Reflexology for symptom management among women with advanced breast cancer

The aim of this RCT was to examine symptom responses resulting from a home-based reflexology intervention delivered by a friend/family caregiver to women with advanced breast cancer undergoing chemotherapy, targeted, and/or hormonal therapy.

Patient-caregiver dyads (N = 256) were randomized to 4 weekly reflexology sessions or attention control. Caregivers in the intervention group were trained by a reflexology practitioner in a 30-min protocol. During the 4 weeks, both groups completed telephone symptom assessments using the M. D. Anderson Symptom Inventory. Those who completed at least one weekly call were included in this secondary analysis (N = 209). Each symptom was categorized as mild, moderate, or severe using established interference-based cut-points. Symptom response meant an improvement by at least one category or remaining mild. Symptom responses were treated as multiple events within patients and analysed using generalized estimating equations technique.

Reflexology was more successful than attention control in producing responses for pain with no significant differences for other symptoms. In the reflexology group, greater probability of response across all symptoms was associated with lower number of comorbid condition and lower depressive symptomatology at baseline. Compared to odds of responses on pain (chosen as a referent symptom), greater odds of symptom response were found for disturbed sleep and difficulty remembering with older aged participants.

Adjusted odds ratios (ORs) of symptom responses for reflexology arm versus control (adjusted for age, number of comorbid conditions, depressive symptoms at baseline, and treatment type: chemotherapy with or without hormonal therapy versus hormonal therapy alone)
Symptom                                 OR      (95% CI)        p value
Fatigue                                    1.76      (0.99, 3.12)       0.06
Pain                                         1.84      (1.05, 3.23)       0.03
Disturbed sleep                         1.45      (0.76, 2.77)       0.26
Shortness of breath                   0.58      (0.26, 1.30)       0.19
Remembering                           0.96      (0.51, 1.78)       0.89
Lack of appetite                        1.05      (0.45, 2.49)       0.91
Dry mouth                               1.84      (0.86, 3.94)       0.12
Numbness and tingling              1.40     (0.75, 2.64)        0.29
Depression                              1.38      (0.78, 2.43)       0.27

The authors concluded that home-based caregiver-delivered reflexology was helpful in decreasing patient-reported pain. Age, comorbid conditions, and depression are potentially important tailoring factors for future research and can be used to identify patients who may benefit from reflexology.

This is certainly one of the more rigorous studies of reflexology. It is well designed and reported. How valid are its findings? To a large degree, this seems to depend on the somewhat unusual statistical approach the investigators employed:

Baseline characteristics were summarized by study group for outcome values and potential covariates. The unit of analysis was patient symptom; multiple symptoms were treated as nested within the patient being analyzed, using methodology described by Given et al. [24] and Sikorskii et al. [17]. Patient symptom responses were treated as multiple events, and associations among responses to multiple symptoms within patients were accounted for by specifying the exchangeable correlation structure in the generalized estimating equations (GEE) model. The GEE model was fitted using the GENMOD procedure in SAS 9.4 [25]. A dummy symptom variable with 9 levels was included in the interaction with the trial arm to differentiate potentially different effects of reflexology on different symptoms. Patient-level covariates included age, number of comorbid conditions, type of treatment (chemotherapy or targeted therapy with or without
hormonal therapy versus hormonal therapy only), and the CES-D score at baseline. Odds ratios (ORs) and their 95% confidence intervals (CIs) were obtained for the essential parameter of study group for each symptom.

Another concern is the fact that the study crucially depended on the reliability of the 256 carers. It is conceivable, even likely, I think, that many carers from both groups were less than strict in adhering to the prescribed protocol. This might have distorted the results in either direction.

Finally, the study was unable to control for the possibly substantial placebo response that a reflexology massage unquestionably provokes. Therefore, we are not able to tell whether the observed effect is due to the agreeable, non-specific effects of touch and foot massages, or to the postulated specific effects of reflexology.