JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 83, NO.

21, 2024

ª 2024 THE AUTHORS. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN

COLLEGE OF CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER

THE CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/).

A Technology-Assisted Web Application

for Consumer Access to a
Nonprescription Statin Medication
Steven E. Nissen, MD,a Howard G. Hutchinson, MD,b Kathy Wolski, MPH,a Karol Watson, MD,c
Seth S. Martin, MD, MHS,d Erin D. Michos, MD,d William S. Weintraub, MD,e Melanie Morris, BS,b Leslie Cho, MD,a
Luke Laffin, MD,a Douglas Jacoby, MD,f Christie M. Ballantyne, MD,g Jan Ekelund, MSC,b Filip Birve, BSC,b
Venu Menon, MD,a Michelle Strzelecki, RN,a Paul M. Ridker, MDh

ABSTRACT

BACKGROUND Although statins reduce adverse cardiovascular outcomes, less than one-half of eligible patients
receive treatment. A nonprescription statin has the potential to improve access to statins.

OBJECTIVES This study sought to assess concordance between clinician and consumer assessment of eligibility for
nonprescription statin treatment using a technology assisted self-selection Web application (Web App) and evaluate
effect on low-density lipoprotein cholesterol (LDL-C) levels.

METHODS This study was a prospective actual use 6-month study to evaluate use of a Web App to qualify participants
without a medical background for a moderate-intensity statin based on current guidelines. Participants entered demo-
graphic information, cholesterol values, blood pressure, and concomitant medications into the Web App, resulting in 3
possible outcomes: “do not use,” “ask a doctor,” and “OK to use.”

RESULTS The study included 1,196 participants, with a median age of 63 years (Q1-Q3: 57-68 years); 39.6% were
women, 79.3% were White, 11.7% were Black, and 4.1% had limited literacy. Mean LDL-C was 139.6
28.3 mg/dL and
the median calculated 10-year risk of atherosclerotic cardiovascular disease was 10.1% (Q1-Q3: 7.3%-14.0%). Initial Web
App self-selection resulted in an outcome concordant with clinician assessment in 90.7% (95% CI: 88.9%-92.3%) of
participants, and 98.1% (95% CI: 97.1%-98.8%) had a concordant final use outcome during treatment. Mean percent
change in LDL-C was 35.5% (95% CI: 36.6% to 34.3%). Serious adverse events occurred in 27 (2.3%) participants,
none related to the study drug.

CONCLUSIONS In this actual use study, a technology-assisted Web App allowed >90% of consumers to correctly
self-select for statin use and achieve clinically important LDL-C reductions. (Technology-Assisted Cholesterol Trial in
Consumers [TACTiC]; NCT04964544) (J Am Coll Cardiol 2024;83:2080–2088) © 2024 The Authors. Published by
Elsevier on behalf of the American College of Cardiology Foundation. This is an open access article under the CC BY-NC-
ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

A dministration of statins to reduce low-

density lipoprotein cholesterol (LDL-C) rep-
resents an essential component of strategies
to improve cardiovascular outcomes.1 However, less
than one-half of primary prevention patients eligible
under current guidelines receive treatment.2 Many
factors contribute to undertreatment of patients at
risk, including reluctance of eligible patients to seek

Listen to this manuscript’s

audio summary by From the aCleveland Clinic Coordinating Center for Clinical Research, Cleveland, Ohio, USA; bAstraZeneca, Wilmington, Delaware,
Editor-in-Chief USA; cWomen’s Cardiovascular Services, UCLA Health, Los Angeles, California, USA; dJohns Hopkins School of Medicine, Balti-
Dr Valentin Fuster on more, Maryland, USA; eMedStar Health Research Institute, Georgetown University, Washington, DC, USA; fPennsylvania Hospital,
www.jacc.org/journal/jacc. Philadelphia, Pennsylvania, USA; gBaylor College of Medicine, Houston, Texas, USA; and the hBrigham and Women’s Hospital,
Boston, Massachusetts, USA.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’
institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information,
visit the Author Center.

Manuscript received March 11, 2024; accepted March 19, 2024.

ISSN 0735-1097 https://doi.org/10.1016/j.jacc.2024.03.388

JACC VOL. 83, NO. 21, 2024 Nissen et al 2081
MAY 28, 2024:2080–2088 Consumer Access to Nonprescription Statin Medication

medical care and limited access to health care in some copy of the trial database was transferred to ABBREVIATIONS

communities.3,4 This problem represents a major C5Research for independent statistical anal- AND ACRONYMS

public health concern that has persisted for decades ysis. This study was conducted in accordance
ASCVD = atherosclerotic
despite extensive efforts to educate patients and cli- with ethical principles in the Declaration of cardiovascular disease
nicians about the importance of controlling choles- Helsinki and international ethical guidelines
C5Research = Cleveland Clinic
terol to reduce the risk of cardiovascular morbidity from the Council for International Organiza- Coordinating Center for Clinical
and mortality. Multiple past attempts to address this tions of Medical Sciences. Advarra served Research

problem through over-the-counter (OTC) statins as the central Institutional Review Board and DFL = drug facts label

were unsuccessful in achieving regulatory approval approved the informed consent form prior to LDL-C = low-density

due to concerns about inappropriate self-selection participant enrollment. lipoprotein cholesterol

by consumers for whom statins could be unnecessary OTC = over the counter
STUDY POPULATION. The study enrolled
5-9
or unsafe. individuals who responded to advertise- TASS = technology-assisted
self-selection
SEE PAGE 2089 ments for a study about heart health or
Web App = Web application
lowering cholesterol. Men or women 20 to 75
The current study sought to address this problem years of age were eligible for participation. Because
through a novel technology-assisted self-selection the Web App was displayed in English, all partici-
(TASS) tool embedded within a Web application (Web pants had to be able to read and understand English.
App) that allows consumers to qualify for nonpre- Participants were excluded from the study if they
scription access to low-dose rosuvastatin (5 mg), were currently employed by a health care–related
a widely used moderate-intensity statin with a company or were ever employed as a health care
well-documented safety record. The TASS program professional (physician, nurse, nurse practitioner,
was developed as medical software that prompts physician assistant, or pharmacist). Women of child-
participants to enter demographic information and bearing potential were excluded unless they were
medical, laboratory, and blood pressure data to assess using an acceptable method of birth control for the
eligibility for treatment with a moderate-intensity duration of the study. Participants were required to
statin. 1 The TACTiC (Technology Assisted Self- have the ability to receive email throughout the
Selection in Consumers) study was an actual use study. Additional inclusion and exclusion criteria are
evaluation designed to determine whether use of described in the protocol (available in the
a Web App would allow consumers to make appro- Supplemental Appendix).
priate self-selection and ongoing treatment decisions.
STUDY DESIGN AND PROCEDURES. This was a
Actual use studies assess the safety and efficacy
of a prescription drug under conditions resembling 6-month, single-arm study designed to evaluate the

nonprescription use.10 The study compared concor- safety and efficacy of the use of a Web App to qualify

dance between participant and clinician assessments participants for initial and ongoing treatment with a

of eligibility for initial and ongoing treatment, statin when used in an unsupervised environment. A

the effect of this therapeutic strategy on LDL-C schematic describing the study design is shown in

levels, and compliance with dosing and retesting Supplemental Figure 1. The study was conducted

recommendations. entirely using virtual visits with no in-person contact.

Individuals who responded to an advertisement and
METHODS met the initial study requirements were sent an email
with a link to the Web App for treatment eligibility
STUDY ORGANIZATION AND OVERSIGHT. The study assessment. Several examples of the screens used by
was designed by the sponsor, academic investigators, participants to enter their medical data are shown in
and Concentrics Research, a contract research orga- Supplemental Figures 2 to 11.
nization specializing in consumer health care. The The Web App incorporates the 2018 Cholesterol
study was reviewed by the U.S. Food and Drug Treatment Guidelines using the pooled cohort equa-
Administration and overseen by an academic advi- tions to determine 10-year cardiovascular risk and
sory group composed of experts in lipid management, includes key components of a proposed drug facts
epidemiology, and public health. The Cleveland label (DFL) for nonprescription rosuvastatin
Clinic Coordinating Center for Clinical Research (Supplemental Figure 12).1 To determine eligibility for
(C5Research) served as the academic coordinating treatment, participants entered demographic and
center and independently reviewed participant out- medical information including total cholesterol, high-
comes to determine suitability for treatment with a density lipoprotein cholesterol, LDL-C, triglycerides,
statin. At the conclusion of the study, a complete and systolic and diastolic blood pressure, and when
2082 Nissen et al JACC VOL. 83, NO. 21, 2024

Consumer Access to Nonprescription Statin Medication MAY 28, 2024:2080–2088

F I G U R E 1 Prescreen Outcomes and Flow of Participants Following Enrollment

133,744 responded to
advertisement

110,712 excluded
47,910 did not want to participate
61,189 failed 1 or more inclusion or exclusion criteria
38,125 health care professional or works in health care
20,200 outside age range for participation
7,935 unable to be contacted by site personnel (no home
address, email address, or Internet access)
3,127 received investigational therapy within 12 months
1,677 resided in Alaska
949 cannot read, speak, or understand English
84 unable to conduct interview in a private location

23,032 sent Web App link

12,624 completed Web

App assessment

10,332 received “Do Not Use” outcome for

1 or more reasons
5,669 taking cholesterol-lowering drug
2,192 low ASCVD risk score
1,273 labs/blood pressure outside range
481 no risk-enhancing factor
467 ASCVD present
347 age outside range
217 high ASCVD risk score
279 cyclosporine or warfarin use
36 liver disease
3 pregnancy

2,292 received an “OK to Use”

or “Ask a Doctor” outcome

127 did not confirm physician approval

223 did not create an account
746 did not purchase the study drug

1,196 signed informed consent

(self-selection population)

7 withdrawals
1 screen failure

1,188 received the study drug

(actual use ITT population)

11 not treated,
34 withdrew from the study
returned all study drug

1,177 took the study drug 1,154 completed final

(Safety Population) use assessment

ASCVD ¼ atherosclerotic cardiovascular disease; ITT ¼ intention to treat; Web App ¼ Web Application.
JACC VOL. 83, NO. 21, 2024 Nissen et al 2083
MAY 28, 2024:2080–2088 Consumer Access to Nonprescription Statin Medication

indicated, any of the risk-enhancing factors cited in

T A B L E 1 Baseline Characteristics (N ¼ 1,196)
the guidelines. If individuals did not have their lab-
Sex at birth
oratory or blood pressure values prior to taking the
Male 722 (60.4)
initial TASS assessment, the Web App or study staff at
Female 474 (39.6)
the call center provided options for testing that Age, y 63 (57-68)
included local pharmacies, in-store clinics, a home <50 y 86 (7.2)
test kit, or a physician office. Study staff were not 50-65 y 675 (56.4)
permitted to arrange or pay for laboratory or blood >65 y 435 (36.4)

pressure testing, which was voluntary for the Self-identified race or ethnicity
White 949 (79.3)
individual.
Hispanic 38 (3.2)
Prescreening of participants using the TASS Web
Black 140 (11.7)
App could result in 1 of 3 possible outcomes: “do not Asiana 34 (2.8)
use,” “ask a doctor,” or “OK to use.” Because this was South Asianb 14 (1.2)
an actual use study, participants were only enrolled if Other or not reported 21 (1.8)
they received an outcome of “OK to use” or an “ask a Education level
doctor” and indicated in the Web App that a physician Less than high school 0 (0)
Completed high school or completed 62 (5.2)
confirmed that they could be treated. The drug facts
graduate development test
label describes conditions or concomitant medica- Some college or technical school 243 (20.3)
tions that would result in an “ask a doctor” because of Graduated college or technical school 885 (74.0)
increased risk of an adverse reaction such as use of Unavailable 6 (0.5)
colchicine. A previous study confirmed that con- Limited literacyc 49 (4.1)

sumers could use the Web App to obtain a correct “do Employment status
11 Retired 502 (42.0)
not use” outcome.
Full-time 485 (40.6)
Upon completion of the initial TASS assessment,
Part-time 105 (8.8)
participants were directed by the Web App to contact
Unemployed 43 (3.6)
the call center for an initial virtual visit. Participants Student 4 (0.3)
were instructed to bring verification of the diagnostic Unknown or other 57 (4.8)
data they entered in the Web App. Verified diagnostic Prior cardiovascular disease
values could include results obtained prior to Myocardial infarction 1 (0.1)

completing the initial TASS assessment if the data Stroke 8 (0.7)

Surgical procedure or stent 3 (0.3)
were obtained within 12 months and had a name that
Peripheral arterial disease 1 (0.1)
matched the participant’s driver’s license or other
Family history of cardiovascular disease
photo identification. If the participant did not have Father or brother before age 55 y 138 (11.5)
verified data, they were required to have a study- Mother or sister before age 65 y 79 (6.6)
mandated test and offered the option of testing or- Hypertension 470 (39.3)
dered by their health care provider or receiving a Diabetes 73 (6.1)

home test kit. Participants were not required to pay Laboratory values and blood pressure
Total cholesterol, mg/dL 221.1
33.0
for study-mandated tests. At the initial visit,
Low-density lipoprotein cholesterol, mg/dL 139.6
28.3
informed consent was obtained, diagnostic tests were
High-density lipoprotein cholesterol, mg/dL 53.5
15.1
verified, and the participant was interviewed by a
Triglycerides, mg/dL 132.0 (93.0-186.5)
clinician who was unaware of the information the Systolic blood pressure, mm Hg 130.0 (121.0-141.0))
participant entered into the Web App assessment. Diastolic blood pressure, mm Hg 80.0 (74.0-87.0)
The clinician obtained a medical and medication Clinician 10-y risk of ASCVD, % 10.1 (7.3-14.0)
history and used only verified laboratory and blood
Values are n (%), median (Q1-Q3), or mean
SD. Data are for the intention-to-treat population
pressure data to complete an independent Web App consisting of all participants who met self-selection criteria. aChina, the Philippines, Vietnam, or
assessment. Participants were unaware of the results Korea. bIndia, Pakistan, or Bangladesh. cLimited literacy is seventh/eighth grade reading level and
below.
of the clinician assessment. At this initial visit, liter- ASCVD ¼ atherosclerotic cardiovascular disease.
acy was assessed using the REALM (Rapid Estimate of
Adult Literacy in Medicine) score.12
All participants who qualified for treatment based 45- or 90-day supply of rosuvastatin. During treat-
on their initial Web App assessment were eligible to ment, participants were instructed to reorder medi-
continue into the treatment phase of the study. Po- cation to allow them to complete the 6-month study.
tential participants were required to create a Web App They could order a 45- or 90-day supply with each
account and make an initial purchase of either a reorder but had to take an abbreviated Web App
2084 Nissen et al JACC VOL. 83, NO. 21, 2024

Consumer Access to Nonprescription Statin Medication MAY 28, 2024:2080–2088

independent clinical team at the academic coordi-

T A B L E 2 Co-Primary Outcomes
nating center (C5Research) reviewed the outcomes
Outcome for Initial Technology-Assisted Self-Selection to determine if the participant’s responses in the
Outcome (N ¼ 1,196)a Number of Participants 95% CI (%) Web App were appropriate when factors were
Correct self-selection 1,085 (90.7) 88.9-92.3 considered that were not known to the study clinician
Concordant “OK to use” 960 (80.3) NA
conducting the interview. If mitigating factors
Concordant “ask a doctor” 47 (3.9) NA
were present, the overall outcome was classified as a
Concordant outcome following 78 (6.5) NA
review with clinician adjudicationb correct or concordant outcome following clinical
Incorrect self-selection 111 (9.3) NA review.

Outcome for Assessment During 6-mo Treatment Period STUDY OUTCOMES. Three co-primary outcomes
Outcome (n ¼ 1,154) c
Number of Participants 95% CI (%) were prespecified. The first co-primary outcome was
Correct use during treatment period 1,132 (98.1) 97.1-98.8 the percent of participants whose initial self-
Concordant “OK to use” 836 (72.4) NA selection was concordant with the clinician assess-
Concordant “ask a doctor” 14 (1.2) NA ment. The second co-primary outcome was the
Concordant “do not use” 84 (7.3) NA percent of participants whose final use assessment
Concordant outcome following review 198 (17.2) NA
was concordant with the clinician assessment. The
with clinician adjudicationb
Incorrect treatment period use 22 (1.9) NA third co-primary endpoint was the percent change
from baseline in LDL-C with treatment. Prespecified
Change in LDL-C at Final Visit
secondary endpoints included the percentage of
Number Exceeding Mean
20% Prespecified LDL-C Change
participants who voluntarily retested their LDL-C,
Outcome (n ¼ 1,063)d Threshold (95% CI) (%) compliance with warnings based on the proposed
Mean percent change from baseline in LDL-C 898 (84.5) 35.5 (36.6 to 34.3) DFL for nonprescription rosuvastatin, and compli-
ance with dosing.
Values are n (%) unless otherwise indicated. Expanded tables reporting more details about outcomes available in
Supplemental Tables 1 and 2. aPopulation consists of all participants who signed the informed consent. bDis- STATISTICAL ANALYSIS. The statistical analysis plan
cordant clinician and participant outcome, but independent clinical review determined that the participant’s
outcome was appropriate. cPopulation who completed second visit or had sufficient information to determine the is provided in the Supplemental Appendix. For the
final use outcome. dOf the 1,188 participants who entered the treatment phase, 1,063 had a verified LDL-C retest. first and second co-primary endpoints, concordance
Does not include 125 participants; 124 were not required to have an LDL-C retest based on their final outcome
status and 1 withdrew from the study. between the participant and clinician for the initial
LDL-C ¼ low-density lipoprotein cholesterol; NA ¼ not applicable. and final use outcomes respectively, formal testing
was performed of the null hypothesis (H0) (P #85%)
vs the alternative hypothesis (H1) (P >85%), where
assessment to help ensure continued appropriateness P represents the percentage of concordant outcomes
for treatment. During the treatment phase, partici- in the analysis population. For the first co-primary
pants were directed through the Web App and emails outcome, the analysis plan prespecified that the null
to undergo retesting of their LDL-C prior to a second hypothesis would be rejected if the lower bound of
visit. At the end of the 6-month period, or sooner if the 2-sided 95% exact binomial (Clopper-Pearson) CI
the participant received a “do not use” outcome from was >85%. The planned sample size of 1,220 partici-
the Web App, participants were scheduled for their pants provided 90% power for a lower bound of
second and final virtual visit. At this visit, partici- concordance of 85% assuming a correct self-selection
pants were required to bring verification of their LDL- rate of 88.2%. For the second co-primary outcome,
C values entered into the Web App or undergo a the analysis plan prespecified that the null hypothesis
study-mandated test. A clinician, unaware of the would be rejected if the lower bound of the lower
participant’s final Web App outcome, conducted bound of the 95% exact binomial CI was >85%. An
another medical and medication history and per- estimated sample size of 500 participants provided
formed an independent final use Web App 90% power to demonstrate a lower bound of 85%
assessment. assuming a true correct final use outcome of 90%. For
Following the virtual visits, an independent clin- the third co-primary endpoint, the percent change
ical coding team reviewed the participant and clini- from baseline in verified LDL-C, the null hypothesis
cian outcomes to determine if they were concordant. (H0) (mean reduction in LDL-C is #15%) would be
In those situations in which the participant and rejected in favor of the alternative hypothesis (H1)
clinician had the same “OK to use,” “ask a doctor,” or (mean reduction in LDL-C is >15%) if the upper bound
“do not use” outcome, the overall initial and final use of the 2-sided 95% CI for the mean percent change
outcomes were classified as correct or concordant. If from baseline in verified LDL-C at the second visit
the participant and clinician outcomes differed, an was lower than 15%.
JACC VOL. 83, NO. 21, 2024 Nissen et al 2085
MAY 28, 2024:2080–2088 Consumer Access to Nonprescription Statin Medication

RESULTS
T A B L E 3 Secondary Outcomes

CHARACTERISTICS OF THE STUDY POPULATION. n Compliant 95% CI (%)

The flow of participants through the study is shown in Compliance with LDL-C retesting
Overall population 1,188 995 (83.8) 81.5-85.8
Figure 1. Baseline characteristics of participants are
Eligible for continuous treatmenta 910 845 (92.9) 91.0-94.4
reported in Table 1. In response to advertising,
Compliance with dosing n Compliant (%)
133,744 individuals were assessed for eligibility. Of Number evaluableb 1,146 95.0 (Q1-Q3: 82.9-98.9)
these, 110,712 did not continue to the Web App or Dosing persistencec n Persistent
failed the prescreen inclusion/exclusion criteria. Of Eligible for continuous treatment 910 896 (98.5)
the remaining 23,032 that were sent a Web App link, Compliance with proposed drug n Compliant
facts label warnings
12,624 completed the Web App assessment and 10,332
Stop used 0 NA
received a “do not use” outcome. Of the remaining
Do not usee 59 47 (79.7)
potential participants, 1,196 signed an informed con- Ask a doctor before usef 24 20 (83.3)
sent and comprised the overall study population.
Most participants were 50 to 65 (56.4%) or Values are n (%) or median (Q1-Q3), unless otherwise indicated. aParticipants who qualified to receive treatment
at all reassessments. bPopulation with data from returned pill bottles or the eDiary to assess compliance. Dosing
>65 (36.4%) years of age. Self-identified race or compliance was assessed based on the number of pills taken relative to the intended duration of exposure for the
participant. cParticipants who qualified to receive treatment at all reassessments and purchased at least 180 days
ethnicity included 79.3% White, 11.7% Black, and
of treatment. d“Stop use” warnings were those related to severe muscle symptoms, myopathy or rhabdomyolysis,
3.2% Hispanic individuals. Participants who gradu- new onset liver disease, or pregnancy. eThe “do not use” warnings were concomitant use of prescription lipid-
lowering medications, warfarin or cyclosporine, and cardiovascular adverse events. Of the 12 participants who
ated college or technical school comprised 74.0% of were noncompliant, 3 had a history of atherosclerosis, 4 were taking an omega-3 fatty acid, 4 were taking another
the population and 4.1% had limited literacy (sev- statin, and 1 was taking cholestyramine. fThe “ask a doctor” warnings were kidney disease, use of colchicine or HIV
medications, and excessive alcohol use. Of the 4 participants who were noncompliant, 3 were taking an HIV
enth/eighth grade reading level or below). Retired medication and 1 had kidney disease.
individuals comprised 42% of the population and Abbreviations as in Table 2.

40.6% were employed full-time. Mean LDL-C was

139.6
28.3 mg/dL, median systolic blood pressure
was 130 mm Hg (Q1-Q3: 121-141 mm Hg), and median outcomes, 89 had a correct final use outcome and 8
clinician-calculated 10-year risk of atherosclerotic withdrew from the study. Fourteen remained incor-
cardiovascular disease (ASCVD) was 10.1% (Q1-Q3: rect for the second co-primary endpoint. The third co-
7.3%-14.0%). primary outcome, mean change in LDL-C was 35.5%
(95% CI: 36.6% to 34.3%) prior to the second
CO-PRIMARY OUTCOMES. Co-primary outcomes are visit with 84.5% of participants exceeding 20%
reported in Table 2. The initial self-selection by par- LDL-C lowering.
ticipants resulted in an overall concordance between The final LDL-C was 88.1
25 mg/dL, an absolute
the participant and clinician in 90.7% (95% CI: 88.9%- reduction of 51.4
29 mg/dL or 1.33 mmol/L.
92.3%). An “OK to use” outcome was concordant be- SECONDARY OUTCOMES. Prespecified secondary
tween the participant and clinician in 80.3% and an outcomes are reported in Table 3. The secondary
“ask a doctor” outcome was concordant in another outcome, compliance with retesting of LDL-C, was
3.9%. An additional 6.5% were deemed concordant documented in 83.8% of the overall population and in
after independent review by academic clinicians at 92.9% of the 910 individuals who qualified for treat-
C5Research. Further details describing the indepen- ment at all reassessments. Among these 910 in-
dently adjudicated outcomes and categories of dividuals, 896 (98.5%) purchased sufficient
incorrect selection are reported in Supplemental medication to last for the duration of the 6-month
Table 1. For the second co-primary outcome, 98.1% treatment period. Compliance with dosing based on
(95% CI: 97.1%-98.8%) of participants had final use pill count for returned bottles was 95.0% (Q1-Q3:
assessments concordant with the clinician during the 82.9%-98.9%). No participants received a “stop use”
6-month treatment period, including concordance warning. A “do not use” warning was received by 59
between the participant and clinician in 72.4% with participants, and 47 discontinued nonprescription
an “OK to use” outcome, 1.2% with “ask a doctor” rosuvastatin. Of the 12 who continued therapy, 3 were
outcomes, and 7.3% with a “do not use” outcome. participants with a history of atherosclerosis who
Another 17.2% of outcomes were adjudicated as were not treated with a cholesterol-lowering medi-
concordant by independent clinicians at C5Research. cation at the start of the study, and the remaining 9
Further details describing the independently adjudi- were taking a concomitant cholesterol- or
cated outcomes and categories of incorrect selection triglyceride-lowering medication. Four were taking
are reported in Supplemental Table 2. Of the 111 par- an omega-3 fatty acid medication, 4 were taking a
ticipants who were incorrect on their initial statin, and 1 was taking cholestyramine. A total of 24
2086 Nissen et al JACC VOL. 83, NO. 21, 2024

Consumer Access to Nonprescription Statin Medication MAY 28, 2024:2080–2088

participants received an “ask a doctor” warning, and

T A B L E 4 Adverse Events (N ¼ 1,177)
20 spoke with a clinician. Of the 4 who did not, 3 were
Participants with any adverse event 623 (52.9)
on HIV medications and 1 had kidney disease.
SARS-CoV-2 test positive 113 (9.6)
ADVERSE EVENTS. Adverse events observed during Arthralgia 82 (7.0)
the trial are reported in Table 4. Adverse events Headache 64 (5.4)
leading to study drug discontinuation occurred in Pain in extremity 51 (4.3)

7.1%, with none leading to withdrawal from the trial. Myalgia 48 (4.1)
Cough 45 (3.8)
The occurrence of myalgias was low, reported by 48
Fatigue 45 (3.8)
(4.1%) participants. There were 40 serious adverse
Back pain 41 (3.5)
events in 27 participants distributed across multiple Diarrhea 41 (3.5)
organ system classes, with none deemed related to Muscle spasms 39 (3.3)
the study drug by the principal investigator Adverse events leading to discontinuation of study 84 (7.1)
medication
(Supplemental Table 3). Discontinuation of the study
Musculoskeletal and connective tissue disorders 36 (3.1)
drug occurred in 3 (2 due to myocardial infarctions
General disorders and administration site conditions 13 (1.1)
and 1 due to stroke) because they met “do not
Gastrointestinal disorders 9 (0.8)
use” criteria. Nervous system disorders 9 (0.8)
Cardiac disorders 3 (0.3)
DISCUSSION Other disorders 16 (1.4)
Leading to withdrawal from trial 0
This phase 3 actual use study compared concordance Serious adverse event 27 (2.3)
between clinicians and participants of guideline Gastrointestinal disorders 6 (0.5)

eligibility for nonprescription rosuvastatin using a Cardiac disorders 5 (0.4)

Nervous system disorders 3 (0.3)
novel TASS approach that allowed potential con-
Neoplasms, benign or malignant 4 (0.3)
sumers to enter their medical data into a Web App to
Othera 12 (1.0)
qualify for initial and ongoing treatment. The first co-
Leading to discontinuation of study medication 3 (0.3)
primary outcome, correct initial self-selection, was Serious adverse event leading to deathb 1 (0.1)
confirmed in 90.7% (95% CI: 89.9%-92.3%) of partic-
ipants. For the second co-primary endpoint, in those Values are n (%). The safety population includes all participants who received the
study drug but does not include 11 participants who returned all study drug. aThe
who entered the 6-month treatment period, 98.1% complete listing of all serious adverse events is reported in Supplemental Table 3.
Participants may have had more than 1 serious adverse event. bDeath due to
(95% CI: 97.1%-98.8%) demonstrated appropriate use.
trauma.
The third co-primary outcome, mean percent change
from baseline in LDL-C assessed through retesting,
was 35.5% (95% CI: 36.6% to 34.3%) (Central commonly untreated. In a cross-sectional study using
Illustration). The secondary endpoints showed that the National Health and Nutrition Examination
participants were largely adherent to dosing, LDL-C Survey, 42.7% of adults with an LDL-C of 160 to
retesting and warnings contained in the DFL. 189 mg/dL were not receiving a cholesterol-lowering
Adverse events were uncommon, and none led to medication.14 Undertreatment is more prevalent in
withdrawal from the trial. The overall results met women, racial and ethnic minority groups, and those
prespecified criteria for efficacy for all 3 co-primary without health insurance. 13 These groups may not be
endpoints, demonstrating that consumers can receiving health care from a provider with prescribing
appropriately self-select for statin treatment with the authority. 3 In the setting in which ASCVD remains the
assistance of a dedicated Web App, potentially leading cause of death in the United States, the
allowing expanded access to this important class of undertreatment of adults with elevated LDL-C rep-
medications. resents a lost public health opportunity and a moti-
The current findings have potentially important vation to consider novel approaches to therapy,
public health implications. Many contemporary including direct consumer access.15
studies demonstrate that a minority of patients Increasing access through nonprescription avail-
eligible for statins based on current guidelines are ability of statin treatment could potentially help
receiving treatment. 2,13 In a registry of 140 primary reduce the societal burden of ASCVD. Several previous
care, cardiology, and endocrinology practices, only efforts to achieve regulatory approval for OTC statins
36% of guideline-eligible primary prevention patients failed, primarily because of concerns about the ability
were receiving statins. 12 The prevalence of elevated of consumers to appropriately self-select for treat-
LDL-C levels in the U.S. population has declined ment. These development programs relied on the DFL
during the past 2 decades but remains high and to guide consumers in self-selection, which resulted in
JACC VOL. 83, NO. 21, 2024 Nissen et al 2087
MAY 28, 2024:2080–2088 Consumer Access to Nonprescription Statin Medication

C ENTR AL I LL U STRA T I O N Technology-Assisted Self-Selection for Nonprescription Statin

Compare Participant Compare Participant

Reorders Drug and
Completes Web and Clinician Home Drug and Clinician Final
Retakes Web App
App Assessment Initial Assessments Delivery Assessments
Assessment
(Initial Self-Selection) (Final Reselection)

Self Selection 6 Month Use Period Co-Primary Endpoints

Cholesterol Levels Have you ever had Results

... any of the following? ...
Crestor OTC is not right for you
Yes No
Total Cholesterol 129 You should talk to a doctor
mg/dL Heart attack
because the following are signs
that you may need a stronger
LDL Cholesterol 34 Stroke medicine available only by
mg/dL prescription:
An operation or
procedure on your heart • Heart attacks
HDL Cholesterol 69 mg/dL • Strokes
Peripheral artery • Procedures of the heart
disease (PAD) • Peripheral artery disease
See Sample Lab Report
I Don’t Know (PAD)
Are you sure everything is correct? Unfortunately, you are not
eligible to continue with the
Yes, everything No, I want to change Confirm and
study. Thank you for your
is correct something continue
interest in participating.

Co-Primary Endpoints
1) Self-selection by participants resulted in an outcome concordant with clinicians in 90.7%
2) Final use assessments concordant with clinicians for 98.1% of participants
3) Mean percent reduction in LDL-C was 35.5% (51 mg/dL)

Nissen SE, et al. J Am Coll Cardiol. 2024;83(21):2080–2088.

(Top) Sequence of trial activities. (Middle) Three representative Web application screens presented to participants during the self-selection process. (Bottom) Results
of the study for the 3 co-primary endpoints. LDL-C ¼ low-density lipoprotein cholesterol; OTC ¼ over the counter.

appropriate use in only 53% to 72% of trial partici- chronic asymptomatic condition. Adherence to statin
pants.5-9 Unfortunately, many consumers do not read treatment also represents a public health concern. 4 In
16
and understand labels for OTC medications. In a label the current study, direct delivery of nonprescription
comprehension study of OTC lovastatin, only 1% of statin achieved 95% adherence based on pill count
consumers were correct on all label criteria specified with 98.5% treatment persistence over a 6-month
for appropriate self-selection.17 The complexity of period in those eligible for continuous therapy. The
decision making in selecting patients for OTC statin resulting 35.5% reduction in LDL-C (51.4 mg/dL or
treatment is not easily addressed through a DFL, which 1.33 mmol/L) is large enough to effectively reduce the
cannot readily incorporate the 10-year ASCVD risk incidence of ASCVD events based on the Cholesterol
score to help guide purchase decisions.18 Treatment Trialists Collaboration. 19
The current trial demonstrates that a novel Web- STUDY LIMITATIONS. It remains uncertain whether
based approach to treatment can help manage these the TASS approach to self-selection can result in long-
concerns by guiding participants through the self- term adherence. Other limitations of the trial
selection process and delivering the medication included the restriction to participants who could
directly to consumers rather than through OTC access. read English, had access to the Internet, and were
If approved, this approach would represent the first willing to use technology. Applicability to people who
treatment made available without a prescription for a cannot read and understand English would need to be
2088 Nissen et al JACC VOL. 83, NO. 21, 2024

Consumer Access to Nonprescription Statin Medication MAY 28, 2024:2080–2088

demonstrated. Other approaches to provide safe ac- conduct this study. Dr Hutchinson, Mr Ekelund, Mr Birve, and Ms
Morris are employees of AstraZeneca.
cess to a statin would be required for those who are
less willing or unable to use technology.
ADDRESS FOR CORRESPONDENCE: Dr Steven E.
CONCLUSIONS Nissen, Cleveland Clinic JB-20, 9500 Euclid Avenue,
Cleveland, Ohio 44195, USA. E-mail: [email protected].
In this actual use study of a TASS tool for access to
PERSPECTIVES
nonprescription rosuvastatin, 90.7% of participants
correctly self-selected for statin use based on current
guidelines and 98.1% demonstrated correct use dur- COMPETENCY IN SYSTEMS-BASED PRACTICE:
ing the trial. Treated participants had a high level of Use of a Web-based application to determine eligi-
adherence and achieved clinically meaningful re- bility for nonprescription rosuvastatin can facilitate
ductions in LDL-C. self-selection for treatment and achieve substantial
reduction in LDL-cholesterol levels.
FUNDING SUPPORT AND AUTHOR DISCLOSURES
TRANSLATIONAL OUTLOOK: Regulatory author-
The trial was funded by AstraZeneca and coordinated by AstraZeneca,
ities should evaluate the safety and efficacy of this
Concentrics Research (a contract research organization) and
the Cleveland Clinic Coordinating Center for Clinical Research technology while future studies assess long-term
(C5Research). Ms Wolski, Dr Watson, Dr Martin, Dr Michos, Dr treatment adherence and clinical outcomes.
Weintraub, Dr Cho, Dr Laffin, Dr Jacoby, Dr Ballantyne, Dr Menon, Ms
Strzelecki, and Dr Ridker have received support from AstraZeneca to

REFERENCES

1. Grundy SM, Stone NJ, Bailey AL, et al. 2018 8. U.S. Food and Drug Administration, Merck 2024. https://www.cdc.gov/nchs/data/databriefs/
AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/ Research Labs. Joint meeting of the Nonpres- db456.pdf
APhA/ASPC/NLA/PCNA guideline on the manage- cription Drugs Advisory Committee and Endo-
16. Catlin JR, Brass EP. The effectiveness of
ment of blood cholesterol: a report on the Amer- crinological and Metabolic Drugs Advisory
nonprescription drug labels in the United States:
ican College of Cardiology/American Heart Committee, Mevacor NDA 21-213, lovastatin. 2000.
insights from recent research and opportunities
Association Task Force on Clinical Practice Guide- Accessed April 5, 2024. https://wayback.archive-it.
for the future. Pharmacy (Basel). 2018;6(4):119.
lines. J Am Coll Cardiol. 2019;73:3168–3209. org/7993/20170404111228/https://www.fda.gov/
https://doi.org/10.3390/pharmacy6040119
ohrms/dockets/ac/00/slides/3622s1.htm
2. Kim CJ, Sussman JB, Mukamal KJ, Eades M,
17. Shay L. SELECT Label Comprehension Studies
Anderson TS. Trends in primary prevention statin 9. U.S. National Library of Medicine. ClinicalTrials. Mevacor Daily. Presented at: Joint Meeting of the
use by cardiovascular risk score from 1999 to gov. Actual Use Trial of Atorvastatin Calcium Nonprescription Drugs Advisory Committee and the
2018: a repeated cross-sectional study. Ann Intern 10 mg. NCT01964326. Accessed April 5, 2024. Endocrinologic and Metabolic Drugs Advisory Com-
Med. 2023;176:1684–1688. https://doi.org/10. https://clinicaltrials.gov/ct2/show/NCT01964326 mittee. December 13, 2007. Accessed April 5, 2024.
7326/M23-1915
10. U.S. Food and Drug Administration. Drug https://wayback.archive-it.org/7993/2017040405
3. Tolbert T, Drake D, Damico A. Key facts application process for nonprescription drugs. 0408/https://www.fda.gov/ohrms/dockets/ac/07/
about the uninsured population. Accessed April 5, Accessed April 5, 2024. https://www.fda.gov/ slides/2007-4331s1-05-FDA-Shay.ppt
2024. https://www.kff.org/uninsured/issue-brief/ drugs/types-applications/drug-application-process- 18. Joint session of the Nonprescription Drugs
key-facts-about-the-uninsured-population/ nonprescription-drugs Advisory Committee and the Endocrinology and
4. Bradley CK, Wang TY, Li S, et al. Patient- 11. Nissen S, Hutchinson H, Wang T, et al. Tech- Metabolic Drugs Advisory Committee for NDA 21-
reported reasons for declining or discontinuing nology-assisted self-selection of candidates for 213 proposing over-the-counter Mevacor TM
statin therapy: insights from the PALM registry. nonprescription statin therapy. J Am Coll Cardiol. (lovastatin) 20 mg. Accessed April 5, 2024. https://
J Am Heart Assoc. 2019;8:e011765. 2021;78:1114–1123. https://doi.org/10.1016/j.jacc. www.federalregister.gov/documents/2004/12/09/
2021.06.048 04-26995/joint-meeting-of-the-nonprescription-
5. U.S. Food and Drug Administration. Joint drugs-advisory-committee-and-the-endocrinologic-
meeting of the Nonprescription DFLs Advisory 12. Davis TC, Crouch MA, Long SW, et al. Rapid
and-metabolic
Committee and Endocrinological and Metabolic assessment of literacy levels of adult primary care
Drugs Advisory Committee, Pravachol NDA 21- patients. Fam Med. 1991;23:433–435. 19. Mihaylova B, Emberson J, Blackwell L, et al.
198, pravastatin sodium [PowerPoint slides]. Cholesterol Treatment Trialists’ (CTT) Collaborators.
13. Navar AM, Wang TY, Li S, et al. Lipid man-
Accessed April 5, 2024. https://wayback.archive- The effects of lowering LDL cholesterol with statin
agement in contemporary community practice:
it.org/7993/20170404111314/https://www.fda. therapy in people at low risk of vascular disease:
results from the Provider Assessment of Lipid
gov/ohrms/dockets/ac/00/slides/3622s2.htm meta-analysis of individual data from 27 randomised
Management (PALM) registry. Am Heart J.
trials. Lancet. 2012;380(9841):581–590. https://
6. Melin JM, Stuble WE, Tipping RW, et al. 2017;193:84–92. https://doi.org/10.1016/j.ahj.
doi.org/10.1016/S0140-6736(12)60367-5
A Consumer Use Study of Over-the-Counter 2017.08.005
Lovastatin (CUSTOM). Am J Cardiol. 2004;94: 14. Sayed A, Navar AM, Slipczuk L, et al. Preva-
1243–1248. KEY WORDS cholesterol, consumers,
lence, awareness, and treatment of elevated LDL
technology
cholesterol in US adults, 1999-2020. JAMA Car-
7. Brass EP, Vassil T, Replogle A, et al. Can con-
diol. 2023;8:1185–1187. https://doi.org/10.1001/
sumers self-select for appropriate use of an over-
jamacardio.2023.3931
the-counter statin? The self evaluation of lova- A PPE NDI X For an expanded Methods section
statin to enhance cholesterol. Treatment Study. 15. Xu J, Murphy SL, Kochanek KD, Arias E. Mor- and supplemental tables and figures, please see
Am J Cardiol. 2008;101:1448–1455. tality in the United States. 2021. Accessed April 5, the online version of this paper.