Control Charts and Injury/Illness Analysis Presentation to the spring 2008 UW System Safety and Loss Prevention meeting, River Falls, WI

Control Charts and Injury/Illness Analysis

Presentation to the spring 2008 UW System Safety and Loss Prevention meeting, River Falls, WI

Overview

• What are we analyzing? What is missed?• Control chart review• Data for monthly employment• Questions for you– What is useful? What is not?– What else might be useful?

What are we looking at?

Near miss

First aid

LT and/or Med

FatalityDangerous occurrence

Sea of potential incidentsfrom people, machines, contaminants, procedures,

etc.

What are we looking at?

• Lagging indicator– Retrospective, little predictive value

• Only those events that are reported and recorded– Currently have no data to suggest how much is

unreported• Tracking failures, implies loss avoidance • However, it is one of the baseline measures of

outcomes

Overview


Control Chart ReviewUsing historic program data

• Monthly data (roughly even sample sizes)– Suggested minimum number of months, 25 to 60– For most campuses, the monthly count of lost

workday cases will be too low to yield meaningful results—use all recordable incidents

• Run chart

Historical dataGraphical record of incidents

07/2

002

09/2

002

11/2

002

01/2

003

03/2

003

05/2

003

07/2

003

09/2

003

11/2

003

01/2

004

03/2

004

05/2

004

07/2

004

09/2

004

11/2

004

01/2

005

03/2

005

05/2

005

07/2

005

09/2

005

11/2

005

01/2

006

03/2

006

05/2

006

07/2

006

09/2

006

11/2

006

01/2

007

03/2

007

05/2

007

07/2

007

09/2

007

11/2

007

01/2

008

0

5

10

15

20

25

30

35

40

UW-MilwaukeeAll claims & incidents (LT/HD/Med/Incd)

Incident/ claim count

Control limits

• Derived from the average of the baseline data• Typically set at ± 3 standard deviations• Data point outside of this is highly unlikely to

be result of random variation– There should be an assignable cause

Creating the control chartAdd the average and ±3 standard deviation

07/2

002

09/2

002

11/2

002

01/2

003

03/2

003

05/2

003

07/2

003

09/2

003

11/2

003

01/2

004

03/2

004

05/2

004

07/2

004

09/2

004

11/2

004

01/2

005

03/2

005

05/2

005

07/2

005

09/2

005

11/2

005

01/2

006

03/2

006

05/2

006

07/2

006

09/2

006

11/2

006

01/2

007

03/2

007

05/2

007

07/2

007

09/2

007

11/2

007

01/2

008

0

5

10

15

20

25

30

35

40


Incident/ claim count

UCL

Avg count 7/02--6/07

LCL

Out of control data

• Data outside of control limits• Other out of control trends• Effect on setting the baseline

Out of control data

• Definition of a trend for control charts– 1 point outside the control limits– 2 out of 3 points two standard deviations above/

below average– 4 out of 5 points one standard deviation

above/below average– 7 points in a row all increasing/decreasing– 7 points in a row all above/below average– 10 out of 11 points in a row all above/below average

Expected vs. unexpected data07

/200

209

/200

211

/200

201

/200

303

/200

305

/200

307

/200

309

/200

311

/200

301

/200

403

/200

405

/200

407

/200

409

/200

411

/200

401

/200

503

/200

505

/200

507

/200

509

/200

511

/200

501

/200

603

/200

605

/200

607

/200

609

/200

611

/200

601

/200

703

/200

705

/200

707

/200

709

/200

711

/200

701

/200

8

0

5

10

15

20

25

30

35

40


Incident/ claim countUCL+2 stan-dard devia-tion+1 stan-dard devia-tionAvg count 7/02--6/07

Overview


What about seasonal employment variations?

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC0

5000

10000

15000

20000

25000

30000

UW-Madison Headcount, CY2005

by month



5000

10000

15000

20000

25000

30000


by month


500

1000

1500

2000

2500

3000

3500

UW-Stevens Point Headcount, CY2005

by month


JAN FE

BMAR

APRMAY

JUN JU

LAUG SE

POCT

NOVDEC

010000200003000040000500006000070000

UW System Headcount, CY2005

by month


5000

10000

15000

20000

25000

30000


by month


500

1000

1500

2000

2500

3000

3500

UW-Stevens Point Headcount, CY2005

by month

Adjusting for variable sample size01

/200

5

02/2

005

03/2

005

04/2

005

05/2

005

06/2

005

07/2

005

08/2

005

09/2

005

10/2

005

11/2

005

12/2

005

01/2

006

02/2

006

03/2

006

04/2

006

05/2

006

06/2

006

07/2

006

08/2

006

09/2

006

10/2

006

11/2

006

12/2

006

01/2

007

02/2

007

03/2

007

04/2

007

05/2

007

06/2

007

07/2

007

08/2

007

09/2

007

10/2

007

11/2

007

12/2

007

-

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

UW-MilwaukeeRate of incidents/med/LT/HD per 200,000 hours

Incident rate per 200K hrsUCLAverageLCL

Overview


Plan for OSLP action, FY09Injury/Illness Analysis

• Listen to your ideas for what we should look at– At your campus– Across UW System campuses– Comparison to others


• Listen to your ideas for what we should look at• Prioritize– Long-term, on-going metric– Short-term, hot topics


• Listen to your ideas for what we should look at• Prioritize• Each month, develop one new data aspect– Finalize STARS report format to update system-

wide C-chart on quarterly basis.– Look at age vs. incident rate.– Look at causes for summer vs. academic year.


• Listen to your ideas for what we should look at• Prioritize• Each month, develop one new data aspect• Report to campuses every 4 months on

findings– For up to one year– Revisit usefulness May ‘09

Resources

• http://www.uwsa.edu/oslp/safety/uwsres/presentations.htm

• http://www.hanford.gov/safety/vpp/vppage.htm• Safety Metrics: Tools and Techniques for Measuring Safety

Performance, Christopher Janicak, 2003, Government Institutes

• Accident Prevention Manual for Business and Industry: Administration and Programs, Krieger and Montgomery, 1997, National Safety Council

• Quality Control and Industrial Statistics, 4th Ed., Acheson Duncan, 1974, Richard D. Irwin, Inc.

Documents

Control Charts and Injury/Illness Analysis Presentation to the spring 2008 UW System Safety and Loss Prevention meeting, River Falls, WI