Overview: This work addresses differences in developmental trajectories of ADHD between females and males. In children ADHD is diagnosed at about a 3:1 ration between males and females. This drops off to close to a 1:1 ratio in ADHD in adults. This shift could be explained by a few things, later onset in females, greater persistence of symptoms in females possibly since males tend to display more hyperactive symptoms which can drop off over time. The authors though, state this could be an artifact of the way symptom burden is collected and diagnoses are made ie jumping off chairs is no longer really a thing in adulthood. Inattention symptom burden is more relevant in adulthood and there are less sex differences seen in this type of burden. Inattention my be particularly more burdensome in adulthood when individuals no longer live with parents and are more responsible for their own lives. Several posited explanations but importantly testing developmental trajectories of symptom burden as opposed to diagnostic status can help clarify some of these questions. Evidence from growth mixture modeling which the current study uses is mixed on whether there are sex differences in grouped trajectories of development. The current study attempts to address this by specifically focusing on sex as a moderator of these trajectories.
Methods: Participants - 1,571 individuals (761 female) a representative sample of the population not a clinical sample. Data was collected from participants from age 7-15 Measures - teacher reports of behavior from the social behavior questionnaire (SBQ) collected The measure includes four inattention items and four hyperactivity/impulsivity items. Items have a 5-point response scale which runs from never to very often. Teachers were the same for ages 7-9 then switched however there were no teacher shift effects. Analyses- SBQ items were used to create inattention and hyperactivity latent variables with the four items loading well onto the respective symptom burden variable and ADHD being specified as an oblique factor model of the two. Next, latent growth curve models were fit to the data to test for sex differences in the average trajectories of inattention and hyperactivity symptom burden. Residual convenience between ratings from different teachers were freely estimated. Growth mixture models were then fit to hyperactivity and inattention separately models were fit for 1-6 classes. K-1 models were rejected if K fit better using LMR test. Models were picked for best fit and then descriptively tested.
Results: Hyperactivity - A multi-group latent growth curve model with linear and quadratic growth and no cross-group equality constraints fit well to the hyperactivity/impulsivity scores. The preferred models were significantly different between males and females. The best model for males was three-class model with quadratic growth three classes could be characterized as ‘low stable’, ‘high stable’, and ‘high increasing’. Model fit was a little less straight forward for the female growth but the best fitting model ended up being a three-class model with both linear and quadratic growth the three classes could be characterized as ‘low stable’, ‘high stable’, and ‘concave’. Inattention - Similar to hyperactivity A multi-group latent growth curve model with linear and quadratic growth and no cross-group equality constraints fit well to the inattention scores and fit was significantly worse when it was constrained as equal across both groups suggesting sex differences in best fitting models. For males the best fitting model was two-class model with linear and quadratic growth with two classes could be characterized as ‘low stable’ and ‘high stable. for females the preferred model was a three-class linear and quadratic growth model with three classes ‘high decreasing’, ‘moderate stable’ and ‘low stable’.
Thoughts: Since this is a broad not just clinical sample the low stable and high stable classes are probably the DX/ NO DX versions of groups. It is interesting that for hyperactivity there was a model for each sex that reflected an increase in symptom burden (high increasing for males and concave for females) since these trajectories cap at 15 this still leaves the open suggestion of the author’s that this would taper off in both groups into adulthood. For inattention male symptom burden trajectories still seemed high than females with high decreasing seen for females and not males. Overall this speaks maybe more to decrease in male symptom burden as opposed to females increasing into adulthood. Overall low stable groups were big probably reflecting low normative burden. The big question to me here is what is the role of puberty and hormonal change in these trajectories - the authors bring up an interesting organizational-activational hypothesis with females being more sensitive to ‘activiational’ hormones. The female trajectories such as the concave shape do suggest that influences of puberty extend to hyperactivity and externalizing symptoms.