From Concept to Rapid Visualization - A Data Analytics Case Study

From Concept to Rapid Visualization – a Data Analytics

Case StudyGregory Wozniak, PhD

Director of Outcomes Analytics

Health Outcomes Group

American Medical Association

The Journey

Objectives

• Share experiences and lessons learned from a multi-stakeholder, multidisciplinary collaboration in exploring the use of visual analytics to transform data into information – various AMA units & Mikan Associates

• To analyze patterns and trends in antihypertensive medication (AHM) use, the AMA and CDC contracted with IMS Health to acquire state level data on AHM fills and spending for 2009 - 2013

• Present the case study of developing an interactive mapping application to assess variation in AHM prescribing behavior among physicians and non-physicians

The Toll of CVD and HTN

• Cardiovascular disease causes one-third of all deaths in the United States

• About 70 million American adults (29%) have high blood pressure

• In 2010, hypertension (HTN) was identified as the underlying or contributing cause in more than 360,000 deaths

• High blood pressure costs the nation $46 billion each year, including the cost of health care services, medications to treat high blood pressure, and missed days of work

• Despite the potential to prevent or manage HTN through diet and lifestyle modifications, most need antihypertensive medications (AHM) to control their blood pressure

5

Hypertension or elevated BP = systolic BP ≥140 mmHg or their diastolic BP was ≥90 mmHg

Hypertension control = BP not elevated (BP < 140/90 mm Hg)

Key reasons for poor BP control

• patients lack awareness of their hypertension status• ineffective pharmacologic management - treatment inertia, and inadequate health

care system adoption and consistent implementation of evidence-based guidelines• poor patient adherence to their AHM therapy regimens due to numerous individual

and environmental factors

Hypertension – A Major Risk Factor in CVD

CONFIDENTIAL

Hypertension Awareness,

Treatment and Control, NHANES

2007-2012

7

AMA Focus on Improving Health Outcomes

One of 3 focus areas under new strategic plan (2012)• Improving health outcomes

• Shaping new delivery and payment models

• Accelerating change in medical education

Improving health outcomes: Long-term goals• Prevent heart disease, stroke and type 2 diabetes

• Improve health outcomes for these conditions

AHM Visualization Project• The initial concept was to expand the AMA IHO web presence:

static visualizations of state-level data on AHM fills and spendingbe public facingresource for healthcare providers, public payers, and state health departmentssupplement AMA IHO research and publication agendas

• The analytic objective was to provide actionable insights and data on AHM cost burden and opportunities for improved AHM adherence:associations between AHM fills or spending and improvement in hypertension

outcomesvariation and trends in AHM prescribing behaviorphysician peer-group comparisonspotential AHM cost savings through expanded generic Rx programs and formulary

review

Mikan Associates – Adaptive Iterative Methodology

9

Data Data Model ETL

Visualization MockupsConnected Dashboards

Final SolutionBusiness Objectives

Feedback• Define KPIs

Iterative Feedback to Ensure Solution Provides Value

Adaptive Iterative Methodology

• Begin with Business Objectives

• Identify key metrics

• Use mockup dashboards for iterative Visualization development

• Continuous feedback from users

• Mockup data first to confirm data model for source data

• Connect mockup dashboards to final data source

11

12

Planned Data Visualization

Fill counts or spending by state, medication class, and year, by

• Brand/Generic

• Prescriber Type

• Payment Type

• Gender

• Age Group

• New vs. Refill

13

Planned Data Visualization

14

Fill counts or spending by state, medication class, and year, per 1,000

• Adults• Hypertensive Adults• Physicians• Non-Physicians

Status

• Building beta versions of the applications provided valuable first use cases and user experience feedback for refining the application

• Usability testing of the pre-production version is ongoing with publishing the interactive dashboards to the web in QII-2015

• Data update – 2014 data, new data structure

.

Questions?

Gregory Wozniak, PhD

[email protected]