GM Function curves in CP

[Marie Raven]

I am probably getting way in the weeds, but I am trying to understand case based exam questions 35-37. First of all, I think there is a typo because the Hanna 2008 reference led me to this article titled “Reference Curves for the Gross Motor Function Measure: Percentiles for Clinical Description and Tracking Over Time Among Children With Cerebral Palsy” (https://academic.oup.com/ptj/article/88/5/596/2742409?login=false) which is from 2008. When I checked in Campbell, Figure 22.4 in the 6th edition references “Hanna SE, Rosenbaum PL, Bartlett DJ, et al. Stability and decline in gross motor function among children and youth with cerebral palsy aged 2 to 21 years. Dev Med Child Neurol. 2009;51:295–302.” This article appears to describe the gross motor curves intended to be referenced in question 37.

In the 2009 Hanna article, the authors state: “Table I shows mean GMFM-66 by GMFCS and age. Although this simple summary uses crude age bands, and does not account for the multiple observations per child, it suggests that the nonlinear trend in GMFM-66 over age may differ by GMFCS level. The trend in means shows no evidence of peak and decline for Levels I and II, whereas there is a suggestion of such a decline for levels IV and V. The trend for level III is not clear.” Yet they go on to statistically model that it is possible to account for declines in scores for Level III and show this derivation in the chart that we’re all supposed to reference. Not sure I understand that, because those curves don’t match the means presented in Table 1, but I’m thinking the utility of trying to understand that further in the context of passing the PCS exam is not necessary.

What I would like to understand better is applying confidence intervals, SD, SEM, and Minimal Detectable Difference to data. Hanna’s Table 1 shows SD and 95% CIs. For children at Level III Hanna’s data shows an increase in SD with increasing age bands. The CIs also begin to overlap and are not as different as the 2-6 and 6-9 age bands. In the absence of whatever statistical model Hanna applied to the data to create the curves, the raw data for Level III does not appear to show significantly different results from ages 9-12, 12-16, and 16-21 where there is a dip and slight recovery of the average scores.

1. Why does the answer explanation for question 35 focus on the difference in the change of score (+2.59) from assessment 5-6 (53.09-52.85= +.24) in relation to the difference in change from assessment 6-7 (55.92-53.09= +2.83) as a means of measuring the change in the outcome measure, and not just the difference in results from assessments 6 & 7 (the 2.83 value)?
2. Regardless of whether we are using 2.59 or 2.83, this value is better than the SEM of 1.17, which should indicate that our score has a higher reliability. SEM is not something we can use on its own to determine if our result is showing real change though, correct?
3. To determine change that is not attributable to error, we need to exceed MDD (which we calculate using SEM) or non-overlapping CI intervals correct?
4. Given the example in question 37, I was looking for an answer option that indicated there was an expected plateau in the child’s skills (not that he needed more therapy). Given that we have previously determined that he made measurable change, but not enough to exceed the threshold for measurement variability or other error, why would we say that he is showing a positive trend compared to others in the Level III classification?

[Jessica Lewis]

Hi Marie!

All great questions.
1. You could look at it either way. As long as you identify that there was an increase in score, that’s the important thing.
2. Correct, the score is reliable, but still does not exceed the threshold for test error. The CI and SEM need to both be taken into account when you are looking at everything and making decisions about change scores.
3. Correct!
4. For this one, the 2008 article was the one we intended to reference. The 2009 article can be a helpful resource, too. If you do consider the 2009 article, you see that a decline in skills for children in GMFCS level III occurs around 7 years, 11 months. The child in this case is 9 years old and is still making improvements in score and percentile. It is trending in the right direction even though we can’t say it is real change. The next score, if he returns to clinic, will be the most telling to determine if he continues to trend in a positive direction or if his scores start to plateau. For this one you are picking the best answer and identifying the positive trajectory in score.

[Helen Carey]

Yes, great questions! The SEM can be a bit confusing. Some researchers use the SEM equivalent to the MDC/MDD, but the preferred calculation for the MDC/MDD is MDC (95) = 1.96 × √ 2 × SEM where 1.96 is the z score associated with the 95% confidence level and SEM includes the pooled standard deviation (SD) and the ICC. Therefore, the SEM is included it the calculation, but not the absolute value. Relative to CP, we know that children in GMFCS level III are more variable in their motor performance compared to levels I and II. Therefore, it is not a surprise to see higher SEMs for outcome measures in level III compared to levels I and II.

We hope these responses help in your journey through the weeds!
Helen

[Author]

Posts