Big Data Health Analytics is Personalizing Care
Download as PPSX, PDF
0 likes492 views
Big Data Health Analytics is Personalizing Care Chronic disease management and preventing hospital readmissions are two examples of how big data health analytics can personalize care. Wearable sensors and cloud computing allow for round-the-clock monitoring of chronic obstructive pulmonary disease patients, improving their condition and medication adherence. Analyzing clinical notes, claims, pharmacy and other data allows hospitals to predict patients at risk of readmission and reduce rates through proactive follow up care. Big data from various sources has potential to further personalize healthcare.
Big Data Health Analytics is Personalizing Care
- 1. Big Data Health Analytics is Personalizing Care ANITA T SALINAS anitatsalinas@gmail.com www.linkedin.com/in/anitatsalinas
- 2. Big Data Health Analytics is Personalizing Care Harnessing high data volumes & varieties with analysis Analyzing todays high data volumes, multiple sources and types of data to discover patterns that improve care for an individuals specific needs 2 examples Chronic disease management Hospital readmissions
- 3. 1. Chronic Disease Management Chronic Obstructive Pulmonary Disease (COPD) 72 year old Larry doesnt get enough oxygen COPD = Bronchitis + emphysema 3rd highest cause of death in US
- 4. 1. Real-time, Round-the-Clock Monitoring Wearable + Cloud + Big Data Health Analytics Solution Sensor collects continuous biometrics stream Thousands of data points/minute to cloud Self-monitoring, Caregivers monitor remotely Analyze symptoms, history over time Trigger automatic alerts, set thresholds Result Improve patients every day condition Improve medication adherence Avoid more acute symptoms & care Respiratory Rate Heart Rate/Variability Sleep Staging/Quality Calorie Burn/Steps Skin Temperature Fall Detection/Severity
- 5. 2. Preventing Hospital Readmissions Who wants to go to the hospital? Who wants to go again? Not patient: One of the most dangerous places in the world Harvard Professor, MD Not plan: Average cost >$1700/day, average length of stay = 5 days Not provider: CMS penalties for readmissions - $227M, 2225 hospitals in 2013
- 6. 2. U Pitt Medical Center Readmission Risk Score Big Data + Predictive Health Analytics Solution Combine Dr. notes, claims, pharmacy, clinical data Analyze for patterns, correlations Predict patients needing post-discharge urgent care Assign readmission risk score at discharge Result Assign proactive follow up care for high-risk patients Reduced readmissions by 37%
- 7. Big Data Health Analytics is Personalizing Care These are just 2 examples. Consider personalized care possibilities of analyzing a combination of big data from Physiology Clinical Pharmacy Claims Public Data / Demographics Social media / Behavioral Genomics Genealogy
Editor's Notes
- #2: Healthcare. Its an exciting time in Healthcare today. What with the unprecedented transformation we are in the midst of, coupled with advances in technology such as BDA. In no other vertical can BDA improve the lives of so many on such a personal level. Lets review 2 examples of how personalized care is being fueled today by BDA.
- #3: BDA: harnessing todays high volumes and varieties of data. data from different sources such as EHRs and caregiver notes, wearables and sensors. Data of different types such as structured and unstructured. And data of humongous sizes for example, sensors are collecting data constantly. Its then combining and parsing these mounds of data to discover and analyze patterns that can be applied to improve an individuals qualify of life and care. The 2 examples I will discuss are chronic disease management with 1 out of every 2 people in the US living with chronic disease by 2020 according to Health IT News - and hospital readmissions. http://www.informationweek.com/healthcare/electronic-health-records/7-big-data-solutions-try-to-reshape-healthcare/d/d-id/1107893
- #4: My 72 yo friend Larry doesnt get enough oxygen Coughs, shortness of breath, mucus, wheezing, hard to sleep Walks to the convenience store 4 blocks for cigs, has to stop 2x He has COPD - Mix of bronchitis and emphysema 3rd highest cause of death in US Has oxygen tank, meds (bronchodilators), inhalers Doesnt always take his meds, get his refills, use his inhaler
- #5: Philips Healthpatch - Cloud HealthSuite Digital Platform - eCareCompanion, eCareCoordinator Monitor Apps Solution http://www.engadget.com/2014/10/13/philips-copd-sensor/ http://www.usatoday.com/story/tech/2014/10/13/philips-gives-copd-patients-a-lifeline-with-new-gadget/17171535/ http://www.vitalconnect.com/overview Record patient biometric data relevant for COPD then push data to cloud HealthSuite Digital Platform using the patient's personal mobile device. using HealthPatch sensor the cloud-based service allows both patients and doctors an unprecedented level of access to realtime, round-the-clock health data. Record physical activity and inactivity, respiratory function, heart rhythm, and heart rate variability data retrievable via Philips eCareCompanion and eCareCoordinator BDA: Collect BD in cloud analyze diagnose combo of factors Result Alert: take inhaler, get on oxygen, take a break, use/get meds Alert: caregiver to check in Auto refill meds
- #6: Who wants to go to the hospital? Who wants to go again? Not the patient or family One of the most dangerous places in the world Harvard MD & Professor Ashish K Jha 180K americans in the medicare community alone die every year in hospital due to med errors preventable harm http://kaiserhealthnews.org/morning-breakout/medical-mistakes-2/ Not the plan Aetna: average cost >$1700/day, average length of stay, 5 days http://www.aetna.com/voluntary/hospital.html Not the provider CMS penalizing 3% Medicare reimbursements in 2015. 2225 hospitals, $227M in 2013. (penalties based on Medicare patients for heart attack, heart failure, pneumonia.) http://kaiserhealthnews.org/news/readmission-penalties-medicare-hospitals-year-two/ UPMC incented to understand why, how to prevent
- #7: Solution BD combo unstructured Dr. notes, claims, pharmacy, EHRs Analyzed: Looking for patterns, correlations Able to Predict patients needing urgent care after discharge WW in notes, low drug adherence in past, # of prior admissions in past year, length of stay >4 days, admitting service (ER or elective) Mom in notes = low risk Assign readmission risk score Result Do intervention, assign a care manager, follow up calls, RPM Reduced readmissions by 37%
- #8: These are just 2 examples. Consider what can be analyzed with the combo of structured and unstructured data about Physiology, Clinical/EHR, Pharmacy, Claims, Med Adherence Propensity, Public Data, Genomics, Genealogy Its an exciting time in healthcare - BDA applied for healthcare scenarios is improving quality of patient care on a personal level