Rasmussen s s Performance-based Actions. Errors in Radiotherapy. One Example of Error Analysis in Radiotherapy. Errors. Bruce Thomadsen Shi-Woei Lin

Size: px

Start display at page:

Download "Rasmussen s s Performance-based Actions. Errors in Radiotherapy. One Example of Error Analysis in Radiotherapy. Errors. Bruce Thomadsen Shi-Woei Lin"

Louise Robinson
6 years ago
Views:

Errs in Radiotherapy Rasmussen s s Perfmance-based Actions Bruce Thomadsen

l Systematic Errs: Usually one mistake tucked into the procedure Affects

Often found in Process Audit Must be rooted out l Rom Errs: Happen on a

(because of creativity) One Example of Err Analysis in Radiotherapy l We

1 Errs in Radiotherapy Rasmussen s s Perfmance-based Actions Bruce Thomadsen Shi-Woei Lin University of Wisconsin - Madison Slides Bruce Thomadsen Errs l Systematic Errs: Usually one mistake tucked into the procedure Affects all, a large class of patients. Often found in Process Audit Must be rooted out l Rom Errs: Happen on a per-patient basis May be caught through QM Will never be eliminated (because of creativity) One Example of Err Analysis in Radiotherapy l We did a study of brachytherapy errs based on all misadministrations repted to the NRC. l We perfmed several analyses of the events.

Set chamber Set source Caveat l The NRC does NOT keep any recds of physician errs in diagnosing prescribing. That, they say, would be dictating medicine.

2 Set chamber Set source Caveat l The NRC does NOT keep any recds of physician errs in diagnosing prescribing. That, they say, would be dictating medicine. l The only data is on deviations from prescriptions. Analysis l We constructed a process tree f the procedure. l We constructed a fault tree f the procedure. l F each event, we: Contacted the principal got the facts. Constructed a root-cause analysis tree. Marked the position of the failure on the fault process trees. Classified the events using three taxonomies. Unintended Area (P) HRV (R) Identification not crect (P) HRI (R) Execution is erroneous The step size Fail to identify the (parameter) was err wrong The nmal The dosimetrist The step size stayed at The physician dosimetrist who who check was Requirement f a the wrong size when just missed the check the plan left not familiar with manual entry the physicist moved it err with an the program with mouse emergency enough (R) Manual variability (S) External (P) OK (R) Procedure is (R) Spontaneous Interference (R) Excessive dem on increct human variability The computer knowledge/training (S) Inadequate Search would not (S) Bounded Rationality Behavi transfer the file at that size (P) TM (R) Infmation not seen The arrow key sought The physicist did rotate through (S) Distraction not notice it different sizes in the step field (P) TD The physicist not He was familiar with the interrupted program? Example Root-cause- analysis Tree Measurement Calibration 2 Reconstruction Planning quality Application assurance Hardware operation Applicat Software operation Procedures leading to an LDR Patient treatment crect Crect target anatomy Crect applicat placement LDR Brachytherapy Process Tree : Placement followed by dosimetry Procedure Source loading Patient Crectness identification Applicat Satisfaction check Source verification Identity Consistence Placement Fiducials Dummies 4 Anatomy Geometry Completion Identification Crect selection Per localization Applicat Per fixation algithm Data entry Entering data Duration calculation in computer Optimization Targeting 2 Specification facts strength date Prescription Calculation of strength Crect films Interpretation Calibration facts clinical Protocol 2 source stage Image quality integrity Reading Anatomic Identification infmation Placement in well Limitations of Insertion in algithms carrier Dosimetry Dummies Image quality calculation Recd Source Geometry selection source facts patient strength data Recd Setup input Localization Dose / time Source calculation preparation termination Source loading Source fixation Time recding Removal calculation Moniting Successful 9 treatment duration executed Dose Other Tx infmation infmation Recding Emergency response Source Radiation count survey Post Tx moniting Source removal Removal preparation Removal time verification (P) OK (R) Excessive dem on knowledge/traning (S) Incomplete rule (S) Bounded Rationality (R) Distraction from other person

Process Tree Analysis l F HDR By far the most common step with failure was entering the treatment distance, usually not changing the default value.

3 Process Tree Analysis l F HDR By far the most common step with failure was entering the treatment distance, usually not changing the default value. Almost all steps in treatment unit programming delivery had some errs. Dose specification accounted f several errs. The only problems with source calibration were in entering the calibration data into the treatment planning computer. Process Tree Analysis 2 l F LDR (placement followed by dosimetry) Errs in four steps accounted f most of the events:» Selection of the sources,» Loading of sources into the applicat,» Using the required units when entering data into the computer,» Fixing the sources in the applicat, applicat in the patient. Most steps in Source Loading, Dose/time calculation, termination had errs. Process Tree Analysis 3 HDR Fault Tree Prescription err Fractionation failure Accounting err Wrong C Go to Page 2 applicat used A Go to Page 2 B Go to Page 2 l F LDR (Dosimetry followed by placement), errs occurred only in source preparation (usually dering), source delivery (usually a failure to monit). 44/44 Deviation from adequate treatment 43/44 Wrong dose distribution site planning failure 3/44 3/44 implementation failure /44 Wrong patient selected 3/44 Err in treatment D Go to Page 3 planning Verification err 3/44 3/44 Applicat positioning G Go to Page 6 err 2/44 Applicat connection H Go to Page 7 err programming I Go to Page 7 2/44 failure Wrong patient treated /44 delivery err 4/44 J Go to Page 8 /44 Failure to identify patient terminated prematurely L Go to Page Page

F 7/44 Dose calculation err Incompatible facts f calibration dose calculation Wrong incompatible units Wrong patient's data used Increct data transfer /44 Increct data entry 2/44 Wrong dose Wrong

(increct optimization) Inconsistent step size Wrong chart referenced Physician's err Err in transfer (transcription) Data transposition Interpretation err Increct entry 2/44 Physician's err

Deep are Broad 3/44 Source strength /44 err E Dosimetry Err Dose calculation err F 7/44 Go to Page 5 Programming err Wrong calibration 3/44 Wrong source data Wrong data (wrong decay fact) Calibration

4 F 7/44 Dose calculation err Incompatible facts f calibration dose calculation Wrong incompatible units Wrong patient's data used Increct data transfer /44 Increct data entry 2/44 Wrong dose Wrong location 2/44 of dose distribution Dose specification to 2/44 wrong points Wrong dwell positions activated Increct dwell times entered (manual, not optimized) Increct shape of dose distribution (increct optimization) Inconsistent step size Wrong chart referenced Physician's err Err in transfer (transcription) Data transposition Interpretation err Increct entry 2/44 Physician's err Inaccurate 2/44 source position entry QM failure Entry err Inappropriate marker Marker in wrong position Err in specification Increct marker used Some Parts of the Tree Some Parts of the Tree are Deep are Broad 3/44 Source strength /44 err E Dosimetry Err Dose calculation err F 7/44 Go to Page 5 Programming err Wrong calibration 3/44 Wrong source data Wrong data (wrong decay fact) Calibration err Failure of verification 3/44 Erroneous strength f source data Failure of verification 3/44 Wrong data fmat (US/Euro) Increct entry Wrong units Measurement err Calculation err Err in data entry 3/44 Failure to enter alter data (unit default) Wrong source in device Discrepency in strength between device planning system Algithm err Acceptance testing err Page 4 Software err Software version incompatibility File cruption QM failure Page 5 Crupt file Summary from HDR Fault Tree Tabulation l 2/3 errs in delivery 4% of errs due to default value f distance not being changed. Applicat shifting in patient was only other frequent problem in delivery. l /3 errs in treatment planning 6% of errs in calculation, but of various types. 7% of errs due to increct source strength entry. l Almost all events had failures in verification. Summary from LDR Fault Tree Tabulation 2 l /4 errs in treatment planning 5% of errs in calculation,» % due to incompatible units. 8% of errs due to increct source strength entry. l Again, almost all events had failures in verification.

8% because the physician placed the applicat increctly. Analysis Based on Taxonomies l Taxonomies, as you have heard, are dered ganized classifications.

5 Summary from LDR Fault Tree Tabulation l 3/4 errs in delivery % because the patient removed the sources the staff didn t t notice crect. 2% because the sources were never placed in the applicat crectly. 4% because the wrong source strengths were used. 8% because the physician placed the applicat increctly. Analysis Based on Taxonomies l Taxonomies, as you have heard, are dered ganized classifications. l They often can give insight into the nature of the errs occurring. l While we looked at several, developed our own, we will just present two today. Rasmussen s What Happened Pathway Rasmussen s Why It Happened Pathway

6 Rasmussen s Why It Happened Pathway Detection 22 Identification 2 Goal Target 3 Task 8 Procedure What 2 Execution 5 Manual variability 4 Topographic disientation Rasmussen Human Err Model (HDR) 3 Stereotype takeover 7 Stereotype fixation Familiar pattern not recognized 3 Fgets isolated act Mistakes alternatives 3 Other slip of memy 4 Familiar association shtcut How 3 Infmation not seen 9 Infmation assumed 5 Infmation misinterpreted 6 Side effects not adequately considered Other 2 Distraction from system Interferring task 3 Distraction from other person Excessive physical dem 2 Excessive dem on knowledge Instruction increct 2 Operat incapacitated 5 Spontaneous human variability Why Other, specify Taxonometric Analysis From Rasussen s Model: What l F both HDR LDR, noticing the problem was the most significant variable. F HDR, mostly the problem was identifying the problem using verification procedures in place (either they were not adequate not perfmed). F LDR, mostly there were no procedures in place to look f problems. l F both HDR LDR, the next ranking failure was in the execution of procedures, l Followed by the procedures being wrong. Taxonometric Analysis 2 From Rasussen s Model: How l From both HDR LDR, the single most common failure is manual variability. This is probably an artifact of the model, which expects a human reactions to a plant problem. This reflects that the initiating events in medicine is usually some person s s action. l Grouped, Stereotype responses come close. l Infmation not seen, assumed misinterpreted also was significant; f HDR they fmed the dominant failure modes.

Taxonometric Analysis 3 From Rasussen s Model: Why l These categies were not codable f many events. l The most common classification was the catchall Spontaneous human variability.

7 Taxonometric Analysis 3 From Rasussen s Model: Why l These categies were not codable f many events. l The most common classification was the catchall Spontaneous human variability. l Excessive dem on knowledge was significant, particularly f HDR, which is me technical. l Interfering tasks were also imptant in HDR, which in me intensive at a given time. SMART Pathway van der Schaaf et al. 9% SMART Human Err Model (Pinball Method) SMART s Suggested Actions 8% 7% 6% HDR LDR 5% 4% 3% 2% % % External (Technical) Design Construction Materials External (Organizational) Knowledge transfer Protocols Management priities Culture External (Human behavi) Knowlede (Knowledge based) Qualifications Codination Verification Intervention Moniting Slips Tripping Patient related failure Unclassifiable

8 Taxonometric Analysis 4 From the SMART model: The results are very similar f both LDR HDR. l By far, the dominant failure mode was Verification failure,, followed by Intervention (which was sced if someone just goofed). l Inadequate Protocols (i.e., procedures) were imptant, particularly in LDR. Also in LDR, lack of Moniting was a common problem. Taxonometric Analysis 5 From the SMART model: (Continued) l Of about equal imptance, Knowledge transfer (training), Management priities (lack of staffing), Culture (disregard f safety procedures) each showed up as imptant. l Design was a common problem, as as was noted by all the other analyses. Taxonometric Analysis 6 l Very few of the events involved knowledge- based errs. l While the taxonomies tested did give useful infmation, they obviously did not match the medical setting well. Overall Conclusions. Evaluation of a medical procedure using risk analysis provides insights. 2. Failure to consider human perfmance 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 operat to make mistakes that resulted in the erroneous treatments. Of particular danger were those situations where equipment malfunctions fce operats to perfm functions usually executed automatically by machines. Entry of data in terms of units other than those expected by a computer system also accounted f several events.

9 Overall Conclusions 2 3. HDR brachytherapy events tended to happen most with actions having the least time available. 4. LDR brachytherapy, the most hazardous steps in the procedure entailed: selecting the crect sources to place in the patient, setting the sources in place properly in the patient keeping them in place. These events mostly result from lack of attention at critical times. Overall Conclusions 3 5. Many events followed the failure of persons involved to detect that the situation was abnmal, often even though many indications pointed to that fact. 6. Once identified, the response often included actions appropriate f nmal conditions, but inappropriate f the conditions of the event. Overall Conclusions 4 7. Lack of training (to the point that persons involved underst principles) 8. Lack of procedures covering unusual conditions likely to arise ( sometimes, just routine procedures) frequently contributed to events. 9. New procedures, new persons joining a case in the middle also present a hazard. 7/46 evaluable in LDR. 2/38 evaluable in HDR. Overall Conclusions 5. Most of the events suffered from ineffectual verification procedures, a failure noted by all three taxonomies. F the most part, improved quality management would serve to interrupt the propagation of errs by individuals into patient events.

Observations on Common Causes of Events l Failures in medicine parallel those in industry. l Errs don t t just happen from a single cause, but are surrounded by complicating situations.

into categies: l Rom errs in a patient treatment Few calculation errs (where much of QA falls) Frequent errs when treatments are odd (e.g., odd angles used in the wrong direction) Not uncommon following a change in prescription mid- course.

10 Observations on Common Causes of Events l Failures in medicine parallel those in industry. l Errs don t t just happen from a single cause, but are surrounded by complicating situations. l Distraction (due to pressures other assignments) l Rushing (due to pressures lack of staffing) l Lack of communication (between parties) Analysis of External-beam Events The events fall clearly into categies: l Rom errs in a patient treatment Few calculation errs (where much of QA falls) Frequent errs when treatments are odd (e.g., odd angles used in the wrong direction) Not uncommon following a change in prescription mid- course. Not checking patient set-up after pause interruption. Analysis of External-beam Events (continued) l Systematic errs Errs in commissioning calibration (note: the errs themselves are rom, but propagate as systematic). Errs in fmulae Errs in data entry use of increct units Usually there has been no verification check of the data (strange, that we now always check a single patient s s calculation, sometimes several times) Commonalty in Most Events The persons involved often fall into traps, set by the practice environment, respond like human beings.

Steven Sutlief, PhD UC San Diego February 13 th, 2015

Corrective Actions Steven Sutlief, PhD UC San Diego February 13 th, 2015 Objectives By the end of this presentation, the listener should gain A vocabulary to discussing and thinking about corrective actions,