Jakes S, Stephens S D
Institute of Laryngology and Otology, London.
Br J Audiol. 1987 Nov;21(4):259-72. doi: 10.3109/03005368709076418.
Multivariate statistical techniques were used to re-analyse the data from the recent DHSS multi-centre masker study. These analyses were undertaken to three ends. First, to clarify and attempt to replicate the previously found factor structure of complaints about tinnitus. Secondly, to attempt to identify common factors in the change or improvement measures pre- and post-masker treatment. Thirdly, to identify predictors of any such outcome factors. Two complaint factors were identified; 'Distress' and 'intrusiveness'. A series of analyses were conducted on change measures using different numbers of subjects and variables. When only semantic differential scales were used, the change factors were very similar to the complaint factors noted above. When variables measuring other aspects of improvement were included, several other factors were identified. These included; 'tinnitus helped', 'masking effects', 'residual inhibition' and 'matched loudness'. Twenty-five conceptually distinct predictors of outcome were identified. These predictor variables were quite different for different outcome factors. For example, high-frequency hearing loss was a predictor of tinnitus being helped by the masker, and a low frequency match and a low masking threshold predicted therapeutic success on residual inhibition. Decrease in matched loudness was predicted by louder tinnitus initially.
多变量统计技术被用于重新分析近期卫生和社会保障部多中心掩蔽器研究的数据。这些分析有三个目的。第一，澄清并尝试复制先前发现的耳鸣相关投诉的因素结构。第二，尝试识别掩蔽器治疗前后变化或改善措施中的共同因素。第三，识别任何此类结果因素的预测指标。确定了两个投诉因素；“痛苦”和“侵扰性”。使用不同数量的受试者和变量对变化措施进行了一系列分析。当仅使用语义差异量表时，变化因素与上述投诉因素非常相似。当纳入测量其他改善方面的变量时，识别出了其他几个因素。这些因素包括；“耳鸣得到缓解”、“掩蔽效果”、“残余抑制”和“匹配响度”。确定了25个在概念上不同的结果预测指标。这些预测变量因不同的结果因素而有很大差异。例如，高频听力损失是耳鸣得到掩蔽器缓解的一个预测指标，低频匹配和低掩蔽阈值预测残余抑制治疗成功。最初耳鸣响度较大可预测匹配响度降低。
