When more conventional approaches lead to inconsistent or inexplicable findings, researchers sometimes find that assessing the variability of an outcome can provide a valuable perspective. But researchers still have a great deal to learn about how much variability is too much, or too little, with regard to lower ext remity issues.
One of the areas in which variability measures may have the greatest value is in the assessment of postural sway after concussion. Traditional measures of postural sway, such as the Balance Error Scoring System, can be useful for differentiating concussed athletes from healthy athletes, but only within a few days of injury. A means of detecting more subtle impairments in postural sway is needed, it seems, to effectively minimize the risks associated with returning to play before a concussed athlete is completely healed.
Researchers at Cincinnati Children’s Hospital are targeting this problem with what’s known as complexity science, which is essentially the study of the variability of variability. In over-simplified terms, complexity science involves plotting the standard deviations associated with postural sway (the first level of variability) and then mathematically describing the shape of the curve defining those standard deviations over time (the second level of variability).
As it turns out, this type of complexity analysis can differentiate concussed youth athletes from their healthy counterparts a mean of 49 days postconcussion, according to results presented earlier this month at the American Physical Therapy Association’s Combined Sections Meeting in Las Vegas. First author Catherine Quatman-Yates, DPT, PhD, received the Excellence in Research Award from the Sports Physical Therapy Section for the work.
Postural sway was significantly less variable in the concussed patients, which suggests an inability to adequately respond to perturbations. Too much variability in postural sway, as is seen in the elderly and other patient populations, is also problematic.
But the term “too much variability” means something quite different when it comes to amputation rates in patients with diabetes. A growing body of research indicates that these amputation rates vary quite widely within the US—sometimes by geographic region, sometimes by payer, sometimes even within the same healthcare system (see “Variable amputation rates in patients with diabetes”).
Diabetes experts lament this variability for two reasons. First, it suggests some patients are receiving a lower quality of care than they deserve. And second, what diabetes experts would really like to see with regard to amputation rates is a variability of zero—meaning the complete eradication of all amputations through effective ulcer healing and limb salvage efforts.
An amputation-rate variability of zero is an admirable goal, and one that’s certainly easy to understand in mathematical terms. But amputation is not always the least desirable outcome, even in patients with diabetes, and does not always reflect substandard care.
Patients are variable, and so are the factors that affect their lower extremity care. The key is to determine the actual relationship between variability and patient health, and to find that ideal middle ground between too much variability and too little.
But that’s often easier said than done. There’s a reason they call it complexity science.