performance improvement (pi) and financial impact · existence is justified by providing...

Performance Improvement (PI) and Financial Impact

Providing Sustainable

Patient-Centered Care

Agenda

• Challenges in healthcare today

• PI and it’s application in healthcare

• PI and financial analysis

Learning Objectives

• Explain the importance of a PI Model to guide learning

• List key problems that PI can address to drive sustainability in healthcare

• Describe how organizational structure and managerial accounting are impacted by PI

Professional Background • 7 years in Automotive Manufacturing

– Ford and Goodyear

– Production, engineering, financial analysis

• 3 years in Electronics Manufacturing

– Rockwell Automation

– PI, engineering management

• 3 years in Healthcare

– Norton Healthcare

– PI

Strengths

Proximity to US Customers

Truck Market Dominance

Weaknesses

Push Flow

Brand / Model Proliferation

High Labor Cost

Opportunities

Leverage Worker Votes

Struggling EUR Brands for Sale

Threats

Transplant Operations

NAFTA

Lax Foreign Reg’s

Internal

External

Favorable Unfavorable

PI efforts over next 2 decades: Lead Time

Reduction Brand / Model

Rationalization Workflow

Efficiency

Industry Need Analogy: 90’s US Auto

US Auto in the 90’s had internal weaknesses that could be addressed to counter external threats

Strengths

Highly Educated Workforce

Outpatient Service Growth

Weaknesses

Controllable Volume Variation

Unnecessary Practice Variation

High Labor Cost

Opportunities

M&A / Partnerships

EBP Sharing

Threats

Retail Markets

VBP / ACA

Higher Deductibles

Internal

External

Favorable Unfavorable

Potential PI Efforts:

Scheduling Optimization

Practice Standardization

Workflow Efficiency

How do we improve without trade-off?

Industry Need: US Healthcare Today

US Healthcare today is facing similar challenges to sustainability that the US Auto Industry faced in the 90’s

Why PI in Healthcare? Healthcare is becoming more competitive

As competition increases, there is greater pressure for an organization to justify its existence

Existence is justified by providing sustainable value

Sustainable value is achieved through PI

PCAST: President’s Council of Advisors on Science and Technology BETTER HEALTH CARE AND LOWER COSTS: ACCELERATING

IMPROVEMENT THROUGH SYSTEMS ENGINEERING

www.whitehouse.gov/ostp/pcast

Find under Documents & Reports (May 2014 Report)

Fact Sheet Recommendations from PCAST

• Recommendation 1: Accelerate the alignment of payment incentives and reported information with better outcomes for individuals and populations.

• Recommendation 2: Accelerate efforts to develop the Nation’s health-data infrastructure.

• Recommendation 3: Provide national leadership in systems engineering by increasing the supply of data available to benchmark performance, understand a community's health, and examine broader regional or national trends.

• Recommendation 4: Increase technical assistance (for a defined period—3-5 years) to health-care professionals and communities in applying systems approaches.

• Recommendation 5: Support efforts to engage communities in systematic healthcare improvement.

• Recommendation 6: Establish awards, challenges, and prizes to promote the use of systems methods and tools in health care.

• Recommendation 7: Build competencies and workforce for redesigning health care.

What is PI?

Performance Improvement (PI) is the practice of enhancing the ability of an organization to learn so that it grows value for its customers with the greatest speed.

CREATE new knowledge SHARE new knowledge

Why Lean and Six Sigma?

Lean Waste Elimination

Standardized Work

Flow

Customer PULL

Six Sigma Standardization, Variation

Defects – Miss Elimination

Optimization

Process Control

SPEED STABILITY & ACCURACY

“We will be the region’s most comprehensive, strongest, and preferred

health care organization, setting the standard for quality and caring.”

What is Value?

An activity that changes the state of the patient or information to meet customer needs and add value (something a customer would be willing to buy / perceive value in having)

Those activities that take time or resources, but do not directly meet customer requirements or add value

Value Added Activity Non-Value Added Activity

Examples of Waste Type of Waste Definition Healthcare Examples

Defects Time spent creating, detecting, and resolving errors

Wrong dose administered to a patient

Overproduction Doing more than the patient needs or doing it sooner than needed

Unnecessary diagnostic procedures

Waiting Waiting for the next activity to occur Insufficient bed volume / turnover post surgery

Not Using Talent Waste incurred due to ignoring employees, their ideas, and their potential / development

Nurses caught in daily workarounds and not engaged in improvement

Transportation Unnecessary movement of the patient, specimens, or materials

Cath lab being far from the ED

Inventory Excessive amounts of supplies Drugs on hand beyond their shelf life

Motion Unnecessary movement by employees Lab techs walking miles per shift in a poor layout

Extra Processing Excessive activity in a process step Unused form data

Sources of Waste Source Definition Healthcare Examples Graphical Depiction

Process Design Waste that is inherent in the processes we do today.

- Long distances between the ED and CT Scan (motion) - Error opportunities resulting in extra inspection (defects / extra processing)

Unevenness Waste that is generated by variation of patient volume or supply delivery

- Poor coordination with EMS resources causing waiting in ED (waiting) - Infrequent ordering of supplies resulting in high inventories (inventory)

Overburdening Waste that results from insufficient capacity

- Poor planning with changing community needs resulting in diversion of patients to other facilities or rushing service (transportation / defects)

ED

CT Scan

Lab 100 ft

NHC DMAIC Improvement Process

PROCESS

MEASUREMENT

ASSESSING PROCESS

PERFORMANCE

*

*

PROCESS

STEPS

FOLLOW EVIDENCED-

BASED PROCESS

PROCESS

IMPROVEMENT

MODEL

DMAIC Process Overview

Phase Outcome

Define - Problem Defined - Key process metric(s) identified (id by customers)

Measure - Understand Current Process - List of inputs that might be causing the problem - Accurate baseline data (KPOMs)

Analyze - Process waste identified - Prioritized list of critical barriers and root cause(s)

Improve - New process identified - Actions needed to deploy new process + pilot

Control - Hard-wired process w/ hand-off to business

owner - Celebration!

Project Overview • Project Description: In 2011, 24 ED Stroke Patients were

treated with rt-PA out of a total Stroke Patient population of 558 patients (4.5%). Of the 24 patients, only 4 (16%) had a door-to-needle time of 60 min or less. Timely use of rt-PA benefits the patient in the following ways: – Drug Effectiveness: 33% probability of a higher-score recovery

outcome vs. no treatment

– Increased Benefit: Increased probability of favorable outcome at 3 months post-event as OTT (Last known well to rt-PA administration) decreased showing statistical significance (p=0.005<<0.05)

• Project Scope: All stroke patients eligible for rt-PA. – Start: Patient arrival at ED / End: Administration of rt-PA

• Project Goal: Achieve median DTN time of 60 min or less.

DMAIC

Current State VSM DMAIC

What we wanted to know: How does the process flow?

What we found: Most of the process time is spent waiting after the patient is done in Radiology and there are opportunities to improve.

• Process observations were made for each process step at each hospital

• Process was timed (stopwatch, checklists, timestamps)

• Current State VSM was assembled based on findings

ICU (NH or NBH)

General Public

EMS

ED Wait29.0 min (95%)

3.0 min (Median)

6.9 min (Average)

EMS

Transfer Pt12.0 min (95%)

0.0 min (Median)4.8 min (Average)

Transfer Pt1.08 min (NAH)0.5 min (NBH)X min (NSH)Y min (NH)

= 1

Sign-in

= 1

Triage

= 3

Stroke Response

= 2

CT / CTA

= 2

Admin rt-PA

Pt Info

= 1

Patient Registration

Stroke Order

EMR System

Records Management

Register Pt

Pt ID #

Pt ID bracelet

Add’t Wait15.8 min (95%)


Wait on Lab &CT Results

Wait on Either:56.9 min (95%)


N = 44 (includes 3 no waits)

Wait on Lab Result:56.9 min (95%)


N = 8

Wait on CT Results35.5 min (95%)


N = 33

Transfer Pt1.08 min (NAH)0.5 min (NBH)X min (NSH)Y min (NH)

Wait on Tx Decision

(includes CTA Wait)60.7 min (95%)


N = 9 (Total rt-PA Pts in Study)

CTA TAT:82.0 min (95%)


N = 6

Pure Wait on Tx Decision(All tests in)

41.7 min (95%)17.7 min (Median)16.7 min (Average)N = 8 (one had MRI)

# Admin = 1

Avg. Time = 2.00 min

# RN = 1


# RN = 1

# PCA = 1

# ED MD = 1


Call Ahead Order for CT

# RN = 1

# CT-Tech = 1

CT Avg. Time = 2.92 min

CTA Avg. Time = 4.83 min

# RN = 2


With the exception of the ED Wait queue, 95% wait times are really the maximum non-outlier data point of the set.

2.00 min 2.17 min

6.9 min

3.33 min

4.8 min 1.08 min

7.75 min

6.3 min 1.08 min 19.3 min

4.33 min

51.2 min

Expected DTN = 110.24 min

NVA Time = 90.66 min

VA Time = 19.58 min

# Admin = 1

CT Order

Lab Order

CT Order

CTA Order

= 3

Review CT

# Radiologist = 1

# ED MD = 1

# Neurologist = 1

CTA Order (as needed)

rt-PA Order

= 2

rt-PA Decision

# ED MD = 1

# Neurologist = 1

Issue Order for rt-PAto be administered

rt-PA Door-to-Needle (DTN) ProcessCurrent State: Data Range from 12/01/11 to 03/31/12

For N = 4 (4 of 8 had rt-PA)Draw to Lab: 23.2 min (Average)

CBC: 5.3 min (Average)INR: 20 min (Average)

CMP: 25.8 min (Average)

Post CT Wait for CTA Order:87.8 min (95%)


N = 6

Process Step: CTA Scan

Aspect Question NBH NSH NH NAH

Systems

How is the impact to

DTN known? (Is it

quantifiable?) - Acute stroke checklist follows patient

- Not known/checklist not with patient

-manually write completion time

-Checklist not with patient - could check

MSTAT for recorded times -Unknown-checklist not always with patient

Pathways

Who is to do the Process

Step(s)?

(How clear is this?)

-CT tech

-Room 1 only

-CT tech

(CT1 preferred due to faster processing time -

10 mins vs 30 mins)

-CT techs

-2 scanners - 129 Seimens preferred for CTA -CT tech

Connections

How does he / she know

when service is

needed?

(How is this signaled?)

-Order from MD (requested based on

discussion with neurology)

-Currently creatinine result is awaited before

performing exam

(CTA done when patient has weakness on

one side and/or contraindications to tpa)

-Radioligist notifies ED MD of CT result

-ED MD consults with neurologist to

determine if CTA is necessary

(Outsourced to VRC at night)

-ISTAT creatinine done in ED - no delay

waiting for result

(table weight limit CT 1 650 #'s/ CT 2 450 #'s)

-Call from ED

-Can see order in computer

-Call from ED MD/neuro

-Can be verbal order

-Consent obtained if possible-use implied

consent if not

Activities


what is needed?

(How is this

determined?)

-Standard process

-CT tech training

-Standard process

-CT tech training

-Contrast dose 1 cc/pound-max 100cc's

-Implied consent used if not able to obtain

consent from patient/family -Same

-Standard procedure

-CT tech training/annual competencies

-Same level skill training all shifts

-Contrast load calculated per protocol

Activities


it is done correctly?

(How is this verified?)

-IV checked with saline flush (blown IV could

delay)

-View images real time to see if contrast is

flowing

-Monitor for adverse reaction real time

-Must have internet for transmitting images

(Re-visit ED MD/neurology dialogue

structure)

(Pts at low risk for CIN; investigating NOT

waiting for result to do exam)

-Check computer to verify no contrast allergy

-20g IV necessary to inject contrast

(CT 1 allows test bolus of saline prior to

contrast)

-Audible alarm for high pressure during

injection

-Images visually checked real time to verify

contrast was delivered

-Long distance to radiology - patient should

stay in dept. until determination is made for

CTA

(Have had issues with internet connectivity

at night - event actio plan?)

-Same with these additions:

-Do not use implied conset. Will wait for ED

MD to sign before completing exam

-Only manualy flush available to check IV

-No internet connectivity issues reported

-Radiologist on back up call

-CT tech review images real time for contrast

flow; contacts radiologist if needed

-Alternative contrast options posted in

radiology for contrast allergies

-Machine does test bolus to check IV; high

pressure alarm

-Daily QA's done; biomed does preventive

maintenance on machines; not posted

anywhere

-

Comments

Impact of Top 3 Priorities DMAIC

What we wanted to know: What is the impact of our efforts?

What we found: Focusing on just the Top 3 improvement priorities alone will address over 70 min of the 110 min expected DTN Time.

= 2

Admin rt-PA

Wait on Lab &CT Results

Wait on Either:56.9 min (95%)


N = 44 (includes 3 no waits)

Wait on Lab Result:56.9 min (95%)


N = 8

Wait on CT Results35.5 min (95%)


N = 33

Wait on Tx Decision

(includes CTA Wait)60.7 min (95%)


N = 9 (Total rt-PA Pts in Study)

CTA TAT:82.0 min (95%)


N = 6

Pure Wait on Tx Decision(All tests in)

41.7 min (95%)17.7 min (Median)16.7 min (Average)N = 8 (one had MRI)

# RN = 2


19.3 min

4.33 min

51.2 min

Expected DTN = 110.24 min

NVA Time = 90.66 min

VA Time = 19.58 min

For N = 4 (4 of 8 had rt-PA)Draw to Lab: 23.2 min (Average)

CBC: 5.3 min (Average)INR: 20 min (Average)

CMP: 25.8 min (Average)

Post CT Wait for CTA Order:87.8 min (95%)


N = 6

#1 Tx Decision

#2 Lab Label

#3 rt-PA Box

Over 40 improvement

options found!

#1: Tx Decision Roadmap DMAIC

What we wanted to know: How is delay in the Tx Decision addressed?

What we found: This is the EBP and NHC Neurologist accepted Tx Criteria. Following this and the broader process for Acute Stroke will reduce delays in Tx.

Stroke call from ED

Process Step to continue Tx Decision

evaluation

Give rt-PA

Contraindication – STOP and Do NOT

give rt-PA

Does CT Scan show Hemorrhage or other

contraindication?Yes

No

Does CT Scan show previous

stroke?

Gather patient data to determine when last stroke occurred

Last stroke within last 3

mos?

Do NOT give rt-PA and consider other

Tx optionsYes

Yes


Tx options

No

No

Does initial history review show any

contraindications?


Tx optionsYes

No

Is bleeding abnormality, thrombocytopenia, or any

anticoagulant use suspected from history?

Do lab results show abnormal values for INR,

H&H, or platelets?Yes


Tx optionsYes

Verify Last Known Well

Last Known Well > 3 hrs ago?

No No

BloodPressure NOT

Controlled> 185 / 110

No

Yes


Tx optionsYesGive rt-PA No


Tx options

Process Map for rt-PA Tx DecisionNorton Healthcare

• Worked with Stroke Neurologists to develop a shared view of risk (Contraindication to Tx)

• Mapped with input from EBP research and Stroke Neurologists (red is contraindication)

• Minimizes delay between receiving required results and issuing an order

• Included in related physician onboarding documentation

DTN Time Comparison DMAIC

What we wanted to know: How will we control DTN Time?

What we found: We have an on-going measurement system, processes in place, and a system team committed to continuously improving.

07/15/2012

06/26/2012

05/20/2012

03/05/2012

01/14/2012

12/01/2011

11/12/2011

10/03/2011

07/29/2011

05/27/2011

02/17/2011

200

150

100

50

0

Date

DTN

Tim

e (

min

)

_X=59.3

UCL=100.5

LCL=18.1

Pre-Project Project

1

DTN Time (min) by Project Stage All solutions implemented with

negligible cost!

PI Impact to Culture

Good Hero

Reactive

Fire fighter

Broken Processes

Fragmented efforts

Great Leader

Proactive

Permanent fire preventer

Efficient, integrated processes

Aligned efforts

The PI Journey

PI Opportunities in Healthcare 1. Redesign Practice Care Settings – Increase Prevention

2. Optimize Scheduling of Procedures – Stabilize Demand

3. Increase Accuracy of Productivity – Optimize Staffing

4. Iterative Redesign of Work – Continually Improve

A significant amount of the nation’s cost-of-care concerns can be addressed by incorporating these strategic objectives

Opportunities mostly focus on addressing operational concerns rather than clinical concerns

Operational vs. Clinical

The science in healthcare has traditionally been on the clinical side, but there is opportunity to have operational excellence

Core Clinical: Activities that involve Dx, Tx, or Prevention

Shared Territory: Where clinical factors have operational implications or where operational limitations influence clinical decisions

Supporting Operational: Activities that are not part of Dx, Tx, or Prevention

Many activities are considered operational: - Scheduling - Staffing - Allocation of Equipment - Capital Planning

Redesign Practice Care Settings

P1-3. Enhance Patient & Provider

Relationship (Assessment,

Dialogue, etc.)

P1. Leadership (Standard Performance

Metrics, Aligned Direction, Stakeholder

Engagement)

P1. Staffing

(Reliable & Efficient Workflows, Balanced to

Volume)

P1. Clinical Design

(EBP Protocols, CEP)

P2. Volume Balancing

(Panel Sizing, Scheduling)

Redesigning practice care settings, starting with primary care, drives preventive care efforts while testing concepts

Workflow Redesign

Eliminating unnecessary work, rebalancing assignments, and redesigning the Provider – MA flows have improved capacity

Reception Patient 2

Desk Nurse/LPN

MA

Provider

Check-out/Scheduling Patient 1

Patient 2

Patient 2

Provider Patient 1

MA Provider Patient 2

MA Patient 1

Provider Patient

MA

Room 1

Room 2

2

3a

3b

5a

5b

4

16

i ii

• Redefined work roles to balance work load

and move interruption away from the MA

• MA now able to be in a cyclical workflow

with provider for higher throughput

Optimize Scheduling of Procedures

Total Excluding

Weekends Current State

Simulated

(~15% Vol. Increase)

OR minutes 87,685 101,520

Actual LH 10,471 10,891

Target LH 9,663 11,188

Prod. Index 92.29% 102.72%

OR min per LH 8.25 9.32

Avg Daily OR min 4,384 5,076

Reducing daily scheduling variation allows for increased utilization and efficiency

>10% Productivity gain possible by

reducing CV from 0.36 to 0.07

Increase Accuracy of Productivity • Productivity standards are based on benchmarks

– Internally modified to better reflect current process – Do not identify true process potential (theoretical capacity)

• Benchmarks are based on peer group performance – Disguises waste from unevenness, hides potential

Hospital Nursing Surgery

Pulmonary

Services Imaging Cardiology

Hospital

Total

#1 30.51 1.68 2.34 0.42 1.05 35.70

#2 (21.91) 3.81 2.57 3.00 6.77 (2.20)

#3 (15.94) (3.24) 0.30 (1.92) (1.48) (15.39)

#4 (3.63) (0.67) 3.00 1.21 1.56 6.12

#5 (28.89) (3.52) (1.48) (2.34) (1.10) (32.58)

System (39.86) (1.94) 6.72 0.37 6.80 (8.36)

Variable Paid FTE Opportunity

(Based on Consultant Productive Standard Recommendation)

Opportunity to

apply

improvement in

practices too

0.67 FTE opportunity was

found to be closer to a

5.00 FTE opportunity

More accurate measurement of performance allows for gaps to be better understood and addressed

Iterative Redesign of Work

Current processes are continually challenged in a structured manner through procedures for increased efficacy

• Operational procedures include plans for reaction to abnormal events – Stop the process to analyze and fix

– Set new standards (inspect for what you expect)

• Take every opportunity to encourage finding problems and solving them – Can’t fix what you can’t see

– Everybody should have a specific role to play

• Structure reports & accounting for identifying system level problems with drilldown

Report Example

Data can be arranged for higher power analysis that identify where further studies / engagement is most beneficial.

Provider

Name

Total Var from AVG

by Px: OR Supply $

Surgeon A $210,798

Surgeon B $66,470

Surgeon C $34,064

Surgeon D $20,434

Procedure

Total Var from AVG

by Px: OR Supply $

Count

over AVG

Count of

Cases

Px 1 $53,971 30 33

Px 2 $52,467 30 32

Px 3 $36,088 11 11

Px 4 $25,944 25 25

Px 5 $15,238 7 7

Px 18 $0 0 1

Px 19 -$11 0 1

Grand Total $210,798 138 173

Provider Name

Total Var from AVG

by Px: OR Supply $

Count

over AVG

Total Var from AVG

by Px: OR Implant $

AVG Var from AVG

OR Minutes by Px

AVG Var from

AVG LOS by Px

AVG Var from

AVG SEV by Px

Count of

Cases

Surgeon A $52,467 30 $201,192 28 (2.1) (0.6) 32

Surgeon B $26,847 22 $139,927 33 (2.4) (0.8) 22

Surgeon X ($19,051) 0 ($6,009) 75 0.7 0.1 30

Surgeon Y ($24,108) 7 $20,423 (43) (0.2) 0.2 71

Surgeon Z ($40,117) 1 ($365,280) (6) (0.3) 0.1 64

Managerial Accounting

Segment reporting provides more insight than traditional allocation of costs.

• Traditional allocation spreads indirect costs to business units via assumed cost driver

– Cost driver may not be highly correlated

– Costs do not necessarily go away if unit is closed

• Segmented income statements only assigned traceable (direct) fixed costs to business units

– Traceable costs disappear if unit is closed

– Common costs are lumped into a cost pool

– Structure allows drilldown to see performance gaps

Segment Reporting Example: Step 1

The Medical Group has a low segment margin that is due to relatively high variable patient care expense in Primary Care.

Example Only Healthcare, Inc.

01/01/14 - 03/31/14

Total Company Hospital Medical Group

Unrestricted revenue $85,000,000 $70,000,000 $15,000,000

Variable expenses:

Variable patient care expenses $66,500,000 $55,000,000 $11,500,000

Other variable expenses $9,000,000 $7,000,000 $2,000,000

Total variable expenses $75,500,000 $62,000,000 $13,500,000

Contribution margin $9,500,000 $8,000,000 $1,500,000

Traceable fixed expenses $4,500,000 $3,500,000 $1,000,000

Division segment margin $5,000,000 $4,500,000 $500,000

Common fixed expenses not

traceable to individual divisions $5,000,000

Net operating income $0

Divisions

Medical Group Division

01/01/14 - 03/31/14

Total Division Specialty Primary

Unrestricted revenue $15,000,000 $6,000,000 $9,000,000

Variable expenses:

Variable patient care expenses $11,500,000 $4,000,000 $7,500,000

Other variable expenses $2,000,000 $500,000 $1,500,000

Total variable expenses $13,500,000 $4,500,000 $9,000,000

Contribution margin $1,500,000 $1,500,000 $0

Traceable fixed expenses $500,000 $200,000 $300,000

Service line segment margin $1,000,000 $1,300,000 ($300,000)


traceable to individual services $500,000

Net operating income $500,000

Service Lines

Segment Reporting Example: Step 2

Addressing the challenges for Primary Care Office B would significantly impact the company’s net operating income.

Primary Care Service Line

01/01/14 - 03/31/14

Total Service Office A Office B Office C

Unrestricted revenue $9,000,000 $2,000,000 $4,000,000 $3,000,000

Variable expenses:

Variable patient care expenses $7,500,000 $1,600,000 $3,450,000 $2,450,000

Other variable expenses $1,500,000 $300,000 $700,000 $500,000

Total variable expenses $9,000,000 $1,900,000 $4,150,000 $2,950,000

Contribution margin $0 $100,000 ($150,000) $50,000

Traceable fixed expenses $200,000 $50,000 $100,000 $50,000

Service line segment margin ($200,000) $50,000 ($250,000) $0


traceable to individual lines $100,000

Net operating income ($300,000)

Practices

Some factors to investigate further for Office B: • Revenue integrity (e.g. proper coding of visits and supporting documentation) • Compensation structure (e.g. NP to MD ratio) • Practice efficiency (e.g. visits per provider per day and use of tests / supplies)

Segment Reporting Impact

Segment reporting can drive the business to be structured with better clarity.

• Only have common costs where necessary – Examples: CEO compensation, corporate office lease

• Make common costs traceable where possible – Example: Accounting staff of 20 broken out to cover

business units

– Supporting departments can still have central leadership to ensure standardization

• For more see managerial accounting text – Garrison, R. H., Noreen, E. W., & Brewer, P. C. (2008).

Managerial Accounting 12e. New York, New York: McGraw-Hill/Irwin.

Summary

• Challenges to healthcare sustainability (SWOT)

• PI to minimize waste & grow value

• Transforming into a proactive culture is a journey

• Great opportunities to improve operations

• Need to align reporting and structure

Getting Started

• Find a PI Mentor

• Identify a project need

– Culturally manageable

– Strategically important

• Get a Senior Leader Champion

• Pick a PI Methodology (e.g. DMAIC) and follow it