Statistical factor analysis and cluster analysis in the etiology of climacteric symptoms.

Abstract

Factor analysis and cluster analysis were applied to a set of 17 climacteric symptoms data obtained from 194 premenopausal and postmenopausal women. Six distinct factors were extracted and the women were divided into 7 groups by hierarchical cluster analysis in terms of their factor scores. Only one representative factor was identified in 3 groups out of the 7. These 3 groups including 87 women were tentatively designated as vasomotor symptom, neurotic, and sensory disturbance groups. These results confirmed statistically that at least 3 simple etiological factors exist among climacteric symptoms.climacteric symptoms ; factor analysis ; cluster analysisAt the First International Congress on the Menopause held at La Grande Motte, France in 1976, etiology of climacteric symptoms was defined as follows : 1) hypothalamic-autonomic imbalance, 2) true estrogen deficiency, 3) psychological factors, and 4) socio-cultural factors (titian and Serr 1976). Climacteric symptoms are, thus, so various as to require different therapies, based on the factors involved in their etiologies. However, only fragmental approaches such as those from the viewpoint of endocrinologic (Chakravarti et al. 1979), psycho somatic (Polit and Larocco 1980), socio-cultural, and other aspects have been undertaken in order to elucidate the etiology of the symptoms. Such complexity may be a reason why the etiology of climacteric symptoms, as a whole, still remains obscure. Therefore, we have carried out a series of studies with the aim of developing a comprehensive approach to the etiology of such symptoms ; factor analysis and cluster analysis were applied to a set of symptoms data and the

Clinical and endocrinological features of statistical clusters of women with climacteric symptoms.

“…One hundred and ninety-four premenopausal and postmenopausal women with climacteric symptoms in the present study were the same subjects that had been already subjected to statistical analysis of their symptoms in the previous report (Abe et al 1984). One hundred and eighteen premenopausal women were at 37 to 55 years of age (45.22 + 5.60, mean +s.D.)…”

“…As described in the previous report (Abe et al 1984), numerical data regarding the severity of the symptoms was analyzed by factor analysis, yielding six distinct extracted factors, and the women were divided into 7 groups by hierarchical cluster analysis in terms of their factor scores (Table 2). However, 2 of the 7 groups, group 2 (heart failure-like symptoms group with insomnia) and group 3 (heart failure-like symptoms group with vasomotor-, sensory disturbance-, and autonomic nervous dysfunction-symptoms) were characterized by more than one representative factors, one of which, the heart failure-like symptoms factor, was common to both groups and was also most representative factor in each one of them.…”

“…However, 2 of the 7 groups, group 2 (heart failure-like symptoms group with insomnia) and group 3 (heart failure-like symptoms group with vasomotor-, sensory disturbance-, and autonomic nervous dysfunction-symptoms) were characterized by more than one representative factors, one of which, the heart failure-like symptoms factor, was common to both groups and was also most representative factor in each one of them. Therefore, in order to simplify the clustered groups as possible, these two groups were combined from clinical standpoints, and the original 7 groups of the previous work (Abe et al 1984) were regrouped into 6 groups as follows (Table 2) : Group 1, insomnia group with neurotic symptoms, remained unchanged ; the original group 2 as well as group 3 were combined and reclassified as group 2 to be called heart failure-like symptoms group with insomnia, vasomotor-, sensory disturbance-, and autonomic nervous dysfunctionsymptoms ; the original group 4, the vasomotor-symptoms group, the original group 5, the neurotic symptoms group, the original group 6, the sensory disturbance-symptoms group, and the original group 7, the mild symptoms group, were reclassified as groups 3, 4, 5 and 6, respectively, without any alteration of their designations.…”

“…Among these six new groups of women, mean age, rate of premenopausal women or postmenopausal women against all women including premenopausal and postmenopausal women in each group, and serum concentrations of estradiol, progesterone, folliclestimulating hormone and luteinizing hormone were compared, in order to provide the etiology of climacteric symptoms with endocrinological basis and also to demonstrate the validity of statistical clustering of women with climacteric symptoms, being reported previously (Abe et al 1984). …”

“…In order to develop a comprehensive approach to the etiology of climacteric symptoms, the authors have undertaken a series of studies of these symptoms ; in our previous report (Abe et al 1984), statistical factor analysis and cluster analysis were applied to data consisting of sets of 17 symptoms of climacteric women. Six distinct factors were extracted and the women were divided into 7 groups by hierarchical cluster analysis in terms of their factor scores.…”