The
The application
application of
of error
error reduction
reduction
QA
QA philosophy
philosophy in
in HDR
HDR
brachytherapy
brachytherapy
Bruce Thomadsen
University of
Wisconsin Madison
Learning
Learning Objectives
Objectives
Learning objectives: To understand
1. The problem with the current QA
paradigm,
2. The advantage of the new paradigm,
and
3. The application to HDR Brachytherapy
The
The Problem
Problem with
with
the
the Old
Old Paradigm
Paradigm
z What
is the old paradigm?
z If you can check it, check it!
QM
QM Guidance
Guidance
z
Regulations
– Derives from recommendations (below).
– Machine specs from the FDA.
– Regs on users comes from NRC, if using
radionuclides.
– User regs used to be spotty, based on the
most recent incident.
– Now, the new CRCPD Suggested
Regulations incorporate professional
guidelines.
QA
QA Guidance
Guidance
NCRP - often a little late
z ACR - very sketchy
z ACMP - seem to be out of the business
z IAEA - very much derivative of other
guidance documents, except for two
new reports.(IAEA-TECDOC 1494, Case
z
studies in the application of probabilistic
safety assessment techniques to radiation
sources; The other is not yet released.)
Mostly,
Mostly, QM
QM from
from AAPM
AAPM
z
Many Reports from the AAPM (American
Association of Physicists in Medicine) discuss
QM
z
Frequently, the reports have been
incorporated into laws.
QA
QA Guidance
Guidance
z
The main source currently is AAPM reports.
Report 13: Physical Aspects of Quality Assurance in Radiation Therapy (1984)Radiation
Therapy Committee Task Group #24, with contribution from Task Group #22; 63 pp.
Report 24: Radiotherapy Portal Imaging Quality (1987) Radiation Therapy Committee
Task Group #28; 29 pp
Report 41: Remote Afterloading Technology (1993) Remote Afterloading Technology
Task Group #41; 107 pp.
Report 46: Comprehensive QA for Radiation Oncology (1994) Radiation Therapy
Committee Task Group #40; 37 pp.
Report 47: AAPM Code of Practice for Radiotherapy Accelerators (1994) Radiation
Therapy Task Group #45; 37 pp.
Report 56: Medical Accelerator Safety Considerations (1993) Radiation Therapy
Committee Task Group #35; 15 pp.
Report 59: Code of Practice for Brachytherapy Physics (1997) Radiation Therapy
Committee Task Group #56; 42 pp.
Report 61: High Dose-Rate Brachytherapy Treatment Delivery (1998) Radiation
Therapy Committee Task Group #59; 29 pp.
Report 62: Quality Assurance for Clinical Radiotherapy Treatment Planning (1998)
Radiation Therapy Committee Task Group #53; 57 pp.
Report 83: Quality assurance for computed-tomography simulators and the computedtomography-simulation process (2003); 31pp. Radiation Therapy Committee Task Group
#66
And
And of
of those
those Reports
Reports
z
z
z
z
Okay, the first three are archaic.
This list doesn’t count dosimetry or procedure
reports.
But that leaves a bunch. So far, the TG 40
and TG 45 have frequently been carved in
stone in regulations.
For brachytherapy, Reports 59 (TG 56) and
61 (TG 59) are of interest in particular.
The
The Problem
Problem
z
z
z
z
z
The reports, except for TG 53 (QA for
Computer Planning Systems) are prescriptive.
(TG 53 is comprehensive, and only partly
prescriptive.)
As noted, the premise: If it can be checked,
check it.
That is…everything.
Almost no facility has the personnel to
completely do it all.
And, the QA may not prevent errors!
Another
Another Problem:
Problem: The
The
reports
’t keep
reports can
can’t
keep up
up
z
z
z
Since the reports, new technologies, such as
3D conformal, IMRT, 4D motion-correction
systems, Image-based localization all have
come on line.
All of these have QM needs.
What’s a physicist to do?
The
The New
New Paradigm
Paradigm
The new paradigm approaches the problem
differently, in an organized way.
z Map out (understand) the process
z Determine the potential failures, with
particular attention to human failures.
z Map out the potential failures along with
protections.
z See when to interrupt the propagation of
failures.
TG
TG 100
100
z
z
z
z
AAPM created TG 100 to update TG 40 for
new modalities.
It soon became evident that this was making
a hard situation even worse.
The TG decided to take a different tack.
The new approach would be based on risk
assessment.
IAEA
IAEA Report
Report
One of the recent IAEA reports looked
at risk assessment and concluded that
most effort went into assessing
equipment, while most events stemmed
from human errors.
Human
Human and
and Mechanical
Mechanical
Failures
Failures
z
z
Even when events have followed mechanical
failures, the human response has lead to the
injuries.
Witness that the same failure could happen in
two facilities leading to an event in one and
no problem in the other (for example an
Omnitron source breaking off in a patient).
Risk
Risk Assessment
Assessment
z
z
z
Regulations lately like to be “risk based” or
“risk informed”.
Unfortunately, the regulations have little to
base their risk on except when “something
has happened,” that is, if it has happened in
the past, it must have a high risk.
There are techniques for assessing risk, and
TG 100 is using some of them.
Process
Process Map
Map or
or Tree
Tree
z
z
z
The goal is to understand the process and
how all the steps interrelate.
Many ways to map the process.
One common mapping is a Process Tree.
LDR Brachytherapy Process Tree 1:
Placement followed by dosimetry
Source
loading
Planning quality
assurance
Reconstruction
Application
Hardware operation
Applicator
Patient
identification
Correctness
Software operation
1
Applicator
check
Satisfaction
1
Procedure
correct Correct
target anatomy Correct
applicator
placement
Identity
Placement
Fiducials
Source verification
10
Consistence
Source loading
4
Dummies
Source fixation
Anatomy
Geometry
Identification
Correct
selection
Applicator
fixation
Per
algorithm
1
factors
Targeting
strength
clinical
Protocol
stage
Anatomic
information
Image quality
Reading
Identification
Placement in well
Dummies
Image quality
Record
Set source
1
Record Setup
Measurement
2
Localization
Successful
Treatment
treatment duration
executed
Dose
Other Tx
information information
Correct films Interpretation
Calibration factors
Set chamber
9
Prescription
Calculation of strength
1
Monitoring
Duration calculation
Optimization
2 Specification
date
Calibration
1
Data entry
Entering data
in computer
Removal
calculation
1
Per localization
1
Procedures
leading to an
LDR Patient
treatment
Time recording
Completion
3
Geometry
source
strength
Dose / time
calculation
Limitations of
algorithms
Dosimetry
calculation
8
3
factors
patient
data
input
Recording
Emergency response
2 source
integrity
Source Radiation
count
survey
Insertion in
carrier
1
10 Source
selection
Removal preparation
1
Source
preparation
1 Post Tx
monitoring
Source removal
Removal time
verification
Treatment
termination
The numbers give the failures at that location in a study
EPID imaging Place patient on table
For localization
Align all
3x7x4=84
Verify beam outlines 4x5x5=100
3x7x4=84
b
Align mold marks
2x7x2=28
marks
3x7x4=84
Select beam in
4x5x5=100
record & verify
Image 4x5x5=100
Pt in mold
3x7x4=84
a
Make AP image 1x2x1=2
Set
Make
exposure mu
1x2x1=2
Set gantry
Set field size
Register beam
outline c plan
Set machine 1x2x1=2
Verify clearance
and achievability
2x5x2=20
17
4x5x5=100
3x5x1=15
Make lateral image 1x2x1=2
Set
Make
exposure mu
Set
parameters
Repeat for
each beam
Set gantry
b
Set machine
Verify images are adequate
Day 1 imaging verification
1x2x1=2
Approve patient position
Approve treatment
if good
Reimage if necessary
a
3x6x4=108
Determine patient
Shifts and rotations
2x5x2=20
4x5x5=100
Verify patient
setup
3x6x4=108
Review beam images
Register EPID and
4x5x5=100
pseudoradiograph
Load
Pseudoradiograph
4x5x5=100
Load
EPID
2x4x3=24
2x5x4=40
Review setup images
2x6x4=48
Approval to treat
These numbers for this step in IMRT are from an FMEA
TG
TG 100
100 and
and FMEA
FMEA
z
z
z
TG 100 is performing a FMEA for IMRT and
HDR brachytherapy.
It’s taking a while.
Here is a sample:
Sample
Sample FMEA
FMEA Topic
Topic
Step
Specify images
for target and
structure
delineation, etc
Potential Failure
Modes
Specify use of
incorrect image
set
Potential Cause of
Failure
Ignorance of available
imaging studies
Viz.; wrong phase
of 4D CT selected
for planning;
wrong MR for
target volume
delineation
Miscommunication
Potential Effects of
Failure
Wrong anatomical
model (leading to
systematic geometric
and dosimetric errors)
O
S
D
RPN
Comments
8
8
8
512
4D CT gating.
Wrong anatomical
model (leading to
systematic geometric
and dosimetric errors)
8
9
3
216
Ambiguous labeling of
image sets
Inadequate training
Software error
User error
Specify protocol
for delineating
target and
structure
Incomplete/
incorrect list of
specified
structures and
corresponding
image sets
Ignorance of available
imaging studies
Miscommunication
Ambiguous labeling of
image sets
Lack of explicit
protocol
User error
Risk
Risk Probability
Probability Number
Number
z
z
z
z
z
z
z
O = likelihood of
•O •S •D •RPN
occurrence;
S = severity of the effects of
the failure;
D = likelihood failure would go undetected.
RPN = risk probability number = product of
OxSxD.
Values for O, S, and D between 1 and 10,
(1 = low danger, 10 = high).
In industry, RPN <125, little concern, however, in
medicine, RPN > 20 might warrant some
consideration.
Hazard
Hazard
z
z
z
Going through the exercise makes one
wonder how we ever get a case right.
It also takes a long time.
But it helps direct resources to the greatest
hazard.
Fault
Fault Trees
Trees
z
z
z
z
From the FMEA, construct fault trees.
They start with a failure mode and work
backwards asking, what could have caused
the failure?
Then ask, what could have caused that
failure?
And you keep asking until you reach the end
of your universe (where you no longer can
control the causes).
0
0
HDR Fault Tree
Fractionation
failure
Wrong
applicator used
C
Wrong dose
distribution or
site
or
44/44
Deviation from
adequate
treatment
13/44
Treatment
planning failure
13/44
1/44
1/44
Wrong patient
treated
Treatment
implementation
failure
B
Go to Page 2
Error in
treatment
planning
D
Go to Page 3
Verification
error
3/44
Applicator
positioning
error
G
Go to Page 6
2/44
Applicator
connection
error
H
Go to Page 7
Treatment
programming
failure
21/44
I
Go to Page 7
Treatment
delivery error
J
Go to Page 8
Treatment
terminated
prematurely
L
Go to Page 10
or
Wrong patient
selected
and
4/44
1/44
Accounting
error
and
or
30/44
Go to Page 2
Go to Page 2
13/44
43/44
A
or
0
0
Prescription
error
Failure to
identify patient
0
Page 1
Incompatible
factors for
calibration and
dose calculation
Wrong or
incompatible
units
Wrong chart
referenced
or
Physician's error
Wrong patient's
data used
Incorrect data
transfer
Error in transfer
(transcription)
or
Interpretation
error
1/44 Incorrect data
entry
2/44
2/44
7/44
F
Dose calculation
error
Data
transposition
Incorrect entry
Wrong dose
or
Wrong location
of dose
distribution
or
Inaccurate
source position
entry
QM failure
or
2/44
Dose
specification to
wrong points
Wrong dwell
positions
activated
Entry error
or
Incorrect shape
of dose
distribution
(incorrect
optimization)
Software error
Page 5
Inappropriate
marker
2/44
Marker in wrong
position
Incorrect dwell
times entered
(manual, not
optimized)
Inconsistent step
size
2/44
Physician's error
Some
Some
Parts
Parts of
of
the
the Tree
Tree
are
are
Broad
Broad
Error in
specification
or
or
Incorrect marker
used
Programming
error
Algorithm error
and
Software version
incompatibility
and
Corrupt file
Acceptance
testing error
File corruption
QM failure
Some
Some Parts
Parts
of
of the
the Tree
Tree
are
are Deep
Deep
Wrong units
Calibration error
Wrong
calibration
and
3/44 Source strength
Failure of
verification
E
or
error
Dosimetry Error
3/44 Wrong source
data
Failure of
verification
F
Go to
Page 5
Wrong data
format (US/Euro)
Wrong data
(wrong decay
factor)
and
Incorrect entry
Page 4
Error in data
entry
3/44 Failure to enter
or
or alter data (unit
default)
Wrong source in
device
3/44
7/44
Erroneous
strength for
source data
and
or
Dose calculation
error
Measurement
error
Calculation error
3/44
10/44
or
Discrepency in
strength between
device and
planning system
What
What the
the Fault
Fault Tree
Tree Tells
Tells
z
z
z
The fault tree shows how failures propagate
to become an event.
Interrupting the propagation path can (may)
stop the failure from propagating.
Quality management tools act to intercept the
failure.
Quality
Quality Management
Management
QM consists of many parts (or should).
z Two parts we will discuss today are
quality control (QC) and quality
assurance (QA).
z QC forces quality on the procedure;
z QA demonstrates a level of quality.
z A diagram can help understand the
difference.
z
Organizational
Organizational Difference
Difference
between
between QA
QA and
and QC
QC
Input
1
Input
2
Input
3
Input
4
QC
QC
Process
QA
QC
QC
In
In General
General
z
z
z
QC makes sure the inputs to a process are
correct before they enter.
QA looks at the output and evaluates if it is
correct.
Of course, most processes outputs become
inputs for other processes, so one process’s
QA becomes another’s QC.
Example:
Example:
Calculation
Calculation Error
Error
Tree
Tree with
with QM
QM
Error in Input
Data
Error in QC
Error in Data
Entry
Error in QC
Error in
Calculation
Error in
QA
Error in
Calculational
Algorithm
Error in QC
Error in
Prescription
Error in QC
Priorities
Priorities for
for QM
QM Tools
Tools
Forcing functions and constraints
• Interlock
• Barriers
• Computerized order entry with feedbac k
Automation and computerization
• Bar codes
• Automate monitoring
• Computerized verification
• Computerized order entry
Protocols, standards, and information
• Check off forms
• Establishing Protocol / Clarify Protocol
• Alarms
• Labels
• Signs
• Reduce s imilarity
Independent double check systems and other
redundancies
• Redundant mea surement
• Independent review
• Operational Checks
• Comparison with standards
• Increase monitoring
• Add status check
• Acceptance te st
Rules and policies
• External Audit
• Internal Audit
• Priority
• Establishing / Clarify Communication Line
• Staffing
• Better Scheduling
• Mandatory Pauses
• Repair
• PMI (Preventive Maintenance In spection)
• Establish and Perform QC and QA (Hardware and
Software)
Education and Information
• Training
• Experience
• Instruction
Environment
• (Environmental Controls) Sound Control
• (Environmental Controls) Cleaning
• (Environmental Controls) Neatening
• (Environmental Controls) Isolation
• (Environmental Controls) Visual Control
• Environmental Design
[Rank the effectiveness of tools based on the suggestion
of ISMP]
Frequency
Frequency for
for QM
QM
z
z
QC — every time a procedure is performed
QA — with a period such that the worse
possible conditions for which the QA
screens would produce no harm.
Comparison
Comparison of
of QA
QA and
and QC
QC
z
z
z
z
z
QC procedures often require more resources
than QA (to cover large numbers of inputs) but
failures detected by QC less costly to correct.
Relying on QA can add time to a procedure
since failures detected often require repeating
the process.
QA and QC work best together.
If your QA picks up a lot of errors, you need to
move resources to the QC;
If your QA does not pick up occasional errors, it
may be time to eliminate it.
Application
Application of
of the
the Principle:
Principle:
Annual
Annual Linac
Linac Calibration
Calibration
z
z
z
The annual calibration takes several days to
complete.
If everything checks out, the effort was mostly
wasted – that is, it could have been spent
checking things with a higher likelihood of
failure.
If some problem was found, how long had it
been wrong and shouldn’t it have been found
earlier?
Application
Application to
to HDR
HDR Brachytherapy
Brachytherapy
z
z
Recognize that all events have been due to
human performance failures.
Many have been initiated or complicated by
machine failures.
HDR
HDR Machine
Machine Failure
Failure
Examples
Examples
z
Omnitron source breaking off in a patient
Developed into an event in one facility but not in
another.
HDR
HDR Unit
Unit Problems
Problems Leading
Leading
to
to Human
Human Failure
Failure
z
Treatment program card writer malfunction
led to manual programming error.
Programming
Programming Error
Error
z
During the programming, the physicist
entered 260 second for one dwell position
instead of 26. Seconds
7
4
8
5
2
9
6
3
1
=
0
Calculator
.
7
4
8 9
5 6
1 2 3
0
HDR Unit
.
2 3
5 6
8 9
0
#
*
Phone
1
4
7
HDR
HDR Unit
Unit Problems
Problems Leading
Leading
to
to Human
Human Failure
Failure
z
z
z
Treatment program card writer malfunction
led to manual programming error.
Change in length of source and transfer tubes
led to delivering the treatment to the wrong
location.
Default setting for the length caused many
treatments to the wrong location.
HDR
HDR QM
QM -- Commissioning
Commissioning
z
Commissioning
– Measure lengths and compatibility of transfer
tubes and applicators.
– Test the x-ray markers and distance rulers.
– Know where the first dwell position lies in all
applicators and the appropriate lengths to use
(particularly for interstitial cases).
– Make sure that the units for source strength are
what will be entered.
HDR
HDR QM
QM –– Daily
Daily Unit
Unit Checks
Checks
Pay particularly attention to those aspects
that:
z Have had reported problems in the past, or
z Would likely lead to an event if a failure
occurred.
Low priority items are those that never have
caused a problem, or if they fail, would
have little impact.
HDR
HDR QM
QM –– Daily
Daily Unit
Unit Checks
Checks
z
High Priority
–
–
–
–
–
z
Distance
Emergency off
Door interrupt
Transit time
Radiation detectors
Low Priority
– Calibration
– Treatment Interrupt.
HDR
HDR QM
QM –– Source
Source Change
Change
z
High Priority
–
–
–
–
z
Those high-priority checks from the daily QA,
Calibration of the source,
Entry of the source strength into computers, and
Check of the x-ray markers for damage.
Low Priority (required, however)
– Radiation survey around the room,
– Transfer tube length measurement (particularly if
use long markers).
Source
Source Change
Change QA
QA
z
z
The calibration of the source and the entry of
the calibration into the treatment planning and
treatment unit computers should be checked
since they affect all patients.
A review by someone else forms the best
check.
HDR
HDR QM
QM –– Per
Per Patient
Patient
z
z
Almost all events occur in this arena.
Items to check
–
–
–
–
z
Applicator function,
The treatment plant,
The treatment unit program,
Connections between the applicator and the unit.
The treatment plan is best checked by
someone other than the person who
generated it.
Things
Things to
to Check
Check on
on aa Plan
Plan
Things that would affect the treatment
z The prescribed is filled:
– Correct dose to the
– Correct location
z
Shape of the dose distribution achieves desire
– Prescribed isodose surface covers target volume
– Dose distribution has appropriate homogeneity
z
z
z
Critical organs remain below tolerance
Plan correctly transferred to program
Plan contains no errors
Independent
Independent Verification
Verification
z
z
z
Key word: Independent!
Can be a second person
Can it be running on a second computer?
– Yes, if the input is completely new.
– No, it the information from the first is simply
entered into the second (this just checks that the
algorithms work the same - something that should
have been done at acceptance testing).
z
The errors we are looking for are in the
inputs, not in the computer’s calculation.
Check
Check Treatment
Treatment Times
Times
for
for Consistency
Consistency
z
z
There are several methods in the literature:
see Thomadsen, Achieving Quality in
Brachytherapy.
Das has developed a set of checks for any
application: Das RK, Bradley KA, Nelson IA,
Patel R, Thomadsen BR. Quality assurance
of treatment plans for interstitial and
intracavitary high-dose-rate brachytherapy.
Brachytherapy. 5(1):56-60, 2006
Position
Position Verification
Verification
• Radiographs with dummies that
include the transfer tubes will verify
source positions.
Applying
Applying the
the Approach
Approach
z
z
A study applying the new paradigm looked at
all reported HDR misadministrations.
Some of the conclusions follow.
Findings
Findings Applying
Applying the
the Approach
Approach 11
1.
2.
Evaluation of a medical procedure using risk
analysis provides insights.
Failure to consider human performance in the
design of equipment led to a large fraction of the
events reviewed.
• While the equipment per se did not fail, the design
facilitated the operator to make mistakes that resulted
in the erroneous treatments.
• Of particular danger were those situations where
equipment malfunctions force operators to perform
functions usually executed automatically by machines.
• Entry of data in terms of units other than those
expected by a computer system also accounted for
several events.
Findings
Findings Applying
Applying the
the Approach
Approach 22
3.
4.
5.
HDR brachytherapy events tended to happen
most with actions having the least time
available.
Many events followed the failure of persons
involved to detect that the situation was
abnormal, often even though many indications
pointed to that fact.
Once identified, the response often included
actions appropriate for normal conditions, but
inappropriate for the conditions of the event.
Findings
Findings Applying
Applying the
the Approach
Approach 33
6.
7.
8.
Lack of training (to the point that persons
involved understand principles) and
Lack of procedures covering unusual
conditions likely to arise (and sometimes, just
routine procedures) frequently contributed to
events.
New procedures, or new persons joining a
case in the middle also present a hazard.
Findings
Findings Applying
Applying the
the Approach
Approach 44
9.
10.
Most of the events suffered from ineffectual
verification procedures. For the most part,
improved quality management would serve
to interrupt the propagation of errors by
individuals into patient events.
Important contributors to events are:
a) Rushing due to lack of staffing or equipment
problems
b) Insufficient information
Conclusion
Conclusion
z
z
z
We can no longer manage quality by doing
everything we can think of.
We have to assess risk and pick and choose.
The techniques are not hard, but require
training and practice.