Baylor Scott & White Health, Temple, TX, United States; Texas A&M School of Public Health, College Station, TX, United States.
Cooperative Studies Program Coordinating Center, Department of Veterans Affairs, Perry Point, MD, United States.
Arch Gerontol Geriatr. 2018 Jan;74:184-190. doi: 10.1016/j.archger.2017.10.014. Epub 2017 Oct 28.
To improve understanding of survival among very elderly male patients with surgically repaired hip fractures, this study applied classification techniques to multiple chronic conditions (MCC) then modeled survival by latent class. Veterans Health Administration (VHA)'s electronic medical records on male inpatients age 85-100 years (n=896) with hip fracture diagnosis and repair were used. MCC defined by Charlson and Elixhauser disorders, medications, demographic covariates, and 5 years follow-up survival were included. Latent Class Analysis (LCA) identified three classes based on patterns of MCC, medications, and demographic covariates: Low-comorbidity (16%), High-longevity (55%), and High-comorbidity (29%). Overall, survival censored at 5 years post-op averaged 717days. The Low-comorbidity group was more likely to be Hispanic, less disabled per VHA determination of eligibility for care, with less risk of postoperative emergency department (ED) visit, and taking no prescription medications. The High-longevity group had longer survival. The High-comorbidity group had more MCC, more prescription medications and shorter survival than the other two groups. Accelerated failure time (AFT) modeled associations between MCC and 5-year survival by class. In AFT models, fewer days until first postoperative ED visit was significantly associated with survival across the three classes. About one in male hip fractured veteran patients over the age of 85 had high levels of MCC and ED use and experienced shorter survival. Hip fracture patients with MCC may merit enhanced post-discharge management. Close investigation targeted to MCC and hip fractures is needed to optimize clinical practices for oldest-old patients in community healthcare systems as well as VHA.
为了更好地理解接受髋关节骨折修复手术的超高龄男性患者的生存情况,本研究应用分类技术对多种慢性疾病(MCC)进行了分析,然后通过潜在类别模型对生存情况进行建模。研究使用了退伍军人健康管理局(VHA)的电子病历,其中包括 85-100 岁男性髋关节骨折诊断和修复患者(n=896),并记录了 MCC 定义的 Charlson 和 Elixhauser 疾病、药物、人口统计学协变量和 5 年随访生存情况。潜在类别分析(LCA)根据 MCC、药物和人口统计学协变量的模式将患者分为三类:低合并症(16%)、高长寿(55%)和高合并症(29%)。总体而言,术后 5 年的生存情况平均为 717 天。低合并症组更有可能是西班牙裔,根据 VHA 确定的护理资格标准残疾程度较低,术后急诊就诊的风险较低,且不服用处方药。高长寿组的生存时间更长。高合并症组的 MCC、处方药物更多,生存时间比其他两组更短。加速失效时间(AFT)通过潜在类别模型对 MCC 与 5 年生存情况的关联进行了建模。在 AFT 模型中,术后首次急诊就诊的天数与三组的生存情况显著相关。大约每 100 名 85 岁以上的髋关节骨折退伍军人患者中就有 1 人患有严重的 MCC 和 ED 就诊,他们的生存时间较短。患有 MCC 的髋关节骨折患者可能需要加强出院后的管理。在社区医疗系统以及 VHA 中,需要针对 MCC 和髋关节骨折进行更深入的研究,以优化针对最年长患者的临床实践。