Human Error Assessment and Reduction Technique (HEART)
Also known as: HEART
The Human Error Assessment and Reduction Technique (HEART), developed by Jeremy Williams in 1988 for the nuclear industry, is a structured method for assessing the probability of human error in safety-critical tasks and identifying error reduction strategies. Unlike scales that measure subjective experience (workload, situational awareness), HEART is an analytical tool combining expert judgment, task analysis, and empirical error rates to quantify task-specific error probability and guide human factors interventions in high-stakes operations.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use HEART to assess error probability and prioritize safety interventions in high-stakes domains: nuclear power plant operations, aviation (preflight checks, emergency procedures), surgery, anesthesia, emergency medicine, military command-and-control. Ideal for tasks where human error has severe consequences (injury, equipment damage, financial loss) and where quantitative risk assessment (e.g., Probabilistic Safety Assessment) is required. Also use to evaluate proposed design or procedural changes: estimate current-state error probability, implement a change (e.g., add a checklist, automate a calculation), and re-estimate to project improvement. Less suitable for tasks with ambiguous or immeasurable error (e.g., creative writing, relationship counseling) or low-consequence errors (e.g., typos in email).
Strengths & limitations
- Quantitative risk assessment: Yields a numerical error probability, enabling prioritization of risk-reduction efforts and regulatory compliance documentation.
- Identifies specific error sources: EPC analysis reveals which contextual factors (workload, training, procedures, communication) are the primary drivers of error; targets interventions precisely.
- Empirically grounded: Base error rates and EPC multipliers are derived from historical data, incident reports, and expert consensus, not arbitrary scales.
- Actionable for system redesign: HEART naturally leads to remediation (improve training if training is a high-impact EPC, redesign procedures if procedural clarity is the issue).
- Applicable across domains: Successfully used in nuclear, aviation, medical, and offshore energy sectors; methodology is generalizable.
- Subjective expert judgment: Base rate classification and EPC multiplier assignment depend on analyst expertise and interpretation; different analysts may yield different estimates for the same task.
- Limited empirical validation of multipliers: Some EPC multipliers (e.g., 'inadequate training' ×2.0) are expert-consensus values with sparse empirical validation; real-world multipliers may differ.
- No quantification of error severity: HEART estimates error probability (will an error occur?), not severity (if it occurs, how bad is it?). Risk = probability × severity; HEART addresses only the numerator.
- Task dependency and human factors complexity: Error probability for a task depends on countless factors (operator fatigue, motivation, environment, recent training); capturing all via a set of multipliers is reductive.
- Requires expert analyst: Not a self-administered scale; requires human factors professional to conduct analysis correctly, limiting accessibility.
- Database of generic tasks may not fit novel tasks: If your task doesn't clearly match a generic task type in the HEART database, base-rate selection becomes ambiguous.
Frequently asked
How do I choose the right base error rate for my task?
Consult the HEART database or published tables and select the generic task type that most closely matches your task in terms of: (1) skill level required (routine vs. complex), (2) task frequency (routine vs. infrequent), (3) nature of cognitive demand (memory, problem-solving, perception). If multiple categories seem plausible, calculate estimates for each and report the range, or consult a domain expert (e.g., experienced pilot for aviation tasks). Document your choice and reasoning explicitly.
What if my task has EPCs not in the standard HEART list?
The standard HEART provides ~11 common EPCs; your task may have domain-specific factors. Identify additional EPCs from task analysis, incident reports, or expert interviews. Assign multipliers using expert judgment, analogical reasoning ('This factor is similar to [standard EPC], so use a similar multiplier'), or literature evidence if available. Clearly document any non-standard EPCs and justify multiplier values. Conduct sensitivity analysis to show whether the additional EPCs materially change the error estimate.
Can HEART account for operator expertise and training level?
Yes, through the 'training adequacy' EPC and base-rate selection. A well-trained, expert operator performing a routine task may have base rate ≈0.001; a poorly trained novice on the same task may have base rate ≈0.05 or higher. Additionally, 'inadequate training' and 'insufficient experience' are standard EPCs with assigned multipliers. When comparing different operator groups (expert vs. novice), run separate HEART analyses with appropriate base rates and training EPCs for each.
How do I validate HEART estimates against real-world error data?
Collect empirical error data (from incident reports, near-miss databases, or prospective observation) for the task of interest. Compare observed error rate to HEART estimate. If observed >> estimated, the analysis likely underestimated EPCs; revisit multipliers and expand EPC list. If observed << estimated, possible explanations: (1) the task is performed more carefully than modeled, (2) EPCs are less prevalent than assumed, or (3) undetected errors mask the true rate. Large discrepancies warrant analysis revision and expert consultation.
Should I report a point estimate or a range for error probability?
Always report a range. Given uncertainty in base-rate selection and EPC multipliers, calculate a low estimate (conservative base rate, smaller multipliers) and high estimate (aggressive base rate, larger multipliers). For example: 'Error probability is estimated at 0.025 (range 0.015–0.040).' This transparency allows stakeholders to assess robustness of the estimate and make informed decisions. If a decision hinges on whether error probability is below a regulatory threshold (e.g., <0.01), test whether the entire range satisfies the threshold; if not, additional risk reduction is required.
Sources
- Williams, J. C. (1988). A data-based method for assessing and reducing human error to improve operational performance. In IEEE Fourth Conference on Human Factors and Power Plants (pp. 436-450). IEEE. DOI: 10.1109/hfpp.1988.27540 ↗
How to cite this page
ScholarGate. (2026, June 3). Human Error Assessment and Reduction Technique (HEART). ScholarGate. https://scholargate.app/en/human-factors/human-error-assessment
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- NASA Task Load IndexHuman Factors↔ compare
- Operator Performance Assessment ScaleHuman Factors↔ compare
- Situational Awareness Rating TechniqueHuman Factors↔ compare
- Workload ProfileHuman Factors↔ compare