Study design and setting
This study is a randomized, two-armed, parallel group clinical trial. The coordinating office is at Tsukuba University of Technology, Japan. Gynecologists in other medical facilities introduce the coordinating office to patients who meet the eligibility criteria and confirm they do not meet any exclusion criteria.
When patients show interest in trial participation, the gynecologists send an introduction form to the coordinating office by facsimile. The coordinating office calls the introduced patient to decide the date on which to meet to explain the trial, and the patient visits the office to be given information verbally on the trial and reads the written trial description. After the meeting, the patient submits the consent form by hand, facsimile or mail upon deciding to register for the trial. Accrual started on 12 October 2012, and the term for patient registration for the trial is within 2 years from this date. Sixty participants are planned to be recruited.
Ethical consideration and study registration
The study protocol was designed in accordance with the World Medical Association’s Helsinki Declaration [17] and the Ethics Guidelines for Clinical Research of the Ministry of Health, Labor, and Welfare of Japan [18]. This trial was approved by the Medical Ethics Committee of Tsukuba University of Technology, Japan, on 27 September 2012 (Approval No. 5) and was registered with the UMIN Clinical Trials Registry as application UMIN000009097 on 12 October 2012.
Eligibility criteria
The inclusion criteria are (1) histologically confirmed uterine cervical, endometrial, ovarian, fallopian tube or peritoneal cancer in the past; (2) no recurrence of such cancer for more than 3 years since receiving standard medical treatment; (3) older than 20 years of age at the date of registration for the trial; (4) patient’s doctor deems the patient to be eligible for the trial; and (5) receipt of written informed consent from the patient to participate in the trial. The exclusion criteria are (1) active infections; (2) serious concurrent disease of the heart, liver or kidney, etc.; and (3) severe mental disorders.
Randomization
After finishing registration, the patients are allocated by block randomization by the coordinating office to continuous Anma massage therapy or no Anma massage therapy, except for a single Anma massage therapy intervention at the end of the trial period (Figure 1). Allocation adjustment factors are not set in the trial because of insufficient evidence at present regarding which factors affect the effectiveness of Anma massage therapy.
Protocol treatment
Interventions
Anma massage therapy group
This group receives treatment by continuous Anma massage therapy according to the trial protocol. The protocol treatment is completed when the patient finishes receiving the final 40-min Anma massage session of a total of eight sessions given once weekly over a 2-month intervention period. Anma massage therapy techniques consist of standard versions of Japanese massage, mainly kneading with lesser amounts of stroking and pressing. Anma massage is performed through clothing, with the intensity of stimulation applied being within each person’s range of comfort. On a massage table, a full-body Anma massage excluding the face, head and abdomen is performed following the procedure described in detail in our previous studies [16, 19, 20], with a focus on the specific locations where patients want to improve physical symptoms. A therapist provides all massage sessions to avoid differences in technical capabilities. This therapist has a national massage practitioner license and over 20 years of experience and was the therapist involved in providing Anma massage treatment sessions in our previous studies [16, 19, 20].
No Anma massage therapy group
This control group is followed by their medical doctors as usual and do not receive the continuous Anma massage sessions. Patients visit the coordinating office and have a 40-min relaxing chat with a massage therapist while lying on a massage table, but without receiving a massage. Two months later they return to the office to receive a single 40-min Anma massage session.
End points
The primary end point is the severity of subjective physical complaints that cancer survivors report in daily life as measured using a VAS. To assess the severity of subjective physical complaints, a sheet of paper (width 100 mm × height 40 mm) is given to the participant, and it is explained that the left edge of the paper represents no complaint and the right edge the most serious complaint that the participant can imagine. The participant records the seriousness of the degree of the subjective complaint at that time as a check mark on the paper. The length from the left edge to the check mark is measured and considered the VAS score.
The secondary end points of the study are as follows. (1) The score on the Profile of Mood State–Brief Japanese Version (POMS) questionnaire is used to assess immediate psychological effects. (2) The score on the Hospital Anxiety Depression Scale–Japanese version (HADS), a 14-item scale developed to assess anxiety and depression for particular cancer patients. (3) The score on the European Organization for Research and Treatment of Cancer QLQ-C30–Japanese version (EORTC QLQ-C30) scale, which was developed to assess health-related QoL, particularly in cancer patients. The subscales are global health status, functional scales (physical functioning, role functioning, emotional functioning, cognitive functioning and social functioning) and symptom scales (fatigue, nausea and vomiting, pain, dyspnea, insomnia, appetite loss, constipation, diarrhea and financial difficulties). (4) The score on the Measure of Adjustment to Cancer–Japanese version (MAC) questionnaire, because some studies [21, 22] have reported a correlation between adjustment of cancer, QoL and mental health. Because some communication between the patient and massage therapist is commonly built into massage sessions, psychological adjustment to cancer (coping style to cancer) may change through such communication during continuous Anma massage therapy. (5) The level of 8-hydroxydeoxyguanosine and catecholamine (noradrenaline, adrenaline and dopamine) in urine is measured. (6) Levels of cortisol, chromogranin A and s-IgA in saliva are measured.
Data collection
Data collection is performed on three occasions in the Anma massage group: before and after the first Anma massage session (measurements A and B) and before the last (eighth) Anma massage session (measurement C) (Figure 1). Data collection is conducted on four occasions in the control group: before and after the relaxing chat session (measurements D and E) and before and after the single Anma massage session (measurements F and G) at 2 months (Figure 1).
To assess the effects of continuous Anma massage sessions, changes in all end points between measurements A and C are compared with those between measurements D and F.
In addition, to confirm immediate changes following a single Anma massage session, which is the secondary objective of the trial, changes in VAS score, POMS score, and levels of 8-hydroxydeoxyguanosine, catecholamine (noradrenaline, adrenaline and dopamine), salivary cortisol, chromogranin A, and s-IgA between measurements A and B are compared with those between measurements D and E. Such changes between measurement F and measurement G are also considered as additional evidence to confirm the immediate effects of a single Anma massage session.
Sample size determination
In our preliminary study [16], the mean VAS score before the first Anma massage session in cancer survivors was 40.4, which was reduced to 19.6 before the last (eighth) Anma massage session. The mean VAS score difference during 2 months was −20.8 (standard deviation (SD) 19.6). We assume for the no Anma massage group that the mean difference will be unchanged (that is, zero) and that its SD will be the same as that of the Anma massage group (that is, 19.6). To test these differences between the two groups, with a 5% type I error rate and 90% power, a sample size of 14 patients per group is required. For sensitivity analysis, we also consider the following situations: the mean VAS score differences in the no–Anma massage group will be −5, –7.5 and −10, in which case sample sizes of 24, 33 and 49 patients per group, respectively, are required for 90% power, and 18, 25 and 37 patients per group, respectively, are required for 80% power. Therefore, the planned sample size will be 30 participants per group (60 participants in total). A sample size of 30 patients per group will also account for possible data loss or dropouts.
Statistical analysis
Analysis of pretreatment characteristics and efficacy analysis will be performed according to the modified intention-to-treat principle and will include all participants who receive at least one Anma massage session in the Anma massage therapy group and one relaxing chat in the no–Anma massage group. If necessary, we might also perform sensitivity analyses, adding all eligible patients who are introduced by medical doctors to the coordinating office but are not registered.
The primary end point is VAS score improvement over the 2-month study period. For primary analysis, we will use the analysis of covariance to compare the mean changes in VAS score over the 2 months between the Anma massage group and the no–Anma massage group, adjusting for the baseline VAS score and age. If we find other significant prognostic factors, we will use them as additional adjusting factors. We will also use a two-sample t-test with Satterthwaite’s approximation to compare the mean changes. We will also test the mean changes in VAS score for each group using a paired t-test. For these mean VAS score changes and their differences, we will also calculate two-sided 95% confidence intervals (CIs) to evaluate the clinical effects.
For the secondary end points, we will not consider multiplicity issues. In all analyses, categorical variables will be described in terms of frequency and percentage. The distributions of continuous variables will be described using means, SDs, medians, and minimum and maximum values. A two-sample t-test or paired t-test will be used to detect differences in continuous variables. Pearson’s χ2 test with continuity correction will be used to test differences in categorical variables. If these data are found not to be normally distributed, we will use in their place the Wilcoxon signed-rank test or the Mann-Whitney U test, respectively. All reported P values will be two-sided. All significance levels will be set at 0.05, and reported CIs will be 95%. In principle, we will use the available-case analysis for missing outcomes. SAS statistical software (version 9.3 or later; SAS Institute, Cary, NC, USA) will be used for all analyses.
