This document discusses the use of predictive analytics in healthcare. It describes how predictive analytics uses data and statistics to analyze massive amounts of patient information to predict outcomes. This can help with readmissions, triage, emergency care, detecting patient decompensation, and adverse events. Challenges to implementing predictive analytics in healthcare electronically include testing models, oversight, data quality, and ensuring interoperability between systems. When done correctly, predictive analytics has the potential to improve patient health and lower healthcare costs.

1. UNCHEALTHCARESYSTEMUNCHEALTHCARESYSTEM
The Use of Predictive
Analytics in Health Care
Kathryn Morbitzer, PharmD
Wednesday, April 27th, 2016

2. 1
UNCHEALTHCARESYSTEM
The Use of Predictive Analytics in Health Care
• Data Analytics in Health Care
• Types of Data Analytics in Health Care
• Predictive Analytics
• Uses of Predictive Analytics in Health Care
• Implementing Electronic Health Care
Predictive Analytics
Agenda

3. 2
UNCHEALTHCARESYSTEM
Data Analytics in Health Care
In today’s data-driven age, healthcare is transitioning
from opinion-based decisions to informed decisions
based on data and analytics.
Analyzing the data reveals trends and knowledge
that may run contrary to our assumptions causing
a shift in ultimate decisions that in turn will better
serve both patients and healthcare enterprises.
Kolker K. Strata Big Data. 2013: https://www.google.com/?gws_rd=ssl#q=predictive+analytics+in+healthcare+ppt&start=0

4. 3
UNCHEALTHCARESYSTEM
Types of Data Analytics in Health Care
https://www.gartner.com

5. 4
UNCHEALTHCARESYSTEM
Predictive Analytics
• Predictive analytics uses technology and statistical methods to
search through massive amounts of information, analyzing it to
predict outcomes for individual patients
https://www.healthcatalyst.com/predictive-analytics-healthcare-lessons/

6. 5
UNCHEALTHCARESYSTEM
Uses of Predictive Analytics in Health Care
• Readmissions
• Use an algorithm to predict which patients are likely to be
readmitted to the hospital
• Tailor intervention to individual patient
• Ensure patients actually receive the precise interventions intended
for them
• Monitor specific patients after discharge to find out if having
problems before they decompensate
• Ensure low ratio of patients flagged for intervention to patients who
experience a readmission
Bates D. Health Affairs.2014;33(7):1123-1131

7. 6
UNCHEALTHCARESYSTEM
Uses of Predictive Analytics in Health Care
• Triage
• Estimating risk of complication when patient first presents
to hospital
• Assists in managing staffing and bed resources
• Anticipates need for a transfer to appropriate unit
• Requires detailed guideline that clarifies how algorithm
will inform care
Bates D. Health Affairs.2014;33(7):1123-1131

8. 7
UNCHEALTHCARESYSTEM
Uses of Predictive Analytics in Health Care
• Emergency Care
Janke A. Ann Emerg Med. 2016;67:227-236

9. 8
UNCHEALTHCARESYSTEM
Uses of Predictive Analytics in Health Care
• Decompensation
• Often before decompensation there is a period in which
physiological data can be used to determine whether
patient at risk for decompensation
• Host of technologies available to monitor patients who are in ICU,
general care, nursing homes, or at home
Bates D. Health Affairs.2014;33(7):1123-1131

10. 9
UNCHEALTHCARESYSTEM
Uses of Predictive Analytics in Health Care
• Adverse Events
• Predict which patients are at risk of adverse events
• Renal Failure
• Combine analytics with data about exposures to specific medications
and measures of kidney function
• Infection
• Example: Monitoring and interpreting changes in heart rate variability
for detection of major decompensation in infants with low birthweights
before emergence of infection
• Adverse Drug Events
• Predict which patients may suffer an adverse drug event and detect
patients who are in early stages of event
Bates D. Health Affairs.2014;33(7):1123-1131
Suresh S. Pediatr Clin N Amer. 2016;63:357-366

11. 10
UNCHEALTHCARESYSTEM
Implementing Electronic Health Care
Predictive Analytics
• Implementation of electronic health care predictive analytics is
needed to aid in real-time, point-of-care decision making
• Challenges include:
• Testing model in real-world setting under appropriate supervision
• Appropriate oversight of implementation
• Stakeholder engagement
• Appropriate patient privacy and consent policies
• Data quality assurance
• Broad implementation of the model in a health care setting
• Interoperability of health care technology platforms
• Transparency within health care systems
• Long-term challenges
• Medical education and training
• Sustainability
Amarasingham R. Health Affairs.2014;33(7):1148-1154
Cohen I. Health Affairs.2014;33(7):1139-1147

12. 11
UNCHEALTHCARESYSTEM
Conclusions
• Predictive analytics has the potential to use the
power of big data to improve health of patients and
lower cost of health care
• Multiple models exist that will benefit from use of
predictive analytics
• Several challenges need to be overcome to take
advantage of full use of predictive analytics
Bates D. Health Affairs.2014;33(7):1123-1131

13. UNCHEALTHCARESYSTEMUNCHEALTHCARESYSTEM
The Use of Predictive
Analytics in Health Care
Kathryn Morbitzer, PharmD
Wednesday, April 27th, 2016