循环系统病例死亡风险分级研究

目的探讨病例死亡风险分级方法,为提高医院综合诊治和管理水平提供方向,为医疗质量评价提供参考。方法以9008例循环系统住院病例为例,以主诊断转归为目标变量,应用卡方自动交互作用检测法CHAID进行分组,并对急性心肌梗死组分年度进行初步评价。结果样本共计分为35个不同死亡风险的组,归入零风险组、超低风险组、低风险组、高风险组和超高风险组五级。其中急性心肌梗死组,低风险组和超高风险组各年度死亡率不同,说明利用零风险组、超低风险组和低风险组患者群死亡率进行基础医疗质量评价更为合理和敏感,而利用高风险组和超高风险组患者群死亡率可进行医院危重救治能力和综合管理水平方面的评价。结论可将病例进行死亡风险分级,并用于科间、院间或不同时段医疗质量评价。     

冠心病患者痰热互结证CHAID决策树识别模式的研究

目的:建立冠心病不稳定性心绞痛患者临床常规检测指标对痰热互结证的识别模式。方法:选取2010年4月至2011年4月多中心收集的冠心病不稳定性心绞痛患者411例的基本资料、中医四诊信息及临床常规检测指标进行归一化处理后,采用CHAID决策树方法从90个临床常规检测指标中自动提取痰热互结证的识别规律。对其中212例患者进行痰热互结证识别模式的外验证。结果:Cl离子、缩短分数、RDW-CV、血常规RBC、D-Ⅱ聚体、CK-MB、PTA和BUN共8个属性指标经筛选后进入决策树识别模型。该模型对411例患者的测试结果显示:敏感度为75.0%,特异度为86.9%,检验准确率为86.1%。外验证模型缺失RDW-CV,模型识别准确率为85.8%,敏感度为89.5%,特异度为85.5%。结论:临床常规检测指标经CHAID决策树方法筛选后,可以直观、清晰的进行冠心病不稳定性心绞痛患者痰热互结证的识别,自动归纳识别规律,在中医证型-生物学指标对应模式的数据挖掘中具备一定的优势。     

Relation between pressure, friction and pressure ulcer categories: a secondary data analysis of hospital patients using CHAID methods

Background

According to the latest pressure ulcer definition provided by the EPUAP and NPUAP pressure and shear are named as factors causing pressure ulcers. Empirical evidence suggests that pressure forces in combination with shear seem to be primarily responsible for deeper tissue injuries leading to category III or IV pressure ulcers. Superficial frictional forces seem to cause skin lesion resembling category II pressure ulcers.

Objectives

The objective of this study was to explore the empirical relationships between friction forces and category II pressure ulcers and between pressure forces and categories III and IV pressure ulcers.

Design

A secondary analysis of data from six German annual hospital pressure point prevalence studies.

Settings

161 Hospitals of all specialties and categories throughout Germany.

Participants

28,299 Adult hospital patients. The average age was 65.4 (SD 17.0) years. Female participation was 55.0%.

Methods

For the classification of the sample regarding pressure ulcers as a dependent variable and the Braden scale items as predictor variables, Chi-square Automatic Interaction Detection (CHAID) for modelling classification trees, controlled for age, has been used. CHAID analysis was performed for category II pressure ulcers and categories III/IV pressure ulcers separately.

Results

7.5% (95% CI 7.2–7.8) of the hospital patients had “Friction & Shear” problems according to the respective Braden sale item. 5.4% (95% CI 5.1–5.6) were “Completely immobile” according to the Braden scale item “Mobility”. The category “Problem” of the item “Friction & Shear” was the strongest predictor for category II pressure ulcers. Categories III/IV prevalence was 1.9%. Compared to all other Braden scale items there was the strongest association between being completely immobile and deeper categories III/IV pressure ulcers.

Conclusions

Based on a large sample of patients from multiple centres throughout Germany results indicate, that there is a strong relationship between friction forces and superficial skin lesions and between pressure forces and deeper categories III and IV PUs. This indicates that there might be different aetiologies causing different wounds. Given, that both superficial and deep ulcers have different aetiologies the validity of the current PU definition and classification is questionable, because ulcers due to maceration and excoriation are excluded from this classification system.     

不同决策树算法建立DRGs模型的差异

目的 通过分析不同决策树算法的数据差异,探讨最为适合建立DRG模型的算法.方法 择2007年1月-2012年9月病案首页第一诊断为呼吸系统疾病的12984病案,分别采用CART、CHAID和E-CHAID三种决策树算法建立DRGs模型,并比较模型间的差异.结果 CART算法建立的DRGs模型包括12个DRGs组,使用了6个分组因素,最重要的分组因素为疾病严重程度,模型的Risk值为0.449.CHAID算法建立的DRGs模型包括20个DRGs组,只使用了5个分组因素,分组因素中不包括有无输血,最重要的分组因素为有无手术,模型的Risk值为0.448.E-CHAID算法建立的DRGs模型包括15个DRGs组,使用了6个分组因素,最重要的分组因素为疾病严重程度,模型的Risk值为0.445.结论 通过比较后可认为E-CHAID算法要比其他两种算法更适合于建立DRGs模型.     

Impact of self-treated hypoglycaemia in type 2 diabetes: a multinational survey in patients and physicians

Abstract

Objectives:

The objectives were to estimate the prevalence of self-treated hypoglycaemia in patients using basal insulin analogues; identify demographic, treatment related and behavioural risk factors; and describe patient and physician responses to these events.     

综合医院急诊病人满意度影响因素分析

目的探讨综合医院急诊病人满意度的影响因素。方法以4所综合医院的503位急诊病人为调查对象,采用现场调查的方式。结果有效问卷487份,有效回收率为96.8%,应答率为96.2%。结论经CHAID方法分析急诊病人满意度的影响因素,发现低年龄、过低收入或过高收入、给医护人员额外报酬的人群满意度低。同时,发现满意度高的病人会多次来该院就诊。     

线外病人发生率的影响因素分析

目的探讨影响浙江省住院病人线外病人发生率的因素。方法用CHAID（Chi-squared Automatic Interaction Detector）方法找出影响线外病人发生率的因素，并分析各因素的重要程度及其之间的相互关系。结果性别、年龄、婚姻、费用类别、入院情况、出院情况、有无手术、有无次要诊断、有无抢救等特征显著不同，它们组成重要性不同的5个层次，并将研究人群分为线外病人发生率不同的24个部分。结论性别、年龄、婚姻、费用类别、入院情况、出院情况、有无手术、有无次要诊断、有无抢救是线外病人发生率的主要影响因素，应对其加以重视，以有效地控制线外病人发生率。     

Análisis de la supervivencia específica y el control local mediante un análisis de partición recursiva en pacientes con carcinomas de cabeza y cuello

Introduction and objectivesRecursive partitioning analysis (RPA) is a technique that allows prognostic classification in oncological patients. The aim of the present study is to analyse by means of an RPA a cohort of patients with squamous carcinomas of the head and neck (SCHN).Methods5,226 SCHN were retrospectively analysed with an RPA, considering the specific survival and local control of the disease as dependent variables. A cohort of patients was used for the creation of the classification model, and another cohort was used to carry out its internal validation.ResultsConsidering specific survival as a dependent variable we obtained a classification tree with 14 terminal nodes that were grouped into 5 categories, including as partition variables the local and regional extent of the tumour, and the location of the tumour. When considering the local control of the disease as a dependent variable we obtained a classification tree with 10 terminal nodes that were grouped into 4 categories, including as partition variables the local extension and location of the tumour, the type of treatment performed, the age of the patient, and if it was a first tumour or a subsequent neoplasm. The validation study confirmed the prognostic capacity of the models developed with the RPA. One of the advantages of the RPA is that it allows the identification of groups of patients with specific behaviour.ConclusionRPA is shown to be an effective technique for the prognostic classification of patients with a SCHN.