Evolving E-Health Business Processes Around Accessible Data Warehouses Background information for demonstrations January 24, 2007.

Evolving E-Health Business Processes Around Accessible Data Warehouses

Background information for demonstrations

January 24, 2007


Agenda

• Overview of ORNEC • Overview of project and team• Discussion of major activities and accomplishments

– Simulation of Ottawa Hospital Data Warehouse and environment

– Business Intelligence prototype – Infection control data mart– Business Process Modeling – Discharge process– Requirements engineering – Quality indicator survey– Others

• Challenges – Obtaining access to real data/real users


Description of ORNEC

• Ontario Research Network for Electronic Commerce– Partnership between 4 universities and 40 corporations– Funded in part by the Ontario government

• Interdisciplinary research involving:– Information and communications technologies– Business and administration– Law and ethics

• Goals– Creation of scientific knowledge, business models, and best practices– Training of highly qualified academic and business leaders– Transfer of knowledge and innovation

• Research Themes– E-Government, E-Commerce Transactions, E-Learning and Collaborative

Environments, E-Governance, E-Health, and others• http://www.ornec.ca/


Project Description

• Evolving E-Health Business Processes Around Accessible Data Warehouses

• 2-year project (Jan. 2006 – Dec. 2007)• Funding: ~ $525,000, with half this amount in in-kind contributions

• Goals– Process improvement to take advantage of e-technologies and DW– Having good methodologies to describe, analyze, evolve, manage,

support, and automate DW-oriented, e-health processes– Promoting the access to DWs and managing changes– All goals consider privacy, confidentiality, quality, and consent, as

well as heavy legacy (and often manual) processes and regulatory environments


Project Investigators

• Daniel Amyot (Principal Investigator), Assistant ProfessorSchool of Information Technology and Engineering (SITE), uOttawa

• Doug Angus, ProfessorSchool of Management, and Director, Ph.D. program in Population Health, uOttawa

• Alan Forster, Associate ProfessorDepartment of Medicine, uOttawa, and Scientist, Clinical Epidemiology, Ottawa Health Research Institute

• Michael Weiss, Assistant ProfessorSchool of Computer Science, Carleton University

• Liam Peyton, Assistant Professor School of Information Technology and Engineering (SITE), uOttawa


Project Partners

• Provides Cognos 8 (most tools)– with setup, training, and consulting

• Contacts: Rupert Bonham-Carter, Gerry Leavy

• Provides (discounted) Adaptive Server Enterprise and IQ– with setup and training

• Contacts: Dan Murphy, Ahmadou Monfopa

• Provides DOORS and FocalPoint– with training and support

• Contacts: Frank J. Araby, Chris Sibbald


Students, Staff and Collaborators• Students involved in this project

– Saeed Behnam, PhD Computer Science– Pengfei Chen, MSc Computer Science – Sepideh Ghanavati, MSc Systems Science– Jason Kealey, MSc Computer Science – Sarah Musavi, Masters in Health Administration– Gunter Mussbacher, PhD Computer Science– Alireza Pourshahid, MSc E-business Technologies– Jean-François Roy, MSc Computer Science– Pierre Seguin, MSc. Computer Science– Bo Zhan, MSc Computer Science

• uOttawa staff and collaborators– Jacques Sincennes, System analyst– Greg Richards, Cognos Professor of Performance Management

• TOH collaborators– Cameron Keyes, Director (Acting), Decision Support– Richard Ciavaglia, Decision Support– Laurie Strano, Decision Support– Sylvain Paquette, Consultant– Josée Blackburn, Research Assistant, OHRI


Research Focus

• Develop new methods to model, analyze, and evolve business goals (why) and business processes (what/who/when/where) based on the use of goals, scenarios, and aspects, and adapted to DW-oriented e-Health services.

• This will in particular lead to suitable ways of exploiting the DW for trends and goal-driven decision support, and allow us to determine how the right data can be made available by the right individuals in the chain of care at the right level of detail, and how this data can best be accessed.


Project Tasks

• Study goal-driven and quality-driven decision making in an e-health context– How to map goals to business processes to data requirements in a

DW, and goals to reports or analytics – Goal example: “improve patient safety at a teaching hospital”

• Study the suitability and relevance of recent developments in requirements engineering, decision support, and business intelligence

• Study how best to combine goal-, scenario-, and aspect-oriented modeling for process modeling and requirements engineering

• Study and model relevant processes– For example: providing data to the DW, managing changes in the

DW requirements, secure access for various stakeholders, etc.• Replication of TOH work environment for lab study• Prototyping support for some of the processes using the lab facilities


Description of DW Project @ TOH

• Multiphase project • Collaboration between OHRI scientists, TOH Information

Systems• Phase 1 – Building the DW for researchers

– Funded through a CFI grant– Nearing completion

• Phase 2 – Using DW for administrative purposes– Investigation now underway– Identify key users– Access through a BI toolkit


Combining Healthcare and IT

Examples of Clinical / Healthcare Issues• Nosocomial infections• Drug costs• Bed utilization• Improving the safety of the discharge process

Information Technology Issues• Data needs, integration, quality, access, and reporting

Information Technology Opportunities• Data Warehouse (for integration and access)• Business Intelligence tools (for reporting)• Requirements Engineering (for needs and surrounding processes)


Agenda





Data Warehouse “Simulation”

• Intelligent Data Warehouse Lab (U. Ottawa)– CFI grant with IBM provides:

• Powerful servers and software• 10 Terabyte Storage Area Network (SAN)

– ORNEC project with Ottawa Hospital provides• Cognos 8 BI, Metrics Studio, Sybase IQ + ASE• Telelogic DOORS and FocalPoint, jUCMNav

• Same database schema as The Ottawa Hospital– Test data generator– Apply for an anonymous data extract


Demonstration / Learning Vehicle

• Students, Researchers (and Ottawa Hospital) have access to – Configuring and designing the environment– Hands-on training and mentoring (from Cognos)– Courses– Creation of sample applications and processes

• Antibiotics Tracking• Discharge process


Integrating Data Warehouse

DataWarehouse

Data Extraction & Transformation

Data Extraction & Transformation

Data Marts End-User Access

OperationalSystems

OperationalFeedback


Agenda



– Business Intelligence prototype – Infection control data mart

– Business Process Modeling – Discharge process– Requirements engineering – Quality indicator survey– Others

• Challenges – Obtaining access to real data/real users• Next steps


Data Mart Extract – For Infection Control


Performance Management PortalKey Metrics

Antibiotics Tracking

Campus Dashboard

Important Links

News Feed


Drill intoGeneral Campus


Drill into Surgery

Drill into Most Prescriptions


Drill into metric…

Metric History

BalancedScorecard


Report Authoring

Dimensional Model


Performance Management Infrastructure

• Data Mart “cubes” or “extracts”– Multi-dimensional snapshot with drill up and drill down– Pre-packaged security roles– Ethics and privacy review

• Performance Management Portal– Dashboards, flexible end-user tools for reporting, exploration, and

metrics

• Operational Integration– Data collection, data quality– Timely effect reports support decision making and track targets,

Service-Level Agreements (SLAs)– Business process improvements, transformations


Assessment Framework Tied to Operational Systems, Performance MGT & Data Warehouse Strategy

Business Systems & Processes

Use Case Maps Goals

Tasks

Performance Mgt Systems & Processes

DataWarehouse

PIQ measures the effectiveness of Reports to measure effectiveness of Organization in meetings its goals.

Stakeholders

Reports PIQ


Assessment Framework Tied to Operational Systems, Performance MGT & Data Warehouse Strategy

• Identify stakeholders, goals, tasks and use case maps to capture requirements and specify operational systems

• Identify reports needed to measure attainment of goals, inform tasks

• Measure importance of reports (related to goals)• Measure quality of reports (related to tasks, goals)• Measure penetration of reports (related to stakeholders, goals)• Measure effort, cost, timeliness, scalability, reliability etc. of data

collection, report creation, and distribution (effectiveness and efficient)

• Performance MGT/Data Warehouse strategy and implementation defined and driven by Reports & PIQ


Agenda





Overview of Approach

Health Care Services

Business Intelligence

Data Warehouse

Process Goals

Which reports to generate?

What data to collect?


Goal: Patient Safety

Discharge Process

Governance Process

Medical Management


Approach: BPM

Collect

Monitor

Redesign


Process Design and Evolution

• Identify goals and indicators (GRL)• Model the process (UCM)• Monitor process execution (DW)• Generate data mart (DM) and reports (BI)• Redesign process (redesign patterns)


jUCMNav:Goal model

editor/analyzer(GRL)


jUCMNav:Process modeleditor/analyzer

(UCM)


Goals and Indicators


Monitoring and Redesign


DischargeProcess

Actors

Subprocess

StartEnd


Discharge to Other Process


Dictate Process

Indicators

• Delay between dictation and transcription time• Delay between discharge and dictation time• Percentage of patients that are delayed over

three months (one month, one week)• Percentage of incomplete dictations


Report: Time Lag in Campuses


Drilling Down General Campus…


Report: Time Lag in Services


Drilling Down General Medicine


Drilling Down Guimarães Rosa, João


Predefined Reports




Agenda





Quality Indicator Survey

• Identified 78 quality of care indicators. – These indicators consist of a patient population and a treatment. Examples:

• Patients with atrial fibrillation should be prescribe warfarin unless there is an important contraindication

• Patients hospitalized due to an asthma exacerbation should be treated with beta-agonists

• 14 participants from TOH– Asked to determine which indicators were the most important

• A quality of care indicator is important if:a) the patient population is large (i.e. diagnosis/condition is highly prevalent)b) the treatment is highly effective and easily available to most patientsc) there are few patients in whom the treatment is contraindicated; and,d) in your role as an attending physician on CTU, you frequently treat this

population.• Survey performed with the help of a new Web-based tool we developed

– Answer collection and prioritization• Analysis of actual data vs. goal satisfaction using clustering• Reports (on-line and PDF) generated via Cognos BI tools


Online Survey ToolAllows for securequestionnaires to

be filled and answers (indicators, requirements, etc.)

to be prioritized


Some Survey Results

1. Quality indicators reflecting treatment decisions were rated higher than those reflecting investigation decisions– Example: For myocardial infarction, internal medicine physicians

felt it was more important to prescribe ASA than to order lipid profiling

2. Good agreement on the main indicators for Civic/General campuses– Top 5 very similar

3. Not all diseases surveyed have important indicators– For instance, pneumonia is very common yet its current indicators

scored second last

We also received feedback to improve the survey tool itself.


Survey: Next Steps

1. Go back to the doctors to discuss and validate the results

2. Design appropriate data marts for the top priorities

3. Create portals for these indicators

4. Get access to the real data and deploy the portals/reports


Overview of Other Activities: Using DWs

• Literature survey on uses and challenges of DW in the health sector – Report available (S. Musavi)

• Joined the Health Data Warehouse Association (HDWA)– Excellent networking opportunity– Attended their annual conference in June (J. Blackburn)

• External peer survey– Experience from peer organizations in HDWA and Canada – Approved by the TOH Research Ethics Board.

• Comparative study– Approaches for implementation of DW in publicly-funded healthcare

organizations (S. Musavi)– Canadian Blood Services and The Ottawa Hospital– Look at effectiveness of current reports (G. Richards)

• Coming soon: pharmacy technician to clean data in DW– Drug frequencies, routes, names…

http://www.hwda.org/

http://www.hwda.org/


Overview of Other Activities: Compliance

• Modelled governance process to access patient data via the DW– Goals/processes documented with jUCMNav (S. Ghanavati)

• Linked to Personal Health Information Protection Act (PHIPA)– Establish compliance and check compliance as the law and

the business process evolve over time.– Integration of model with Telelogic DOORS

GRL- High Level-Softgoals, and Goals

GRL- Detail-Softgoals, Goals, Tasks and

Actors

Use Case Maps

Law and Legislations Documents

Policies and Procedure Documents

GRL- High Level-Softgoals, and Goals

GRL- Detail-Softgoals, Goals, Tasks and Actors

1- High Level Traceability Link

2- Detail Traceability Link

Hospital Privacy Laws


Overview of Other Activities: Tools

• Tool support for business goal/process models– Improved jUCMNav tool substantially: editing, analysis, export (J. Kealey)– Improved the User Requirements Notation itself to explore aspect-oriented

modelling (G. Mussbacher)• Might ease the description and analysis of evolving goals / processes

• Business Intelligence tools– Training of students and partners on Cognos BI tools – Created tool to generate fake but representative data to simulate existing

DW (B. Zhan)– Performance evaluation of Cognos BI tools

• Heavy usage of DW, growing/evolving DW, etc.– New descriptor tool, based on SAS (A. Forster)

• Could be used as a preprocessor for BI tools (e.g. Cognos)– New graduate course on the use of databases for measuring healthcare

quality• To be offered for the first time in January 2007 (A. Forster)


Agenda





For Further Information

• Daniel AmyotAssistant ProfessorSITE, University of Ottawa(613) 562-6800 ext. [email protected]

• Web site:http://cserg0.site.uottawa.ca/twiki/bin/view/EHealth/WebHome

Evolving E-Health Business Processes Around Accessible Data Warehouses Background information for demonstrations January 24, 2007.

Documents

ehealth processes

advantage of e

business models

business leaders

population health

manual processes

uottawa slide

year project