2. BASIC CHALLENGES IN ADOPTING PREDICTIVE
ANALYTICS
Business Challenges
Data Challenges
Infrastructure Challenges
Recruiting Challenges
3. 39%
45%
60%
Business Challenges
Money Issues
Staff with sufficient Analytical Skills
Insufficient Skills
45%
40%
36%
51%
Data Challenges
Data Integration Issue
Managing Data Quality
Larger Volume of Data Generated
Mentioning Data Integration Issue By Larger
Companies
4. 44%
41%
23%
Recruiting Challenges
Staffs in Predictive Model Management
BI Technology Development Skills
Third Party Consultants
26%
10%
37%
37%
31%
Infrastructure Challenges
Infrastructure Manageability
Fully Meeting The Requirement
Executive Perspective on Meeting Analytics Technology Need
Custmer Facing Employees Less Satisfied
Dissatisfied Data Scientists and Advanced Analytics Staffers
6. 1. Data volume –
• According to estimates, the volume of business data worldwide, across all companies, doubles
every 1.2 years. Insofar these volumes are an indication of change predictive models need to keep
up or they can risk becoming obsolete.
• If a model requires too much data to find a pattern, it may be too complex to provide reliable
predictions.
2. Data quality –
• Data needs to be useful in order to produce the desired results.
• Must be automatically analyzed and manipulated in order for it to be safe for vigorous modeling.
Predictive analytics solutions need to address two key modeling objectives in order to be successful –
they need to balance both accuracy and stability in order to yield high data quality.
3. Model complexity –
• Data miners need to keep their predictive models as simple as possible.
• Businesses need to take a top-down modeling approach to achieve the best results – simpler models
are less prone to changes in the data and are more robust.
• In the long run, they will prove to be much more cost effective and quicker to implement
7. 4. Model usability –
• Businesses need to be able to use models in order for them to be effective. Otherwise
data mining is just an expensive hobby.
• The end goal, the desired improvement in outcomes, must be clear in one’s mind when
developing models.
• This goes back to the model factory, versus laboratory.
• Businesses cannot afford scarce resources to do anything but delivering actional
intelligence.
• IDC estimates that by 2020,business transactions on the internet, business-to-business
and business-to-consumer, will reach 450 billion per day. Those
transactions demand better decisions and many of those decisions will require high
quality predictions.
8. 1. What’s the challenge?
• The ability to quantify the value of available information within the context of an organization, a
department or business function. At the outset, issues range from working out how to derive benefits
from information and in which context such data might be relevant.
• Prioritize those activities, as views may differ sharply between what IT considers important versus
what the rest of the business does.
• Challenge of determining how to perform and operationalize analytics, while taking the impact on
processes and behaviors into account.
2. Why now?
• Organizations now embrace data as a fourth factor of production, alongside capital, people and
materials. They use it to help sharpen their business performance by differentiating their offerings,
uncovering new opportunities and minimizing their risk exposure.
• Predictive analytics is not brand new, but the technologies that help firms make sense of their data
have only recently become available. This is allowing firms to uncover and exploit patterns in their
historical data, identifying both risks and opportunities ahead.
• Businesses can use data to look forward, rather than at past performance.
• Leading organizations increasingly recognize that predictive analytics can deliver more than just
customer insight; it can also have a positive impact on compliance, security, fraud detection and risk
management
DEEP DIVE:
9. 3. How does this affect you?
• Many businesses report a disconnect between their desire to capitalize on data and their ability to do so. It becomes even more imperative
for business and IT to develop a joint model and terminology for valuing information, which is directly linked to the organization’s key
performance indicators.
• Efficient analytics-enabled business processing can help boost revenues, reduce risks and increase agility. But getting this right often
demands that traditional IT and operational roles, structures and culture adapt to a new way of working, through the introduction of
specialist positions, such as data scientists.
• The adoption of analytics also brings with it new risks. Traditional levels of comfort around data quality, privacy, intellectual property and
reputation management must evolve.
• In an era of consumerization , those organizations able to monitor and predict customer behavior and preferences closely, without crossing
the line on privacy, can gain significant advantage.
4. What’s the fix?
Bringing technology-driven analytics to bear on a project involves several key steps:
• Understand the problem and address it in a way that it becomes clear which insights need to be discovered through predictive analytics.
• Collect the information needed to tackle the problem. This demands an analysis of which data is most needed, what is already available
and where any key gaps lie, along with an assessment of data quality and a sense of where missing data might be sourced.
• Perform the analytics, using mathematical algorithms to help uncover patterns within the data. These findings need to be translated back
to the business problem to help interpret the outcomes in the most useful context.
• Act on the findings — even if they imply a major shift — by adapting processes and behaviors to capitalize fully on the transformative
potential of predictive analytics.
• As such steps become embedded , one should develop a portfolio of analytics projects. With limited resources, the most important initiatives
need to be pushed to the front, charting the likely impacts a given project might have.
10. 5. What’s the bottom line?
At a high level, predictive analytics can help companies to:
• Move from a retroactive and intuitive decision-making process to a proactive data-driven one
• Build models that more closely predict future real-world scenarios and their related problems and
opportunities
• Uncover hidden patterns and relationships in the firm’s data
• More specifically, information-led companies will be able to sharpen their competitive edge.
There are numerous examples of what this might deliver:
• Attracting more valuable and loyal customers
• Charging prices closer to the market rate
• Ensuring more focused and relevant marketing campaigns
• Running more-efficient and less-risky supply chains
• Ensuring the best product or service quality levels
• Ensuring highly individualized customer service
• Guaranteeing a deep understanding of how process performance drives financial performance
11. ISSUES EG. 1 : ROTATING MACHINERY
MAINTENANCE
Involves several inter-related disciplines
Not adequately addressing predictive maintenance
12. 1 ) Involves several inter-related disciplines:
• data acquisition,
• database technology,
• data mining, and
• visualization.
2 ) Not adequately addressing predictive maintenance -
which is hard to do for the toolbox providers. As seen in the diagram, they are
only a small, but critical piece of the overall puzzle
• Even within the red box, there is only so much a toolbox developer can address.
This is because they want to keep their software as generally applicable as
possible, whereas the end user applications tend to be very segmented.
• The point is that for a toolbox provider, developing these required
customizations may not be very profitable, and as a result this critical part of
the application goes undeveloped.
• When this happens, collecting evidence for financial justification (which is
really the crux of the matter for increased adoption) becomes
impossible. Custom analytics and reporting development will therefore be a
big part of effective integration of these new technologies.
13. EG 2 : HEALTH CARE
Lack of participation of data scientists
Analytics
Execution Difficulties
Incorporation
Interdisciplinary design teams
Addressing the needs of clinicians within their workflow
Research opportunities
Privacy
Factors for Readmission
Non updated Data
Technology adoption
14. 1. Analytics:
There are many challenges in data collection and management, as arising from incomplete,
heterogeneous, incorrect, or inconsistent data.
2. Privacy :
Becomes critical as organizations collaborate and share data to enable an integrated EMR.
• There is need to preserve privacy while sharing knowledge and to retain most of the utility of
aggregated data for purposes of predictive modeling and population studies, while adhering to
privacy constraints.
• The privacy-utility tradeoff becomes even more critical to understand when data from multiple
sources is integrated, potentially leading to unintended consequences or leaks.
15. 3. Technology adoption: at two levels, namely –
a) Adoption –
• The first is due to a lack of understanding of its usefulness, and lack of alignment with strategic
objectives.
• Clinicians are incredibly busy, and there have been few incentives for them to explore how analytics
can make them more effective and efficient.
• Privacy, and sensitivity issues associated with patient information also limit the data access and
sharing capabilities.
b) System usage -
Clinicians face a variety of challenges while interfacing with electronic medical records on a day-
to-day basis. Some usability challenges include –
• The lack of support for information integration,
• Data heterogeneity,
• Too many options for each field,
• Lack of visualization of summary information,
• Sharing information in a privacy-safe manner, etc.
• clinicians often resort to unstructured notes due to the lack of time and system usability. This
leads to natural language processing challenges and to data errors.
• The difficulty in primary usage of EMRs leads to a low confidence in their value, which results
in reduced interest in their secondary, and potentially very powerful purpose, predictive
clinical analytics.
16. 4. Lack of participation of data scientists:
The acceptance and integration of skill sets by Data Scientists has been slow in healthcare as a whole, with the possible
exception of payers, although there is a growing realization of the need.
5. Research opportunities:
• Incorporating domain knowledge and real world evidence to address data quality issues will be crucial to help improve the
effectiveness of follow-up predictive modeling efforts.
• Feature selection techniques will be necessary, as most data in healthcare is of very high dimensions.
• Smart ensemble methods will play a crucial role in incorporating evidence from different sources.
• Finally, privacy-aware and knowledge-preserving data collaboration techniques are essential for successful collaboration and
for the integration of data from different sources.
• One of the most important challenges is in harmonizing data elements across data collection systems.
6. Addressing the needs of clinicians within their workflow:
• Solutions should focus on making systems smarter and easy for physicians to work with.
• For example, a system that summarizes key points in a text format and answers questions (like IBM's Watson or the clinical
"Siri") for clinicians as they look at patient records, will let physicians focus on patient care rather than data recording and
data analysis.
• Useability and effectiveness will drive the adoption of technology
17. 7. Interdisciplinary design teams:
It is important to build a more cohesive and collaborative interdisciplinary team comprising of
• Clinicians,
• Data scientists,
• Biostatisticians,
• Epidemiologists,
• Policy makers,
• Legal, etc., to design effective solutions and avoid dissatisfaction and lack of belief in the efficacy of such systems.
8. Incorporation - to incorporate these technologies into mission critical processes –
( to complete successfully every single step of the value chain ) -
• Data collection,
• Data storage,
• Data preparation,
• Predictive modeling,
• Validated analytic reporting,
• Providing decision support and
• Prescriptive tools to realize value.
18. 9. Execution Difficulties :
• Difficulty to connect to the right data at the right time,
• To deliver the right results to the right stakeholder within the actionable time interval where the right decision can make a difference,
• To incorporate the predictions and prescriptions into an effective automated process that implements the right decisions.
• It is an overworked IT department dealing with outdated and inadequate hardware and storage technologies, trying to manage the
“prevention of IT” given these other challenges.
• Sometimes there are challenges integrating diverse data sources like –
structured data in relational databases on premise,
information that needs to be accessed in the cloud or
from internet-based services,.
with unstructured textual information stored in distributed file systems MODELING PROCESS
To improve the accuracy of predictive modeling, developers may
take an approach known as “supervised learning,” in which the
outcome is known ahead of time and is used to “train” an
algorithm.
• But in healthcare, many important kinds of patient outcomes
are not captured as structured data. Without outcomes data
to train the algorithm, it’s difficult to apply a supervised
learning model.
19. • The data may be there, the technologies to do useful things with those data exist ,but the two cannot readily be connected. ( It is generally
acknowledged that data preparation consumes about 90% or more of the effort in analytic projects.)
10. Non updated Data (The health status of patients ) -
• After discharge may not be available unless patients fill out functional status surveys at specified intervals.
• Also, the follow-up on most patients after discharge or between office visits is limited or nonexistent. As a result, only the data generated in the
EHR during a visit or an episode of care may be available.
• Diagnoses, lab values, medications, and vital signs from these encounters appear in a data warehouse, but they don’t reflect the time period
between visits, which would show how the patient fared between visits or episodes.
• Even the episodic data are frequently not structured in the EHR, partly because some providers don’t enter them in the ubiquitous pull-down
check boxes.
• Paid claims data, always include diagnostic and treatment codes. Moreover, claims data show the services and prescriptions that patients
received from providers outside an organization or network.
But claims have a built-in lag time, so they’re not very good for predicting what might happen in the near future.
Furthermore, claims are not precise enough to describe in detail what has been done for the patient in various care settings.
• For most analytic purposes, organizations rely on a combination of clinical and claims data, if they have access to the latter.
BUT TO MAKE THE BEST USE OF PREDICTIVE ANALYTICS, HEALTHCARE ORGANIZATIONS MUST BUILD DATA WAREHOUSES CAPABLE
OF
AGGREGATING,
NORMALIZING AND
CLEANING UP THIS DATA AND
PRESENTING IT IN A FORMAT THAT IS EASY TO USE IN REPORT GENERATION.
20. 11. Factors for Readmission – ( Solution is action grounded in
clinical experience & not in Analytics )
• Low patient literacy,
• Poor understanding of discharge instructions,
• Failure or inability to make an appointment with a primary
care doctor,
• And lack of communication between inpatient and
ambulatory providers.
21. PREDICTIVE SOLUTIONS TO THE CHALLENGES
Companies must first be able to-
• Find technologies that deliver good decision support, and
• Know how to use them effectively.
• Training staff with advanced analytics skills.
• Finally, IT executives need to address IT shortcomings in -
infrastructure
data integration and
quality management
for advanced BI techniques to gain mainstream acceptance.