphia. Correspondence: Andrew M. Peterson, PharmD, BCPS, Philadelphia College of Pharmacy, University of the Sciences, 600 S. 43rd Street, Philadelphia, PA 19104. Fax: 215-59EH3586. E-mail: [email protected].
Vol. 41, No.2
MarcblApril2001
tar stores in cyberspace, sometimes through alliances with existing online pharmacies.4 lndependent pharmacies, especially those that serve a particular population of patients (e.g., infertility patients, HIV-positive patients),5 are also establishing a Web presence. Many do not accept new prescriptions online, but instead cater to their existing clientele by providing medication refills.6 Lifestyle drugs, such as Viagra, appear to represent an exception to this trend, and mail order pharmacies are using the Web as a new avenue for expanding their market. Not surprisingly, many interested organizations have expressed increasing concern about patients' access to pharmaceuticals via the Web. Their concerns include verification of prescription integrity and assurance of appropriate patient-provider communication. The National Association of Boards of Pharmacy (NABP), for example, has taken steps to assure the public of the safety and integrity of Web pharmacy sites. 7 The presence of the Verified Internet Pharmacy Practice Site (VlPPS) seal on a Web site assures consumers that the online pharmacy is certified to dispense medications anywhere in the United States and has met a rigorous 17-point set of criteria that addresses confidentiality,
Joumal of !be American I"hannaceudcaI A&!iodatIoo
205
RESEARCH
Internet Pharmacies
quality assurance, and licensure. 8 While privacy is a global concern for the Internet, it is a particularly compelling issue for sites that provide medical information and access to drugs. Patients often erroneously believe that when they access a Web site, their privacy is protected. To ensure the security of a transaction (i.e., to prevent the information from being read by someone other than the intended recipient), Web sites-in particular, sites that transmit financial informationoften use secure socket layer (SSL) technology. Sites using SSL technology may be identified by a dialog box or a lock icon at the bottom of the screen. Often, the addition of an "s" at the end of the "http//:" ("https//:") prefix on a Web site address, or uniform resource locator (URL), indicates a secure site. Some sites use cookies to track visits. A cookie is a unique number that is automatically assigned to the hard drive of the visitor's computer when he or she first enters a Web site. The cookie is then used to track when the user returns and which pages he or she visits. Cookies are generally transparent; because many Web browsers enable cookies by default, the visitor is typically unaware that they are being used. Although users may believe cookies confer anonymity, they in fact become uniquely identifiable customer IDs when a visitor registers at the site. 9 The California HealthCare Foundation9 reviewed the policies and practices of 21 health-related Web sites, including several well-known online pharmacies, such as drugstore.com and Planetrx.com. Their results showed that visitors to health-related Web sites are not anonymous: Many of the sites use cookies, profiling, and banner ads to collect information about individuals without their knowledge or consent. Furthermore, they found that many sites lacked privacy policies or failed to follow their stated policies. Patients' use of the Web as a means of buying medications and accessing associated information is controversial. Among the issues under discussion are product availability, privacy, verification of prescription authenticity, validity and reliability of drug information, and secure methods of payment. The present study was the first to examine whether different types of pharmacy practice sites provide different levels of service.
Objectives The purpose of this study was to determine whether differences in the provision of pharmacy services exist among different types of Internet pharmacies. To this end, this study compared selected services and product availability among different types of pharmacies purporting to allow online prescription ordering.
Methods Four types of Web pharmacies were defined: (1) chain pharmacy extensions (e.g., Walgreens.com); (2) independent pharmacy
206
Joumal of the American Phannaceutical Association
extensions (e.g., villageapothecary.com); (3) mail order pharmacies extending their service through the Web (e.g., Merck-Medco.com); (4) online pharmacies-those pharmacies that sell medications exclusively online (e.g., Planetrx.com). WebFerret (ZDnet), a metasearch engine, was used to identify pharmacies with a Web presence. Examples of search engines used in this version of WebFerret include goto.com and altavista. com. Search terms used were "Internet pharmacy" and "Internet pharmacist." Microsoft Internet Explorer 4.0 was used as the Web browser. The automatic cookie acceptance function was disabled and the secure transmission notification function was enabled. All pharmacies obviously based outside the United States, and duplicates from different search engines, were removed from the sample. Only those sites that allowed electronic ordering, via email or online application, of prescription medications were included in the final analysis. Sites allowing only refills of existing prescriptions or providing only a telephone number for the patient to call to request a prescription were excluded. A data collection instrument was developed containing questions and/or statements pertaining to various aspects of Web pharmacies. For each item, investigators could indicate "yes," "no," or "not applicable." There was a space for comments beside each item. Data were collected during a 3-week period from March 22 through April 12, 2000. Two pharmacy students reviewed each site and collected the data, and the primary investigator reviewed all reports and double-checked all information recorded against the original Web site. One aspect studied was the availability of prescription and nonprescription drugs. Particular attention was paid to common controlled substances (CII and em), unstable or dated medications (amoxicillin suspension and progesterone suppositories), and generic and brand name drugs covering a variety of diseases and/or patient populations (e.g., Pre marin tablets, Viagra, hydrochlorothiazide). The availability of written drug and health information from each Web site was investigated. Sildenafil (Viagra-Pfizer) was chosen as an example for determining the reading level of the drug information offered. This product was chosen because it was felt that a large number of sites would be offering it, given its extensive coverage by the media immediately before the study. The study also probed the availability of a health care practitioner via telephone and e-mail and the methods of payment available to the consumer. Other topics covered in the survey were prescription verification, security/privacy, and sites' ease of use. Additional information was obtained on security of information transmission through the use of SSL technology, cookies, banner ads, and the posting of privacy policies throughout the ordering process. Descriptive statistics were developed for the survey questions. The Flesch-Kincaid readability scale (available through Microsoft Word 97) was used to detennine reading level of drug information.
MarchlApril2001
Vol. 41, No.2
Internet Pharmacies
RESEARCH
Table 1. Availability of Medications Through Surveyed Internet Pharmacies (n = 33) Chain Extensions (n = 5) (%)
Medication Alprazolam tablets
Mail Order (n =6) (%)
Online (n = 14) (%)
Independent Extensions (n = 8) (%)
PValue"
100
50
57
75
.25
60
50
64
87
.48
Celecoxib tablets
100
67
79
87
.51
Hydrochlorothiazide tablets
100
67
64
87
.19
Insulin injection
Amoxicillin suspension
100
67
64
87
.47
Isotretinoin cream
80
50
57
87
.19
Percocet tablets
80
50
36
62
.33
Premarin tablets
100
67
64
100
.12
20
67
57
100
.03
Sildenafil tablets
100
100
86
87
.72
Over-the-counter medications
100
60
64
100
.04
Progesterone suppositories
'X 2 test.
Comparisons were made using either the X2 or Fisher's exact test for dichotomous data and the Kruskall-Wallis one-way analysis of variance for nonparametric continuous data. All ex's were set at 0.05. Systat version 7.0 software was used to analyze data.
and 8 (24%) were independent pharmacy extensions. Table 1 shows product availability on the Web sites in the sample. With the exception of progesterone suppositories and overthe-counter (OTC) medications, there was generally little, if any, difference in the availability of products. Independent pharmacy extensions provided consumer access to progesterone suppositories more often than did all others (P = .03). Chain pharmacy extensions and independent phannacy extensions provided access to OTC medications more often than did mail order or online pharmacies (P = .04). Table 2 shows the availability of drug information. Chain pharmacy extensions and online pharmacies provided written drug and health information more often than did the others, although these differences were not significant. Only 8 sites (24%) provided an electronic newsletter for patients, and only 12 sites (36%) provided a drug monograph for sildenafil (again, used to determine reading level of the drug information provided). None of the independent pharmacy extensions had a sildenafil monograph available. The reading level of the information provided ranged from seventh- to ninth-grade based on the Flesch-Kincaid scale.
Results Initially, more than 500 pharmacy sites were identified using WebFerret. Sites that did not allow electronic ordering of medications were excluded, as were sites that only provided links to other sites (e.g., RiteAid.com links to drugstore.com). Many sites were excluded because they were not pharmacies but organizational and/or informational sites on the topic of Internet pharmacies. Seven sites (1 %) were excluded because they allowed only refills of existing prescriptions. Of the more than 500 sites identified, 75 were considered to be U .S.-based pharmacy sites; of these, only 33 (44%) were included in the final analysis. Of the 33 sites selected, 5 (15%) were chain pharmacy extensions, 6 (18%) were mail order pharmacies, 14 (42%) were online pharmacies,
Table 2. Availability of Drug Information on Internet Pharmacy Sites (n = 33) Chain Extensions (n = 5)
Mail Order (n = 6)
Online (n = 14)
PValue
62%
.32"
36%
12.5%
.11"
50%
37%
.19" .47"
.44b
Written drug information
100%
50%
71%
Drug interaction information
80%
33%
80%
17%
Internal health/disease information
Independent Extensions (n = 8)
Electronic newsletter
0%
17%
36%
25%
Grade average of Flesch-Kincaid reading level (sildenafil)
7.1
8
8.8
NA
"x 2 test . bKruskall-Wallis.
Vol. 41, No.2
MarchlApril2001
Joomal of the American PbannaceUticaI ~Iion
207
RESEARCH
Internet Pharmacies
Table 3. Internet Pharmacies' Prescription Verification Methods (n Chain Extensions (n = 5)
Mail Order (n = 6)
= 33)
Online (n = 14)
Independent Extensions (n = 8)
PValue a
Patient supplies own prescription Patient can mail prescription Patient can fax prescription
5 5 5
5 3 5
11 11 10
8 6 8
.38 .1 0 .38
Company contacts prescriber Via phone Via e·mail
3 3 0
4 4 0
8 8
6 6 0
.91 .64
·X 2 test.
Table 4. Internet Pharmacies' Display of Pharmacy License (n = 33) Chain Extensions (n =5)(%)
Mail Order (n=6) (%)
Online (n = 14) (%)
Independent Extensions (n=8) (%)
PValue a
Home page states licensure
20
17
29
0
Informs consumer anywhere
40
17
57
13
.36
Licensed in all 50 states b
20
17
14
0
.42
.42
x
a 2 test. bUnless stated, we assumed extensions were not licensed in all 50 states.
There was no statistical difference among the types of pharmacies in the provision of drug interaction information (P = .11). The majority (n = 29; 88%) of the sites reviewed required patients to provide their own prescription (see Table 3). However, several sites, particularly online pharmacies, either allowed the patient to obtain the prescription from the site's online physician or made provisions for contacting the patient's physician directly. There appeared to be no difference in the mechanism of prescription transmission among pharmacy types. Table 4 shows the number of sites reporting licensure status to visitors. Overall, most sites did not report which states had licensed them to dispense; those that did include that information did not place it conspicuously on the home page. VIPPS-certified sites are by definition licensed to dispense in all states. Privacy and confidentiality policies varied significantly among sites (see Tables 5 and 6). Independent pharmacy extensions
required patients to register far less often than did all others (P < .01). Chain pharmacy extensions posted privacy policies significantly more often than did all others (P < .01). There was no significant difference among the pharmacy types in the number of VIPPS-certified sites. However, VIPPS-certified sites were more likely to have a privacy policy (P = .04, Fisher's exact test). Nor was there any difference in the use of banner ads. Independent pharmacy extensions were less likely to use cookies than were all others (P < .01), and all VIPPS-certified sites used cookies. Fewer than 60% of all sites used SSL technology during the ordering process, and there was no statistical difference among sites in that regard. VIPPS-certified sites were more likely to use SSL technology than were all others (P = .02, Fisher's exact test). Table 7 shows characteristics of the sites' ease of use. The only significant difference was seen in chain pharmacy extensions, which required visitors to complete an average of 10.2 pages of
Table 5. Internet Pharmacies' Use of Tools That Affect Privacy and Confidentiality (n = 33) Chain Extensions (n= 5) (%)
Mail Order (n = 6) (%)
Online (n = 14) (%)
Independent Extensions (n =8) (%)
Patient to register online
100
83
100
25
<.01
Privacy policy present
100
16
64
25
< .01
Privacy policy on every page during ordering
100
100
44
50
<.01
40
16
14
0 13
VIPPS-certified
PValue a
.28 <.01
100
66
79
Site uses banner ads
60
33
64
62
.62
Site uses SSL technology
60
33
50
50
.84
Site uses cookies
SSL = secure socket layer; VIPPS = Verified Internet Pharmacy Practice Site. "x2 test.
208 JoomaI or the American Pharmaceutical Association
MardIIApril2001
Vol. 41, No. 2
~
""
i,
I
I
I
I·
~
I
!
:!:
..,~
~
:::
...~
!"'
www.medprescribe.com www.nofrillspharmacy.com
Preferred Rx
ePrescribe.com
Familymeds
Global Rx.com
Catalog.com
MedPrescribe
No Frills Pharmacy
Mail order
Online
Online
Online
Online
Online
Online
5 No No
Yes
No
Yes
Yes
www.ppsrxbymail .com
Postal Prescription Services
Mail order
No
No
No
No
Yes
www.pharmacystore .com
The Pharmacy Store
Mail order
3
3 No
www.merck-medco.com
Merck Medco
Yes
No
Yes
Mail order
3 No Yes (Yes)
No Yes
Yes
Yes (3)
No Yes
Yes
www.at-cost-drugs.com
At-Cost-Drugs
2
No
No
No
No
www.pro-mark- pharmacies.com www.rpspharmacy.com
Pro Mark Pharmacies
AARP Pharmacy
Online
Online www.rxuniverse .com safenetpharmacy.com www .safewebmedical.com starkscripts.com
RX Universe
SafeNet Pharmacy
SafeWeb Medical
StarkScripts. com
Online
Online
Online
Online
Service
www.planetrx.com
Planet Rx .com
Online
catalog.com/highcom
www.globalrx.com
familymeds.com
www.eprescribe.com
www.preferredrx.com
www.abeewell.com
No
Mail order
12
No
ABeeWell Pharmacy
2 2
4 4
7 4
8
Yes (Yes) Yes (No) No Yes (No) No Yes (Yes)
Yes No No No No Yes
Yes Yes No Yes Yes Yes
No No Yes (1) Yes (4) Yes (1) No
Yes No No No Yes
Yes Yes Yes Yes Yes Yes
2 3
3 3
No (No) No Yes (Yes) Yes (No)
No No No No
Yes No Yes Yes
Yes (3) No Yes (1) Yes (0)
No Yes No
Yes Yes Yes
Yes
Yes
Yes
9 Yes (Yes) No Yes No
No
Yes
Yes
2 No No Yes Yes (2)
Yes
Yes
Yes (5)
5 No Yes (No)
No No
Yes Yes
Yes (3)
No Yes
Yes
Yes
Yes (Yes)
Mail order
8
Yes (Yes)
Yes
Yes No
Yes (2)
12
Yes
Yes (Yes)
7
No
Yes (Yes) 12
Yes
No Yes (Yes)
Yes
Yes
VIPPSCertified
No. of Pages to Complete to Order Medication
Yes
Yes
No
Yes (1)
Yes (2)
Uses Cookies
Privacy Policy Present (on Every Page)
No
www.walgreens.com
Walgreens
Chain extension
Yes
No
Yes
Yes
Uses SSL
Uses Banner Ads (No . on Home Page)
No
www.cvs.com
CVS/ Pharmacy
Chain extension
Yes
www.e-pharmacy.com
Eckerd.com
Chain extension
Yes
Yes
www.drugstore.com
www.drugemporium.com
URL
Drugstore.com
Drug Emporium
Name
Chain extension
Chain extension
Type
Patient Required to Register
Table 6. Classification of Web Sites and Related Privacy and Confidentiality Issues
::c
~
~
U)
Cfj
(b.
Q) ()
3
Q)
.,
~
"ll
~
::J
5" Cti ..,
RESEARCH
Internet Pharmacies
N
N
N
information to order a prescription versus fewer than 4 pages for all other sites (P < .01). Little difference was seen in the availability of health care professionals among the sites (see Table 8). More than twothirds provided a toll-free number for contacting a health care professional. None provided telephone access to a nurse. Notably, only the online pharmacies provided telephone access to a physician. Finally, there was no apparent difference in payment method among the sites (see Table 9). More than 75% of the sites accepted insurance or credit cards, and fewer than 50% of the sites accepted checks or money orders.
N
N
"0 Z
,
Z
Z
>-
'"
Z
Z
o Z
Z
o
o Z
o
o
Q)
Z
"0
II)
en;:
~ 'f _
o
o
o
o
Z
II)
o
o
Z0
o
o
Z
Z
Z
>u
Z
0
Z
(J) II) (J):i
~
g
0
o
o Z
Z
o
Z
Z
(J)
o Z
Z
o
o
o
Z
o
Z
Z
Z
:JU
... II)
C
-
II)
C 0 Cl C . ro ro 0 e..
CD Z II) ~-;;E
'"
II)
>-
>-Q)
0 :J
(J) II)
o
Z
>-
o Z
Z
o
Z
o
o
Z
~
E
ro ~
o
8E
0
tJ - '
tJ
~ ~ ~....
X o
~ ~
E '0.
II)
._>..
.E. . 8
'f
.!!i
~
-E,
(J) II)
o
o Z
Z
E ~
888 ~ ~ ~ E ~ E ....
E o E
>-
Z
Z
Z
(J) II)
o
o
Z
o
o
o
>-
E
E
ro
II)
....
ro
J:. 0.
J:. '00.
~
~
J:. 0.
~
'(jj
Cl
J:.
0
~
E ~
~
E
ro
~i 1 iii f 111 x o >al tJ ro
II)
E ro
Z
i:L II)
E ro
J:. J:.
~
(J)
Q)
->- >- -t: >>t tJ ro tJ tJ ro II) ro e.. ro
OlE
= ro Q)
...
OJ:.
C "0 0 ~ .Q)
II)
C
C ~o 0.
E
:J
ro
Ol ....
OJ:.
(J)
C
E ...
'(jj II)
.- ro
o.~II)
Q)
"Ox II)
E
...
...
C
C "0 0 II)
~
'(jj
II)
II)
o.c
"Ox
E
II)
C ~ C
o.~II)
ro
E ....
...Je..
...
...C
C
C
Q)
~.2 ~ .~ -g
II)
>tJ
-" ro roJ:.
~J:.
e.. ue.. Oe.. we..
...c
Q)
J:.
II)
"0
II)
C
0
Q) . -
o.c
II)
Q)
o.~ Q) Q)
~ ~ ~ ~ "x E
21 0 Journal of the American Phannaceutical Association
II)
~ ro
(J) ~ :... ro II) E ClJ:. Qi J:. ... ~ (5 (J) ro
tJ
II)
::: 0.
>
OJ:.
>e..
Discussion This study revealed differences among pharmacy providers on the Web. The most interesting, and disturbing, finding was the variation in privacy and confidentiality policies. Chain pharmacy extensions appeared to be most sensitive to this issue, and the chain sites in this study had taken steps to alert ~ users to their privacy policies during the ordering process. Vi VIPPS-certified sites were more likely than noncertified sites Q) 'E to have a privacy policy in place and displayed throughout the ~ ordering process. This is not surprising given that a privacy 0.. ~ policy is a criterion for VIPPS certification. Noteworthy, E though, was the use of cookies by VlPPS-certified sites. As ~ discussed earlier, cookies may belie the idea of confidentiality 0.. ~ and privacy accorded VIPPS-certified sites. ~ The more than 10 pages of information chain pharmacies C required users to complete during the ordering process com"0 ~ pared unfavorably with other types of sites. This finding has ~ implications for patients' perceptions of quick access; con~ sumers who desire rapid access to medications may find chain g: sites more onerous to navigate and tend to use sites with easier 5> access but fewer privacy protections. ...o Another difference found was in product availability, with ~ .Q the independent pharmacy extensions being more likely to disOJ \3 pense a specialty product, such as progesterone suppositories. g Future studies should investigate specialty sites as a separate ~ E category (e.g., fertility, HIV) to determine how their products .2 and services differ from those of other Web sites. Also, mail § order and online pharmacies did not tend to sell OTC agents to ~ the same extent that the chains and independents did. This § finding was not unexpected given that the former focus on Qj selling prescription medications for chronic conditions. >.!!1 Most of the drug information provided was written at an ~ t eighth-grade reading level. Surprisingly few sites provided an g electronic newsletter, in contrast to the 16 of 21 sites found to ~ do so by the California HealthCare Foundation. 2,9 Electronic ~ newsletters afford opportunities to increase pharmacy business ~ by providing timely information about drug products to con~ sumers. This difference may be due to the sampling frame. ()
MardlIApril2001
Vol. 41, No.2
Internet Pharmacies
RESEARCH
Table 7. Ease of Use of Internet Pharmacy Sites (n - 33) Chain Extensions (n = 5) Average number of steps required to order a new prescription
Mail Order (n - 6)
Online (n 14)
Independent Extensions (n 8)
PValue
10.2
3.5
4 .0
2.5
< 01 8
New prescription link on home page (%)
60
75
93
87
.28 b
Refill prescription link on home page (%)
100
100
64
87
.14b
20
17
36
0
.26 b
Consultation link (%)
aKruskal-Wallis. Post-hoc analysis revealed on ly difference existed between chain extension and others. No significant differences were f ound among mall order. online. and independent pharmacies . b X? test.
Table 8. Availability of a Health Care Professional on Internet Pharmacy Sites (n = 33) Chain Extensions (n =5)(%) Nurse Pharmacist
Mail Order (n = 6)(%)
Online (n = 14) (%)
Independent Extensions (n =8) (%)
PValue 8
0
0
0
0
NA
100
100
79
94
.45
0
.25
0
21
Professional available via telephone
100b
100
93
87
.72
Available via toll-free number
100
100
64
75
. 19
Available via e-mail
100
100
100
88
.27
0
Physician
"x2
test . bOf those sites with a health care professional available.
The California HealthCare Foundation viewed not only online pharmacies but also general health-related information and service sites, which may be more likely to provide informational newsletters to their customers. Access to health care professionals differed slightly among the groupings. Chain pharmacy extensions were most likely to have around-the-clock access to a pharmacist via telephone, but e-mail access was similar across all groupings. Only the online pharmacies had a physician available by phone, most likely to increase business through online prescribing. The Institute for the Future2 predicted that tomorrow's health care consumer will demand access to good information with high customer ser"ice. These consumers will also "expect to be able to interact with their health care providers and payers by Internet and e-mail" as well as in person or by telephone. Web pharmacies that do not provide access to a pharmacist around the clock, or at least bye-mail, may lose customers. Most of the sites accepted insurance and credit cards. However, many sites did not accept checks, money orders, or Internet accounts as means of payment. This, too, is disturbing in that only 33% to 60% of the sites in any category accepting insurance or credit card transactions used SSL or encryption technology to maintain the integrity of the transmission. Of note, the VIPPScertified sites were more likely to use SSL technology to secure the transmission than were the other sites. In this vein, the consumer must be wary of divulging sensitive information, such as
Vol. 41, No.2
MarchlApril2001
that contained in insurance or credit card transactions, and should check sites for SSL technology and privacy policies before using these methods of payment.
Limitations One limitation of the study was the sampling frame. The key words "Internet pharmacy" and "Internet pharmacist" may have limited the sample. In addition, the freeware metasearch engine, WebFerret, would have missed some of the sites not registered in its database of nine search engines. In contrast, the nonfreeware cousin, WebFerretPro, has more than 33 search engines in its database and would likely have yielded a larger sample of sites to review. Also, several sites eliminated as foreign based on their names (e.g., mexicanpharmacy.com) may have had U.S.-based mirror sites. Another limitation was the categorization of the sites into chain pharmacy extension, mail order, online, and independent pharmacy extensions. These categories appear to be wholly separate; however, the definition of a chain extension may vary in that some independent sites with five or more actual stores could be considered a local chain. In this sample, chain pharmacy extensions were defined as having several stores engaging in multistate commerce. No independent site in this sample had multi state stores.
Journal of the American I'bannac:eUtIaI As!Iodation
211
RESEARCH
Internet Pharmacies
Table 9. PaylTtent Methods Accepted by Surveyed Internet Pharmacy Sites (n = 33) Chain Extensions (n=5) (%)
Mail Order (n=6) (%)
Online (n = 14) (%)
Independent Extensions (n=8)(%)
PValue
Insurance
100
66
71
87
.44
Credit card
100
66
100
100
.50
Money order
20
50
43
50
.72
Check
20
50
36
50
.68
0
0
7
0
.71
Net account
Similarly, the sample sizes for the chain extension and mail order categories were quite small, leaving many of the cells with expected values of less than five. Actual values greater than one are adequate for X2 analysis, but only if no fewer than 20% of the cells have low expected values. In this study, most of the X2 analyses did not meet this criterion, possibly overestimating the statistical significance. Where appropriate, Fisher's exact test was used. To improve the validity of the statistics, the categories could have been combined to increase the frequency, but no natural combination could have been used without a loss of information. In this small study, basic statistical analyses were performed only to gain a sense of trends. Future analyses should ensure adequate sample size in each category. No attempt was made to determine whether the sites adhered to their policy. This aspect was extensively reviewed by the California HealthCare Foundation, which found inconsistencies between the privacy policies posted and the actual practices of health-related Web sites.
Conclusion This study revealed inconsistencies among Internet pharmacy providers with respect to privacy of information, security of electronic transmissions, and availability of drug information for consumer use. VIPPS certification aids in minimizing these problems; however, there is room for improvement even for these sites. These issues must be addressed through the development and enforcement of appropriate technologic solutions and of laws and regulations that protect consumers.
212
Joumal or the American PhannaceuticaJ Association
The author declares no conflicts of interest or financial interests in any product or service mentioned in this article, including grants, employment, gifts, stock holdings, or honoraria. No part of this paper has been presented in any form at a medical or pharmacy meeting. The author would like to acknowledge Darshan Kulkarni, Barbara Fisher, and Mary St. Amour for their help in data collection and site identification, and Abbey Cohen for her review of the original manuscript.
References 1. Health on the Net Foundation (HON). 5th HON survey on the evolution of Internet use for health purposes. Available at www.hon.ch/Survey/ ResultsSummary_oct_nov99.html. Accessed April 25, 2000. 2. Mittman R, Cain M. The Future of the Internet in Health Care: Five Year Forecast. Oakland, Calif: California HealthCare Foundation; January 1999. 3. Silberg WM, Lundberg GO, Musacchio RA. Assessing, controlling, and assuring the quality of medical information on the Internet: caveat lector et viewor-Iet the reader and viewer beware . JAMA. 1997;277:1244-5. 4. Ukens C.lnternet pharmacies. Drug Topics; May 17, 1999:63-70. 5. Latner AW. Pharmacists on the Internet: how to attract customers. Pharm Times. 1998:64:78. 6. Landis NT. Virtual pharmacies boast easy access, privacy safeguards. Am J Health Sys Pharm. 1999;56:1174,1177-9. 7. Marwick C. Several groups attempting regulation of Internet Rx. JAMA. 1999;281:975-£. 8. Paulsen M. New NABP program combines criteria, inspections to certify online pharmacy quality. JAm Pharm Assoc. 1999:39:870. 9. California HealthCare Foundation. Privacy: Report on the Privacy Policies and Practices of Health Web Sites. Oakland, Calif: California HealthCare Foundation; January 2000.
MarchlApril2001
Vol. 41, No.2
A Survey of Selected Internet Pharmacies in the United States Andrew M. Peterson
Objective: To determine whether differences in the provision of pharmacy services exist among different types of Internet pharmacies. Design: Survey of selected pharmacies with a presence on the Internet. Data were abstracted onto a data collection form for further analysis. Data collection was limited to 3 weeks. Setting: U.S.-based Internet pharmacies that allow patients to purchase prescription medications online. Pharmacies were identified using a metasearch engine with the search terms "Internet pharmacy" and "Internet pharmacist." Intervention: Survey. Main Outcome Measures: Comparisons of availability of 10 commonly used products representing a variety of product categories, prescription verification methods, and privacy issues; and determinations of site navigability, drug information and provider access, and payment methods. Sites were categorized as "chain pharmacy extensions," "mail order pharmacies," "independent pharmacy extensions," and "online pharmacies." Results: Thirty-three sites were reviewed. There was significant variation among the four types of pharmacies selling prescriptions over the Internet. Most pharmacies provided all of the drugs in the survey. Patients were required to provide their own prescription at 88% of the sites, and 75% of sites used mail or fax to verify prescription integrity. More than 50% of sites had privacy policies posted, and 64% used cookies. Chain pharmacy extensions required completion of an average of 10.2 pages to order drugs versus 2.4 to 4 pages for all other site types. Drug information was written at an eighth-grade reading level at 36% of the sites. More than two-thirds of the sites provided a toll-free telephone for a health care professional. Nearly 80% of the sites accepted health insurance, and 95% accepted credit cards; however, only 40% used a secure transmission mechanism for patient or payment information. Conclusion: Internet pharmacies provide varying levels of service. Policies regarding the use of the Internet for obtaining medications should focus on improving the privacy of consumer information and ensuring the secure transmission of financial information. JAm Pharm Assoc. 2001;41:205-12.
The provision of health-related services via the Internet has grown tremendously over the past several years.' Health sites on the World Wide Web provide a broad range of products, as well as clinical and diagnostic services, and connect patients directly with health care providers. Providing rapid, 24-hour access to information is a primary function of the Web. In fact, this convenience has often been considered the driving force behind the growth of the Web in the health care sector. 2 However, the occasional display of inconsistent or inaccurate infonnation may be detrimental to patients. 3 Pharmacies are a growing presence on the Web. Many chain pharmacies have developed an extension of their brick-and-rnorReceived August 4, 2000, and in revised form December 1, 2000. Accepted for publication December 12, 2000. Andrew M. Peterson, PharmD, BCPS, is assistant professor of clinical pharmacy, Philadelphia College of Pharmacy, University of the Sciences in Philadelphia. Correspondence: Andrew M. Peterson, PharmD, BCPS, Philadelphia College of Pharmacy, University of the Sciences, 600 S. 43rd Street, Philadelphia, PA 19104. Fax: 215-59EH3586. E-mail: [email protected].
Vol. 41, No.2
MarcblApril2001
tar stores in cyberspace, sometimes through alliances with existing online pharmacies.4 lndependent pharmacies, especially those that serve a particular population of patients (e.g., infertility patients, HIV-positive patients),5 are also establishing a Web presence. Many do not accept new prescriptions online, but instead cater to their existing clientele by providing medication refills.6 Lifestyle drugs, such as Viagra, appear to represent an exception to this trend, and mail order pharmacies are using the Web as a new avenue for expanding their market. Not surprisingly, many interested organizations have expressed increasing concern about patients' access to pharmaceuticals via the Web. Their concerns include verification of prescription integrity and assurance of appropriate patient-provider communication. The National Association of Boards of Pharmacy (NABP), for example, has taken steps to assure the public of the safety and integrity of Web pharmacy sites. 7 The presence of the Verified Internet Pharmacy Practice Site (VlPPS) seal on a Web site assures consumers that the online pharmacy is certified to dispense medications anywhere in the United States and has met a rigorous 17-point set of criteria that addresses confidentiality,
Joumal of !be American I"hannaceudcaI A&!iodatIoo
205
RESEARCH
Internet Pharmacies
quality assurance, and licensure. 8 While privacy is a global concern for the Internet, it is a particularly compelling issue for sites that provide medical information and access to drugs. Patients often erroneously believe that when they access a Web site, their privacy is protected. To ensure the security of a transaction (i.e., to prevent the information from being read by someone other than the intended recipient), Web sites-in particular, sites that transmit financial informationoften use secure socket layer (SSL) technology. Sites using SSL technology may be identified by a dialog box or a lock icon at the bottom of the screen. Often, the addition of an "s" at the end of the "http//:" ("https//:") prefix on a Web site address, or uniform resource locator (URL), indicates a secure site. Some sites use cookies to track visits. A cookie is a unique number that is automatically assigned to the hard drive of the visitor's computer when he or she first enters a Web site. The cookie is then used to track when the user returns and which pages he or she visits. Cookies are generally transparent; because many Web browsers enable cookies by default, the visitor is typically unaware that they are being used. Although users may believe cookies confer anonymity, they in fact become uniquely identifiable customer IDs when a visitor registers at the site. 9 The California HealthCare Foundation9 reviewed the policies and practices of 21 health-related Web sites, including several well-known online pharmacies, such as drugstore.com and Planetrx.com. Their results showed that visitors to health-related Web sites are not anonymous: Many of the sites use cookies, profiling, and banner ads to collect information about individuals without their knowledge or consent. Furthermore, they found that many sites lacked privacy policies or failed to follow their stated policies. Patients' use of the Web as a means of buying medications and accessing associated information is controversial. Among the issues under discussion are product availability, privacy, verification of prescription authenticity, validity and reliability of drug information, and secure methods of payment. The present study was the first to examine whether different types of pharmacy practice sites provide different levels of service.
Objectives The purpose of this study was to determine whether differences in the provision of pharmacy services exist among different types of Internet pharmacies. To this end, this study compared selected services and product availability among different types of pharmacies purporting to allow online prescription ordering.
Methods Four types of Web pharmacies were defined: (1) chain pharmacy extensions (e.g., Walgreens.com); (2) independent pharmacy
206
Joumal of the American Phannaceutical Association
extensions (e.g., villageapothecary.com); (3) mail order pharmacies extending their service through the Web (e.g., Merck-Medco.com); (4) online pharmacies-those pharmacies that sell medications exclusively online (e.g., Planetrx.com). WebFerret (ZDnet), a metasearch engine, was used to identify pharmacies with a Web presence. Examples of search engines used in this version of WebFerret include goto.com and altavista. com. Search terms used were "Internet pharmacy" and "Internet pharmacist." Microsoft Internet Explorer 4.0 was used as the Web browser. The automatic cookie acceptance function was disabled and the secure transmission notification function was enabled. All pharmacies obviously based outside the United States, and duplicates from different search engines, were removed from the sample. Only those sites that allowed electronic ordering, via email or online application, of prescription medications were included in the final analysis. Sites allowing only refills of existing prescriptions or providing only a telephone number for the patient to call to request a prescription were excluded. A data collection instrument was developed containing questions and/or statements pertaining to various aspects of Web pharmacies. For each item, investigators could indicate "yes," "no," or "not applicable." There was a space for comments beside each item. Data were collected during a 3-week period from March 22 through April 12, 2000. Two pharmacy students reviewed each site and collected the data, and the primary investigator reviewed all reports and double-checked all information recorded against the original Web site. One aspect studied was the availability of prescription and nonprescription drugs. Particular attention was paid to common controlled substances (CII and em), unstable or dated medications (amoxicillin suspension and progesterone suppositories), and generic and brand name drugs covering a variety of diseases and/or patient populations (e.g., Pre marin tablets, Viagra, hydrochlorothiazide). The availability of written drug and health information from each Web site was investigated. Sildenafil (Viagra-Pfizer) was chosen as an example for determining the reading level of the drug information offered. This product was chosen because it was felt that a large number of sites would be offering it, given its extensive coverage by the media immediately before the study. The study also probed the availability of a health care practitioner via telephone and e-mail and the methods of payment available to the consumer. Other topics covered in the survey were prescription verification, security/privacy, and sites' ease of use. Additional information was obtained on security of information transmission through the use of SSL technology, cookies, banner ads, and the posting of privacy policies throughout the ordering process. Descriptive statistics were developed for the survey questions. The Flesch-Kincaid readability scale (available through Microsoft Word 97) was used to detennine reading level of drug information.
MarchlApril2001
Vol. 41, No.2
Internet Pharmacies
RESEARCH
Table 1. Availability of Medications Through Surveyed Internet Pharmacies (n = 33) Chain Extensions (n = 5) (%)
Medication Alprazolam tablets
Mail Order (n =6) (%)
Online (n = 14) (%)
Independent Extensions (n = 8) (%)
PValue"
100
50
57
75
.25
60
50
64
87
.48
Celecoxib tablets
100
67
79
87
.51
Hydrochlorothiazide tablets
100
67
64
87
.19
Insulin injection
Amoxicillin suspension
100
67
64
87
.47
Isotretinoin cream
80
50
57
87
.19
Percocet tablets
80
50
36
62
.33
Premarin tablets
100
67
64
100
.12
20
67
57
100
.03
Sildenafil tablets
100
100
86
87
.72
Over-the-counter medications
100
60
64
100
.04
Progesterone suppositories
'X 2 test.
Comparisons were made using either the X2 or Fisher's exact test for dichotomous data and the Kruskall-Wallis one-way analysis of variance for nonparametric continuous data. All ex's were set at 0.05. Systat version 7.0 software was used to analyze data.
and 8 (24%) were independent pharmacy extensions. Table 1 shows product availability on the Web sites in the sample. With the exception of progesterone suppositories and overthe-counter (OTC) medications, there was generally little, if any, difference in the availability of products. Independent pharmacy extensions provided consumer access to progesterone suppositories more often than did all others (P = .03). Chain pharmacy extensions and independent phannacy extensions provided access to OTC medications more often than did mail order or online pharmacies (P = .04). Table 2 shows the availability of drug information. Chain pharmacy extensions and online pharmacies provided written drug and health information more often than did the others, although these differences were not significant. Only 8 sites (24%) provided an electronic newsletter for patients, and only 12 sites (36%) provided a drug monograph for sildenafil (again, used to determine reading level of the drug information provided). None of the independent pharmacy extensions had a sildenafil monograph available. The reading level of the information provided ranged from seventh- to ninth-grade based on the Flesch-Kincaid scale.
Results Initially, more than 500 pharmacy sites were identified using WebFerret. Sites that did not allow electronic ordering of medications were excluded, as were sites that only provided links to other sites (e.g., RiteAid.com links to drugstore.com). Many sites were excluded because they were not pharmacies but organizational and/or informational sites on the topic of Internet pharmacies. Seven sites (1 %) were excluded because they allowed only refills of existing prescriptions. Of the more than 500 sites identified, 75 were considered to be U .S.-based pharmacy sites; of these, only 33 (44%) were included in the final analysis. Of the 33 sites selected, 5 (15%) were chain pharmacy extensions, 6 (18%) were mail order pharmacies, 14 (42%) were online pharmacies,
Table 2. Availability of Drug Information on Internet Pharmacy Sites (n = 33) Chain Extensions (n = 5)
Mail Order (n = 6)
Online (n = 14)
PValue
62%
.32"
36%
12.5%
.11"
50%
37%
.19" .47"
.44b
Written drug information
100%
50%
71%
Drug interaction information
80%
33%
80%
17%
Internal health/disease information
Independent Extensions (n = 8)
Electronic newsletter
0%
17%
36%
25%
Grade average of Flesch-Kincaid reading level (sildenafil)
7.1
8
8.8
NA
"x 2 test . bKruskall-Wallis.
Vol. 41, No.2
MarchlApril2001
Joomal of the American PbannaceUticaI ~Iion
207
RESEARCH
Internet Pharmacies
Table 3. Internet Pharmacies' Prescription Verification Methods (n Chain Extensions (n = 5)
Mail Order (n = 6)
= 33)
Online (n = 14)
Independent Extensions (n = 8)
PValue a
Patient supplies own prescription Patient can mail prescription Patient can fax prescription
5 5 5
5 3 5
11 11 10
8 6 8
.38 .1 0 .38
Company contacts prescriber Via phone Via e·mail
3 3 0
4 4 0
8 8
6 6 0
.91 .64
·X 2 test.
Table 4. Internet Pharmacies' Display of Pharmacy License (n = 33) Chain Extensions (n =5)(%)
Mail Order (n=6) (%)
Online (n = 14) (%)
Independent Extensions (n=8) (%)
PValue a
Home page states licensure
20
17
29
0
Informs consumer anywhere
40
17
57
13
.36
Licensed in all 50 states b
20
17
14
0
.42
.42
x
a 2 test. bUnless stated, we assumed extensions were not licensed in all 50 states.
There was no statistical difference among the types of pharmacies in the provision of drug interaction information (P = .11). The majority (n = 29; 88%) of the sites reviewed required patients to provide their own prescription (see Table 3). However, several sites, particularly online pharmacies, either allowed the patient to obtain the prescription from the site's online physician or made provisions for contacting the patient's physician directly. There appeared to be no difference in the mechanism of prescription transmission among pharmacy types. Table 4 shows the number of sites reporting licensure status to visitors. Overall, most sites did not report which states had licensed them to dispense; those that did include that information did not place it conspicuously on the home page. VIPPS-certified sites are by definition licensed to dispense in all states. Privacy and confidentiality policies varied significantly among sites (see Tables 5 and 6). Independent pharmacy extensions
required patients to register far less often than did all others (P < .01). Chain pharmacy extensions posted privacy policies significantly more often than did all others (P < .01). There was no significant difference among the pharmacy types in the number of VIPPS-certified sites. However, VIPPS-certified sites were more likely to have a privacy policy (P = .04, Fisher's exact test). Nor was there any difference in the use of banner ads. Independent pharmacy extensions were less likely to use cookies than were all others (P < .01), and all VIPPS-certified sites used cookies. Fewer than 60% of all sites used SSL technology during the ordering process, and there was no statistical difference among sites in that regard. VIPPS-certified sites were more likely to use SSL technology than were all others (P = .02, Fisher's exact test). Table 7 shows characteristics of the sites' ease of use. The only significant difference was seen in chain pharmacy extensions, which required visitors to complete an average of 10.2 pages of
Table 5. Internet Pharmacies' Use of Tools That Affect Privacy and Confidentiality (n = 33) Chain Extensions (n= 5) (%)
Mail Order (n = 6) (%)
Online (n = 14) (%)
Independent Extensions (n =8) (%)
Patient to register online
100
83
100
25
<.01
Privacy policy present
100
16
64
25
< .01
Privacy policy on every page during ordering
100
100
44
50
<.01
40
16
14
0 13
VIPPS-certified
PValue a
.28 <.01
100
66
79
Site uses banner ads
60
33
64
62
.62
Site uses SSL technology
60
33
50
50
.84
Site uses cookies
SSL = secure socket layer; VIPPS = Verified Internet Pharmacy Practice Site. "x2 test.
208 JoomaI or the American Pharmaceutical Association
MardIIApril2001
Vol. 41, No. 2
~
""
i,
I
I
I
I·
~
I
!
:!:
..,~
~
:::
...~
!"'
www.medprescribe.com www.nofrillspharmacy.com
Preferred Rx
ePrescribe.com
Familymeds
Global Rx.com
Catalog.com
MedPrescribe
No Frills Pharmacy
Mail order
Online
Online
Online
Online
Online
Online
5 No No
Yes
No
Yes
Yes
www.ppsrxbymail .com
Postal Prescription Services
Mail order
No
No
No
No
Yes
www.pharmacystore .com
The Pharmacy Store
Mail order
3
3 No
www.merck-medco.com
Merck Medco
Yes
No
Yes
Mail order
3 No Yes (Yes)
No Yes
Yes
Yes (3)
No Yes
Yes
www.at-cost-drugs.com
At-Cost-Drugs
2
No
No
No
No
www.pro-mark- pharmacies.com www.rpspharmacy.com
Pro Mark Pharmacies
AARP Pharmacy
Online
Online www.rxuniverse .com safenetpharmacy.com www .safewebmedical.com starkscripts.com
RX Universe
SafeNet Pharmacy
SafeWeb Medical
StarkScripts. com
Online
Online
Online
Online
Service
www.planetrx.com
Planet Rx .com
Online
catalog.com/highcom
www.globalrx.com
familymeds.com
www.eprescribe.com
www.preferredrx.com
www.abeewell.com
No
Mail order
12
No
ABeeWell Pharmacy
2 2
4 4
7 4
8
Yes (Yes) Yes (No) No Yes (No) No Yes (Yes)
Yes No No No No Yes
Yes Yes No Yes Yes Yes
No No Yes (1) Yes (4) Yes (1) No
Yes No No No Yes
Yes Yes Yes Yes Yes Yes
2 3
3 3
No (No) No Yes (Yes) Yes (No)
No No No No
Yes No Yes Yes
Yes (3) No Yes (1) Yes (0)
No Yes No
Yes Yes Yes
Yes
Yes
Yes
9 Yes (Yes) No Yes No
No
Yes
Yes
2 No No Yes Yes (2)
Yes
Yes
Yes (5)
5 No Yes (No)
No No
Yes Yes
Yes (3)
No Yes
Yes
Yes
Yes (Yes)
Mail order
8
Yes (Yes)
Yes
Yes No
Yes (2)
12
Yes
Yes (Yes)
7
No
Yes (Yes) 12
Yes
No Yes (Yes)
Yes
Yes
VIPPSCertified
No. of Pages to Complete to Order Medication
Yes
Yes
No
Yes (1)
Yes (2)
Uses Cookies
Privacy Policy Present (on Every Page)
No
www.walgreens.com
Walgreens
Chain extension
Yes
No
Yes
Yes
Uses SSL
Uses Banner Ads (No . on Home Page)
No
www.cvs.com
CVS/ Pharmacy
Chain extension
Yes
www.e-pharmacy.com
Eckerd.com
Chain extension
Yes
Yes
www.drugstore.com
www.drugemporium.com
URL
Drugstore.com
Drug Emporium
Name
Chain extension
Chain extension
Type
Patient Required to Register
Table 6. Classification of Web Sites and Related Privacy and Confidentiality Issues
::c
~
~
U)
Cfj
(b.
Q) ()
3
Q)
.,
~
"ll
~
::J
5" Cti ..,
RESEARCH
Internet Pharmacies
N
N
N
information to order a prescription versus fewer than 4 pages for all other sites (P < .01). Little difference was seen in the availability of health care professionals among the sites (see Table 8). More than twothirds provided a toll-free number for contacting a health care professional. None provided telephone access to a nurse. Notably, only the online pharmacies provided telephone access to a physician. Finally, there was no apparent difference in payment method among the sites (see Table 9). More than 75% of the sites accepted insurance or credit cards, and fewer than 50% of the sites accepted checks or money orders.
N
N
"0 Z
,
Z
Z
>-
'"
Z
Z
o Z
Z
o
o Z
o
o
Q)
Z
"0
II)
en;:
~ 'f _
o
o
o
o
Z
II)
o
o
Z0
o
o
Z
Z
Z
>u
Z
0
Z
(J) II) (J):i
~
g
0
o
o Z
Z
o
Z
Z
(J)
o Z
Z
o
o
o
Z
o
Z
Z
Z
:JU
... II)
C
-
II)
C 0 Cl C . ro ro 0 e..
CD Z II) ~-;;E
'"
II)
>-
>-Q)
0 :J
(J) II)
o
Z
>-
o Z
Z
o
Z
o
o
Z
~
E
ro ~
o
8E
0
tJ - '
tJ
~ ~ ~....
X o
~ ~
E '0.
II)
._>..
.E. . 8
'f
.!!i
~
-E,
(J) II)
o
o Z
Z
E ~
888 ~ ~ ~ E ~ E ....
E o E
>-
Z
Z
Z
(J) II)
o
o
Z
o
o
o
>-
E
E
ro
II)
....
ro
J:. 0.
J:. '00.
~
~
J:. 0.
~
'(jj
Cl
J:.
0
~
E ~
~
E
ro
~i 1 iii f 111 x o >al tJ ro
II)
E ro
Z
i:L II)
E ro
J:. J:.
~
(J)
Q)
->- >- -t: >>t tJ ro tJ tJ ro II) ro e.. ro
OlE
= ro Q)
...
OJ:.
C "0 0 ~ .Q)
II)
C
C ~o 0.
E
:J
ro
Ol ....
OJ:.
(J)
C
E ...
'(jj II)
.- ro
o.~II)
Q)
"Ox II)
E
...
...
C
C "0 0 II)
~
'(jj
II)
II)
o.c
"Ox
E
II)
C ~ C
o.~II)
ro
E ....
...Je..
...
...C
C
C
Q)
~.2 ~ .~ -g
II)
>tJ
-" ro roJ:.
~J:.
e.. ue.. Oe.. we..
...c
Q)
J:.
II)
"0
II)
C
0
Q) . -
o.c
II)
Q)
o.~ Q) Q)
~ ~ ~ ~ "x E
21 0 Journal of the American Phannaceutical Association
II)
~ ro
(J) ~ :... ro II) E ClJ:. Qi J:. ... ~ (5 (J) ro
tJ
II)
::: 0.
>
OJ:.
>e..
Discussion This study revealed differences among pharmacy providers on the Web. The most interesting, and disturbing, finding was the variation in privacy and confidentiality policies. Chain pharmacy extensions appeared to be most sensitive to this issue, and the chain sites in this study had taken steps to alert ~ users to their privacy policies during the ordering process. Vi VIPPS-certified sites were more likely than noncertified sites Q) 'E to have a privacy policy in place and displayed throughout the ~ ordering process. This is not surprising given that a privacy 0.. ~ policy is a criterion for VIPPS certification. Noteworthy, E though, was the use of cookies by VlPPS-certified sites. As ~ discussed earlier, cookies may belie the idea of confidentiality 0.. ~ and privacy accorded VIPPS-certified sites. ~ The more than 10 pages of information chain pharmacies C required users to complete during the ordering process com"0 ~ pared unfavorably with other types of sites. This finding has ~ implications for patients' perceptions of quick access; con~ sumers who desire rapid access to medications may find chain g: sites more onerous to navigate and tend to use sites with easier 5> access but fewer privacy protections. ...o Another difference found was in product availability, with ~ .Q the independent pharmacy extensions being more likely to disOJ \3 pense a specialty product, such as progesterone suppositories. g Future studies should investigate specialty sites as a separate ~ E category (e.g., fertility, HIV) to determine how their products .2 and services differ from those of other Web sites. Also, mail § order and online pharmacies did not tend to sell OTC agents to ~ the same extent that the chains and independents did. This § finding was not unexpected given that the former focus on Qj selling prescription medications for chronic conditions. >.!!1 Most of the drug information provided was written at an ~ t eighth-grade reading level. Surprisingly few sites provided an g electronic newsletter, in contrast to the 16 of 21 sites found to ~ do so by the California HealthCare Foundation. 2,9 Electronic ~ newsletters afford opportunities to increase pharmacy business ~ by providing timely information about drug products to con~ sumers. This difference may be due to the sampling frame. ()
MardlIApril2001
Vol. 41, No.2
Internet Pharmacies
RESEARCH
Table 7. Ease of Use of Internet Pharmacy Sites (n - 33) Chain Extensions (n = 5) Average number of steps required to order a new prescription
Mail Order (n - 6)
Online (n 14)
Independent Extensions (n 8)
PValue
10.2
3.5
4 .0
2.5
< 01 8
New prescription link on home page (%)
60
75
93
87
.28 b
Refill prescription link on home page (%)
100
100
64
87
.14b
20
17
36
0
.26 b
Consultation link (%)
aKruskal-Wallis. Post-hoc analysis revealed on ly difference existed between chain extension and others. No significant differences were f ound among mall order. online. and independent pharmacies . b X? test.
Table 8. Availability of a Health Care Professional on Internet Pharmacy Sites (n = 33) Chain Extensions (n =5)(%) Nurse Pharmacist
Mail Order (n = 6)(%)
Online (n = 14) (%)
Independent Extensions (n =8) (%)
PValue 8
0
0
0
0
NA
100
100
79
94
.45
0
.25
0
21
Professional available via telephone
100b
100
93
87
.72
Available via toll-free number
100
100
64
75
. 19
Available via e-mail
100
100
100
88
.27
0
Physician
"x2
test . bOf those sites with a health care professional available.
The California HealthCare Foundation viewed not only online pharmacies but also general health-related information and service sites, which may be more likely to provide informational newsletters to their customers. Access to health care professionals differed slightly among the groupings. Chain pharmacy extensions were most likely to have around-the-clock access to a pharmacist via telephone, but e-mail access was similar across all groupings. Only the online pharmacies had a physician available by phone, most likely to increase business through online prescribing. The Institute for the Future2 predicted that tomorrow's health care consumer will demand access to good information with high customer ser"ice. These consumers will also "expect to be able to interact with their health care providers and payers by Internet and e-mail" as well as in person or by telephone. Web pharmacies that do not provide access to a pharmacist around the clock, or at least bye-mail, may lose customers. Most of the sites accepted insurance and credit cards. However, many sites did not accept checks, money orders, or Internet accounts as means of payment. This, too, is disturbing in that only 33% to 60% of the sites in any category accepting insurance or credit card transactions used SSL or encryption technology to maintain the integrity of the transmission. Of note, the VIPPScertified sites were more likely to use SSL technology to secure the transmission than were the other sites. In this vein, the consumer must be wary of divulging sensitive information, such as
Vol. 41, No.2
MarchlApril2001
that contained in insurance or credit card transactions, and should check sites for SSL technology and privacy policies before using these methods of payment.
Limitations One limitation of the study was the sampling frame. The key words "Internet pharmacy" and "Internet pharmacist" may have limited the sample. In addition, the freeware metasearch engine, WebFerret, would have missed some of the sites not registered in its database of nine search engines. In contrast, the nonfreeware cousin, WebFerretPro, has more than 33 search engines in its database and would likely have yielded a larger sample of sites to review. Also, several sites eliminated as foreign based on their names (e.g., mexicanpharmacy.com) may have had U.S.-based mirror sites. Another limitation was the categorization of the sites into chain pharmacy extension, mail order, online, and independent pharmacy extensions. These categories appear to be wholly separate; however, the definition of a chain extension may vary in that some independent sites with five or more actual stores could be considered a local chain. In this sample, chain pharmacy extensions were defined as having several stores engaging in multistate commerce. No independent site in this sample had multi state stores.
Journal of the American I'bannac:eUtIaI As!Iodation
211
RESEARCH
Internet Pharmacies
Table 9. PaylTtent Methods Accepted by Surveyed Internet Pharmacy Sites (n = 33) Chain Extensions (n=5) (%)
Mail Order (n=6) (%)
Online (n = 14) (%)
Independent Extensions (n=8)(%)
PValue
Insurance
100
66
71
87
.44
Credit card
100
66
100
100
.50
Money order
20
50
43
50
.72
Check
20
50
36
50
.68
0
0
7
0
.71
Net account
Similarly, the sample sizes for the chain extension and mail order categories were quite small, leaving many of the cells with expected values of less than five. Actual values greater than one are adequate for X2 analysis, but only if no fewer than 20% of the cells have low expected values. In this study, most of the X2 analyses did not meet this criterion, possibly overestimating the statistical significance. Where appropriate, Fisher's exact test was used. To improve the validity of the statistics, the categories could have been combined to increase the frequency, but no natural combination could have been used without a loss of information. In this small study, basic statistical analyses were performed only to gain a sense of trends. Future analyses should ensure adequate sample size in each category. No attempt was made to determine whether the sites adhered to their policy. This aspect was extensively reviewed by the California HealthCare Foundation, which found inconsistencies between the privacy policies posted and the actual practices of health-related Web sites.
Conclusion This study revealed inconsistencies among Internet pharmacy providers with respect to privacy of information, security of electronic transmissions, and availability of drug information for consumer use. VIPPS certification aids in minimizing these problems; however, there is room for improvement even for these sites. These issues must be addressed through the development and enforcement of appropriate technologic solutions and of laws and regulations that protect consumers.
212
Joumal or the American PhannaceuticaJ Association
The author declares no conflicts of interest or financial interests in any product or service mentioned in this article, including grants, employment, gifts, stock holdings, or honoraria. No part of this paper has been presented in any form at a medical or pharmacy meeting. The author would like to acknowledge Darshan Kulkarni, Barbara Fisher, and Mary St. Amour for their help in data collection and site identification, and Abbey Cohen for her review of the original manuscript.
References 1. Health on the Net Foundation (HON). 5th HON survey on the evolution of Internet use for health purposes. Available at www.hon.ch/Survey/ ResultsSummary_oct_nov99.html. Accessed April 25, 2000. 2. Mittman R, Cain M. The Future of the Internet in Health Care: Five Year Forecast. Oakland, Calif: California HealthCare Foundation; January 1999. 3. Silberg WM, Lundberg GO, Musacchio RA. Assessing, controlling, and assuring the quality of medical information on the Internet: caveat lector et viewor-Iet the reader and viewer beware . JAMA. 1997;277:1244-5. 4. Ukens C.lnternet pharmacies. Drug Topics; May 17, 1999:63-70. 5. Latner AW. Pharmacists on the Internet: how to attract customers. Pharm Times. 1998:64:78. 6. Landis NT. Virtual pharmacies boast easy access, privacy safeguards. Am J Health Sys Pharm. 1999;56:1174,1177-9. 7. Marwick C. Several groups attempting regulation of Internet Rx. JAMA. 1999;281:975-£. 8. Paulsen M. New NABP program combines criteria, inspections to certify online pharmacy quality. JAm Pharm Assoc. 1999:39:870. 9. California HealthCare Foundation. Privacy: Report on the Privacy Policies and Practices of Health Web Sites. Oakland, Calif: California HealthCare Foundation; January 2000.
MarchlApril2001
Vol. 41, No.2