An Audio-Recorded Hypnosis Intervention for Chronic Pain Management in Cancer Survivors

Abstract

This pilot study evaluated the feasibility, acceptability, and potential efficacy of a 4-week hypnosis audio-recording intervention in cancer survivors with chronic pain. Forty participants were randomly assigned to treatment (n = 21) or wait-list (n = 19) conditions. Pain intensity ratings were lower at Week 4 for both groups. The effect size for pain reduction in the treatment group was d = 0.25 from baseline to 4 weeks, and the interaction effect (Time × Group) was F = .024; η²_p = .001. The small interaction effect may be due to the availability of only one recording and large variability in dose. Qualitative data indicated that the intervention’s benefits included participation in self-care, improved relaxation, and an opportunity to focus on oneself in a positive way. Further efficacy testing of an audio-recording intervention in a fully powered clinical trial is warranted.

Introduction

Unrelieved pain and its physical, psychological, and social sequelae affect more than a third of the nearly 16 million cancer survivors in the United States (Miller et al., 2019; Van den Beuken-van Everdingen et al., 2016). The incidence and severity of chronic pain in cancer survivors is related to cancer type and level of cancer invasiveness, cancer treatments received, and time elapsed since completing treatment (Green et al., 2011; Van den Beuken-van Everdingen, 2012). It is also associated with the presence of nonpain comorbidities, such as anxiety, depression, fatigue, and sleep disturbance (Avis et al., 2017; Syrjala et al., 2014).

Chronic pain management, by definition, is challenging. Pharmacotherapy (e.g., most often analgesics and antiinflammatory agents) has historically been the primary treatment, but its long-term effectiveness can be limited due to loss of efficacy, patient tolerance, and significant adverse effects such as sedation and constipation (Ballantyne, 2017; McNicol et al., 2003; Nafziger & Barkin, 2018). Cancer survivors who have completed curative treatment and have no evidence of ongoing disease but who continue to experience significant pain, could potentially benefit from nonpharmacologic pain management approaches. Nonpharmacologic interventions such as hypnosis have demonstrated efficacy for many chronic pain conditions and rarely have adverse effects (Majeed et al., 2018; Montgomery et al., 2013; Stoelb et al., 2009; Tick et al., 2018).

Hypnosis

Hypnosis has been defined as a state of “consciousness involving concentration” and focused attention that is thought to enhance response to suggestion (Elkins et al., 2015). Hypnotic treatment can facilitate greater control over symptoms such as pain and also increase an individual’s sense of well-being (Jensen & Patterson, 2014; Wortzel & Spiegel, 2017). Hypnosis may improve pain by diminishing anxiety through relaxation and/or by directly influencing the neurophysiological activity that underlies the subjective experience of pain (Jensen, 2011). A hypnotic session for pain management typically involves a hypnotic induction (e.g., suggestions for increased focus of attention and relaxation), symptom-specific suggestions (e.g., for pain reduction or changes in the meaning of sensations), and posthypnotic suggestions for the maintenance of the benefits that occurred during the session, as well as for practicing self-hypnosis outside of formal sessions (Jensen, 2011).

Hypnosis may be delivered in person by a clinician, via audio recordings, or a combination of the two. Delivery by a clinician allows for the development of a therapeutic clinician-patient relationship and tailoring of the hypnosis session to the patient’s clinical presentation (e.g., pain cause, type, and location), hypnotic skills, and pain management goals. Both treatment tailoring and therapeutic rapport are thought to contribute to the beneficial outcomes of hypnosis treatment (Barabasz & Barabasz, 2006; Lankton, 2017; Sheehan, 1980; E. B. Spiegel et al., 2020). Prerecorded audio recordings do not allow for the development of clinician-patient rapport, nor allow for tailoring of treatment. Thus, they would be expected to be less effective than in-person treatment with an experienced clinician. On the other hand, access to an experienced clinician can be challenging—there are limited numbers of clinicians who are trained and experienced in the use of hypnosis for pain management. Moreover, not every individual who could benefit from this treatment approach has the resources to be able to afford this treatment. Prerecorded audio recordings, while perhaps being less effective, could potentially provide a low-cost and more accessible method of delivering hypnosis.

Consistent with these ideas, two meta-analyses concluded that audio recordings’ effects on pain intensity are less than those that occur with in-person hypnotic treatment (Kekecs et al., 2014; Tefikow et al., 2013). On the other hand, the potential efficacy of scripted (i.e., nontailored) hypnosis treatments for reducing abdominal pain related to gastrointestinal disorders has been reported by Palsson and colleagues (Palsson et al., 2006; Palsson & van Tilburg, 2015). In one nonrandomized pilot study, participants (19 treatment, 57 matched standard care controls) with irritable bowel syndrome (IBS) and related abdominal pain were asked to listen once to a different 30-minute audio recording at home every other week for a total of seven recordings over 12 weeks. Beginning in the second week, they were asked to listen daily to the same 13-minute hypnosis recording until the end of the 12-week study period (Palsson et al., 2006). Responders were defined as those who experienced 50% or greater reduction in IBS symptom severity measured by the IBS Severity Scale. Fifty-three percent of the treatment group and 26% of the control (usual care) group participants were classified as treatment responders. The 53% response rate for the treatment group participants was lower than previous studies in this same population that used in-person delivery of the same hypnosis scripts (Palsson & van Tilburg, 2015).

Hypnosis and Cancer-Related Pain

Hypnosis has demonstrated efficacy in reducing acute pain in individuals with cancer. In the adult cancer population, an RCT demonstrated a large effect size (d = 0.82) for a brief 15-minute hypnosis intervention in reducing acute postsurgical pain compared to nondirect empathic listening for women (N = 200) undergoing an excisional biopsy or lumpectomy for the diagnosis of breast cancer (Montgomery et al., 2007). Hypnosis also reduced acute oral pain related to mucositis in the bone marrow transplant population (Syrjala et al., 1992). In that RCT, patients (N = 45) were randomized to cognitive behavioral training (n = 11), hypnosis delivered by a therapist prehospitalization that was recorded for use by the patient during hospitalization (n = 12), therapist contact control condition (n = 12), and usual care control condition (n = 10). The hypnosis group had the lowest pain intensity score at 3 weeks posttransplant with a large effect size (d = 1.28) for women (n = 6) and a smaller effect size (d = 0.18) for men (n = 6).

For chronic pain management in the adult cancer population, research shows that hypnosis in combination with other treatments can reduce pain severity (Butler et al., 2009; Mendoza et al., 2017; D. Spiegel & Bloom, 1983), while promising evidence exists from two pilot trials that hypnosis alone reduces chronic pain in cancer survivors. In one of these studies, using a case-series design, a clinician delivered four to five weekly self-hypnosis training sessions to 4 women receiving breast cancer treatment and 4 breast cancer survivors (Jensen et al., 2012). Standardized hypnosis scripts for management of pain, fatigue, sleep, or hot flashes were used. The women were also asked to listen to at least one audio recording of the different hypnosis sessions daily. Significant pre- to posttreatment decreases in pain intensity, fatigue, and sleep problems were found. Although the effect size for pre- to posttreatment pain intensity was very large (d = 1.47) at the end of the treatment period for the study participants, there was no control condition in this study.

A second trial, using an RCT with a crossover design, tested the efficacy of audio-recorded self-hypnosis instructions versus pharmacologic treatment (analgesics) alone in a sample of 61 adults who had cancer-related pain (Ebell, 2008). The participants in this study were randomly assigned to 4 weeks of either hypnosis audio recordings or pharmacologic treatment only, after which they participated in the second treatment condition. Analysis of data from the initial 4 weeks of treatment, which represent a 2-arm RCT, showed a greater and statistically significant reduction in pain intensity in the hypnosis condition, relative to the pharmacologic treatment alone condition. However, the effect size associated with this treatment effect was not reported, nor were pre- and postpain intensity mean values. Furthermore, although all of the study participants had a history of cancer, it was not clear how many of the participants were in active cancer treatment and how many were cancer survivors. To build on these two studies, further research with large sample sizes and rigorous study designs is needed to establish the efficacy of hypnosis audio recordings for reducing chronic pain in cancer survivors.

In summary, research suggests that although audio recordings of hypnosis sessions may be less effective than in-person hypnosis treatment (even when the scripts used for the in-person treatment are not tailored for each participant individually), audio recordings of hypnosis treatment may still be beneficial for pain management. The advantages of low cost and convenience make the use of audio recordings a potentially viable complementary approach to pharmacotherapy for reducing chronic pain severity in cancer survivors. This intervention also has the potential to empower cancer survivors by engaging them more actively in the management of their pain in the context of their daily lives. Research is therefore needed to evaluate this potential. The current pilot study was designed as a first step for such research. Specifically, we sought to evaluate the feasibility, acceptability, and preliminary efficacy of a standardized hypnosis recording for reducing chronic pain in cancer survivors.

Method

Participants

The study was conducted at the Seattle Cancer Care Alliance—a National Cancer Institute-designated comprehensive cancer center—in Seattle, Washington, USA. The participants were adult cancer survivors who had no evidence of disease and who had completed active cancer treatment. A convenience sample of 40 cancer survivors with chronic pain that was deemed cancer-related were recruited for the study. Participant inclusion criteria included: (a) having moderate or higher pain intensity on average during the last week (3 or greater on a 0-to-10 pain intensity numeric scale); (b) having completed active cancer treatment other than hormonal or maintenance therapy 3 months or longer ago; (c) being 18 years or older of age; (d) having functional fluency in English; and (e) being able and willing to participate in the study.

Study Design

This was a pilot randomized controlled feasibility trial with participants randomly assigned to 4 weeks of either listening to a hypnosis recording daily or a wait-list control condition. Those assigned to the wait-list condition listened to the recording for 4 weeks from Weeks 5 through 8. Evaluating the feasibility of a single hypnosis recording (as opposed to a set of different hypnosis recordings) is a simple approach for determining how a single set of the same suggestions might be beneficial to cancer survivors with chronic pain.

Study measures were administered via telephone interview by a research assistant. Measures were administered at baseline, at Week 4, and at Week 8 (end of study). Each participant also participated in a structured qualitative interview with the research assistant by telephone at Week 8. The research assistant was blind to study-group assignment until the structured interview was conducted. Participants completed at home the following investigator-developed questionnaires: (a) a pre- and postindividual-session questionnaire to obtain ratings of current pain intensity and anxiety prior to and after listening to the recording, level of motivation to use the intervention, and the frequency (“dose”) of listening to the recording; and (b) a daily diary questionnaire to record average pain intensity and anxiety levels, and other treatments for pain and anxiety throughout the 8-week study period.

Measures

Demographic questionnaire.

An investigator-developed questionnaire was used to collect data on gender, age, race/ethnicity, educational level, marital history, and cancer diagnosis.

Primary and secondary study outcomes.

Pain intensity (primary outcome) was measured using the Patient Reported Outcome Measurement Information System (PROMIS; PROMIS-29 v.2.0). The item that assesses pain intensity does so with an average of the past 7 days (0-to-10 numeric scale) with 0 = No pain and 10 = Worst pain imaginable. Anxiety (secondary outcome) was measured using the PROMIS-29 v.2.0 four anxiety items (Items: In the past 7 days…I felt fearful, …I found it hard to focus on anything other than my anxiety, …My worries overwhelmed me, and …I felt uneasy). Response options for the PROMIS Anxiety items are Never, Rarely, Sometimes, Often, and Always. Other secondary outcomes measured by the PROMIS-29 v.2.0 were depression, fatigue, pain interference (effect of pain on cognitive, social, physical, and recreation activities), and sleep disturbance. Per standard scoring procedures, the raw scores for all of the PROMIS scales (i.e., all except for the 0-to-10 numeric scale) were transformed to a T score using available tables, which have a mean of 50 and standard deviation of 10 in a U.S. general sample (Northwestern University, 2020). PROMIS measures have strong support for their reliability and validity (Bartlett et al., 2015; Cella et al., 2010, 2019; Cook et al., 2016). All outcome measures were administered at baseline, Week 4, and Week 8.

Daily diary.

A 9-item investigator-developed daily diary questionnaire was used to assess pain and anxiety intensity on average in the past 24 hours (0–10 numeric scale), and interventions (both medication and other methods) used to control pain and anxiety in the past 24 hours. Participants were asked to complete the diary each day of the study period for a total of 56 days.

Structured interviews.

Qualitative structured interviews with participants at Week 8 were used to evaluate intervention acceptability. Participants were asked about their experience with using the hypnosis recording and if it had an impact on their pain or other symptoms. In addition, they were asked about study logistics, such as difficulties in using the intervention and completing the daily diary; and if they would recommend participating in the study to a friend. Finally, treatment group participants were asked if they used the optional hypnosis recordings that were provided for use during Weeks 5 to 8 of the study. Phone interviews were conducted by the research assistant and audio recorded. The interview recordings were then transcribed verbatim by the research assistant. The interview transcripts were reviewed by the first author (LE) who performed a thematic analysis to identify key thoughts regarding acceptability of the intervention. The first author coded these data by identifying key themes and direct quotes that exemplified these themes.

Study feasibility.

Feasibility of the study was assessed by maintaining a record of the number of individuals who were invited to participate in the study, number of individuals who declined participation and reasons for declining participation, and number of participants who did not complete the study.

Intervention and Study Protocol

Hypnosis treatment.

The script used for the 15-minute hypnosis audio recording was based on scripts published in a therapist manual for hypnotic management of chronic pain (Jensen, 2011). It included a relaxation induction followed by suggestions for relaxation and comfort, as well as posthypnotic suggestions for permanence of the benefits experienced with the recording and for self-hypnosis practice. (See Appendix for key excerpts from the script.) The script specified that the listener was to respond only to those suggestions they find helpful. A digital recording was made by the principal investigator (LE) who read the script verbatim and uploaded the recording to an MP3 player. Five additional recordings were created using other standardized hypnosis scripts for chronic pain (Jensen, 2011; Yapko, 2014). These scripts’ suggestions addressed in order: (a) decreased pain unpleasantness, (b) sensory substitution, (c) hypnotic analgesia, (d) deep relaxation, and (e) feelings of comfort. These additional recordings were made for the treatment group to use during Weeks 5 through 8 as an optional part of the study and were placed in a folder on the MP3 player marked “Do not use until instructed.” Treatment group participants were told that they could use these recordings during study Weeks 5 through 8. The participants assigned to the hypnosis treatment condition met with the PI, who presented an overview of hypnosis and what to expect during the study. This discussion was followed by a 15-minute hypnosis session that entailed having the PI read aloud the primary hypnosis script (i.e., the one that participants had access to during the first 4 weeks of the study) while the participant sat in a chair with his or her eyes closed. The next day, the participant was asked to begin listening to a recording of that same script at least once per day for 4 weeks. Participants were called weekly by the PI to remind them to listen the recording and to complete the pre- and posttreatment individual session questionnaires and the daily diary. At Week 5, they were told they could listen to the other hypnosis recordings in any order and as often as they wanted.

Wait-list condition.

The wait-list participants received weekly calls from the PI during study Weeks 1 to 4 to remind them to complete the daily diary assessing daily pain and anxiety levels and pain and anxiety relief interventions used in the past 24 hours. They received the same intervention during Weeks 5 to 8. They listened to the primary hypnosis recording daily after meeting with the PI for the hypnosis session.

Study procedures.

Clinical staff at the survivorship clinic identified potential participants with chronic pain (greater than 3 on a 0-to-10 pain intensity numeric scale) by reviewing survivor surveys completed by patients prior to being seen at the clinic (see Figure 1). Names and contact information for eligible patients who completed a survey during the past 5 years were given to the PI who contacted patients by phone to describe the study. If the patient was interested in participating, he or she was then screened for eligibility by asking them to report their current pain intensity, their perceived cause of pain, and whether they were still receiving cancer treatment. If found to be eligible and interested in participating, verbal consent to participate was then obtained. Then an appointment was scheduled within the next 3 days for completion of the baseline questionnaires by telephone with the research assistant. The consent form was mailed to the potential participant to sign and return to the PI by mail. Eligible participants who returned a signed consent form were randomized to the treatment or wait-list condition using a SPSS randomization program to yield condition assignments, which were placed in opaque, sequentially numbered sealed envelopes by the research assistant. Once the participant completed the baseline measures, the PI opened the appropriate envelope, called the participant, and scheduled a hypnosis session at the cancer center for within 5 days after baseline measures for the treatment condition and within 5 days after Week 4 measures for the wait-list condition. At the end of the study, all participants returned their pre- and postindividual-session questionnaires and daily diary by mail in a provided prepaid envelope. Participants were able to keep the MP3 player.

Figure 1.

Flow of Participants in the Study

Data Analysis

Descriptive statistics were used to describe the sample’s demographic and clinical characteristics. Means and standard deviations for the primary and secondary outcomes at each assessment point were computed. Effect size estimates (Cohen’s d) of the intervention on the primary outcome were calculated from baseline to Week 4 and baseline to Week 8 for each group. To evaluate the efficacy of the intervention from baseline to Week 4, a series of repeated measure analyses of variance (ANOVAs) were performed with both primary and secondary outcome measures as the dependent variables, and time (baseline and Week 4) and group (treatment and wait-list control) as the independent variables. We did not conduct an analysis of covariance (ANCOVA) to control for potential confounding variables (e.g., age, sex, type of cancer) due to the small sample size. One ANOVA was performed for each of the outcome variables. The primary statistic of interest was the effect size (η²_p) for Time × Group interaction, because it provides a preliminary estimate of the effect regarding the differences in outcome between the treatment and wait-list condition over time. For ANOVAs, interaction effects are interpreted as small (η²_{p =} .01 - .05), medium (η²_p = .06 - .13), and large (η²_p ≥ .14; Cohen, 1988). To estimate the within-subjects effects of having access to the audio recordings on pain intensity for the wait-list control group, we compared the difference between the 4-week measure (pretreatment) to the 8-week measure (posttreatment) using a paired t-test. The treatment dose was determined by the mean number of times the participants reported that they listened to the hypnosis recording during the treatment phase. Finally, to further understand the participants’ pain experience during the study, we compared qualitative and quantitative data for those participants who had a clinically significant reduction in pain intensity (i.e., a 2-point or more decrease in pain intensity from pre- to posttreatment on the 0-to-10 pain intensity numerical scale; Farrar et al, 2001).

Results

Sample Demographic and Clinical Characteristics

The sample consisted primarily of non-Hispanic, Caucasian females, most of whom had some college and were married or living with a partner (see Table 1). The mean age was 52.7 years (SD 13.1, range 21–76). The most frequent cancer diagnosis was hematologic cancer followed by breast cancer. No statistically significant differences in demographic or clinical characteristics were found between groups.

Table 1.

Baseline Demographic and Clinical Characteristics

	Treatment (n=21)		Wait-list Control (n=19)		Total (N=40)
	M (SD)	Range	M (SD)	Range	M (SD)	Range
Age (in years)	49.1 (13.06)	21–65	56.42 (12.39)	28–76	52.67 (13.10)	21–76
	n (%)		n (%)		n (%)
Gender
Female	16 (76)		15 (79)		31 (78)
Male	5 (24)		4 (21)		9 (23)
Ethnicity
Non-Hispanic/Latino	20 (95)		17 (90)		37 (93)
aRace
White	17 (81)		17 (90)		34 (85)
Black	1 (5)		1 (5)		2 (5)
American Indian	0 (0)		3 (16)		3 (8)
Asian	3 (14)		0 (0)		3 (8)
Hawaiian/Pacific Islander	1 (5)		1 (5)		2 (5)
Education
High School Graduate	2 (10)		2 (11)		4 (10)
Some College	12 (57)		7 (37)		19 (48)
Bachelor’s Degree	3 (14)		2 (11)		5 (13)
Some Graduate School	1 (5)		2 (11)		3 (8)
Graduate Degree	3 (14)		6 (32)		9 (23)
Marital Status
Never Married	5 (24)		5 (26)		10 (25)
Married/Living with Partner	13 (62)		10 (53)		23 (58)
Divorced/Separated	3 (14)		4 (21)		7 (18)
Cancer Diagnosis
Hematologic	7 (33)		7 (37)		14 (35)
Breast	7 (33)		4 (21)		11 (28)
bOther	4 (19)		2 (11)		6 (15)
GYN	1 (5)		4 (21)		5 (13)
Sarcoma	2 (10)		2 (11)		4 (10)

^aMore than one race may be identified by study participants

^bOther cancer diagnoses include brain, prostate, rectal, and salivary

Feasibility

The first study aim was to test the feasibility (recruitment, attrition, adherence, logistical issues) of delivering the hypnosis intervention to cancer survivors with chronic pain. The study was open for accrual for 12 months (January to December 2016). We extended study recruitment to the cancer center’s Survivorship Day held in June 2016 and the pain clinic in order to meet our target enrollment of 40 participants. We were also approached by 3 cancer survivors who enrolled in the study after learning about it on the clinicaltrials.gov website.

We used the CONSORT participant flow diagram to summarize study recruitment screening, eligibility, allocation, and attrition (see Figure 2). During the accrual period, we attempted to screen a total of 108 cancer survivors who were referred to us or contacted us. Seventy-five cancer survivors consented to be screened, and 65% (n = 49) of these met the study inclusion criteria. Of these, 82% (n=40) consented to participate. We enrolled these 40 participants (31 female, 9 male) into the study, and 21 were randomized to the treatment group and 19 to the wait-list control group. Of the 40 participants, 95% (n = 38) completed the 4-week measures and 75% (n = 30) completed the 8-week measures. One treatment group participant completed only the baseline questionnaires, did not attend the hypnosis session, and was lost to follow-up. Five wait-list control group participants did not use the recorded hypnosis intervention during Weeks 5 to 8; 2 withdrew from the study at Week 4 due to health issues and/or feeling too busy to participate, and 3 were lost to follow-up. Two wait-list control group participants discontinued listening to the audio recording (in Weeks 5 to 8) due to being tired of hearing the same script (n = 1) and due to feeling anxious when listening to the primary script (n = 1). We also had 1 treatment group participant who experienced cognitive dissonance while listening to the recording. She found the script’s imagery of sending her pain away in a box incongruent with her Buddhist philosophy of being present and accepting her pain. She was invited to change the image while listening to the recording so that it was consistent with her beliefs. This was acceptable to her and she continued listening to the hypnosis recording daily.

Figure 2.

CONSORT Participant Flow Diagram

Also, a modification that was made to the study protocol was the use of Skype for delivery of the in-person hypnosis session for some participants. Because we had 4 participants who were unable to travel to the cancer center due to geographical distance, the PI met with the participants by Skype to discuss and provide the initial hypnosis session.

Acceptability

The second study aim was to evaluate the acceptability of the recorded hypnosis intervention among cancer survivors with chronic pain. Structured interviews were used to achieve this aim. Interviews were completed by 75% (n = 30) of the participants. Fifty percent (n = 15) of the interviewees reported experiencing pain reduction, while 97% (n = 29) of the participants described experiencing relaxation, calmness, and/or improved sleep after they started listening to the hypnosis recording. The worst aspect of the hypnosis recording reported by 27% (n = 8) of the participants was the script. The most common complaint was that it was boring to hear the same script every day; a few others (13%, n = 4) reported that they found the specific wording of the script bothersome. Nine of the sixteen treatment group participants (56%) who participated in the structured interview reported that they listened to one or more of the five additional optional recordings during Weeks 5 to 8.

Acceptability of the intervention was further supported by 93% (n = 28) of the participants who said they would recommend the study to others with chronic pain. However, many of these participants said they would not recommend it to everyone; that is, they thought that they would only recommend it to someone who is both open to hypnosis and able to focus for 15 minutes. Finally, an important insight from the interview findings was that a number of participants reported that participation in the trial gave them permission to take care of themselves, gave them an opportunity to focus on themselves, and resulted in them feeling more relaxed (see Table 2).

Table 2.

Excerpts and Codes from the Structured Interview Qualitative Data

Q1. What was the hypnosis experience like?	Code
I think I was surprised to find that before doing the hypnosis study, I rarely took a set amount of time to just be with myself. One of the benefits that I really got from this study is that when I ran the recording, that I actually stopped and found time to just focus on myself.	Self-care
It was really interesting.. kind of slowly learning how to control the, the panic part of the pain coming on. After just you know, going through the sessions for well probably the first week, after that, the pain would start but I felt reassured. Yeah, I didn’t panic and therefore cause it to escalate because I knew that the breathing and the visualization would at the very least halt it and in most cases kind of back it off.	Control pain
It gave me a sense of control. It allowed me to take 12–14 minutes and just stop everything without being um that the pain would be so severe. It became a portal to relax.	Control pain
I never felt that I was hypnotized at all at any time. I thought that it was great for the visualizing and, and thinking about it…. Did it do anything for me other than give me 15 minutes of space, of working on healing myself? Yes it did.	Self-care
I have already that predisposition to be acceptable and open to have your mind take, take over and work with and improve and so my experience was awesome. It helped me a lot.	Mind-body connection
Q2: Did you find that hypnosis affected your pain?	Code
….it doesn’t make the pain go away completely, but it relaxed me a little bit to focus on something other than the pain…	Temporary pain relief
It did. I think it allowed me to relax and just be more present with it….and just taking the time to focus on my breathing and my body during that time I would notice the other stuff before and after and then after I had done it for a while, I would actually stop at times during the day when I felt particularly in an amount of pain and just do a couple minutes of self-hypnosis, and that was really helpful too.	Pain management skills
Oh, pain! I don’t know. I don’t know that it really affected it. Maybe it just reshaped it into something that, that maybe it was a little more under my control.	Control pain
Immediately after the hypnosis, after the 15 minutes…..I felt a really great relief of pain and it would last for a few hours, and then you know, and then the pain would start a little bit again….it helped me a lot and the pain has gotten less since I started the study.	Lasting pain relief
I don’t think it really affected the pain at all. I think again, it put you in a little bit calmer mood, you know, in that way I think it was somewhat effective, but I didn’t think it was you know, very effective at all in reducing the actual pain level.	Ineffective pain relief
Um, there were a couple times that afterward, I wasn’t sure if my pain felt a little less because the hypnotism had actually helped or if it was because my anxiety had been up before too, but since it had lowered my anxiety, when my anxiety is down and I am able to control that, then I am not panicking as much over the pain and so I don’t know if that’s what made it feel like a little less because it was only a tiny bit and so I’m just not sure which of those it was. But still even if it only helped me notice my pain while I was listening to it, that’s still a tool that can I keep to use to try because that’s 15 minutes of relief. You know, if I’m ever having a really really bad day, 15 minutes of relief is a big deal.	Temporary pain relief
It, it reduced it. As far as well physically I mean, it was right there, I went from one and a half tablets of Oxycodone….. to one..	Medication reduction

Efficacy

The third study aim was to explore potential efficacy of providing cancer survivors with chronic pain the opportunity to listen to a single audio recording of hypnosis for reducing pain at 4 weeks, compared to a wait-list condition. Table 3 presents the means and standard deviations for the primary outcome (pain intensity) and secondary outcomes (anxiety, depression, fatigue, pain interference, and sleep disturbance) at baseline, Week 4, and Week 8. As can be seen, baseline levels of the study participants’ mean T scores were greater than the normalized score of 50 for all secondary outcomes except for sleep disturbance.

Table 3.

Means and SDs for Primary and Secondary Outcomes

Treatment Condition	N	Baseline Mean (SD)	N	Week 4 Mean (SD)	N	Week 8 Mean (SD)
Primary Outcome
Pain Intensity
Intervention	21	6.62 (1.47)	20	6.15 (2.18)	16	6.06 (2.14)
Wait-List	19	6.89 (1.49)	18	6.61 (1.38)	14	6.57 (1.79)
aSecondary Outcome
Anxiety
Intervention	21	57.62 (10.45)	20	59.23 (9.04)	16	56.12 (7.13)
Wait-List	19	58.00 (6.65)	18	54.84 (9.28)	14	53.63 (11.37)
Depression
Intervention	21	55.61 (8.68)	20	56.47 (7.57)	16	52.40 (5.62)
Wait-List	19	55.33 (6.61)	18	52.57 (8.64)	14	53.04 (9.48)
Fatigue
Intervention	21	60.17 (8.42)	20	61.01 (6.68)	16	61.10 (7.01)
Wait-List	19	62.97 (9.19)	18	60.64 (7.40)	14	59.76 (8.73)
Pain Interference
Intervention	21	62.43 (5.94)	20	61.69 (5.70)	16	59.86 (6.47)
Wait-List	19	64.74 (5.18)	18	64.56 (5.59)	14	63.34 (5.34)
Sleep Disturbance
Intervention	21	50.47 (3.25)	20	50.52 (3.45)	16	51.97 (3.99)
Wait-List	19	50.83 (4.53)	18	51.12 (5.07)	14	50.35 (4.90)

^aA score of 50 is the mean for the U.S. general population. Higher scores equal more of the symptom.

Primary outcome: Pain intensity.

On average, both the treatment group and the wait-list control group reported small reductions in pain intensity at Week 4, relative to baseline levels. The effect size from baseline to Week 4 (treatment group d = 0.25; wait-list group d = 0.20) and baseline to Week 8 (treatment group d = 0.31; wait-list group d = 0.19) indicates the intervention had a very small effect on the groups’ pain intensity, while the Time × Group interaction effect of the hypnosis intervention on pain intensity from baseline to Week 4 was very small (F= .024; η²_p = .001). For the wait-list control group (n = 14), mean pain intensity decreased from Week 4 to Week 8 (t = 0.62, 95% CI −0.90, 1.61) after using the intervention for 4 weeks, but the effect size was less than small (d = 0.03).

Because this study was not powered to detect statistically significant effects on the primary outcome, we examined clinical significance. We found a clinically important difference (2-point or greater reduction; Farrar et al., 2001) in the pain intensity score from pre- to posttreatment for 10 participants. Of these, a 3-point or greater reduction was achieved by 9% (treatment n = 2, wait-list n = 1) of the study participants and 2-point reduction occurred for 21% (treatment n = 5, wait-list n = 2). Because 3 of the participants who experienced a clinically important difference in pain intensity did not complete the interview at Week 8, we were only able to compare the quantitative and qualitative data for 7 participants. The quantitative and qualitative data matched for 5 participants. That is, both reductions in pain intensity as assessed by the 0-to-10 numeric scale and verbal report indicated a reduction in pain. The 2 participants who verbally reported that the intervention did not reduce their pain were both from the wait-list control group.

Secondary outcomes: Anxiety, depression, fatigue, pain interference, and sleep disturbance.

Little or no improvement in the secondary outcomes was found for the treatment and wait-list conditions as indicated by the mean group scores (see Table 3) from baseline to Week 4 to the end of study. The interaction effect sizes (η²_p) from baseline to Week 4 ranged from .000 to .051 with the treatment’s effect being the largest for anxiety (F = 1.929, η²_p = .051), depression (F = 1.424, η²_p = .038), and fatigue (F = 1.192, η²_p = .032).

Dose effects.

The mean number of days the treatment group (n = 15) listened to the audio recording during Weeks 1 to 4 was 20.53 (SD 6.35). The mean number of days that the wait-list group (n = 13) listened to the audio recording during Weeks 5 to 8 was 22.23 (SD 5.78). The number of days that the recording was listened to by the participants (n = 28) ranged from 7 to 28. We had missing data for the remaining 12 participants, as they did not return the pre- and postindividual-session questionnaires, withdrew from the study, or were lost to follow-up. Common reasons recorded on questionnaires for not using the intervention daily included being too busy (32%), feeling unwell (14%), feeling exhausted (14%), forgetting to listen (14%), or growing tired of listening to the same recording every day (7%).

Discussion

The findings from this pilot study indicate that a study to evaluate the effects of a hypnosis audio recording for chronic pain management was acceptable and feasible. The findings also indicate that the intervention was associated with meaningful pain reductions for some of the study participants. As evident by the percent of eligible cancer survivors who participated in the study, cancer survivors are interested in hypnosis as a way to manage chronic pain. Moreover, the majority of participants were willing to listen to the audio recording for a majority of the treatment days and said they would recommend this study to others.

To our knowledge, this is the first RCT to evaluate the daily use of a single standardized hypnosis recording to help cancer survivors manage chronic pain. Although this pilot study was not powered to detect statistically significant effects on the primary and secondary outcomes, we were able to evaluate clinical significance and found 29% (n = 10) of the study participants who completed the pre- and posttreatment measures (n = 34) experienced a two-point or more reduction in pain intensity from pre- to posttreatment. In comparison, qualitative data at Week 8 indicated that 50% (n = 15) of the interviewees reported a reduction in their pain. For many of these participants, pain reduction lasted from 15 minutes to a few hours after listening to the recording. Being able to “control the pain” was also reported as a benefit of the intervention.

It is possible that the outcome effects might have been greater if more participants had listened to the recording on a daily basis for 4 weeks as recommended. In fact, the dose ranged from 25% to 100% during the 28-day treatment period. One reason for not using the recording for the full treatment period for a few participants was that they became bored with listening to the same recording. Although we were interested in determining if a very simple single recording might be acceptable and feasible, in hindsight, the use of a single recording may have ultimately limited efficacy. Interestingly, though, we had only 1 participant discontinue the study due to boredom, although she completed all of the study measures. Our attrition rate at Week 8 was 25% (n =10) which was less than the 38% attrition rate reported in an RCT evaluating a daily 10-minute hypnosis recording for 14 days with healthy adults (n =56) experiencing stress (Cardeña et al., 2013). Attrition rates in symptom management intervention studies with cancer survivors have ranged from 23% to 36% (Foster et al., 2016; Johannsen et al., 2017; Nakamura et al., 2013; Smith et al., 2019). These interventions included mindfulness-based cognitive therapy (Johannsen et al., 2017) and an online coping skills program (Smith et al., 2019) for chronic pain, a web-based resource for increasing self-efficacy to manage fatigue (Foster et al., 2016), and mind-body interventions for sleep disturbance (Nakamura et al., 2013). Our lower attrition rate may be due to the weekly phone contact that the PI had with participants. It is also possible that participants remained in the study due to other benefits of the intervention. For example, many participants reported that it was beneficial for them to take time every day (in this case, 15 minutes) to relax and focus on taking care of themselves.

In addition to offering participants a variety of recordings, we may have found more beneficial effects of the intervention if the recordings were of varying times and made by different individuals, so that the participants had a choice of voices. Giving participants more options in terms of scripts, voices, and duration of the recordings is a way of “tailoring” treatment to each participant’s needs. Tailoring hypnosis treatment is thought to be essential for effective hypnosis treatment (Jensen, 2017).

Limitations

This study has a number of limitations that should be considered when interpreting the results. One limitation was that only one researcher performed the thematic analysis of the qualitative data, so we were unable to check rater reliability to ensure that the identified themes provided an accurate representation of the participants’ comments about the treatment. In addition, we compared the recorded hypnosis intervention to a usual care condition. This did not allow us to control for attention from the study clinician (i.e., during the first session) or expectancy effects, both of which can influence outcomes. Further, the participants were a sample of convenience who self-selected for study participation and consisted of predominately female non-Hispanic Caucasian participants. The extent to which the findings generalize to the general population of cancer survivors, including individuals who are from minority populations or are men, is not known.

Summary and Conclusions

Despite the study’s limitations, the findings support the feasibility and acceptability of a fully powered clinical trial to evaluate the efficacy of audio recordings for chronic pain management in cancer survivors. That study is currently underway (clinicaltrials.gov: NCT03867760). In this study, a variety of hypnosis recordings are being compared to an attention control condition (a set of relaxation recordings). We are also collecting data on psychological variables to identify mediators of the beneficial effects of hypnosis. An understanding of these mechanisms of action will help in better integrating this intervention into standard care.

We learned several lessons from the pilot study that informed the design of a clinical trial. Because some participants reported that it was monotonous to listen to the same hypnosis recording for 28 days, the intervention was modified. Participants now have access to four different recordings of standardized hypnosis scripts for 3 days each, and they may then listen to any recording daily for the remainder of the 4-week study. In addition, recordings are provided in both a female and male voice. We are using recruitment mailings rather than “cold calls” since some participants in the current study shared that they were hesitant to answer their phone since they did not recognize the caller’s phone number. Furthermore, they are given the option of “opting out” of being contacted by returning a provided postcard. Lastly, we added a second study site to increase the racial and ethnic diversity of our sample.

The goal of this study was to evaluate the feasibility, acceptability, and preliminary efficacy of a recorded hypnosis intervention in reducing chronic pain in cancer survivors. The intervention was feasible and acceptable among the majority of the 40 participants, but the single recording and variability in dose (e.g., number of times the recording was listened to) may have limited the intervention’s effect in reducing pain. However, for a substantial portion of the participants, the reduction in pain was found to be clinically significant. Because this is a low-cost and convenient intervention with the potential to engage and empower cancer survivors in the management of their chronic pain, further study with a modified hypnosis intervention is deemed warranted and is currently underway.

Appendix

Appendix 1. Key Excerpts from the Hypnosis Script

1. Text from the induction: “Now, in a moment, I’m going to count from 1 to 10. As I count each number, take a deep, satisfying breath and then let it all the way out…And, as I count each number, I’d like you to imagine something with me. I’d like you to imagine that, as I count each number, you feel yourself settling down, one level of comfort at a time, into a deeper and deeper experience of comfort and relaxation” (Jensen, 2011, p. 147).

2. Text from the suggestions for relaxation and comfort: “With every breath you take, breathing comfort in and breathing discomfort out, you can wonder how you can be feeling more and more comfortable, right here and now. You may be pleased, of course, but you may also be surprised that it’s so much easier now to simply focus on relaxation and comfort, to simply not pay attention to anything other than your comfort…So much easier to enjoy the relaxing, peaceful comfort of each breath. So simple, so natural, to attend to your breathing” (Jensen, 2011, p. 158).

3. Posthypnotic suggestions: “Any time you want to feel more comfortable, all you have to do is just rest back and take a very deep, very satisfying breath and hold it…All the time you are learning new ways of helping yourself to feel more comfortable…..and these skills are becoming more automatic and effortless each time you practice” (Jensen, 2011, p. 206).