This is the authors' abstract. We don't add key points for this paper.
To the Editor: Nowadays, many clinical trials use health-related quality of life (HRQL) to assess treatment effects. In unblinded trials, this subjective outcome measure may be susceptible to bias. We have noticed in a randomized clinical trial on the effect of spinal cord stimulation (SCS) on reflex sympathetic dystrophy an improvement in HRQL even before the start of treatment. Patients in this study were suffering from chronic reflex sympathetic dystrophy, a disabling disease of unknown pathophysiology, characterized by pain and impaired function. They had not responded to any previous treatment, so their last hope was the pain-relieving effect of SCS. In this procedure, an electrode is positioned in the epidural space on the dorsal aspect of the spinal cord at the level of the nerve roots innervating the painful area; electrical current from the electrode brings about paresthesia, a sensation that suppresses the pain. The patients were randomized into two groups: one group to receive SCS and physical therapy (SCS group; N = 36) and one group to receive physical therapy alone (control group; N = 17). The SCS group underwent a testing period of 7 days, using a test electrode. A complete SCS system was implanted only when the patient reported ≥50% reduction in pain during the testing period (SCS+, N = 24; SCS−, N = 12). Among other parameters, HRQL, by means of a “thermometer” from 0 (worst imaginable health state) to 100 (perfect health), and pain, using a visual analogue scale from 0 (no pain) to 10 (worst imaginable pain) were measured, both before randomization (T0) and then again 1 month later (T1). For SCS+ patients, T1 was the day before implantation, while SCS− patients knew at that point that they would not receive a SCS system. Between T0 and T1, the HRQL and pain scores of the total SCS group improved compared with the control group (Table 1). Trials on SCS cannot be blinded because of the paresthesiae and, therefore, patients know immediately after randomization what treatment they will be receiving. Merely the fact of being assigned to the SCS group was enough to cause an improvement in HRQL and pain intensity. As might be expected, the HRQL improvement was higher in SCS+ patients than in SCS− patients. On the other hand, for pain intensity this difference was not found. Therefore, we argue that the effect of randomization is due to bias rather than to chance. Olsson and colleagues also showed that a deliberate attempt to maximize the expectation effect can influence the perceived clinical course. 1 This finding raises the question of whether pre-randomization values are valid as baseline values in unblinded studies. If, for example, HRQL remains stable at the T1 level following treatment, despite the difference between the SCS group and the control group, it is questionable whether SCS improves HRQL. While T0 values (pre-randomization) are necessary to ensure prognostic similarities between study groups, T1 values (pre-treatment) are more appropriate as baseline values in unblinded trials. Table 1: Differences in Scores on Health-Related Quality of Life and Pain Intensity between T0 and T1Marius A. Kemler Henrica C. W. de Vet
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Kemler et al. (2000) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: