Real-World Evidence: A Better Life Journey for Pharmas, Payers and Patients
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
• Cognizant 20-20 Insights
Real-World Evidence: A Better Life Journey
for Pharmas, Payers and Patients
Using RWE, the entire healthcare ecosystem can use actual
health outcomes to better assess the value of drug treatments
and related services.
Executive Summary Defining RCTs
Randomized control trials (RCTs), the established RCTs as a mechanism for evaluating the efficacy
way for measuring the safety and efficacy of of the drug cannot be replaced (as controlled
drugs, are increasingly being challenged by payers environments are needed to separate the impact
and healthcare providers. These players in the of medical interventions). Two measures are
wellness ecosystem are asking for real-life data to important to define:
validate whether new drugs provide similar safety
and efficacy as indicated by RCT results. • Efficacy is the extent to which an intervention
does more good than harm under ideal circum-
With RCTs, the results are obtained in a highly stances.
controlled environment, over a small set of
patient population, over a short period of time,
• Effectiveness is the extent to which an inter-
vention does more good than harm when
across a handful of highly dispersed clinical
provided under the usual circumstances of
trial centers. Questions are being asked by the
healthcare practice.
Federal Drug Administration (FDA) regarding
the sole dependence on RCTs and the lack of However, to evaluate the cost-efficiency of a drug
real-life supporting evidence by pharmaceuticals in a real-world environment and gauge its impact
companies on the efficacy and effectiveness of on improving the quality of healthcare, RCTs
new drug treatments. This is driving an increase need to be supplemented or followed up with the
in drug recalls, including those of blockbuster comparatively new standard, called real-world
drugs such as Xigris by Eli Lilly, Avandia in 2010 evidence (RWE).
(in Europe, only) by GSK, etc.
When it comes to pharmaceuticals, all wellness
This white paper details our point of view on ways stakeholders today require “evidence.” Patients
the pharma industry can apply real-life data to are looking for a better end result with their
more effectively gauge drug treatment effective- treatment. Providers are looking for data-oriented
ness and efficacy, enabling the entire ecosystem proof that the prescribed drug helps to optimize
of providers, payers and patients to deliver patient treatment, and brings added cost-effi-
“wellness” to all constituents. ciency and better profit margins. Payers (both
cognizant 20-20 insights | april 2015
RCT vs. RWE
Randomized Control Trial Real-World Evidence
Patients are randomized to the Therapy or medications to
treatments; physicians’ patients are determined by
and patients’ choices are doctors’ choices as per the
not considered for selection standard practice.
of the treatment.
Non-adherent patients
Non-adherent can switch the treatment
patients are taken and in such a case are
out of the analysis. likely to remain included.
Experiment is based on Contains heterogeneous patient
an artificially created population reflecting realistic
homogeneous treatment group. scenario. The study is likely to
The purpose here is to establish indicate the effectiveness of the
the efficacy of the medication/therapy. drug/therapy under various conditions.
Figure 1
government and private) are asking providers market access compared with the competition
and manufacturers to prove and promote the through “comparative effectiveness research.”
benefits that they will reimburse for in their However, it is not limited to this application. The
healthcare systems. Regulators, from the overall following analytics applications typically depend
public health and well-being perspective, are also on real-world data.
looking for evidence in a real-world environment.
With all these pressures, pharma manufacturers • Comparative effectiveness research/optimal
are forced to think “evidence” — and the time has treatment algorithm: Compares alterna-
come to think beyond the controlled environment tive treatment regimens, drugs and dosage
of clinical trials. on multiple criteria — cure, adverse effects,
cost, etc. — to arrive at the best treatment
RWE vs. RCT option for particular diseases for particular
But what exactly is real-world evidence? It uses patient profiles defined by demography, family
observational data to generate insight, foresight disease profile and comorbid conditions.
and predictive findings on diseases, products and While historical data mining is important here,
patient populations. Figure 1 depicts the differ- proactive strategy on patient targeting and
ences between RCT and RWE. pharmacovigilance via predictive analytics is
gaining popularity.
RWE Data
So what is real-world data? It is “observational
• Patient adherence study: A related area of
importance is the patient adherence study.
data” — data that involves the information related While for a manufacturer or provider non-
to a patient’s treatment. Typically, this consists of adherence means a lower observed effective-
four revealing data types that have overlapping ness compared with what is expected, for payer
and distinct characteristics — patient claims data and regulator it means increased healthcare
(available typically with payer), patient registries costs. Patient segmentation based on their
(available with provider and payer), electronic chances of becoming non-adherent may help
health records (EHR)/electronic medical records design a targeted educational program. If the
(EMR) and Web/social data. Figure 2 offers a non-adherence reason, as suggested by data
deeper explanation. analytics, is hidden in the drug or treatment
methods itself (i.e., twice-a-day vs. once-a-day
Application of RWE Data
treaments, pills vs. injectibles, invasive vs. lapa-
The straightforward and most imminent applica-
roscopic, etc.), it may also suggest product or
tion of real-world data is defending or improving
treatment innovations that may be needed.
cognizant 20-20 insights 2
The Four Categories of Real-World Data
Patient Claims Patient/Medical Registries EHR/EMR Social Data
• Hospital claims: • Patient registry refers to a collection • The electronic medical • This relates to patient
episode-level of patient/diseases/therapy-related- record is a patient-level interaction on diseases,
information. information collected through the electronic record of health treatment experiences
observational study method of information collected from and side effects. Social
• Provider claims:
patients, physicians and a single provider practice. networking sites such
procedure-level
laboratory tests. An electronic health record as Facebook and Twitter
information.
is a similar concept, but have pages specific to
• Registries are focused on target
• Prescription claims: it goes beyond a single diseases and medications.
populations and are designed to
prescription-level provider practice and data Sites such as Medhelp,
fulfill specific purposes defined a
information. are generally collected from PatientsLikeMe,
priori. For example, they include:
multiple healthcare practices CureTogether, Diabetic
product registries (patients exposed
and hence would follow Connect, Disaboom also
to a particular drug or medical
nationally recognized and allow patients to form
device), health services registries
standardized practice. communities, groups and
(patients with common procedure/
discuss experiences.
clinical intervention/hospitalization) • EHRs are not focused on any
or disease registries (patients with particular product, healthcare
similar diagnosis). services or disease-based
target population.
Data Elements
• Patient claim • Patient reported data: demographic • Patient demographic • Unstructured texts,
information. information, patient reported information. dialogues, sentiments as
outcomes (PROs). expressed by the patients
• Patient • Patient and family disease
or their kin (in cases where
demographic • Clinician reported data: diagnosis, history.
the patient is managed by
information. treatment/drug prescribed,
• Patient physical report. his or her kin).
laboratory/clinical test suggested,
• Consultation details.
follow-up treatment physician rating • Consultation details.
• Hospitalization of effectiveness. • Hospitalization and
details along with
• Laboratory: diagnostic/clinical discharge details.
cost of treatment.
test results. • Patient operative report.
• Diagnoses.
• Procedures/drug
names with doses
and days supplied.
Figure 2
• Outcome-based pricing/contract: Payers in • Clinical research feasibility: Real-world data
both government and private sectors are chal- can evaluate the feasibility of the RCT protocol
lenged by increasing healthcare costs. To coun- by judging whether there actually will be
ter this challenge, they are overly cautious enough patients to be recruited to participate
regarding how they spend constrained budgets in a clinical trial or, from a commercial perspec-
and are seeking return on investment justifica- tive, whether a study can really reflect the
tion in advance of making any outlays. Many are conditions of the target population if the pool
establishing their own therapeutic guidelines is limited. RWE can also help to determine the
over and above what is required by regulation. potential investigator and the potential site,
To land in the right plan and price category, based on the patient composition.
pharma companies need to provide payers with
supporting evidence based on real-life data to RWE-Led Pharma/Payer Partnership
document drug treatment efficiency and cost- As the need for RWE is mutual, manufacturers
effectiveness. and payers are already strengthening their bonds.
cognizant 20-20 insights 3For example, Healthcore, Wellpoint’s analytics is the validation of such predictive accuracy.
division, has been collaborating with AstraZen- Partitioned data sets are used to accomplish
eca since 2011. The objective of the alliance is this. Some of the examples are neural networks,
for each partner to help the other to more effec- regression trees, cluster analyses, support
tively integrate data from disparate sources to vector machines and random forests.
understand evidence-based outcomes. In 2011,
Humana entered into a five-year partnership With the variety of data sources and structures, it
with Pfizer to improve their healthcare delivery is clear that ML techniques are better suited for
systems by improving the quality, outcomes and RWE analysis. These techniques can play a very
costs of treatment. Humana also entered, in important role in RWE’s evolution and also in the
2013, into an agreement to sell its claims data to production of robust RWE in what is known as
Eli Lilly, which will use it to estimate and predict data settings with highly complex mechanisms.
patient outcomes, adherence and overall costs. ML techniques are better suited where insights
are to be gained using unstructured or semi-
New Techniques to Suit New Requirements structured data.
While the origins of RWE are diverse, achiev-
In the case of structured data, however, tradi-
ing RWE is not trivial. That’s because the big
tional approaches are still thought to be more
data sources are scattered and are composed of
effective, but only in statistical inferences and
elements in various structures. For instance, they
also while testing hypotheses. On the other hand,
originate from social media platforms, claims
while it comes to finding critical growth drivers,
data, medical records (electrical) and even pub-
predicting a trend and then determining non-lin-
lished literature. Sometimes it is organized and
earities and interactions, ML techniques are again
structured and sometimes it includes minimal
more effective.
text and other elements (graphics, videos, etc.).
This data can be everything and anything describ- Last, regarding mathematical/statistical tech-
ing in bits and bytes the journey of the patient. niques: At times, it may not be just one technique
but rather a combination of statistical method-
As a result, RWE is vast and varied in its type
ologies, often called “ensemble techniques,” that
and also in its sources. The challenge: If it is so
will provide better accuracy than individual tech-
diverse, how can data be meaningfully analyzed
niques for a particular problem in hand.
for insights that can inform strategic decision-
making? Two ways to glean insight from disparate
Looking Forward
data sources are the classical/traditional
approach and the machine learning approach. RWE provides the answer for understanding
whether any treatment/service/care delivery
• Classical/traditional: It is taken as an axiom- method would perform in the real world as it
atic truth that data in a randomized control trial would in controlled conditions. RWE, along with
is generated by a given probability distribution the data acquired from clinical trials, can provide
curve. In essence, an RCT involves the testing a true picture of what is actually happening among
of different causes and effects. The model used patients. This data can be used to build a better and
is influenced by parameter significance. If a more complete understanding of the diseases and
parameter is not significant, then that particular their patterns, thus providing better healthcare.
factor is removed from the model. Bayes Rule is
one such example as are duration models, prin- With more public databases available, there is
ciple components and many others. a need to use the best analytical approach to
address critical questions. The use of modern
• Machine learning: This is very different from
techniques has already begun. Once we have
the previous method. In this approach, algorith-
empirically-sound results, extra care must be
mic models are used. The data mechanism is
given to project efficacy and effectiveness on the
treated as unknown. Predictive accuracy is the
general population since there can be inherent
focus. And even if the model is not interpretable,
limitations in those results.
predictive accuracy is prioritized. The basis here
cognizant 20-20 insights 4Modern analytical techniques such as machine • Embrace product-agnostic solutions and
learning algorithms can bring radical positive overcome stakeholders’ reluctance to embrace
changes in healthcare, but high-end analytics will advanced methodologies.
not work unless pharma companies:
Once all these hurdles are crossed, there is little
• Make continuous investment in the underlying that can hold back the true power of real-world
tools and processes. data.
• Reach a common objective on which all stake-
holders agree.
References
• “Patient registries: a key step to quality improvement,” ACP Observer, September 2005,
American College of Physicians.
• Registries for Evaluating Patient Outcomes: A User’s Guide, 2nd edition, Gliklich RE, Dreyer NA,
editors, Rockville (MD): Agency for Healthcare Research and Quality (U.S.), September 2010.
• Understanding Health Outcomes and Pharmacoeconomics, Geroge E Mackinnon III.
About the Authors
Yashajit Saha is a Director within Cognizant Analytics Practice specializing in life sciences and healthcare
analytics. He has 15-plus years of analytics experience working with global pharmaceuticals, healthcare,
and banking and financial services companies. He can be reached at Yashajit.Saha@cognizant.com.
Dinesh Kumar Pateria is a Manager within Cognizant Analytics Practice. Focused on life sciences, Dinesh
has nine-plus years of experience in the analytics space with demonstrated expertise across a multi-
plicity of statistical techniques and models. Dinesh holds a Ph.D. in statistics from Indian Agricultural
Research Institute (IARI). He can be reached at DineshKumar.Pateria@cognizant.com.
About Cognizant Analytics
Within Cognizant, as part of the social-mobile-analytics-cloud (SMAC) stack of businesses under our
emerging business accelerator (EBA), the Cognizant Analytics unit is a distinguished, broad-based market
leader in analytics. It differentiates itself by focusing on topical, actionable, analytics-based solutions
coupled with our consulting approach, IP-based nonlinear platforms, solution accelerators and a deeply
entrenched customer-centric engagement model. The unit is dedicated to bringing insights and foresights
to a multitude of industry verticals/domains/functions across the entire business spectrum. We are a
consulting-led analytics organization that combines deep domain knowledge, rich analytical expertise
and cutting-edge technology to bring innovation to our multifunctional and multinational clients; deliver
virtualized, advanced integrated analytics across the value chain; and create value through innovative and
agile business delivery models. www.cognizant.com/enterpriseanalytics.
About Cognizant
Cognizant (NASDAQ: CTSH) is a leading provider of information technology, consulting, and business
process outsourcing services, dedicated to helping the world’s leading companies build stronger business-
es. Headquartered in Teaneck, New Jersey (U.S.), Cognizant combines a passion for client satisfaction,
technology innovation, deep industry and business process expertise, and a global, collaborative work-
force that embodies the future of work. With over 75 development and delivery centers worldwide and
approximately 211,500 employees as of December 31, 2014, Cognizant is a member of the NASDAQ-100,
the S&P 500, the Forbes Global 2000, and the Fortune 500 and is ranked among the top performing and
fastest growing companies in the world. Visit us online at www.cognizant.com or follow us on Twitter: Cognizant.
World Headquarters European Headquarters India Operations Headquarters
500 Frank W. Burr Blvd. 1 Kingdom Street #5/535, Old Mahabalipuram Road
Teaneck, NJ 07666 USA Paddington Central Okkiyam Pettai, Thoraipakkam
Phone: +1 201 801 0233 London W2 6BD Chennai, 600 096 India
Fax: +1 201 801 0243 Phone: +44 (0) 20 7297 7600 Phone: +91 (0) 44 4209 6000
Toll Free: +1 888 937 3277 Fax: +44 (0) 20 7121 0102 Fax: +91 (0) 44 4209 6060
Email: inquiry@cognizant.com Email: infouk@cognizant.com Email: inquiryindia@cognizant.com
© Copyright 2015, Cognizant. All rights reserved. No part of this document may be reproduced, stored in a retrieval system, transmitted in any form or by any
means, electronic, mechanical, photocopying, recording, or otherwise, without the express written permission from Cognizant. The information contained herein is
subject to change without notice. All other trademarks mentioned herein are the property of their respective owners.You can also read
