Error Reduction Model in Radiation TherapyJohn W. Sweet, PhD, Urologic Consultants of S.E. PA, Bala Cynwyd, PA

Edward B. Kline, MS, HealthTronics, Inc. / RadPhysics Services LLC, Austin, TX

The software program proved to

be an effective tool for reducing

errors. Process weaknesses

resulted in most errors of clinical

significance. Action plans showed

improvement in problem areas.

1Data for Centers A & B was

annualized for errors identified

9/09 to 9/10 and 2/06 to 3/08,

respectively.

-5

5

15

25

35

45

55

65

75

85

95

38 38

16 14

3 3 3 1 1 0 0 0

56

0

21

0

01

4

1 3

82

53

5

Frequency

Categories

Frequency of Errors - Pre & Post Tx - Center A1

Post-Tx

Pre-Tx

1 Data was annualized for

errors identified 9/09 to 9/10.

-15

5

25

45

65

85

105

125

145

75

2014 13

6 5 3 3 2 1 0.5 0.5 0 0 0 0

8

0.5

2 3

53

0

142

0.5 0 0 0 0

93

45

92

Frequency

Categories

Frequency of Errors - Pre & Post Tx - Center B2

Post-Tx

Pre-Tx

2 Data was annualized for

errors identified 2/06 to 3/08.

0

5

10

15

20

25

30

35

40

01 1

0 0 0

9

0

8

10 0 0 0 0 0

6

3733

27

25

8

5

0

4

3

21 1 0.5 0.5 0.5 0.5

0

Frequency

Attributes

Frequency of Errors : Attributes of Severity Level 1Centers A & B6

Center B

Center A

6 Data for Centers A & B was annualized

for errors identified 9/09 to 9/10 and 2/06

to 3/08, respectively.

Table 1: Error Rates in Treatment Delivery7,8

This Work This Work

Error Software Software Kline Frass Huang Marks Macklis Patton Margalit

Category Center A Center B et al. et al. French et al. et al. et al. et al. et al.

Per Patient, % 0.32 3.20 1.97 1.2 - 4.7

Per Fraction, % 0.01 0.11 0.44 0.32 0.29 0.5

Per Field, % 0.001 0.001 0.13 0.037 0.18 0.17 0.064

Overall Per

Field, % 0.28 a 0.009 a 0.05 a 0.13 b

7 Treatment delivery means the administration of radiation.

8 Data for Centers A and B was annualized for all Tx errors identified from 9/09 to 9/10 and 2/06 to 3/08, respectively.

a Errors per field in the entire post-Tx delivery process (from initial patient consultation to completion of Tx).b Errors per total Tx units.

Table 2: Error Rates in Entire Treatment Process8

Pre-Tx Post-Tx Pre-Tx + Post Tx

Error Center A Center B Center A Center B Center A Center B

Category 115 errors 145 errors 225 errors 362 errors 340 errors 477 errors

Per Patient, % 37.20 10.10 72.80 25.40 81.80 27.33

Per Fraction, % 1.10 0.34 2.10 0.85 2.40 0.92

Per Field, % 0.14 0.004 0.28 0.01 0.31 0.01

8 Data for Centers A and B was annualized for all pre-Tx and post-Tx errors (all aspects of the treatment process from

registration to completion of treatment) identified from 9/09 to 9/10 and 2/06 to 3/08, respectively.

.

9482

33 31 2616 12 12

6 4

25

27.7%

51.8%

61.6%

70.8%

78.4%83.0%

86.5%90.1% 91.7% 92.8%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

0

50

100

150

200

250

300

Frequency

Frequency & Cumulative % of Errors per SubcategoryCenter A4

4 Data was annualized for all errors (pre-Tx and post-Tx)

collected 9/09 to 9/10.Subcategory

94 82

020406080100120

ElectronicPortal Images

CPT/ICD Codes

139

7159 52 45

24 22 21 17 11

43

27.6%

41.7%

53.3%

63.7%

72.6%77.4%

81.8%85.9%

89.2% 91.4%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

0

50

100

150

200

250

300

350

400

450

Frequency

Frequency & Cumulative % of Errors per SubcategoryCenter B5

5 Data was annualized for all errors (pre-Tx and post-Tx)

collected 2/06 to 3/08.Subcategory

Fig 5B

139 71

0

50

100

150

SimulationNotes

CPT/ICD Codes

Table 3: Likelihood of Occurrence

Infractions of Federal/State Regulations per Patient9

Center A Center B

Category 309 patients 659 patients

CMS Billing, % 26.54 a 5.1 b

State Required QA, % 2.59 0.19

State Required Radiation Safety, % 1.62 0.23

9 Data for Centers A and B was annualized for all data collected 9/09 to 9/10 and 2/06 to 3/08, respectively.

a Approximately 80% of the infractions were caught/corrected at time of charge capture and before exporting to CMS

or insurance company.

b Approximately 50% of the infractions were caught/corrected at time of charge capture and before exporting to CMS

or insurance company.

0

20

40

60

80

100

120

140

160

180

200

Frequency

Frequency of All Errors - Center B

Months

• Learning curve

of MERP startup

• Identification

of errors &

violations

• Improved

process, &

MERP action

plans

implemented

• ROs failing to

complete OTV

consult/sim/Tx

notes timely

• Billing mistakes

• Started new

SRS and HDR

programs

• Increased

patient load

• 2 new RO centers

built, startup

• Physics /staff

stretched

• QA missed, billing,

clinical mistakes

•More physics,

therapists & staff hired

• Improved process thru

procedures & training

• Training &

procedures for

SRS

• Assigned HDR

ownership &

physics schedule

• Penalty for RO

report timeliness

implemented

• Billing training

0

10

20

30

40

50

60

70

80

Frequency

Frequency of All Errors - Center A

MonthMonths

• New center startup

process & MERP

learning curve

• High vol. of patients

• Performance issues w/

prior physicist & CT sim

therapist

•Missed/incorr. billing

• MERP audits-priorwkly physics chart

checks & QA missed

• RO left - images not

timely approved

• 9 locum ROs – Docs

missing/late: OTV, sim

notes, consults

• CBCT/kV imager

malfunctioning-

images not timely

approved

• Patient regist. -

emergency nos.

missing• CBCT/kV imager

fixed-images appr.

• Regist. & CT sim

procedure drafted

• Retraining at reg.

office &CT sim

• Increased onsite 3rd

party support

•MERP action plans

implemented & QIC

meeting tasks compl.

• New physicist-improv.

perform./tasks done

• Billing manual/trging

• 9 locum ROs -image

cks, consults & sim

notes missed

• RO image check

lists/trging started

Medical errors in radiation oncology

are receiving increased public and

government scrutiny. This work

examines the results of implementing

a software-based program to reduce

the overall number of incidents,

adverse events, and regulatory

infractions. Audits validated the

effectiveness of the program.

• Software program was implemented

at radiation oncology centers A and B

over 2 years and 1 year, respectively.

• Used to self-identify, categorize,

evaluate, and correct pre and post-

treatment errors and infractions found

in the overall treatment process.

• Errors are classified based on type,

category, attribute, and significance.

Reports follow root-cause analysis.

Errors are automatically routed to

designated reviewers. Links allow

creation of benchmark procedures.

A total of 1,460 (438 pre-Tx and 1,022

post-Tx) errors were identified at both

centers. Centers A and B experienced

0 vs. 2 medical events and 2 vs. 4 near

misses, respectively. Center B had 7

clinically significant errors, defined as a

single fraction dose difference of > than

10% and weekly dose > than15%.

Purpose/Objective

Materials/Methods

Results

Conclusion

Portal imaging, billing, & QA were problem

areas at Center A (Fig 1).

Patient consults/notes, R&V data entry, & billing

errors occurred most at Center B (Fig 2).

Action plans were effective in reducing errors

in process & performance at Center A (Fig 3).

Root cause analysis worked at Ctr. B (Fig 4).

Both centers had high nos. of billing errors (Fig 5A/5B).

Ctr. B produced more significant errors (Fig 6).

Center B experienced 45 errors in treatment

delivery vs Center A only 1 (CBCT)(Table 1).

Higher error rates at Ctr. A due to startup of

new center w/ high patient volume (Table 2).

Center A startup problems result in charge

capture errors & physics turnover (Table 3).

Fig 1

Fig 2

Fig 4

Fig 3

Fig 4

Fig 5A

Fig 6