BACKGROUND Very little is known about specific trajectories or patterns of falls over time. Using the well-characterized cohort of the Osteoporotic Fractures in Men Study (MrOS), we classified individuals by fall trajectories across age and identified predictors of group assignment based on characteristics at baseline. METHODS Using an analysis sample of 5 976 MrOS participants and 15 years of follow-up data on incident falls, we used group-based trajectory models (PROC TRAJ in SAS) to identify trajectories of change. We assessed the association of baseline characteristics with group assignment using 1-way analysis of variance and chi-square tests. Multivariable logistic regression was used to analyze the outcome of the high-risk fall trajectory groups compared to the low-risk groups. RESULTS Changes in rates of falls were relatively constant or increasing with 5 distinct groups identified. Mean posterior probabilities for all 5 trajectories were similar and consistently above 0.8 indicating a reasonable model fit. Among the 5 fall trajectory groups, 2 were deemed high risk, those with steeply increasing fall risk and persistently high fall risk. Factors associated with fall risk included body mass index, use of central nervous agents, prior history of diabetes and Parkinson's disease, back pain, grip strength, and physical and mental health scores. CONCLUSIONS Two distinct groups of high fall risk individuals were identified among 5 trajectory groups, those with steeply increasing fall risk and persistently high fall risk. Statistically significant characteristics for group assignment suggest that future fall risk of older men may be predictable at baseline.

中文翻译：

---

老年男性的跌倒轨迹：按年龄划分的变化轨迹和未来跌倒风险的预测因子。


背景 关于跌倒随时间变化的具体轨迹或模式知之甚少。使用男性骨质疏松性骨折研究 （MrOS） 的明确特征队列，我们按不同年龄的跌倒轨迹对个体进行分类，并根据基线特征确定了分组分配的预测因子。方法 使用 5 976 名 MrOS 参与者的分析样本和 15 年的事件跌倒随访数据，我们使用基于组的轨迹模型（SAS 中的 PROC TRAJ）来确定变化轨迹。我们使用 1 因子方差分析和卡方检验评估了基线特征与组分配的关联。多变量 logistic 回归用于分析高危跌倒轨迹组与低风险组相比的结局。结果 跌倒率的变化相对恒定或增加，确定了 5 个不同的组。所有 5 个轨迹的平均后验概率相似，并且始终高于 0.8，表明模型拟合合理。在 5 个跌倒轨迹组中，2 个被认为是高风险的，跌倒风险急剧增加的那些和持续较高的跌倒风险。与跌倒风险相关的因素包括体重指数、中枢神经药物的使用、糖尿病和帕金森病的既往病史、背痛、握力以及身心健康评分。结论 在 5 个轨迹组中确定了两组不同的高跌倒风险个体，跌倒风险急剧增加和跌倒风险持续高。分组分配的统计学显着特征表明，老年男性未来跌倒的风险在基线时可能是可预测的。