Data Science for the ASC - Becker's ASC Review D/2_T… · Staff Costs Payroll System Accounting Payroll system Software report Procedure Cost Weekly Supply ... Histogram . Basic

Data Science for the ASC

Joseph Engel, MPS

The Endoscopy Center of Fairfield

Gastroenterology Associates of Fairfield County

Fairfield, Connecticut

Becker’s ASC 22nd Annual Meeting: The Business and Operations of ASCs

Data Science for the ASC

Joseph Engel, MPS

Administrator

Gastroenterology Associates of Fairfield County

The Endoscopy Center of Fairfield

I have no disclosures to report

Today we will cover…

• The Data Revolution

• A Data Analytics Plan for ASCs

• An example of a data analysis using a simple Run Chart

The Data Revolution

• What’s going on with data?

• Terminology

• ASCs lag behind

• What is a data analysis?

The Data Revolution

What’s going on with data?

• An enormous amount of data is collected every day

• Cost of collecting data is dropping rapidly

• Realization that data has value

• The need to manage all this data is growing rapidly

What’s going on with data?

Every 2 days we create as much information as we did from the beginning of time until 2003

Over 90% of all the data in the world was created in the past 2 years.

Every minute we send 204 million emails, generate 1,8 million Facebook likes, send 278

thousand Tweets, and up-load 200 thousand photos to Facebook

570 new websites spring into existence every minute of every day.

The number of Bits of information stored in the digital universe is thought to have exceeded

the number of stars in the physical universe in 2007.

By better integrating big data analytics into healthcare, the industry could save $300bn a

year, according to a recent report – that’s the equivalent of reducing the healthcare costs of every man, woman and child by $1,000 a year.

http://www.smartdatacollective.com/bernardmarr/277731/big-data-25-facts-everyone-needs-know Posted 10/28/14

Data Terminology

• Big Data

• Data Science

• Data Analysis

• Predictive Analytics

• Machine learning

• Data Mining

• Statistics

• Etc.

The Skill Set of the Health Care Manager of the Future

Data Science

Quality Measurement

Negotiation

Compliance Management

Everything Else

ASCs lag behind

Industry and hospitals are swimming in data.

The Question is where is the comparable data for ASCs?

http://www.beckersasc.com/asc-turnarounds-ideas-to-improve-performance/where-is-all-the-asc-data.html

accessed 7/18/2015

What is a Data Analysis?

• Data can be studied in the same way that an astronomer studies objects in the universe or a chemist studies molecules.

• Data is a collection of measurements obtained from observing a phenomenon such as a process.

• The objective of an analysis is to discover the nature and characteristics of the data, describe what one finds and then use the findings to answer questions about that phenomenon.

Create an analytics plan for the

ASC

Think data

Determine the purpose

Brainstorm with staff etc.

Inventory your data

What do you have, what do you need?

Identify who’s involved

Have a central guidebook for your data

Plan organization and storage

Create an analytics plan for the

ASC

List questions

Baseline questions

Questions that might be asked

Outline the data capture process

Learn simple analytical techniques

Don’t hide your data

Use the results

Check out existing analytic programs

Use The Data Science analysis model

Think Data

How can your measure of interest (data) best be tracked and presented?

What is the cost and production relationship?

How much does a unit of production cost?

What works best? e.g. what does your experience show is the best prep?

How can you measure quality?

What’s the best way to manage and protect your data?

How do you share your data?

What is the optimal use of resources?

Purpose

Assess patient experience

Get paid

Satisfy interested parties

Control processes

Measure compliance

Purpose

Measure cost

Measure efficiency and effectiveness

Measure use of resources

Demonstrate how good you are

Predict an outcome

Inventory Your Data – What do you

have now?

JAC AAAHC

Health Department

Surveys

Benchmarking

Budgets

Room usage

Counting meds

Outcomes

Quality measures

Billing data

Logs of all kinds

The data you collect – can it

answer your question?

“The data may not contain the answer. The combination of some data and an aching

desire for an answer does not ensure that a reasonable answer can be extracted from a

given body of data.”

John Tukey (quoted by Jeff Leak)

Useful results require accurate data

Analytics Plan Example

The Surgery Center

Analytics Plan

Data Source Who collects Where it's kept Form Used Purpose Collection Frequency

Schedule time Program Scheduler Software Schedule Software report Set Pt arrival time Each case

Procedure times Procedure Sheet Nurse Pt Chart Paper forms Stats Each case

Staff Costs Payroll System Accounting Payroll system Software report Procedure Cost Weekly

Supply Costs Accounting Accounting staff Acct Software Software report Procedure Cost Monthly

Meds Used Med Log Nurse Head Nurse Office Med Form Ttrack Usage/Order Daily

Patient Survey Patient Outside service Head Nurse Office Pt reply form Measure Satisfaction Per visit

Answer a question with your

data

Are we on time?

What is our cost per procedure?

Are we safe?

Are we cost effective?

Are patients satisfied?

Are we effective?

Learn simple analytical

techniques

Excel

Online templates

Software reports

Use the results

Try to present results graphically

PDSA for using results:

• Plan

• Do

• Study

• Act

Check out existing analytic programs

Two existing quality approaches that can be a source of ideas and techniques or even part

of your analytics plan

Lean

Six Sigma

Lean and Lean Six Sigma in Healthcare

Lean and Lean Six Sigma applications in healthcare require an understanding of how the tools and methodologies translate to the people-intensive processes of patient care. Once applied, the possibilities are endless.

Using Lean Six Sigma, Morton Plant Hospital in Clearwater, FL, improved patient satisfaction over 50%, reduced emergency department length of service by 21%, and recovered over $4 million in cost of quality.

•Lean

Eliminate non-value-added activities and waste. Save time and money.

•Lean Six Sigma

Increase patient satisfaction and reduce errors by improving your processes.

http://asq.org/healthcaresixsigma/

The Data Science Analysis Model

Ask the question

Gather the data

Clean the data

Explore the data

Analyze the data or build an algorithm

Communicate the results

Act

An example of the data

analysis process

Start with a question:

Are we on time?

Analyze the question

“Are we on time?”

Carefully define the parts of the question – establish an “operational definition” for this study. What is the purpose of the study?

What are you really asking?

Who is “we”?

What does “on time” mean in this case?

On time for what?

Refine the Question

“We”

…means all staff, the provider, the room, the patient. Everything that has to be ready to start the case.

“On time”

…means any procedure that starts within 5 minutes of it’s scheduled time.

Refine the Question

The purpose of the study

Identify and measure controllable factors. Separate them from random, uncontrollable

factors.

Identify unforeseen factors that cause some level of variation.

Test interventions such as a catch up period or staff retraining.

Set the center’s target for this measure.

E.g. You might set a standard of one delayed case per day.

Tidy data

What data do we have?

Can it be used to answer this question?

Do you need more data or different data?

How is the data stored and organized?

What has to be done to make it usable?

Is it in a form that your analytical tools can process?

Collect, examine and explore

your data

Collect about 30 data points to start with.

This spreadsheet shows a few lines of the data collected for this example.

Raw Data

Cases Scheduled Cases on Time

9/1/2015 15 10

9/2/2015 15 12

9/3/2015 12 8

9/4/2015 14 13

9/5/2015 15 9

9/6/2015 12 9

9/7/2015 15 9

9/8/2015 10 2

9/9/2015 15 6

9/10/2015 15 7

9/11/2015 12 5

9/12/2015 15 6

Know your tools and techniques so you can analyze the data appropriately

Descriptive Statistics

Run Charts

Control Charts

Fishbone diagrams

Excel

Charting or graphing programs

Report writers

MS Access

Etc.

An analytical starting point:

Descriptive Statistics

Mean

Median

SD

Range

Scatter Plot

Box Plot

Histogram

Basic descriptive statistics

Useful for initial data exploration and getting to know

your data

Typically shows a point in time

Does not easily show random vs external causes of

variation

Basic descriptive statistics

Raw Data

Cases Scheduled Cases on Time

9/1/2015 15 10

9/2/2015 15 12

9/29/2015 15 6

9/30/2015 15 3

Mean 415 243

Median 15 7.5

Max 15 15

Min 10 2

Choose an appropriate tool or method of analysis

To answer the question “How many procedures begin at the time scheduled each day?” you will need a tool that shows the variations in the measure of interest over time.

The measure of interest is the number of cases starting within 5 minutes of the time scheduled each day.

The tool used should separate fluctuations due to chance from fluctuations due to outside causes which you may be able to control.

A simple “run chart” can show this difference

The Run Chart

For this example I will use a run chart template available on the Institute for Healthcare Improvement website.

http://www.ihi.org/resources/Pages/Tools/RunChart.aspx

Simple to construct

Precursor to the more detailed Control Chart tool

A reasonable first step

A Simple Run Chart

Median

Goal

0

2

4

6

8

10

12

14

16

9/1

/201

5

9/2

/201

5

9/3

/201

5

9/4

/201

5

9/5

/201

5

9/6

/201

5

9/7

/201

5

9/8

/201

5

9/9

/201

5

9/1

0/2

015

9/1

1/2

015

9/1

2/2

015

9/1

3/2

015

9/1

4/2

015

9/1

5/2

015

9/1

6/2

015

9/1

7/2

015

9/1

8/2

015

9/1

9/2

015

9/2

0/2

015

9/2

1/2

015

9/2

2/2

015

9/2

3/2

015

9/2

4/2

015

9/2

5/2

015

9/2

6/2

015

9/2

7/2

015

9/2

8/2

015

9/2

9/2

015

9/3

0/2

015

The ASC On Time Cases

http://www.ihi.org/resources/pages/tools/runchart.aspx

Evaluating a Run Chart

What is a “Run”?

One or more consecutive data points on the same side of the median

What is “Common cause variation”

Variation caused by random factors

Indicates that the process is stable and predictable

Leads to a change in the process

What is “Special Cause Variation”

Variation caused by external factors

Indicates that the process has changed

Leads to an investigation of the external cause

Evaluating a Run Chart

http://www.ihi.org/resources/pages/tools/runchart.aspx

Evaluating a Run Chart

http://www.ihi.org/resources/pages/tools/runchart.aspx

Evaluating a Run Chart

http://www.ihi.org/resources/pages/tools/runchart.aspx

Evaluating a Run Chart

http://www.ihi.org/resources/pages/tools/runchart.aspx

Evaluating a Run Chart

http://www.ihi.org/resources/pages/tools/runchart.aspx

Results

Evaluating the Run Chart Results

Number of data

points 30

Number of Runs 6

Trends 2

Shifts 2

Too many too few 11 to 20 runs Too few

Astronomical 0

Control Charts

More advanced tool to study change in measures over time

Get an idea of things with a run chart then move on to a control chart

Templates and instructions are available on the web

Control Charts

Review of what was covered

Like your data and get to know it

Look into programs that make sense to you

Be organized about it

Involve the right people (everyone)

Get the data that fits your question

Keep it simple

Resources

Carey, R. G. Improving Healthcare With Control Charts – Basic and Advanced SPC Methods and Case Studies,

ASQ Quality Press, 2003.

Leak, J. The Elements of Data Analytic Style - A guide for people who want to analyze data.

http://leanpub.com/datastyle Version accessed 2015-03-02

Health care IT news

http://www.healthcareitnews.com/directory/agency-healthcare-research-and-quality-ahrq

Agency for Healthcare Research and Quality

http://qualityindicators.ahrq.gov/

Resources

Oregon Ambulatory Surgery Center Reporting

http://www.oregon.gov/oha/OHPR/RSCH/pages/ambulatory_surgery_reporting.aspx

The home of the U.S. Government’s open data

http://www.data.gov/

American Society for Quality

http://asq.org/index.aspx

Institute for Healthcare Improvement

http://www.ihi.org/about/Pages/default.aspx

Resources

National Associat ion for Healthcare Quality

http://www.nahq.org/about/content/aboutnahqindex.html

Montgomery, Douglas, C. Int roduction to Stat ist ical Quality Cont rol, Sixth Edit ion

978-0-470-16992-6 2009 by John Wiley & Sons, Inc.

http://www.directtextbook.com/isbn/9780470169926

Creating The Analyt ics Plan

http://stevetodd.typepad.com/my_weblog/2012/03/phase-1-innovation-analyt ics-creating-the-analyt ics-plan-.html

Accessed 9/30/2015

Walter A Shewhart , 1924, and the Hawthorne Factory

http://www.ncbi.nlm.nih.gov/pmc/art icles/PMC2464836/

Accessed 10/2/2015

Resources

Engineering Statistics Handbook

http://www.itl.nist.gov/div898/handbook/index.htm

Handbook for Basic Process Improvement – Air University

http://www.au.af.mil/au/awc/awcgate/navy/bpi_manual/handbook.htm

Data Analytics for the ASC

Questions