Kitamura T, Toda M A, Shima S, Sugawara M
National Institute of Mental Health, Japan.
Int J Psychiatry Med. 1994;24(2):149-56. doi: 10.2190/W6MU-XRC6-16KF-3Q88.
The authors examined the variability of the validity of the General Health Questionnaire (GHQ) on two different occasions.
The subjects were 120 pregnant women attending an antenatal clinic of a general hospital in Japan. The GHQ was distributed twice--in the first and third trimesters. They were then interviewed by a psychiatrist blind to the GHQ scores using the standard and the "change" version of the Schedule for Affective Disorders and Schizophrenia (SADS).
Of the 120 women, 108 and ninety-eight completed the GHQ and were successfully interviewed in the first and third trimesters, respectively. Seventeen percent (18/108) and 13 percent (13/98) women were given RDC diagnoses in the first and third trimesters, respectively: They were designated as cases. Despite a satisfactory discriminatory power of the GHQ on the first occasion [1], the validity measures of the GHQ on the second occasion were generally poor. Thus, the sensitivity was 39 percent and specificity 82 percent for the cut-off point of 7/8.
The GHQ should be validated separately when distributed repeatedly to the same subjects.
作者考察了一般健康问卷（GHQ）在两种不同情况下效度的变异性。
研究对象为120名在日本一家综合医院产前诊所就诊的孕妇。GHQ分两次发放——在孕早期和孕晚期。然后由一位对GHQ得分不知情的精神科医生使用情感障碍和精神分裂症日程表（SADS）的标准版和“变化”版对她们进行访谈。
120名女性中，分别有108名和98名在孕早期和孕晚期完成了GHQ并成功接受访谈。分别有17%（18/108）和13%（13/98）的女性在孕早期和孕晚期被给予研究诊断标准（RDC）诊断：她们被认定为病例。尽管GHQ在首次测量时具有令人满意的区分能力[1]，但第二次测量时GHQ的效度指标总体较差。因此，对于7/8的分界点，敏感性为39%，特异性为82%。
当对同一组受试者重复发放GHQ时，应分别对其效度进行验证。
