Masiak Marek, Loza Bartosz
Katedry i Kliniki Psychiatrii AM w Lublinie
Psychiatr Pol. 2004 Sep-Oct;38(5):795-808.
A lot of inconsistencies across dimensional studies of schizophrenia(s) are being unveiled. These problems are strongly related to the methodological aspects of collecting data and specific statistical analyses. Psychiatrists have developed lots of psychopathological models derived from analytic studies based on SAPS/SANS (the Scale for the Assessment of Positive Symptoms/the Scale for the Assessment of Negative Symptoms) and PANSS (The Positive and Negative Syndrome Scale). The unique validation of parallel two independent factor models was performed--ascribed to the same illness and based on different diagnostic scales--to investigate indirect methodological causes of clinical discrepancies.
100 newly admitted patients (mean age--33.5, 18-45, males--64, females--36, hospitalised on average 5.15 times) with paranoid schizophrenia (according to ICD-10) were scored and analysed using PANSS and SAPS/SANS during psychotic exacerbation. All patients were treated with neuroleptics of various kinds with 410mg equivalents of chlorpromazine (atypicals:typicals --> 41:59). Factor analyses were applied to basic results (with principal component analysis, normalised varimax rotation). Investing the cross-model validity, canonical analysis was applied.
Models of schizophrenia varied from 3 to 5 factors. PANSS model included: positive, negative, disorganisation, cognitive and depressive components and SAPS/SANS model was dominated by positive, negative and disorganisation factors. The SAPS/SANS accounted for merely 48% of the PANSS common variances. The SAPS/SANS combined measurement preferentially (67% of canonical variance) targeted positive-negative dichotomy. Respectively, PANSS shared positive-negative phenomenology in 35% of its own variance. The general concept of five-dimensionality in paranoid schizophrenia looks clinically more heuristic and statistically more stabilised.
精神分裂症维度研究中存在诸多不一致之处。这些问题与数据收集的方法学方面以及特定的统计分析密切相关。精神科医生基于阳性症状评定量表/阴性症状评定量表（SAPS/SANS）和阳性与阴性症状量表（PANSS）开展了大量源自分析性研究的精神病理学模型。对平行的两个独立因素模型进行了独特的验证——归因于同一种疾病且基于不同的诊断量表——以探究临床差异的间接方法学原因。
对100例新入院的偏执型精神分裂症患者（根据国际疾病分类第10版，平均年龄33.5岁，年龄范围18 - 45岁，男性64例，女性36例，平均住院5.15次）在精神病性发作期使用PANSS和SAPS/SANS进行评分和分析。所有患者均接受了各种抗精神病药物治疗，氯丙嗪等效剂量为410mg（非典型药物：典型药物比例为41:59）。对基本结果应用因子分析（采用主成分分析、归一化方差最大化旋转）。为研究交叉模型效度，应用典型相关分析。
精神分裂症模型的因素数量从3个到5个不等。PANSS模型包括：阳性、阴性、紊乱、认知和抑郁成分，而SAPS/SANS模型以阳性、阴性和紊乱因素为主。SAPS/SANS仅占PANSS共同方差的48%。SAPS/SANS联合测量优先针对阳性 - 阴性二分法（占典型方差的67%）。相应地，PANSS在其自身方差的35%中共享阳性 - 阴性现象学。偏执型精神分裂症的五维度总体概念在临床上看起来更具启发性，在统计学上更稳定。
