*Dr J. Williamson, School of Complementary Health, Exeter EX1 1ED, UK.
Objective Clinical experience suggests that reflexology may have beneficial effects on the symptoms occurring in menopausal women, particularly psychological symptoms. This study aims to examine that effect rigorously.
Design Randomised controlled trial with two parallel arms.
Setting School of Complementary Health in Exeter, Devon, UK.
Sample Seventy-six women, aged between 45 and 60 years, reporting menopausal symptoms.
Methods Women were randomised to receive nine sessions of either reflexology or nonspecific foot massage (control) by four qualified reflexologists given over a period of 19 weeks.
Main outcome measures The Women's Health Questionnaire (WHQ), the primary measures being the subscores for anxiety and depression. Severity (visual analogue scale, VAS) and frequency of flushes and night sweats.
Results Mean (SD) scores for anxiety fell from 0.43 (0.29) to 0.22 (0.25) in the reflexology group and from 0.37 (0.27) to 0.27 (0.29) in the control group over the course of treatment. Mean (SD) scores for depression fell from 0.37 (0.25) to 0.20(0.24) in the reflexology group and from 0.36 (0.23) to 0.20 (0.21) in the control (foot massage) group over the same period. For both scores there was strong evidence of a time effect (P < 0.001) but no evidence of a time–group interaction (P > 0.2). Similar changes were found for severity of hot flushes and night sweats. In the control group, 14/37 believed they had not received true reflexology.
Conclusion Foot reflexology was not shown to be more effective than non-specific foot massage in the treatment of psychological symptoms occurring during the menopause.
Pictures from Egyptian tombs suggest that foot massage was used as a form of therapy 5000 years ago1. Its introduction into the Western world is generally dated to the beginning of the 20th century when an American ear, nose and throat surgeon, Dr William Fitzgerald, learnt the technique from native American Indians and explored reflex anaesthetic effects of pressure on feet and hands2. As practised today, reflexology is a specific form of foot massage in which it is believed that areas in the feet and hands correspond to the glands, organs and other parts of the body. One controlled study could find no evidence to support this concept3 although two other studies suggest there is a relationship4,5. Practitioners believe that local finger pressure can influence the function of organs encouraging homeostasis and promoting relaxation and the healing response. Precision reflexology is a form of the therapy using lighter pressure and an additional technique in which both hands join two relevant points simultaneously (‘linking’) with the aim of increasing the effectiveness of treatment1. Proposed theories of action include energetic effects, the dispersal of calcium, lactate or uric acid crystals, improvement of blood flow and a relaxant effect on the autonomic nervous system6. Others have suggested that reflexology is simply a method of providing care and attention for patients7.
In order to test the specific effects of reflexology separately from non-specific effects such as empathy, care and attention, an indistinguishable control procedure is ideally necessary. A search for randomised controlled trials of reflexology in Medline and in CISCOM (a specialised database of complementary medicine) in July 1996 located only four8–11. In two studies, the control group received sham reflexology, one in the form of ‘uneven tactile stimulation’ (overly light or very rough) on ‘points considered inappropriate for menstrual problems’9 and the other reflexology ‘to nonspecific zones’10. In a third study8, an attempt at sham treatment was abandoned due to lack of professional agreement concerning whether it might have an effect of its own. Reflexology was therefore compared with a group receiving no additional treatment, as in the fourth study11. Other studies have used reflexology to the wrong zone12 or massage to the calf13 as the control procedure.
Symptoms of the menopause such as vasomotor and urogenital problems14 can be treated effectively with hormone replacement therapy (HRT). Psychological symptoms that coexist may also respond15. However, HRT is recognised to have significant adverse effects14. Concern about possible side effects of drug treatment is one reason why people seek complementary and alternative medicine16 and a likely explanation for its increasing use by the general population17. The first author has observed a striking clinical response in a number of patients with menopausal symptoms during treatment with precision reflexology. We have been unable to find any previously published studies of the effect of reflexology on physical or psychological symptoms occurring during the menopause.
It was decided to undertake a rigorous investigation to determine whether precision reflexology improves the psychological symptoms occurring during the menopause to a greater extent than non-specific foot massage.
A patient-blinded randomised control trial was conducted in which one group received true reflexology and the control group received non-specific foot massage.
Women were recruited for the study by means of notices placed in primary care surgeries. Respondents were screened by the principal investigator by telephone to apply the inclusion criteria (aged between 45 and 60, presence of menopausal symptoms for at least three months) and exclusion criteria (currently taking HRT or psychoactive medication, severe pathology of the feet, previous reflexology treatment, current complementary therapy for menopausal symptoms or history of severe psychiatric illness); at the same time, baseline data were obtained on age and time since last menstrual period. Potential participants were sent a patient information leaflet and asked to consult their general practitioner in order to confirm that the women were menopausal (from medical history alone; hormone concentrations were not measured) and to agree on their suitability for inclusion. Participants were told they ‘would receive one of two different forms of foot massage’. Signed informed consent was obtained from all participants when they attended their first intervention session. The study was approved by the North and East Devon Research Ethics Committee.
After enrolment, each participant was randomised (by telephoning a central office) in two steps, first to one of four reflexologists, then to real or sham treatment. Randomisation lists (with block-sizes of eight between practitioners and four for each practitioner) were prepared from computer-generated random numbers, in order that patients were distributed evenly between the therapists, and each therapist treated equal numbers of patients with real and sham therapy. In an attempt to maintain blinding, therapists were trained not to discuss details of the treatment with the patients (i.e. standardised, limited interaction), although conversation on non-medical matters was permitted. Success of blinding was tested by means of a questionnaire completed by the participants at weeks 2 and 19, in which they were asked ‘Do you believe you were in the group that received true reflexology?’ (possible responses were ‘yes’, ‘no’ or ‘don't know’). The two-week interval was used to check blinding before the patients' responses were likely to be unduly influenced by any change in their condition.
Patients in the reflexology group received precision reflexology treatment using point location according to charts1. Patients in the control group received foot massage using a series of techniques according to a protocol that was predefined in order to provide a routine. This treatment involved no pressure techniques and the practitioners confirmed at the end of the study that all reflexology strokes were avoided in this group. Precise treatment protocols were established beforehand and training was given for both procedures. All patients were treated supine and no treatment was given except to the feet. The same non-perfumed foot cream was used in both groups. Both groups received treatment approximately weekly (intervals of between five and nine days were permitted) for six weeks, followed by one treatment monthly for three months, making a total of nine treatments. Each session lasted for 45 minutes. Participants were not given any self-care advice relating to menopause.
The main outcome measure was the Women's Health Questionnaire (WHQ), a measure of subjective reports of psychological and physical wellbeing of women aged 45–65, which had been validated in a group of women attending a screening clinic18. It was shown to be sensitive to changes in symptoms after various interventions19. The primary focus for this study was the subscores for anxiety and depression, which comprise four and seven items, respectively. Typical questions are ‘I feel anxious when I go out of the house on my own’ and ‘I feel miserable and sad’, to be answered on a four-point scale. The questionnaire was administered at baseline, at the end of week 6, at the time of the final treatment (week 19) and at follow up four weeks later (week 23).
Secondary outcome measures included the remaining seven subscales of the WHQ, which assess somatic symptoms, vasomotor symptoms, memory/concentration, sleep problems, menstrual symptoms, feelings of attractiveness and sexual behaviour. Severity of flushes and night sweats were assessed by an 100 mm visual analogue scale (VAS) for severity from zero to ‘worst imaginable’. Frequency of these symptoms were scored in categories: for flushes, the options for scoring were 0–5, 6–10, 11–15, 16–20 and more than 20; for night sweats, the options were 0–5, 6–10 and more than 10. This item was completed at baseline and weeks 6, 19 and 23.
An additional secondary outcome was the MYMOP score, a validated, self-completed measure of quality of life20. Patients are asked to nominate their two worst symptoms and one activity of daily living that their problem prevents them from doing. These three items and ‘general feeling of wellbeing in the last week’ are rated on six-point ordinal scales. This measure was also completed at baseline and weeks 6, 19 and 23. After the final treatment at week 19, global assessment was recorded by asking ‘How much you think that that the treatment had helped you?’ Available responses were on a six-point scale from ‘major increase in symptoms’ to ‘symptom-free’.
All outcome measures were completed by patients out of the sight of therapists and immediately placed in sealed containers, in order to reduce measurement bias.
Sample size was calculated from the data provided by a pilot study in which seven menopausal women received reflexology and completed the WHQ18. The total subscore for the first 12 questions, each scored 0–3, improved from 14.1 (SD 4.0) to 9.8. In order to be clinically relevant, we would expect this to be twice the response in the placebo group. It was calculated that a total sample size of 80 women would be required to identify this difference with 80% power and an alpha value of 0.05.
Since this study was a test of efficacy, the protocol specified that the analysis would include only those who completed treatment, defined as having had at least seven treatment sessions and completing the WHQ at 19 weeks. The planned primary analysis was therefore the difference in changes between groups from baseline to 19 weeks, by analysis of covariance, with baseline score as the covariate. Follows up scores at 23 weeks were summarised as further information. A second analysis was performed on data from all patients who commenced treatment, replacing missing data from dropouts by carrying forward the last available score (or in one case where data were missing at 19 weeks, using values from 23 weeks) and for other missing values the most recently recorded value. The subscores for anxiety and depression, as commonly employed18, were used in these analyses. This sum dichotomised responses to the relevant individual questions. Clustering effects due to different therapists were not modelled.
In total, 80 women were recruited but four did not commence treatment (see Fig. 1). The mean (SD) age of patients in the reflexology group was 50.8 (2.7) years and in the massage group 51.9 (2.5) years. In the reflexology group, 6 (15%) had menopausal symptoms for less than one year, compared with 8 (22%) in the massage group. Distribution of baseline scores for anxiety and depression are comparable between the groups (Table 1), with a small trend towards higher anxiety in the reflexology group. The assessment of success of blinding of the two groups is shown in Table 2.
Table 1. Baseline group comparisons. Values are given as number (%) of women in each group scoring different baseline values for anxiety and depression.
Reflexology (n= 39)
Foot massage (n= 37)
*One withdrew before the third assessment.
**Two withdrew before the third assessment.
One woman in each group had a baseline score of zero for both anxiety and depression.
Table 2. Responses to the question ‘Do you believe you were in the group that received true reflexology?’ asked on two different occasions during an RCT. Values are given as numbers (%).
At 2 weeks
At 19 weeks
The WHQ scores for 36 women completing the reflexology protocol and 34 who completed the non-specific foot massage protocol, as shown by summary statistics in Table 3, were not normally distributed, in contrast to those claimed for a previous study18. The reflexology group tended to have higher anxiety and depression scores at baseline, but the groups were comparable in age (reflexology group 50.8 (2.7) years, massage group 51.9 (2.5) years) and duration of symptoms (reflexology group 1.8 (0.4) years, massage group 1.8 (0.4) years).
Table 3. Primary analysis: anxiety and depression subscales of WHQ scored by menopausal women completing treatment. Values are given as mean [SD] and median (IQ range).
Reflexology (n= 36)
Foot massage (n= 3)
Median (IQ range)
Median (IQ range)
*Reflexology, n= 35; massage, n= 32, at 23 weeks.
The analysis of the sum of the first 12 questions, as used for the sample size calculation, showed an improvement in the reflexology group from mean (SD) of 14.8 (6.8) at baseline to 9.8 (5.6) at 19 weeks. The massage group improved from 14.0 (5.5) at baseline to 10.0 (5.5) at 19 weeks. Using analysis of covariance with initial score as covariate, the differential effect is −0.54 (95% CI −2.96 to 1.87, P= 0.66), which, although favouring reflexology, is neither statistically nor clinically significant, although the extremes of the confidence interval would be.
Applying an analysis of covariance on anxiety and depression subscores at 19 weeks, with initial score as a covariate and negative differences being in favour of reflexology, the mean differential treatment effect for anxiety was −0.082 (95% CI −0.18 to 0.026, P= 0.135). For depression scores, the mean differential treatment effect was −0.001 (95% CI −0.10 to 0.09, P= 0.979). These results are confirmed by repeated measures analysis of variance using the three scores and multivariate tests, although, given the non-normality of the data, these results are less reliable. For anxiety, time effect P < 0.001 and time–group interaction P= 0.208; for depression, time effect P < 0.001 and time–group interaction P= 0.793. For each score, post hoc testing confirmed that the average scores reduced significantly from baseline to week 6, but not from weeks 6 to 19. Responses to the global questionnaire indicated that 17 (47%) in the reflexology group scored overall improvement compared with 13 (38%) in the nonspecific foot massage group, with no between-group differences.
Intention-to-treat analysis was performed on 39 patients in the reflexology group and 37 in the massage group, using the last recorded value carried forward, essentially confirmed the per protocol analysis. Using repeated measures of analysis of variance for anxiety and depression scores separately, there was strong evidence of a time effect (P < 0.001) but no evidence of a time–group interaction (P > 0.1).
A similar pattern of changes was seen in most other symptom subscores (somatic symptoms, vasomotor symptoms, memory/concentration, sleep problems, menstrual symptoms and feelings of attractiveness) (i.e. fall from baseline to sixth treatment with no further change, data not presented). There were no group differences. Only the scores for sexual behaviour showed a different pattern, with little evidence of change in either group.
Similar patterns of reduction from baseline to the sixth weekly treatment, followed by no further change, are seen for hot flushes and night sweats (Table 4) with no meaningful differences between the groups. There were many missing or unusable responses to the MYMOP questionnaire (see Discussion). Median scores for the nominated two worst symptoms were 4 in both groups at baseline and fell to 3 at 23 weeks.
Table 4. VAS scores for severity of hot flushes and night sweats in menopausal women completing treatment with either active reflexology or control nonspecific foot massage. Values are given as mean [SD] and median (IQ range).
In this first published randomised controlled trial of reflexology for psychological symptoms at the time of the menopause, there were no significant differences between the effects of reflexology and non-specific foot massage control. Both were associated with improvements, but it is impossible to be certain whether the observed changes were due to the specific interventions or to ‘placebo’ effects including the non-specific effects of treatment such as touch, attention and expectation, or simply to time, regression to the mean and so on. Other studies in menopausal women have shown that symptoms may respond, in some cases dramatically, to placebos such as the transdermal patch21 and to psychological treatments, including cognitive–behavioural therapy and relaxation19. There is little published evidence for the effect of touch or massage on menopausal women: one study found that mechanical massage of the abdominal muscles of women was associated with hormonal changes but this was not tested in menopausal women22. The inclusion of an attention control arm to address these questions was considered in designing this study, but was not thought feasible in this setting.
The study does have limitations. The sample size was calculated on the answers to the first 12 questions of the WHQ, whereas the analysis focussed on the more commonly used anxiety and depression subscales. In addition, the placebo effect was probably under-estimated and the required sample was not achieved because of early dropouts. A much larger sample would be required to test whether the small difference between the groups that we observed is statistically significant. Another limitation is the handling of the MYMOP questionnaire that was completed unsupervised (in contrast to the usual instructions) in an attempt to reduce bias in the responses. Several women were not consistent in the worst symptoms they nominated, invalidating much of the data. This was particularly disappointing because this was an opportunity to obtain information about which symptoms women think are most important. Given the diversity of menopausal symptoms, the MYMOP questionnaire would seem to be especially useful. In addition, the categories used to measure the frequency of hot flushes and night sweats were not sensitive enough to distinguish between zero and a small number of attacks. A further limitation of the study is the short duration of follow up. Considering the similarity of changes in the two groups, a six-month follow up questionnaire could have been helpful in identifying any sustained improvement. A follow up was considered at the planning stage but was rejected due to problems of increased administration, cost and likely dropouts.
Baseline mean values for all WHQ subscores tended to be higher than those of women attending a routine ovarian cancer screening clinic18, as expected. This was particularly true for sleep disturbance. The women who dropped out of the study tended to be older than those who remained (mean (SD) age of 53.4 (2.2) years compared with 51.0 (2.6) years, respectively) and also tended to have higher scores for anxiety and depression. One woman who dropped out of the control group recorded scores for anxiety and depression that were increasing over time, which was unusual.
Blinding was of limited success because 14 women (out of 37) in the control group knew that they were not receiving reflexology. This may indicate that, although women were only included if they had had no previous experience of reflexology, some knew or discovered enough to be able to recognise the difference between reflexology and non-specific massage. This highlights the problem of devising an appropriate placebo control for trials of reflexology. However, the lack of blinding does not appear to have prejudiced the women against massage. Previous studies9,10 have not tested the success of blinding, which seems essential.
In conclusion, foot reflexology was not shown to be more effective than non-specific foot massage in the treatment of psychological symptoms of the menopause. The improvements shown may have been due to non-specific (placebo) effects.
The study was funded by a grant from the Foundation for Integrated Medicine, London. This work was conducted at the School of Complementary Health, Exeter, UK. The trial was planned and conducted by J. W., A. W. and E. E. A. H. was responsible for the statistical analysis of data.