Gomez Rapson, Skilbeck Clive, Thomas Matt, Slatyer Mark
School of Health Sciences, Federation University, BallaratVIC, Australia.
Psychology, School of Medicine, University of Tasmania, HobartTAS, Australia.
Front Psychol. 2017 Aug 22;8:1320. doi: 10.3389/fpsyg.2017.01320. eCollection 2017.
Growth Mixture Modeling (GMM) was used to investigate the longitudinal trajectory of groups (classes) of depression symptoms, and how these groups were predicted by the covariates of age, sex, severity, and length of hospitalization following Traumatic Brain Injury (TBI) in a group of 1074 individuals (696 males, and 378 females) from the Royal Hobart Hospital, who sustained a TBI. The study began in late December 2003 and recruitment continued until early 2007. Ages ranged from 14 to 90 years, with a mean of 35.96 years ( = 16.61). The study also examined the associations between the groups and causes of TBI. Symptoms of depression were assessed using the Hospital Anxiety and Depression Scale within 3 weeks of injury, and at 1, 3, 6, 12, and 24 months post-injury. The results revealed three groups: low, high, and delayed depression. In the low group depression scores remained below the clinical cut-off at all assessment points during the 24-months post-TBI, and in the high group, depression scores were above the clinical cut-off at all assessment points. The delayed group showed an increase in depression symptoms to 12 months after injury, followed by a return to initial assessment level during the following 12 months. Covariates were found to be differentially associated with the three groups. For example, relative to the low group, the high depression group was associated with more severe TBI, being female, and a shorter period of hospitalization. The delayed group also had a shorter period of hospitalization, were younger, and sustained less severe TBI. Our findings show considerable fluctuation of depression over time, and that a non-clinical level of depression at any one point in time does not necessarily mean that the person will continue to have non-clinical levels in the future. As we used GMM, we were able to show new findings and also bring clarity to contradictory past findings on depression and TBI. Consequently, we recommend the use of this approach in future studies in this area.
生长混合模型（GMM）被用于研究一组1074名遭受创伤性脑损伤（TBI）的患者（696名男性和378名女性，来自皇家霍巴特医院）抑郁症状群组（类别）的纵向轨迹，以及年龄、性别、严重程度和创伤性脑损伤后住院时长这些协变量如何对这些群组进行预测。该研究始于2003年12月下旬，招募工作持续至2007年初。年龄范围为14至90岁，平均年龄为35.96岁（标准差=16.61）。该研究还考察了这些群组与创伤性脑损伤病因之间的关联。在受伤后3周内以及受伤后1、3、6、12和24个月，使用医院焦虑抑郁量表对抑郁症状进行评估。结果显示有三组：低抑郁组、高抑郁组和延迟抑郁组。在低抑郁组中，创伤性脑损伤后24个月内的所有评估点，抑郁评分均低于临床临界值；在高抑郁组中，所有评估点的抑郁评分均高于临床临界值。延迟组在受伤后12个月时抑郁症状增加，随后在接下来的12个月内恢复到初始评估水平。发现协变量与这三组存在不同的关联。例如，相对于低抑郁组，高抑郁组与更严重的创伤性脑损伤、女性以及较短的住院时长相关。延迟组的住院时长也较短，年龄较小，创伤性脑损伤程度较轻。我们的研究结果表明抑郁随时间有相当大的波动，并且在任何一个时间点处于非临床水平的抑郁并不一定意味着该患者未来仍将处于非临床水平。由于我们使用了生长混合模型，我们能够展示新的研究结果，也能澄清过去关于抑郁与创伤性脑损伤的相互矛盾的研究结果。因此，我们建议在该领域未来的研究中使用这种方法。
