Performance-Based Functional Assessment: An Algorithm for Measuring Physical Capabilities
Kathleen J. Price, Andrew Sears · 2008 · Proceedings of the 10th International ACM SIGACCESS Conference on Computers and Accessibility (Assets '08) · doi:10.1145/1414471.1414511
Summary
This paper presents PB-FACT (Performance-Based Functional Assessment for Computer Technology), a novel tool for objectively and quantitatively measuring an individual's physical capabilities in relation to their use of information technology. The authors argue that current practice in accessibility research — describing participants by medical diagnosis alone (e.g., "has Cerebral Palsy") — is inadequate because people with the same diagnosis can have dramatically different functional capabilities, and people with different diagnoses may have functionally identical abilities when it comes to IT use. Similarly, self-reporting and single-observer assessments are prone to bias. PB-FACT was developed through an iterative process involving a multidisciplinary team including a physiatrist, nurse, ergonomics expert, and HCI professionals. The tool works by having users physically demonstrate everyday motions (grasping a telephone, holding a pen) while wearing a P5 virtual reality gaming glove equipped with finger bend sensors and six-degree-of-freedom motion sensors. The system captures detailed movement data including smoothness, distance, speed, steadiness, and consistency across 14 tasks involving arm, hand, finger, head, and neck motion. Thirty-one participants were recruited, ranging from fully capable individuals to people with conditions including post-stroke, spinal cord injuries, cerebral palsy, Parkinson's Disease, spinal stenosis, dystonia, and multiple sclerosis.
Key findings
The initial validation focused on two tasks — a gripping motion (grasping a tennis ball) and a left-right range of motion task. Forty-five measurement metrics were calculated from raw sensor data, then refined through correlation analysis and collinearity remediation to identify the most predictive variables. The resulting models accounted for 57% to 85% of the variance in observer ratings (the gold standard), with all three validation groups for both tasks producing statistically significant correlations (Pearson R values ranging from 0.752 to 0.923, all p < 0.01). For the grip task, five predictor metrics were identified including pause steadiness and motion smoothness measures. For the range of motion task, three pause-related metrics were most predictive — direction change rate, correlation coefficient, and Pythagorean rate. The strong correlations between PB-FACT's automated measurements and averaged observer ratings demonstrate that performance-based assessment can serve as a viable, objective alternative to subjective evaluation methods.
Relevance
This research addresses a fundamental methodological problem in accessibility research: the lack of standardised, objective ways to describe participants' actual functional capabilities. When studies describe users only by diagnosis, it becomes impossible to compare results across studies, replicate research conditions, or determine which users will benefit from specific assistive technologies. PB-FACT offers a model for moving beyond diagnosis-based descriptions toward a continuum-of-capability approach. For practitioners involved in assistive technology matching, this work highlights why the common trial-and-error process leads to high abandonment rates — without precise capability assessment, the initial technology match is often poor. While the specific hardware used (the P5 gaming glove) is now obsolete, the conceptual framework — mapping IT equipment to interaction devices to operational components to required functional capabilities — remains a valuable model for thinking about accessibility assessment systematically. Modern motion-sensing technology could make this approach far more practical and precise.
Tags: functional assessment · motor disability · assistive technology · physical disability · human-computer interaction · measurement · research methodology · rehabilitation