Things I Didn't Know About the Berg Balance Scale
Last weekend I attended a course about outcomes measures and physical therapy. I already knew PTs didn't use as many outcome measures as they should. Now I have an understanding why that is. One reason is lack of knowledge regarding which one to use. Another is inability to interpret the results. It doesn't matter which measure I use. If I don't know what the results are telling me, I've wasted my time.
Another problem is lack of consistency in the way the measures are given and scored. That's understandable with some of the more complex measures. Take the Stroke Rehabilitation Assessment Measure (STREAM). While the scoring is straightforward, the grading of individual responses is complex. There are three choices - no movement, normal movement and something in between. For the in-between category, there are three choices based on deviation, weakness or both. It also fails to indicate if one is better than the other. You almost need a course on how to use the thing before you try.
Then there is the Berg Balance Scale. It is routinely used in PT. There is even a short form for children. Almost everyone has used it at some point. Before this class I thought everyone pretty much did it the same way. Not. We debated how to score almost every item. For example, when performing the chair-to-chair transfer. How do you decide between assisting the patient (score one) or letting the patient do it independently with a chance of falling (score two)? When tandem standing, which leg goes in front? On which leg does the patient stand for unilateral standing? Can the patient have a practice try on those?
The answer for the transfer one depends upon therapist judgment but it should be judged the same way by everyone in the facility. Because the Berg doesn't specify, the patient can use either leg but the uninvolved limb will score higher than the hemiparetic one. That practically eliminates the ability for the patient to show improvement on that item. It also inflates the score because the uninvolved limb will naturally have more stability.
Therein lies the problem. Suppose I test my patient using the involved leg and a few days later someone else does but uses the involved limb. The score could increase by four to six points but nothing will be different. Now suppose that patient moves to another facility and is tested again by someone who uses the involved leg. Now the score will decrease by four to six points and the patient will appear to have gone backwards when in reality probably changed very little. As a result, the Berg score will not reflect the patient's true ability and will show less progress because of the differing scores by different testers.
The Dynamic Gait Index (DGI) and Functional Gait Index are worse. How does one define unsteadiness while walking? Are the scores valid if the patient requires someone to assist with walking? In reality we have two problems; getting PTs to use the outcome measures consistently and getting everyone to use them in a uniform manner. No one at the course had previous training on any of the measures. Everyone learned what another therapist taught them at some time in the past. One woman realized she'd been doing the DGI wrong for several years because she'd been taught how to do it incorrectly.
I'm going to give an in-service at work about the Berg and DGI. That way we can somewhat improve the inter-tester reliability. I'm going to have a bigger problem with the STREAM. Coming to any sort of consensus on that one will be a challenge.