QA Tools And QI Flashcards
Traceability: Definition
Property of a measurement result which can be related to a reference through a documented unbroken chain of calibrations, each contributing to the measurement uncertainty
Examples of Traceability
Electrometer:
* Needs to have calibration certification prior to use within a department
- Within Australia, ARPANSA can calibrate electrometers for radiotherapy dosimetry
- Calibration is traceable to the Australian Standards of Voltages
Policies for Traceability
- Equipment calibrations must be traceable to a reputable organisation such NATA
NATA = National Association of Testing Authorities, Australia
* Develop policies for traceability when performing tests or measurements
Inter-comparison: Definition
Systematic comparison of measurements, performance, or output between different machines, techniques, or facilities to ensure consistency and accuracy in treatment delivery
Help identify discrepancies, standardise practice and improve treatment accuracy
Process performed on simple equipment that cannot be calibrated
Inter-comparison: Example
Thermometer and Barometer
Up to 8% of the Australian Clinical Dosimetry Service (ACDS) recommendations stem from simple errors in thermometer and barometer calibrations
* Includes up to 4 degrees error in thermometer readings and up to 1% in barometer readings –> errors feed directly into beam output calibration
Mercury Thermometer can be utilised as a backup to confirm temperature shown on Digital Thermometer
Importance of Assessing Room Temperature
Overheating of the treatment machine –> leads to alternation of dose output
Physics utilise an electrometer to determine room temperature on a monthly basis
Phantom: Definition and Common Examples
These are physical objects, often designed to mimic human tissue –> used for calibration and testing
Common Examples:
1. Water Phantoms
2. Solid Water Phantoms
3. Anthropomorphic Phantoms
Dosimeters: Definition and Common Examples
Instruments used to measure radiation dose
Common Examples:
1. Ionisation Chambers
2. Diode Detectors
3. TLD’s
Ionisation Chamber: Definition and Use
Devices used to directly measure the ionisation of air or a liquid by radiation
Highly accurate and commonly used for reference dosimetry
Electrometers: Definition
Tool used to measure and verify the electrical charge produced by ionisation chambers and other detectors
Help to ensure accurate dose measurements
Portal Imaging Devices: Definition
Devices used to capture x-ray images of the patient’s treatment area before or during treatment to verify the alignment of the treatment beam
Quality Improvement: Definition
Quality Improvement is a systematic approach to enhancing the quality of products, services or processes within an organisation
Quality Indicator: Definition
Generally thought of as the most appropriate action for patient treatment given a certain disease and stage
Quality Measure: Definition
Quantitative description of the degree of adherence to a quality indicator
Action Limits: Description
Define the degree to which measured quantities in the clinic are allowed to vary without risking harm to the patient
Tolerance Levels: Definition
Define the boundary within which the process is considered operating normally
Three Methodologies for QI in RT
Failure Mode and Effects Analysis (FMEA)
Incident Learning and Root Cause Analysis (IL-RCA)
Statistical Process Control (SPC)
Failure Mode and Effects Analysis (FMEA)
Failure Modes and effects Analysis (FMEA)
Clinical Team develop a process map
Individual steps of the process map are analysed for ways in which the desired outcome of a step may not be achieved –> known as potential failure modes
For each potential failure mode, three components are assessed and assigned a numerical value from zero to ten:
Severity (possible outcome on a patient)
Occurrence (how likely it is that the failure pathway occurs)
Detectability (how likely is it that the failure pathway, once initiated, will not be intercepted)
Numerical values of the three parameters are multiplied together to calculate the risk priority number
By ranking potential failure modes according to risk priority numbers, FMEA’s enable the clinical team to understand where safety and quality issues could arise and their relative priority.
Originally developed by NASA
Use of AI in the FMEA Process
- Data Analysis and Pattern Recognition
* AI can analyse large datasets of historical failure data and identify patterns and trends that may not be immediately apparent to human analysts
* This can help in the identification of potential failure modes and their associated effects - Risk Assessment
* AI can assist in the risk assessment phase of FMEA by calculating risk scores based on historical data, probabilities, and consequences
* This can help prioritise which failure modes to focus on - Recommendations
* AI algorithms can generate recommendations for mitigating or preventing specific failure modes based on historical data and expert knowledge
* Recommendations can help in the development of effective preventive and corrective actions - Automating Documentation
* AI can assist in generating documentation for the FMEA process, including failure mode descriptions, risk assessments, and recommended actions
* Reduces the administrative burden on FMEA teams - Real-time Monitoring
* AI-driven sensors and monitoring systems can provide real-time data on equipment and processes
* AI can analyse this data to detect abnormal conditions and potential failure modes in real-time, allowing for proactive maintenance and intervention - Continuous Improvement
* AI can help in continuously improve the FMEA process itself by analysing historical FMEA data and suggesting refinements to the criteria used for risk assessment
Incident Learning: Definition
Identification of problems in the care delivery process and the subsequent investigation of those problems to uncover and address causal factors and latent conditions for error
Root Cause Analysis (RCA)
Is a process analysis used to identify the underlying causes of system failures
Provides the information needed to solve problems and address these failures
Comparison between FMEA and RCA
FMEA:
* Pro-active
* Deal with hypothetic failure
* Look forward in time
* Aimed at predicting the the adverse outcomes of various human and machine failures, and system states
RCA:
* A reactive process
* Takes places after the error/harm has been done
* Deal with actual failures
* Look backward in time
Statistical Process Control (SPC): Definition
A method to measure and control variability in processes
Analytical decision making tool that employs statistics to measure and monitor a system process
Fundamental concept of SPC is to compare current statistics in a process with its previous corresponding statistic for a given period
Statistical Process Control (SPC): Steps
- Define the Process
- Select Quality Characteristics
- Collect Data
- Create Control Charts (Graphical Tools)
- Determine Control Points
- Collect and Plot Control Charts
- Analyse and Interpret Control Chart
- Take Corrective Action
- Document and Communicate
Statistical Process Control (SPC): Measurable Endpoints
Might include the following:
- Physics-related quality measures (e.g., LINAC output)
- Clinical practice measures (e.g., time from simulation to planning)
- Patient-related measures (treatment breaks)
SPC: Control Charts
Visually tracks data over time to monitor the stability of a process, helping to identify variations that may indicate issues
By plotting data points on a control chart, one can observe whether variations are within normal limits (indicating stable performance) or if they fall outside predefined control limits (suggesting potential problems requiring investigations)
Upper and lower control limits are set to accommodate for expected variability (these are distinct from action limits that are clinically acceptable)
SPC: Control Charts - RT Specific Example
- Data Collection
* Daily dose output measurements are taken with a QA device, and each day’s result is plotted on the control chart - Setting Control Limits
* Based on historical data, acceptable control limits (e.g., +/- 2%) are set around the baseline dose level - Monitoring
* By plotting daily dose values, physicists can quickly see if the Linac is performing within acceptable limits or if there are outliers - Actionable Insights
* If the dose output strays outside control limits, this signals a need for maintenance or recalibration, preventing potential inaccuracies in patient dose delivery.
SPC: Patient-related Measure Example
Indicator/Process:
* Patient comfort during treatment
Measure
* Percent of patients with pain level scores <1
Process Target (%)
* 95
Action Limits (%)
-5
SPC: Clinical Practice-related Measure Example
Indicator/Process
* Clinical Process Efficiency
Measure
* Time from simulation to first treatment
Process Target
* 5 working days
Action Limits
* +1 working days
Other machine parameters which can also be measured
* beam output
* beam energy
* couch travel time
* minimum percentage of open multi-leaf collimators (MLC) during treatment
Can AI assist SPC for RT QA?
Patient Management System
Example: Mosaiq
* Generate care patterns and operational trends
* Use IQ scripts for data analysis
** IQ Scripts **
A set of software tools lets you modify your Elekta onoclogy information system (OIS) to match your clinic’s specific needs
No programming required
Remote Monitoring and Data Management Systems: Definition
These systems allow for continuous monitoring of treatment delivery and dose verification remotely
Survey Metre: Purpose
Can be used as a dose counter for brachytherapy
Hand-held ionising radiation measurement instruments are used to check personnel, equipment and the environment for radioactive contamination
