Measures.Of the 15 patients who started the study, only one indicated definite reluctance to take the pills at the conclusion of the first visit. Fourteen patients were completers, ie, they kept the second appointment. Pill counts revealed that only one of these 14 completing patients had deviated from prescribed dosage by more than one third, having taken 11 of 21 pills.
As shown in Table 1, the average initial Symptom Checklist score for the 14 completing patients per item initially was 1.04, "just a little," with an average improvement per item of 0.43. Thirteen of the 14 patients showed improvement, eight improving by at least 10 raw-score points and 11 by at least 5 points. One patient, whose husband had made a suicidal attempt during the study week, had a final score 10 points higher than at her initial visit. This 41% decrease in symptoms for the population is statistically highly significant and is actually more than the score reduction which has occurred with groups of patients on drugs or placebos in other drug studies we have conducted using the same measurements. (Unpublished results, NIMH-PSC Outpatient Studies.) However, expected length of treatment may have been an issue here. In the other studies duration was not clearly spelled out, whereas in the present study patients were informed they could improve on placebo in one week.
On the Target Symptoms, all 14 completers were improved, the average initial patient score per item of 1.78, approximately "quite a bit," reducing by 0.77 (43%) to an average post-treatment score of 1.01, "just a little." (As would be expected, selecting from the total symptom checklist those items of greater importance to the individual patient gives a higher initial score than for the total list. The higher score and larger improvement as indicated by the target symptom method is representative of our clinical impression.)
The average Patient Overall Change score at completion of the study was 2.07, "quite a bit better," with 13 patients indicating improvement and one patient indicating no change.
On the Doctor Overall Change ratings, the average patient score at the end of treatment was 1.79, closest to "quite a bit better," with all patients improved on this measure.
On the Pathology ratings, the average patient score at the initial visit was 3.79, with a pathology decrease of 1.36. This was a 36% improvement, with 12 patients rated as improved, one as unchanged, and one patient as a point worse.
In summary, there is very strong statistical evidence, on the basis of both doctor and patient ratings, that the completing patients as a group felt considerably improved. Eleven patients were rated as improved on all five measures. One patient was improved on all measures except for the Pathology Score, for which there was no change. One patient improved on all measures except for the Patient Overall Change score, for which there was no change. There was only one patient for whom any measure indicated an increase of symptomatology or pathology; for this patient there was a slight increase of scores on the Total Symptom Checklist and the Pathology rating, and improvement was indicated on the other three measures.
The one patient who did not return for the second appointment was seen by a social worker. Her Symptom Checklist showed a 10-point raw score increase in symptoms, and the Target ratings showed a 5-point increase. She estimated no change on the Overall Change Measure. Other ratings were not obtained.
There were no significant differences in improvement scores of patients seen by one or the other doctor.
Interview Content Evaluation.Table 2 is based on opinions elicited from the patients during the second visit; their opinions of the capsule content are compared with their interpretations of the key factors causing improvement. The average Symptom Checklist change score for each group of opinions is also given. All 14 completing patients are listed, since in the interviews they all indicated there was at least minimal improvement.
A most important finding is that improvement occurred in patients believing placebo was administered, "in spite of" such belief. There was no difference in improvement ratings between those eight patients who believed the pills contained placebo and the six patients who believed an active drug was involved.
It can be seen from Table 2 that only three patients had absolutely no doubts that the capsules contained an inactive ingredient. As a result of the use of the term "sugar pill" three patients, P, Q, and R, wondered during the treatment week if the capsule contained a "special sugar" of some sort.
Table 2 also shows that the five patients who felt without any doubt the pills definitely contained placebo or definitely contained drugs were rated as a group to be more improved than the other patients. This was significant by the Mann-Whitney U-Test (P<0.05).
It can be noted from the table that nine of the 14 patients felt the pill was the major factor in their improvement. It is quite interesting that not only five of the six patients who felt the pills probably contained medicine attributed improvement to the pills, but also four of the eight patients who assumed they probably contained placebo also attributed the major beneficial effects to the pills.
©2004 Lee Crandall Park, M.D.