Bayesian consensus clustering for multivariate longitudinal data

In clinical and epidemiological studies, there is a growing interest in studying the heterogeneity among patients based on longitudinal characteristics to identify subtypes of the study population. Compared to clustering a single longitudinal marker, simultaneously clustering multiple longitudinal markers allow additional information to be incorporated into the clustering process, which reveals co-existing longitudinal patterns and generates deeper biological insight. In the current study, we propose a Bayesian consensus clustering (BCC) model for multivariate longitudinal data. Instead of arriving at a single overall clustering, the proposed model allows each marker to follow marker-specific local clustering and these local clusterings are aggregated to find a global (consensus) clustering. To estimate the posterior distribution of model parameters, a Gibbs sampling algorithm is proposed. We apply our proposed model to the primary biliary cirrhosis study to identify patient subtypes that may be associated with their prognosis. We also perform simulation studies to compare the clustering performance between the proposed model and existing models under several scenarios. The results demonstrate that the proposed BCC model serves as a useful tool for clustering multivariate longitudinal data.     

Effects of tamoxifen on normal breast tissue histological composition: Results from a randomised six-arm placebo-controlled trial in healthy women

Tamoxifen prevents recurrence of breast cancer and is suggested for preventive risk-reducing therapy. Tamoxifen reduces mammographic density, a proxy for therapy response, but little is known about its effects in remodelling normal breast tissue. Our study, a substudy within the double-blinded dose-determination trial KARISMA, investigated tamoxifen-specific changes in breast tissue composition and histological markers in healthy women. We included 83 healthy women randomised to 6 months daily intake of 20, 10, 5, 2.5, 1 mg of tamoxifen or placebo. The groups were combined to “no dose” (0-1 mg), “low-dose” (2.5-5 mg) or “high-dose” (10-20 mg) of tamoxifen. Ultrasound-guided biopsies were collected before and after tamoxifen exposure. In each biopsy, epithelial, stromal and adipose tissues was quantified, and expression of epithelial and stromal Ki67, oestrogen receptor (ER) and progesterone receptor (PR) analysed. Mammographic density using STRATUS was measured at baseline and end-of-tamoxifen-exposure. We found that different doses of tamoxifen reduced mammographic density and glandular-epithelial area in premenopausal women and associated with reduced epithelium and increased adipose tissue. High-dose tamoxifen also decreased epithelial ER and PR expressions in premenopausal women. Premenopausal women with the greatest reduction in proliferation also had the greatest epithelial reduction. In postmenopausal women, high-dose tamoxifen decreased the epithelial area with no measurable density decrease. Tamoxifen at both low and high doses influences breast tissue composition and expression of histological markers in the normal breast. Our findings connect epithelial proliferation with tissue remodelling in premenopausal women and provide novel insights to understanding biological mechanisms of primary prevention with tamoxifen.     

社区“5+1”糖尿病分阶段达标管理对2型糖尿病患者生存质量干预效果分析

目的 分析社区"5+1"糖尿病分阶段达标管理对2型糖尿病患者生存质量的干预效果及其影响因素，为提高患者生存质量提供依据。方法 采用分层整群抽样的方法在山西省、江苏省和宁夏回族自治区选择12个社区卫生服务中心，分别作为干预组（管理方式：社区"5+1"糖尿病分阶段达标管理）、对照组[管理方式：依据《国家基本公共卫生服务规范（2011年版）》的相关要求]，进行为期2年的随访观察。采用面对面问卷调查的方式，收集患者的人口学信息等基本信息；采用健康调查简表（SF-36）对患者在干预前后测量生存质量。采用SAS 9.4软件进行双重差分法以及多重线性回归模型分析。结果 基线时共纳入2 467名研究对象，终末时共1 924人接受了为期2年完整的随访管理。干预后，干预组、对照组患者生理健康维度（PCS）、心理健康维度（MCS）评分变化净差值分别为13.6分、29.8分。多重线性回归分析结果显示，影响患者PCS得分的主要因素有年龄、医保类型、基线PCS得分以及所在地区，影响患者MCS得分的主要因素有年龄、医保类型、基线MCS得分、是否合并高血压以及所在地区。结论 社区"5+1"糖尿病分阶段达标管理对2型糖尿病患者生存质量的干预效果较好。