white paper

Healthcare
Predictive Analytics

Predictive Analytics
In Healthcare
A data-driven approach to transforming care delivery

“The powerhouse organizations of the Internet

Author
Andrew Bartley era, which include Google and Amazon…
Senior Health and Life
Sciences Solution Architect, have business models that hinge on predictive
Intel Corporation models based on machine learning 1.”
The global healthcare industry is transforming, driven by increasing costs and an
aging population. By 2025 the global population is estimated to be 1 billion, with
500 million of those people over the age of 50. Multiple sources including the
World Health Organization and UN predict that in that same year over 70 percent
of illnesses will be chronic conditions. Place that against a backdrop of increasing
global healthcare spend that is expected to reach USD 18.3 trillion by 2030 and
change is a clear imperative2.
These factors are forcing a global healthcare reform to move from volume-based
payment models to outcome-, or value-based models. This shift requires providers
to significantly change the way they operate. They must focus on the delivery of
personalized medicine that is tailored for each patient, while also deepening their
understanding of population health more broadly to better identify and respond
to patterns. Both objectives can only be met by applying advanced analytics to a
healthcare provider’s data.
Table of Contents
A data-driven approach to
transforming care delivery. . . . . . . . 1
Digital transformation means
more data. . . . . . . . . . . . . . . . . . . . . . 2
External pressures to meet the
Quadruple Aims. . . . . . . . . . . . . . . . 2
Government reform and policies .
drive change . . . . . . . . . . . . . . . . . . . 2
Advanced analytics is closer than
you think . . . . . . . . . . . . . . . . . . . . . . . . 2
Building a predictive analytics
strategy: where to start. . . . . . . . . . . 4
People. . . . . . . . . . . . . . . . . . . . . . . . . 4
Process. . . . . . . . . . . . . . . . . . . . . . . . 4
Technology . . . . . . . . . . . . . . . . . . . . 6
Continue your journey . . . . . . . . . . . . 7
2 White Paper | Healthcare Predictive Analytics

Three global trends in healthcare have the potential to 3. Government reform and policies drive change
impact this transformation.
Government policy can influence healthcare providers’ data
1. Digital transformation means more data strategy in various ways. In the US, for example, policies
The digitization of healthcare is resulting in the creation such as the CMS EHR Incentive Program8 and the HiTech Act9
of massive new data sets. One of the most significant have increased investment in healthcare since 2011. This has
examples of this is the Electronic Medical Record (EMR). accelerated the adoption of digital solutions, which in turn
Over 85 percent4 of healthcare organizations in the US has led to a boom in the volumes of data a typical healthcare
have now adopted an EMR system5. Health claims data, organization holds. Six years on, many are reaching the
radiology images and lab results are other common data point at which they have a large enough pool of current and
sets. In the future ‘omics (i.e. genomics) data is expected to historic data to make really meaningful analytics possible.
grow significantly as precision medicine gets increasingly Similar trends can be seen or expected in Europe, where
integrated into standard clinical workflows. government initiatives in the UK, France and Sweden10 are
also spurring investment in this area.
Meanwhile, around the world, governments are placing
"Digitization has allowed us to have a more pressure on providers to demonstrate and track
more meaningful conversation compliance with data protection and other regulations,
often across lengthy processes involving multiple people
around how to use predictive analytics and departments. Rising pressure to curb spending and the
need for improved outcomes has pushed responsibility for
to improve patient outcomes3." managing the risk of poor outcomes and high costs from the
payer to the provider.
Outside of the clinical setting patients generate a significant Advanced analytics is closer than you think
amount of quasi-health data through the use of wearable
health and fitness trackers and health applications. The The ability to meet many of these challenges now forms
proliferation of health-related Internet of Things (IoT) the basis of key performance indicators (KPIs) for many
devices, as well as apps and other health-monitoring healthcare providers. Failure to respond can result in
technologies also generate a huge amount of data every day. competitive disadvantage, a negative impact on bottom line,
reputational damage, unhappy patients, or loss of staff.
Healthcare providers have an unprecedented opportunity to
incorporate these disparate data sources and structures into
new solutions. For example, to understand population health
trends, it is important to use external resources such as social Real-world results: Penn Medicine
media or publicly available government data, in addition to
Penn Medicine operates a network of healthcare
a hospital’s own records. However, 94 percent of hospitals6
facilities in Pennsylvania and southern New Jersey.
are not currently capturing enough information to carry out
Working with Intel, it has created a collaborative data
effective population health analytics.
science platform, based on the open source Trusted
2. External pressures to meet the Quadruple Aims Analytics Platform* to help clinicians make faster,
smarter decisions based on large-scale clinical data
Regulators, staff, patients and peers within the industry and big data.
all have expectations of the healthcare providers that they
come into contact with. The Quadruple Aims7 that each In its first trials of the platform, called Penn Signals*,
organization must commit to addressing are: the organization developed algorithms to help predict
and prevent two of the most common and costly issues
• Improving access to resources and services for patients for hospitals: sepsis and heart failure. It was able to
and clinicians, enabling them to find what they need, correctly identify about 85 percent of sepsis cases (up
when they need it, securely and quickly from 50 percent), and to make these identifications
• Reducing operating costs and ensuring that limited as much as 30 hours before the onset of septic shock
financial resources can be channeled towards the most (as opposed to just two hours prior, using traditional
value-adding priorities while optimizing the bottom line identification methods). This meant clinicians could
deliver treatment sooner, speeding time to recovery for
• Improving treatment outcomes by helping physicians the patient and saving resources for the hospital.
to make diagnoses faster, deliver the most personalized
treatment possible for the patient, and reducing average Meanwhile, Penn Medicine estimates that using
length of stay traditional methods of diagnosis, between 20 and 30
percent of heart failure patients had not been properly
• Optimizing provider satisfaction by ensuring that busy identified. With the predictive model it was able to
staff, working in a stressful environment, are able to identify these patients and give them the education to
focus on delivering patient care and achieving positive self-manage their condition successfully. As a result,
outcomes rather than grappling with complex tools and readmission rates dropped.11
workflows
3 White Paper | Healthcare Predictive Analytics

Figure 1: Stages of analytics maturity and associated technologies

Analytics has already proven helpful here. A recent Intel- analytics uses a variety of statistical and machine learning
commissioned report12 from the International Institute for methods and are honed over time with the addition of new
Analytics found that the highest performers in analytics in data. An example of predictive analytics would be to use
healthcare are using it to help improve patient engagement, historical data from the hospital’s records along with external
population health, quality of care and business operations – sources such as weather forecasts and social media to
areas that map closely to the Quadruple Aims. predict peaks in ER admissions for the purpose of improving
staffing levels
However, for those not already proficient in analytics,
knowing where to start can be tricky. Advanced analytics Prescriptive analytics builds on predictive analytics by
techniques like machine learning and artificial intelligence including a single or set of recommended actions based
(AI) promise ground-breaking insights and the potential for on the prediction. Prescriptive solutions can employ data
great innovation, but they can also be complex and daunting. techniques like simulation and machine learning as well
as cognitive systems. Building on the previous example, a
The good news is that if you are doing any form of business
prescriptive solution would develop a prediction for peaks
intelligence (BI) today, then you are already on the path to
in ER admissions, and then recommend adjustments to
advanced analytics. From there, it is simply a question of
workflow or prioritization to reduce length of stay and ensure
evolution and steady progress.
staff can see as many patients as possible.
Analytics capabilities develop in stages (see Figure 1), each These recommendations can be based on discrete business
building on the one before and adding the ability to handle processes, or in the case of cognitive systems can be based
more complex data and draw more detailed insights in a on previous activities that produced an optimal result.
more timely fashion.
Cognitive analytics uses artificial intelligence (AI)
The first two stages constitute traditional analytics technologies such as machine learning, deep learning and
methods. Descriptive analytics (or BI) reviews data about reasoning or logic systems. It automates decisions using
what has happened and answers questions about past human-like analysis or provides insights and suggestions to
actions. Diagnostic analytics takes this a step further by augments human decisions. This could help create a full view
adding insight into why past events happened as well. of a patient’s state of health, considering everything from
With these types of analytics, you can for instance review EMR records to CT scans and the patient’s fitness tracker
your admissions data from last month and identify that the
number of flu patients increased by 18 percent compared to
What is AI?
the previous month.
AI defines a class of techniques designed to enable
The more advanced stages of analytics bring in capabilities
machines to sense, reason, act, and adapt like humans
that can add significant value to strategic planning, such
do. There are several underlying technologies that
as forecasting, real-time insight and automated decision
make up AI which are shown in figure 2. Think of these
making.
as various tools that can be used to solve different
Predictive analytics analyzes historical data to predict future types of data problems. As your AI projects become
target events. This transition to forward-looking analytics more complex, you will likely need several of these
is an important crossover for an organization from both a technologies as part of an end-to-end solution.
technology and business process perspective. Predictive
4 White Paper | Healthcare Predictive Analytics

that is well recognized14. Complement this by contacting

other organizations, both healthcare and non-healthcare
to identify best practices. The technology and finance
industries are possible places to start.
• Which employees are currently responsible for any BI or
analytics? How will their roles be impacted and how can
you prepare them?
• How will you set expectations within the organization
for what this team is responsible for? Where and how do
handoffs between the data science and engineering teams
occur?
• What new skills will you need in the team in order to make
the most of your new analytics capabilities? Will you train
Figure 2: The technologies behind AI
existing staff or hire in new talent?
feed, and then make proactive, real-time recommendations • What is the career path within the organization? What
to the physician about possible treatment options or risk professional growth opportunities can you provide both
factors, and send alerts to the patient with suggestions on from within and outside of your organization?
how to improve their wellbeing.
• What management structure will you use? Will this
When starting out on your first advanced analytics team fall under IT (CIO), clinical (CMIO), or does it make
initiative, aim to get value from predictive analytics. sense to create a new branch with reporting up to senior
A focused business case with measurable ROI will help leadership?
ensure proper organizational alignment. You can then
move on to prescriptive and/or cognitive analytics • What structure will you use for assigning, tracking,
whenever and however is appropriate for your organization. and evaluating team KPIs? Should this be monthly or
Remember though, that the goal isn’t always full-scale quarterly, and who is on the leadership team that will
cognitive AI – in many cases other types of analytics are assign these objectives?
more suitable. Success lies in finding the right tools for
Process
your organization, not necessarily in making it to the ‘end’
of the maturity journey. In a busy environment like a hospital, clear processes are
essential. Patients move between departments, staff change
Building a predictive analytics strategy: shifts and all the while potentially life-altering decisions must
where to start be made and actioned at speed. There is no margin for error
or confusion. Therefore, any changes that have the potential
Form your predictive analytics roadmap by considering three to impact process must be planned carefully.
key areas: people, process and technology.
1. Data management and governance
People
Until relatively recently, healthcare organizations typically
Building a solid advanced analytics team can be a challenge. had a collection of best-of-breed applications to support
Competition for scarce talent is fierce. Healthcare has a
unique competitive advantage inherent in the purpose of the
work it does. Much like those of us working in and around Real-world results: Montefiore Health System
healthcare, there is purpose in knowing that the work you do Montefiore Health System runs a number of healthcare
contributes to saving lives. Using that advantage is one way facilities in the Bronx, New York. It has created a
to differentiate from other organizations that may be able Semantic Data Lake solution that pulls together data
to offer more compensation or other perks. Make sure there from multiple sources and in many different formats
is a clear tie to the work that the team completes and the so that it can be analyzed holistically. The data ranges
outcomes that it drives. from clinician data, demographic, environmental,
behavioral and wellness research findings and
population demographics to medical imaging files
"Data, and the capability to extract and doctors’ notes. Montefiore has worked with
useful knowledge from data, should Intel, Cloudera and Hadoop to apply predictive and
prescriptive analytics algorithms to its data, and to use
be regarded as a key strategic asset13." machine learning to optimize its insights over time.
In its first implementation of the solution, Montefiore
Just as important is the internal culture for this team. used it to identify patients at risk of death or in need
Practitioners of advanced analytics are problem solvers. of intubation within the following 48 hours, which is
They get energized by the opportunity to work with new the window of opportunity to complete a successful
data and solve new problems. Slotting them into roles that intervention. It achieved accurate prediction at a
require a high amount of routine work can be frustrating, rate of more than 70 percent , helping physicians
resulting in employees becoming disengaged, and increasing deliver treatment earlier to prevent fatal episodes or
costly attrition. There are several good resources about how respiratory failure.15
to build an analytics team – O’Reilly Media provides one
5 White Paper | Healthcare Predictive Analytics

Figure 3: Levels of executive engagement are much higher in the top 10 percent of analytics in healthcare initiatives
different use cases and workflows. Now, most have larger Commitment to invest must come from the top, and a study
EMR systems that pull much of this into one place. While the conducted by the International Institute for Analytics18 shows
vendors of these systems tend to take responsibility for data that the top 10 percent of performers in healthcare analytics
governance and management within their own parameters, have a significantly higher level of executive support and
the evolution of analytics capabilities means healthcare engagement than other organizations (see figure 3). However,
providers now also want to bring in external data types which expectation of a positive ROI is not enough. Early selection
may not be so stringently managed. Properly collating and of analytics projects is critical to building a solid foundation
managing these disparate systems can be a challenge in of support. Analytics teams without proper guidance run the
itself. Add to that the fact that sentiment-based data such as risk of selection projects that are academically stimulating,
social media posts – while extremely valuable – is much more but not well positioned to drive business value. One of the
ambiguous and open to interpretation than an EMR database, best ways to ensure that a project is on the right track is
and the challenge becomes more complex. It’s important that framing it within a defined business need. This need should
you work with your IT department to set rules around how have a clear definition for success, ideally quantified by some
certain data should (and should not) be interpreted by an financial measure, and a clear path to implementation.
analytics system.
Data collection and delivery must also be carefully
controlled, with defined rules and processes to ensure that
"It’s necessary to have a multidisciplinary
as your data grows, it uses standardized characteristics to team, with clinical, analytics, data
maintain consistency and quality. If analyzing your data
management capabilities for the first time, it is recommended science, information technology, and
to use a recognized framework like the Data Management
Maturity (DMM) Model from the CMMI Institute16. Control your
behavior change skill sets available
data fidelity and quality by putting in the time up front to from start to finish17."
define decision models and data sets upon which the system
can learn and build.
Projects scope and strategy must also be considered. As
Finally, ensure your methodologies for conducting analytics
Prashant Natarajan states in his book, Demystifying Big Data
and applying insights are clear and accessible to your user
and Machine Learning For Healthcare, “many use cases,
base. By providing standard approaches, you can empower
while meeting the requirement of ROI and strategic impact,
users, minimize risk and maximize the value you glean from
have such large scope and complexity that success may be
your analytics investments.
put at risk through over-reach”. Meeting these requirements
2. Organizational alignment requires a significant amount of up front work. However the
Building a lasting advanced analytics capability requires tight long-term payoff, being able to show measurable benefit to
alignment to clinical and business stakeholders. Depending the organization, is well worth it.
on the scope of the advanced analytics program, significant Finally, ensure proper cross-functional representation within
investment may also need to be made in infrastructure to the analytics program. It is critical to engage subject matter
collect, store and process large data sets. Ensuring that experts (SMEs) to provide context to the business problem
this investment is available will require demonstrable ROI. being addressed. SMEs can help address changes in data like
6 White Paper | Healthcare Predictive Analytics

Figure 4: The analytics technology stack

when a data collection process evolves, or external factors optimization is key to extracting those game-changing
that may impact the way that model results are used. insights that will get your patients feeling better sooner or
let your busy nurses finish their shifts on time. However, it
Some questions to consider here are:
does mean that your data management and clinical teams
• Who are the executives and key leaders that need to may need to change the way they work and collaborate,
champion the predictive analytics program? What incorporating more time for feedback and iteration to ensure
business results will speak to them? long-term success.
• What are our organization’s stated business goals and • Which aspects of organizational culture will be affected
strategy over the short and medium term? How do and how will you ensure your innovations work with,
proposed uses cases align to those goals? rather than against, it? James Merlino, MD’s book Service
Fanatics describes the significant effort that Cleveland
• Who are the cross-functional stakeholders that will be Clinic undertook to change its internal culture to be more
impacted by the project? Who are the SMEs that should patient-centric. While a different goal, it shows how much
be engaged to provide context and understanding? work can go into influencing a hospital culture.
• When and how are successes shared? What are the key • Who will lead change management efforts, an internal
internal meetings that you can embed team members in to team or external consultant? How might this change over
ensure alignment with the business? time?
• What will the data science group be known for? What are • Who are the formal and informal leaders that can bring
standard deliverables that will define the team throughout support for change initiatives?
the organization?
• How will insights be shared with hospital managers,
3. Change management clinical staff and consultants? What would be the most
Change management is one of the greatest challenges useful format, channel and cadence for them?
that many organizations face as they begin to implement
advanced analytics. Moving from retrospective analysis to • How will you demonstrate the value of any changes to
predictive analysis, requires workflows to be adaptable, those they impact?
sometimes in real-time, based on new data. Clearly this Technology
presents technical considerations, but just as, if not more
important, are the considerations for cultural adoption of this When building an analytics environment, picture it as
new approach. forming four complementary layers (see figure 4), each
including a variety of technologies, depending on your own
Organizations must have a strategy for how to implement organizational needs, legacy systems and preferences, and
change across the business. As your data yields more the type of analytics you need.
insights, how will you adapt existing clinical workflows
to make the most of them? Predictive analytics projects The infrastructure layer: This is the foundation that will
are constantly evolving. Many follow the CRISP-DM enable you to acquire, store and protect your data, and
lifecycle process which lays out six phases from Business to run commercial and open source analytics. A typical
Understanding to Deployment19. This continuous infrastructure layer will support a number of different
7 White Paper | Healthcare Predictive Analytics

frameworks, databases and applications.

Real-world results: Sharp HealthCare
The data layer: Pre-analytics, this layer may have been
Based in San Diego, California, Sharp HealthCare
comprised mostly of internal, siloed databases. As today’s
runs a specialized rapid response team for medical
mix of data is more complex, distributed and unstructured,
emergencies that occur within its hospital. It has
this layer may take the form of a data lake (for example based
developed a predictive model intended to detect
on the Cloudera Enterprise Data Hub*) that pulls together
patient health decline and identify patients at risk for
multiple data streams – from EMR records to clinician’s
requiring the team’s help. It worked with Intel, Cloudera
handwritten notes to fitness tracker feeds – into one holistic
and ProKarma to build the solution, which analyzes
environment. Many organizations will have more than one
a wide range of data – including blood pressure,
data platform, such as a relational database and big data
temperature and pulse rate – from the hospital’s
platform based on technology like Hadoop* or NoSQL*
EMR system, and trains the algorithm over time using
side-by-side, to serve different analytics and reporting
machine learning.
requirements.
In a proof of concept (PoC), Sharp tested the model
The analytics layer: This is the layer that will be subject to
against historical data and found that the model was
the most change. There are a wide range of open source
80 percent accurate in predicting the likelihood of
and commercial solutions available, with new solutions
a rapid response team event within the next hour,
emerging regularly. Digital transformation is still new for
demonstrating the potential to drive real-time clinical
many organizations, so there are many opportunities to
interventions, improve outcomes and enhance resource
innovate around data analysis. The analytics layer should
utilization. 20
support many different types of algorithms. In particular for
predictive analytics, regression, classification, and clustering
algorithms are three of the more common techniques used. Continue your journey
Most analytics platforms will support these algorithm types
‘Analytics’ is not a once-and-done initiative – rather than a
along with many others to meet different needs.
single project, it’s an ongoing, evolving way of working that
Most organizations will have several technologies and/or should adapt with your organization as needs and priorities
solutions built into the analytics layer to support increasingly change. Start with a simple project that will deliver clear
complex data analytics pipelines. Due to the fact that value to the business in a short time frame, and use this initial
much healthcare data is unstructured, data analytics tools success to build support from business users and longer-
like a natural language processing (NLP) solution from term commitment from senior leadership.
Linguamatics* or Deep Learning for image detection can be
Once you have your foundations in place, you can add more
incorporated to analyze specific data sources. These point
analytics capabilities as your technology resources evolve
solutions will generally feed back into the core platform to
and your data sources become richer. As your clinicians and
allow for developing models across all of the relevant data.
other user groups become more familiar with the potential
The applications layer: This layer supports the visualization of analytics, they may also start to propose new use cases
of predictive model results. Open source or commercial themselves.
software provides tailored analytics capabilities that
In this paper we have focused on making the move from BI
support your specific use cases and workflows. Give
to predictive analytics, but there is plenty more potential for
careful consideration to how you will deliver the results of
development once you’ve done that – for example you can
predictive models back to clinical stakeholders. It is highly
look ahead to more automated possibilities that come with
recommended to engage a cross-functional team when
the next stages in the analytics journey, such as AI, deep
deciding the best way to visualize and deliver results. Some
learning and cognitive computing. Some organizations are
questions to ask here are:
already exploring such use cases.
• What data should be delivered in a pull vs. push method?
The key to all this is data. The data that sits in your
• Should results be integrated into an existing application organization today. Make sure your data and your analytics
like the EMR or a new application? strategy are aligned to meet tomorrow’s challenges by
considering these questions:
• What data needs to be displayed to a health worker? How
will you calibrate results to maximize signal vs. noise? 1. What are your data sources today and (how) are
they connected?
• How will interaction with the model results, either positive
or negative, be captured to improve the model over time? 2. What’s your strategy for collecting and managing
your data today and tomorrow?
These layers enable you to take advantage of your existing
data management systems as well as newly implemented 3. What information are you not capturing that you
technologies. Additional technologies like AI and should be?
performance or security solutions can be applied across
4. What will you need historical data for in two years’
the whole stack to accelerate insights and strengthen data
time and are you collecting it now (eg unstructured
protection. Solution providers with expertise in delivering
content like handwritten doctors’ notes, images etc)?
analytics solutions for healthcare organizations, like
ProKarma and NTT Data, can help you define the right mix of 5. What was the last question you were unable to
technologies for you. answer? What would have enabled you to answer it?
8 White Paper | Healthcare Predictive Analytics

Read on to learn more about how Intel technology can help

you create your transformational analytics environment:
Report: Driving Clinical and Operational Performance
Through Analytics
For more information about Intel's healthcare solutions,
and to review the most up to date content visit: intel.com/
healthcare

1 Professor Vasant Dhar, Stern School of Business, New York University, Predictive Analytics: The power to predict who will click, buy, lie, or die, Eric Siegel
2 All statistics in this paragraph are from: Healthcare Disrupted: Next Generation Business Models and Strategies, Jeff Felton and Anne O’Riordan
3 Vindell Washington, MD, National Coordinator for HIT in the US Department of Health and Human Services , Healthcare Executive Magazine, volume 31, Number 5, September/October 2016
4 https://www.cdc.gov/nchs/fastats/electronic-medical-records.htm
5 https://dashboard.healthit.gov/quickstats/quickstats.php
6 Healthcare Executive Magazine, volume 31, Number 5, September/October 2016
7 From Triple to Quadruple Aim: Care of the Patient Requires Care of the Provider, Thomas Bodenheimer, MD and Christine Sinsky, MD, http://www.annfammed.org/content/12/6/573.full
8 https://www.cms.gov/Regulations-and-Guidance/Legislation/EHRIncentivePrograms/index.html?redirect=/ehrincentiveprograms
9 https://www.healthit.gov/policy-researchers-implementers/health-it-legislation
10 France: http://blog.ihs.com/frances-2016-social-security-financing-act-a-figure-is-worth-a-thousand-words
UK: https://www.eiuperspectives.economist.com/sites/default/files/ValuebasedhealthcareUK.pdf
Sweden: http://ki.se/en/lime/value-based-health-care
11 https://www.intel.co.uk/content/www/uk/en/big-data/penn-medicine-trusted-analytics-platform.html
12 Driving Clinical and Operational Performance Through Analytics, International Institute of Analytics, August 2017
13 Data Science For Business: What you need to know about Data Mining and Data-Analytic Thinking, Foster Provost & Tom Fawcett
14 Building Data Science Teams, O’Reilly Media: http://www.oreilly.com/data/free/building-data-science-teams.csp
15 https://www.intel.co.uk/content/www/uk/en/healthcare-it/solutions/documents/montefiore-advance-patient-care-solution-brief.html
16 http://cmmiinstitute.com/data-management-maturity
17 Getting Buy-In for Predictive Analytics in Health Care, Meetali Kakad, MD; Ronen Rozenblum; David Westfall Bates, MD, https://hbr.org/2017/06/getting-buy-in-for-predictive-analytics-in-
health-care
18 Driving Clinical and Operational Performance Through Analytics, International Institute for Analytics, August 2017
19 Machine Learning for Predictive Data Analytics, John Kelleher, Brian Mac Namee, Aoife D’ Arcy
20 https://www.intel.com/content/www/us/en/healthcare-it/solutions/documents/using-machine-learning-and-emr-data-to-predict-patient-decline-case-study.html
Intel technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation. Performance varies depending on system
configuration. No computer system can be absolutely secure. Check with your system manufacturer or retailer or learn more at intel.com/content/www/us/en/architecture-and-technology/
quick-sync-video/quick-syncvideo-general.html
All information provided here is subject to change without notice. Contact your Intel representative to obtain the latest Intel product specifications and roadmaps.
Intel and the Intel logo are trademarks of Intel Corporation in the U.S. and/or other countries.
*Other names and brands may be claimed as the property of others.
© Intel Corporation Printed in USA 0917/FP/CAT/PDF Please Recycle 336536-001EN