program R version 3.1.1. Each number of the randomization sequence was put inside an opaque envelope, which was sealed and handed to the practitioner only when the first session was about to be performed. The blinding was applied to the researchers and the statistician, who ignored into which group the participants had been allocated. 4 www.ee.usp.br/reeusp Effects of auricular acupuncture on chronic pain in people with back musculoskeletal disorders: a randomized clinical trial Rev Esc Enferm USP · 2019;53:e03418 Data analysis and treatment Collected data were inserted into a Microsoft Office Excel® spreadsheet, version 2013, by two independent researchers, and the consistency of the data was tested later. Statistical analysis was carried out using the programs SSPS version 23.0 and BioEstat version 5.0. Data were treated with the intention-to-treat analysis through the repetition of the values of the last evaluation, in accordance with Consort recommendations(17). The variables of the sociodemographic and pain profiles were submitted to chi-square and Kruskal-Wallis tests. In the intergroup evaluation, the Kruskal-Wallis test was applied, followed by the Student-Newman-Keuls test whenever necessary. The intragroup evaluation was performed using the Wilcoxon signed-rank test. A level of significance of 5% was adopted. Inclusion Evaluated for eligibility (n=386) Allocation Final Follow up Randomized (n=110) Placebo group (n=36) Treatment group (n=37) Treatment group (n=36) Discontinued intervention (n=1) Analyzed (n=37)* Placebo group (n=34) Discontinued intervention (n=2) Analyzed (n=36)* Control group (n=26) Discontinued intervention (n=11) Analyzed (n=37)* Treatment group (n=31) Lost to follow-up (n=5) Analyzed (n=37)* Placebo group (n=30) Lost to follow-up (n=4) Analyzed (n=36)* Control group (n=22) Lost to follow-up (n=4) Analyzed (n=37)* Control group (n=37) Excluded (n=276) - Did not answer the call/declined it (n=207) - Did not meet inclusion n=65) - Other reasons – death; disabling disease (n=4) Figure 1 – Flowchart showing the sample screening process. *Intention-to-treat analysis Source: Adapted from Consort (http://www.consort-statement.org/consort-statement/flow-diagram). Ethical aspects The present study was approved by a research ethics committee as per report no. 1.041.266 of 2015 and followed the ethical recommendations of the Brazilian National Health Council according to Resolution 466/12(18). At the end of the investigation, the people in the placebo and control groups had been submitted to the same number of AA sessions than participants in the treatment group. RESULTS Data shown in Table 1 present the comparison among groups according to age, gender, and pain profile. The participants’ characteristics were homogenous, indicating that the randomization process was adequate. The patients recruited from June 2015 to March 2016 were mostly 5 Moura CC, Chaves ECL, Chianca TCM, Ruginsk SG, Nogueira DA, Iunes DH www.ee.usp.br/reeusp Rev Esc Enferm USP · 2019;53:e03418 Table 1 – Sample characterization regarding age, gender, and pain profile in the three studied groups – Alfenas, MG, Brazil, 2016. Variables Treatment group (n=37) Placebo group (n=36) Control group (n=37) p value Age (µ±sd) (kal-Wallis test; b chi-square test; µ: mean; sd: standard deviation; (n=110). women between 40 and 50 years old, who presented a prevalence of recurring pain for around four years. The most affected activity was dressing in the treatment and control groups, whereas people in the placebo group reported similar difficulties regarding dressing and showering. Most participants declared to be anxious and complained of feeling tired and being unable to practice physical activities because of the fear to move. These people were monitored for one month and a half, a period which encompassed the evaluations and intervention. According to the data shown in Table 2, pain intensity and relief did not differ among the examined groups in the initial evaluation, and there was no modification in the performance of the control group over time in the three analyzed variables (pain intensity and relief and impact of pain on daily activities), as expected. However, the treatment and placebo groups had a decrease in the score of pain intensity in the final evaluation. A remarkable aspect found in this set of evaluations which illustrates the limited benefit of the so-called placebo effect is the evidence that the pain intensity got back to pre-session levels in the placebo group, but not in the treatment group, in the follow-up evaluation. Nearly the same behavior was observed for pain relief, which was higher in the treatment group and lower in the placebo group in the final evaluation. Nevertheless, after 15 days of the end of the treatment, pain relief decreased in the treatment group, getting close to the results of the placebo group. Pain relief remained higher in the follow-up evaluation in the treatment and placebo groups in comparison with the results registered in the initial evaluation. Table 2 also brings evidence that pain interferes with the execution of daily activities. Once more, the intragroup analysis revealed an improvement in this parameter after the sessions (final evaluation) in the treatment and placebo groups. Nevertheless,