1. Patient-reported Outcomes as a Source
of Evidence in Off-Label Prescribing:
Analysis of Data From
Jeana Frost, Sally Okun, Timothy Vaughan, James Heywood, and Paul Wicks
Presenter: Parul Seth
University of California, Irvine
INF 295 Topics In Health Informatics
November 6th, 2013
1

2. Background: Patient-reported
Outcomes (PROs)
“The term "patient-reported outcomes
(PRO)" is used to refer to a host of
outcomes that can be provided only by
the patient. Examples of these
outcomes include symptom severity,
perception of daily functioning, feelings
of well being, global impressions of the
impact of treatment on daily life,
satisfaction with treatment, and healthrelated quality of life.”
Source: http://www.ispor.org/meetings/va0502/symposium.asp
2

3. Background: Off-label and Onlabel (FDA Approved)
21% of all prescriptions are for off-label purposes
73% lacked scientific evidence underlying their use
(Radley et al 2006)
3
Off-label prescribing: A source of innovation. But how to have supporting evidence for
off-label uses of a drug?

4. Off-label Use in the News
4
Source:
http://www.usatoday.com/story/money/business/2013/11/04/johnsonand-johnson-fines-drug-marketing/3432955/

5. Problem
 Off-label prescribing although legal in USA can be done
without supporting medical evidence.
 Providers and patients may rely upon their own individuallevel experience to inform clinical practice, which generates
only anecdotal and unstructured data.
 Patients may subject to unnecessary, ineffective, and even
harmful treatments.
 Clinical Trials for testing off-label use are rare, expensive and
time consuming (costs US $15,700 for a phase 1 trial and over
US $26,000 for a phase 3 trial per patient).
5

6. Problem
 Limitations of FDA guidance (2008), “Good Reprint Practices for the
Distribution of Medical Journal Articles and Medical or Scientific
Reference Publications on Unapproved New Uses of Approved Drugs
and Approved or Cleared Medical Devices”

Trade-off between quality of scientific literature & informal acceptability benchmark

Need for counterbalancing evidence from industry- publication bias, significant vs. nonsignificant findings

Off-label medication is frequently the only option for “orphan conditions” or where there is
no “approved treatment”.

Limited applicability of regulations, mostly to visible means of promotion.

Unlikely enforcement of guidance (not legally binding)
6
There is a need for evaluating a new use of a drug in an inexpensive and efficient
manner.

7. Proposed Solution
“A new source of evidence to evaluate off-label use: patientreported outcomes entered via an online community.”
7
Patients, entering outcomes within an online community, could inform how drugs are
working for off-label uses by expanding the available evidence base.

8. Online Community: Health 2.0
8
Source: http://scienceroll.com/2013/05/01/healthcare-infographics-help-start-a-conversation/

9. Proposed Solution
A new source of evidence to evaluate off-label use: patientreported outcomes entered via an online community:
Patients Like Me (PLM).
PROs
Structured Patientreported Data
9

10. “A tool for patients, researchers, and
caregivers that helps users make
treatment
decisions,
manage
symptoms, and improve outcomes.”
200,000+
Patients
7721
Symptoms
10
Interactive treatment report from PLM can suggest usage patterns and effectiveness
rates for drugs across different purposes.

11. Interactive Treatment Report
11

12. Study Objective
To examine the illustrative cases of Amitriptyline and Modafinil –
drugs commonly used off-label across five PLM communities.
“Modafinil (available since 1998) is a
wakefulness-promoting agent for approved
purposes related to sleep disorders, including
narcolepsy, shift-work disorder, and obstructive
sleep apnea.”
“Antidepressants like Amitriptyline (available since
1961) with one FDA approved purpose (depression)
have been highlighted as an important target for
future research into off-label drug use. (14 clinical
trials involving amitriptyline on clinicaltrials.gov)”
12

13. Methods
 Post hoc analysis of the prevalence of on-label versus off-label
use, dosing, and perceived effectiveness and side effects for the
medications.
 Two drugs: Amitriptyline and Modafinil
 Five condition based PLM communities with 53,928 members:
MS (Multiple sclerosis), fibromyalgia/chronic fatigue syndrome,
ALS (Amyotrophic lateral sclerosis), mood disorders (depression,
bipolar disorder, and anxiety disorders), and Parkinson's disease
13

14. Results: Modafinil
 1948 treatment histories: 1316 described current treatment at the time of
analysis
 2% (1316/53928) of all members reported currently taking modafinil.
 Most prevalent use in MS community- 1565 reports for 17,820 members
(6%)
 Purposes were reported in 1755 of 1948 (90%) modafinil treatment
histories.
 Only 34 of 1755 (less than 1%) of members reported taking modafinil for an
approved purpose (narcolepsy and excessive daytime sleepiness resulting
from sleep apnea)
 Modafinil was most commonly used to treat purposes that fall within
“general disorders and administration site conditions” (1277/1755, 73%)
followed by “nervous system disorders” (415/1755, 24%).
 726 treatment evaluations written about modafinil at time of analysis and
383 side-effect reports.
 Most users (532/726, 72%) rated the effectiveness of modafinil in the
highest response categories: either “major” or “moderate”
14

15. Results: Amitriptyline
 1394 treatment reports
 2% (865/53928) of all members reported currently taking amitriptyline.
 Most prevalent use in ALS community- 228 of 4060 (6%) reported earlier use of
the drug, 178 of the 4060 (4%) currently taking the drug. Most common
purpose in ALS was treating excess saliva (88 of 220 -40%).
 Purposes were reported in 1197 of the 1394 (86%) amitriptyline treatment
reports.
 104 of 1197 reports (9%), patients reported taking amitriptyline for the
approved use of depression
 Most commonly, patients reported taking it for insomnia and other sleep
problems (321/1197, 27%) or pain (197/1197, 17%)
 270 side-effect reports, most common, feeling sleepy, dry mouth & weight gain.
 Most users (532/726, 72%) rated the effectiveness of modafinil in the highest
response categories: either “major” or “moderate”
 The effectiveness ratings for off-label purposes were higher than for
depression: 28 of the 70 (40%) respondents taking it for the prescribed purpose
of depression rated it as having either a major or moderate efficacy in
comparison to 273 of 520 (52%) taking it for off-label uses
15

16. Results
 Only less than 1% of nearly 2000 patients taking modafinil
and 9% of nearly 1400 patients taking amitriptyline reported
taking each drug for purposes approved by the FDA.
 The effectiveness for off-label uses was reported to be as
either higher than or comparable to approved
indications.
 Many patients used some of the most common side
effects reported for amitriptyline, including sleepiness, as
their purpose for taking the drug, such as the treatment
of insomnia.
16

17. Conclusions
 Online communities like PatientsLikeMe, offer a unique real-time
approach to understand utilization and performance of treatments
across many conditions.
 Offer ability to reach a large population of users at relatively little
marginal cost.
 Gain experiences directly from patients: researchers can elicit new
types of data not recorded systematically through routine clinical
practice.
 New source of evidence about secondary uses and potentially
identify targets for treatments to be studied systematically in
traditional efficacy trials.
 Internet offers a unique vehicle to collect vast quantities of data far
more effectively than traditional studies permit.
17

18. Challenges & Limitations
 Information Bias: incomplete information reported by
patients.
 Lower-level of Confidence in the data reported
 Unreported comorbidity and confounding (socio-economic
status and health literacy).
 Misrepresentation of identity: false identities
18

19. Developments @
Tool for showing list of trials from ClinicalTrials.gov to the PLM community
19

20. References
 Frost, J, Okun, S, Vaughan, T, Heywood, J, Wicks, P. Patientreported outcomes as a source of evidence in off-label
prescribing: analysis of data from PatientsLikeMe. Journal of
medical Internet research 13, 1 (2011), e6.
 Janssens, A Cecile JW, and Peter Kraft. "Research conducted
using data obtained through online communities: ethical
implications of methodological limitations." PLoS medicine
9.10 (2012): e1001328.
 All images have been searched and downloaded from Google
Images
 http://www.patientslikeme.com/
 https://www.openresearchexchange.com/
 http://www.patientslikeme.com/clinical_trials
20

21. Discussion
21