Factors Predisposing to a Placebo Response in New Out-patients with Anxiety States

¹ Consultant Psychiatrist, Long Grove Hospital, Epsom, Surrey

1. Forty-four new out-patients with an agreed diagnosis of anxiety state were treated by the same doctor for three weeks; the first 29 received placebo and the next 15, sodium amylobarbitone. Progress was recorded weekly on Hamilton's Anxiety State Rating Scale (ASRS); placebo patients also returned self-ratings twice weekly.

2. For both sets of patients it was predicted that age, sex and duration of illness before referral would not be related to response; and those who expressed more confidence in doctors, favourable attitudes to hospital, psychiatry and psychiatric treatment, more complaints about their physical health and were more extraverted, would show the better response.

3. In both treatment samples there is a significant linear correlation between initial and third week ASRS scores; a regression correction was therefore applied to yield individually adjusted final scores. Their validity is indicated by a significant correlation with patients' self-ratings.

4. Among patients treated with placebo, the predictions concerning age, sex, attitudes and extraversion are confirmed. Contrary to expectation, duration of illness differentiates response to a significant extent, patients being ill for less than a year showing the most improvement.

5. The pattern of results for patients treated with sodium amylobarbitone is identical to that for the placebo sample, although only extraversion differentiates response at a statistically significant level.

6. The two samples do not differ significantly on any of the independent and dependent variables. Analysis of the pooled sample provides no further support for the two unconfirmed hypotheses.

7. It is concluded that patients can be independently classified in a way that also differentiates their response to placebo. The question whether the same classification will differentiate their response to active medication cannot be answered by the present data, since there is no indication that a specific drug effect was achieved.

8. The results confirm in one sample those that have usually been reported singly in previous studies.

9. Duration of illness before referral differentiates short-term outcome and has also been shown by others to be a long-term prognosticator. One explanation is that these three variables represent various manifestations of a single constitutional or personality factor.

10. The existence of independently defined sub-groups with different prognoses suggests that patients diagnosed as anxiety states do not constitute an homogenous population.

This article has been cited by other articles:

				T. Johnson Clinical trials in psychiatry: background and statistical perspective Statistical Methods in Medical Research, June 1, 1998; 7(3): 209 - 234. [Abstract] [PDF]