Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 8.
Published before final editing as: Int J Clin Exp Hypn. 2026 Oct 6:1–23. doi: 10.1080/00207144.2026.2741062

Digital interactive media for chronic low back pain management: A feasibility study

Mark P Jensen 1, Aaron P Flaster 1, Andrew Humbert 1, M Elena Mendoza 1, Laurence I Sugarman 2, Owen Gottlieb 2
PMCID: PMC13644849  NIHMSID: NIHMS2181837  PMID: 42836317

Abstract

Chronic pain is common and is associated with significant costs and suffering. Therapeutic hypnosis has shown promise as an effective chronic pain treatment, but access to this intervention is extremely limited. Digital applications that offer therapeutic hypnosis, if found to be effective, could address this access problem. The primary aim of the current paper is to describe the results of two studies (Study 1 = individuals with low back pain; Study 2 = individuals with mixed chronic pain conditions) conducted to evaluate the feasibility of the digital hypnosis interactive media application for providing therapeutic hypnosis. Fifty participants in both studies were randomly assigned to one of two groups: (1) 8 weeks of access to the application or (2) 4 weeks of no access followed by 4 weeks of access. Five feasibility and eight clinical outcomes were assessed. The findings indicate that the application is feasible, and that additional work to develop and evaluate the application is warranted. Both studies were prospectively registered at ClinicalTrials.gov (Identifier: NCT06050083).

Keywords: Hypnosis, hypnotic analgesia, chronic pain, interactive media

Introduction

Chronic pain is a highly prevalent condition experienced by 20% to 30% of the world's population (Breivik et al., 2006; Nahin, 2015; Qureshi et al., 2023). Low back pain is the most common chronic pain problem, with an estimated point prevalence of 12% and an estimated 1-month prevalence of 23% (Hoy et al., 2012). Opioid medications continue to be frequently prescribed to treat chronic pain, despite their marginal efficacy and significant risks for harm (Ellis et al., 2024; Manhapra et al., 2023).

Therapeutic hypnosis has been found to be an effective non-pharmacological treatment for a variety of pain conditions (Adachi et al., 2014), including chronic low back pain (McCauley et al., 1983; Rizzo et al., 2018; Tan et al., 2010; Tan et al., 2015). However, access to therapeutic hypnosis is extremely limited, because there are so few health care providers who are trained in its application, and health insurance plans do not usually reimburse hypnosis treatment.

Digital health interventions (also referred to as eHealth interventions) for pain are growing in number and popularaity (Bartels et al., 2025; Weatherly et al., 2024). Preliminary research supports the efficacy of some, but not all, of the existing digital health interventions for reducing pain intensity and improving psychological function (Weatherly etl al., 2024). This approach could also potentially be used to increase access to therapeutic hypnosis. An initial preliminary 2-arm randomized feasibility study of a prototype of such an application for pain management supported its feasibility (Jensen et al., 2025). Specifically, we found that listening to a 10- or 20-minute therapeutic hypnosis session offered by a prototype application was associated with pre- to post-session medium effect size decreases in pain intensity and improvements in positive affect (specifically, feeling “inspired”). We also found that having access to the prototype for 2 to 4 weeks resulted in medium to large effect size improvements in several pain-related outcomes, including average pain intensity, sleep disturbance, pain interference, pain catastrophizing, and ‒ for the group who had access to 4 weeks of the therapeutic ‒ anxiety. The beneficial effects of having access to the protoype accumulated over time, with the magnitude of baseline to 2-week improvements generally being less than the baseline to 4-week improvements.

The primary aim of the current paper is to describe the results of two studies that further evaluate the feasibility of the digital hypnosis interactive media application adapted from the one evaluated previously (Jensen et al., 2025). We incorporated best practice design features for digital health clinical trials in the current study, such as maximizing racial/ethnic diversity in the study population, making the digital application easy to use and tailored to user needs (to maximize engagement), monitoring use, and assessing a spectrum of clinical outcomes (Bartels et al., 2025; MacLean et al., 2024; Svendsen et al., 2020; Weatherly et al., 2024).

We determined a priori that the application would be feasible if it met five conditions: (1) ≥50% of participants use the application at least once/week during the first four weeks of access to the application; (2) ≥50% of participants who used the application report that they are “Somewhat satisfied” or “Very satisfied” with it; (3) ≥50% of participants who used the application per protocol (i.e., at least once/week on average during the first four weeks of access) report that they would like continued access to the application after completing the study; (4) ≥40% of participants interested in continued access report they are willing to pay for that access; and (5) the average System Usability Score (Lewis, 2018) of the application is ≥68 (Banger et al., 2008). A secondary aim was to estimate the effect sizes associated with application use on several pain-related clinical outcomes, to help identify the outcome domains and measures to assess in a fully-powered clinical trial.

Methods

Overview of Study 1 and Study 2

We conducted two studies to address the project aims. These studies used the same experimental design to evaluate the application's feasibility. The primary difference between the two studies was that the first study (Study 1) recruited individuals with chronic low back pain and the second study (Study 2) recruited individuals with general chronic pain (which included individuals with chronic low back pain and individuals with chronic pain as a secondary condition to another health problem, such as spinal cord injury). None of the participants enrolled in Study 1 participated in Study 2. Both studies aimed to enroll 50 participants. Assuming a 20% dropout rate, this would provide complete data from at least 40 participants in each study.

Both Study 1 and Study 2 used a 2-arm prospective randomized controlled trial design to evaluate the application’s feasibility over an 8-week period. Participants were randomly assigned to one of two groups: application access for a full 8 weeks (Access group) or application access for 4 weeks (Waitlist control group). The Waitlist group did not have access to the application for the first 4 weeks of the study. Starting at week five, waitlist participants could then access the application for the final 4 weeks of the study. The 4-week assessment point was the primary study endpoint. Some of the primary (feasibility) and secondary (clinical) outcome data were collected by the application, and other outcome data were collected using the Research Electronic Data Capture (REDCap) application, hosted by the University of Washington Institute of Translational Health Sciences (ITHS). Participants completed the REDCap self-report measures at three timepoints: baseline (before randomization), 4 weeks after randomization (primary endpoint), and 8 weeks after randomization. Both studies were prospectively registered at ClinicalTrials.gov (Identifier: NCT06050083).

Participants

Study 1 and study 2 participant demographic information are presented in Table 1. Compared to the Study 1 participants, the Study 2 participants were older (mean = 48.9 years vs 39.6 years), identified as female less frequently (48% vs. 72%), reported less pain interference (mean = 64.3 vs. 60.8 on the PROMIS Pain Interference scale [see measures section]), and were less likely to have taken an opioid medication in the past week (6% vs. 24%). The between-group difference in opioid use at baseline was due almost entirely to the larger number of participants in the Access condition who reported that they took an opioid in the past week (10, or 38% of the Study 1 participants in the Access condition) than participants in the Waitlist condition (2, or 8% of the partcipants in the Waitlist condition). Study 2 participants also tended to have higher education and income levels and were more likely to be widowed or widowered (18% reported they were widowed or widowered in Study 2 vs. 2% in Study 1). The demographic and baseline pain characteristics between Access and Waitlist conditions were similar for each study.

Table 1.

Baseline demographic and pain characteristics stratified by study and treatment assignment. All values represent count (percent) unless otherwise noted.

Study 1 Study 2
Overall (n=50) Access (n=26) Wailist (n=24) Overall (n=50) Access (n=26) Wailist (n=24)
Age; mean (SD) 48.9 (18.3) 48.3 (20.8) 49.5 (15.6) 39.6 (14.6) 38.6 (12.7) 40.6 (16.7)
Birth Sex
 Female 37 (74%) 19 (73%) 18 (75%) 24 (48%) 12 (46%) 12 (50%)
 Male 13 (26%) 7 (27%) 6 (25%) 26 (52%) 14 (54%) 12 (50%)
Gender identity
 Female 36 (72%) 19 (73%) 17 (71%) 24 (48%) 12 (46%) 12 (50%)
 Male 14 (28%) 7 (27%) 7 (29%) 24 (48%) 13 (50%) 11 (46%)
 Transgender 1 (2.0%) 0 (0%) 1 (4.2%) 2 (4%) 2 (8%) 0 (0%)
 Non-binary 0 (0%) 0 (0%) 0 (0%) 1 (2%) 1 (4%) 0 (0%)
Race identify
 American Indian/Alaska Native 8 (16%) 5 (19%) 3 (13%) 2 (4%) 1 (4%) 1 (4.2%)
 Asian/Asian-American 3 (6.0%) 1 (4%) 2 (8%) 1 (2%) 1 (4%) 0 (0%)
 Black/African-American 19 (38%) 11 (42%) 8 (33%) 29 (58%) 15 (58%) 14 (58%)
 Hawaiian Native/Pacific Islander 1 (2%) 0 (0%) 1 (4%) 1 (2.0%) 1 (4%) 0 (0%)
 White/Caucasian 26 (52%) 13 (50%) 13 (54%) 20 (40%) 10 (38%) 10 (42%)
 Other 2 (4%) 2 (8%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
 Prefer not to answer 1 (2%) 0 (0%) 1 (4%) 0 (0%) 0 (0%) 0 (0%)
Ethnic Identify
 Hispanic/Latino 11 (22%) 6 (23%) 5 (21%) 4 (8%) 1 (4%) 3 (13%)
 Not Hispanic/Latino 38 (76%) 19 (73%) 19 (79%) 46 (92%) 25 (96%) 21 (87%)
 Prefer not to answer 1 (2%) 1 (4%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Education level
 High school graduate/GED 4 (8%) 2 (8%) 2 (8%) 2 (4.0%) 1 (4%) 1 (4.2%)
 Some college of vocational school 21 (42%) 10 (38%) 11 (46%) 7 (14%) 2 (7.7%) 5 (21%)
 2 year college graduate 9 (18%) 5 (19%) 4 (17%) 3 (6%) 3 (12%) 0 (0%)
 4 year college graduate 9 (18%) 3 (12%) 6 (25%) 10 (20%) 7 (27%) 3 (13%)
 Graduate degree 7 (14%) 6 (23%) 1 (4%) 28 (56%) 13 (50%) 15 (63%)
Worst pain intensity (0–10 NRS) 7.4 (1.4) 7.2 (1.3) 7.6 (1.4) 7.3 (1.0) 7.3 (1.0) 7.3 (1.0)
Average pain (0–10 NRS) 5.8 (1.6) 5.8 (1.4) 5.7 (1.9) 5.9 (1.7) 5.9 (1.7) 5.8 (1.8)
Pain interference (PROMIS) 64.3 (6.9) 64.3 (6.7) 64.4 (7.2) 60.8 (4.9) 60.5 (5.3) 61.0 (4.6)
 Missing 1 1 0 1 1 0
Anxiety (PROMIS) 56.8 (10.7) 58.0 (10.5) 55.5 (10.9) 56.0 (6.4) 57.8 (4.8) 54.0 (7.4)
Sleep interference (PROMIS) 57.9 (8.2) 57.3 (7.9) 58.5 (8.7) 56.2 (8.5) 55.2 (7.4) 57.2 (9.5)
Taking opioids
 No 37 (74%) 15 (58%) 22 (92%) 47 (94%) 25 (96%) 22 (92%)
 Yes 12 (24%) 10 (38%) 2 (8%) 3 (6%) 1 (4%) 2 (8%)
 Prefer not to answer 1 (2%) 1 (4%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Open in a new tab

0–10 NRS: 0 to 10 Numerical Rating Scale; PROMIS: Pateint-Reported Outcomes Information Measurement System; SD: standard deviation

Measures

Feasibility

As noted previously, both studies had five primary feasibility outcome domains: (1) frequency of listening to hypnosis sessions; (2) participant satisfaction with the application; (3) interest in having continued access to the application after study completion; (4) willingness to pay for continued access; and (5) system usability. We set a priori specific criteria for each feasibility outcome for determining if the application was feasible.

Frequency of Access.

The application tracked how frequently participants listened to a session. In this domain, we proposed that the application would be feasible if at least 50% of the Access participants used the application at least once/week. We selected this criterion because clinical hypnosis sessions are usually scheduled to occur once a week.

Satisfaction with the application.

We assessed satisfaction with the application at the 4-week timepoint (i.e., 4-weeks after randomization) using a 5-point global treatment satisfaction scale (“Taking all things into account, how satisfied are you with using the application?”), whose response options ranged from 1 (“Very satisfied”) to 5 (“Very dissatisfied”), or “Prefer not to Answer.” We determined that the application would be feasible if ≥ 50% of the participants who listened to at least one application session rated being either “Somewhat satisfied” or “Very satisfied” with the application.

Interest in Continued Application Access.

We assessed interest in continued access to the application by asking the participants, if the application were available after the study ended, “… how interested would you be in downloading and using this application?” Participants responded to this question using a 5-point Likert scale ranging from 0 = “I would not be interested” to 5 = “I would be extremely interested.” We determined that the application would be feasible if ≥ 50% of the Access group participants who used the application per protocol (i.e., at least 4 times during the four weeks after randomization) rated themselves as being at least “… a little interested” in continuing to use the application.

Willingness to Pay for Access to the Application.

We also assessed participants’ willingness to pay for continued access to the application (after finishing the study). However, after the study began, we discovered that the question had been phrased incorrectly on the study forms. The error was corrected immediately, but most of the Study 1 participants and four of the Study 2 participants had responded to the incorrectly worded question. We therefore present the findings for this question only for the Study 2 participants in the Access group who used the application per protocol. We determined a priori that the application would be feasible if ≥ 40% of the Access group participants who used the application per protocol endorsed a willingness to pay for continued access.

Usability.

Finally, we assessed the usability of the application using the System Usability Scale (SUS; Lewis, 2018), a common 10-item measure of the overall usability of applications (sample item, “I thought the product was easy to use”). We selected a feasibility cutoff score of ≥ 68, which is the average SUS score across a large variety of software products (Banger et al., 2008).

Secondary Outcomes: Changes in Clinical Domains

Study 1 and Study 2 were both feasibility studies and therefore not powered to detect statistically significant effects on clinical outcomes. However, the primary purpose of the application is to improve pain-related quality of life domains, so these two feasibility studies were designed to provide data that could inform the design of a full clinical trial.

As noted previously, a secondary aim of the two feasibility studies was to estimate the application’s effects on several pain-related clinical outcomes. This would help: (1) determine if there were improvements in these outcomes and, if so, help (2) identify (a) the outcome domains to assess in a fully-powered clinical trial, and (b) the best measures to assess those domains in such a clinical trial.

We assessed eight clinical outcome domains: immediate (i.e., pre- to post-session) change in (1) pain intensity and (2) positive mood; change in (3) average and (4) worst pain intensity from baseline (i.e., at randomization) to four weeks after randomization (primary endpoint) and to eight weeks after randomization (secondary endpoint); (5) change in pain interference from baseline to 4-weeks and from baseline to 8-weeks; (6) change in anxiety from baseline to 4-weeks and baseline to 8-weeks; (7) change in sleep disturbance from baseline to 4-weeks and baseline to 8-weeks; and (8) change in opioid use from baseline to 4-weeks and baseline to 8-weeks after randomization.

Pain intensity and positive mood.

We measured the first four clinical outcomes by asking study participants to rate their current pain and positive mood before and after each session (to assess the immediate effects of each session on these outcomes) and their average and worst intensity pain in the past 7 days at baseline, 4 weeks after randomization, and 8 weeks after randomization (to assess the effects of having access to the application on usual and worst pain intensity) using 0–10 Numerical Rating Scales (NRSs), whose endpoints were 0 = “No pain/Not inspired” and 10 = “ Worst imaginable pain/Extremely inspired.” NRSs are commonly used to assess pain intensity in clinical trials and have a great deal of evidence supporting their reliability and validity as measures of pain intensity, as evidenced by their strong associations with other measures of pain intensity, moderate associations with domains hypothesized to be impacted by pain intensity (e.g., pain interference), and sensitivity to the effects of pain treatments (Jensen, 2018). NRSs are also recommended over other rating scales in pain clinical trials (Dworkin et al., 2005). Research also supports the reliability and validity of single-item measures of mood, as evidenced by strong test-retest stability coefficients and strong associations with validated multi-item measures of the same constructs (cf. Youngblut and Casper, 1993; Song et al., 2022; Verster et al., 2023).

Pain interference, anxiety, and sleep disturbance.

Pain interference, anxiety, and sleep disturbance at baseline, week 4, and week 8 were assessed by administering the Patient Reported Outcomes Measurement Information System (PROMIS) 8-item short form scales assessing these domains (PROMIS, 2024). PROMIS measures ask respondents to consider the magnitude or frequency of each symptom in the past seven days listed on the short forms using 5-point Likert scales with the specific response options varying as a function of the item (e.g., for magnitude, “Not at all” to “Very much”; for frequency, “Never” to “Always”). Responses are summed and then transformed to T-scores with a mean of 50 and a SD of 10 in the normative samples. All PROMIS measures are scored such that higher scores indicate higher levels of the domain being assessed (i.e., here, more pain interference, anxiety, or sleep disturbance). Evidence supports the validity and reliability of each of these measures in a variety of clinical populations, including individuals with chronic pain (Bhandari et al., 2022; Chimenti et al., 2021). The baseline internal consistencies (Cronbach's alphas) for these measures were: .95, .95, and .90 for Study 1 and .93, .86. .91 for Study 2 for the PROMIS Pain Interference (8-item Short Form), PROMIS Anxiety (8-item Short Form), and PROMIS Sleep Disturbance (8-item Short Form) scales, respectively.

Opioid use.

Opioid use was assessed by a self-report Yes/No question asking participants if they were taking any opioid medications (with opioid examples listed, such as oxycodone, hydrocodone, and fentanyl).

Description of the Digital Interactive Media Application

The interactive media application is currently web-based, and available for internet enabled devices. It was evaluated in the two feasibility studies described in this paper and is a refinement of an initial prototype developed by the study authors (Jensen et al., 2025). The application provides access to a set of professionally recorded clinical hypnosis sessions.

When Study 1 started enrolling participants, users were able to choose from two voices, one female and one male (“Emily” and “Laurence”). Laurence's voice was the default option, but participants could select Emily's voice as the preferred voice at any time. Study staff encouraged participants to listen to both voices so each participant could determine which voice they found most useful. A third female voice (“Carla”) was added after the trials began, based on the participant ratings of satisfaction with their experiences and the number of individuals who were choosing each voice (see below). At the start of enrollment for Study 2, most of the Study 2 participants had access to all three voices, with Laurence’s voice remaining the default option.

Participants could select from among three session durations: about 5 minutes, 10 minutes, and 20 minutes, although the actual duration of the sessions varies slightly from these times, depending on the content selected. The application offered therapeutic hypnosis sessions organized under three goal categories: (1) pain reduction (referred to as increased “comfort” by the application); (2) improved positive mood (referred to as increased “inspiration” by the application); and (3) improved sleep.

We monitored how frequently participants used the application by downloading a report that included information about frequency of listening to sessions and responses to scales administered within the application before and after each session. The application also monitors how user satisfaction with the sessions by asking them to rate how much the like or dislike each session by asking them rate the sessions on a 0–10 NRS (0 = “I disliked it very much,” 5 = “I feel neutral,” and 10 = “I liked it very much”) immediately after each session.

We sought to improve the application and sessions as the study progressed, without compromising the consistency of data collection, by monitoring responses weekly and adjusting the application as needed based on the results of this monitoring. Three results emerged during the first two weeks of monitoring Study 1 participants' use patterns and responses to the pre- and post-session questions. First, there was a notable difference in satisfaction with and response to the two voice offerings, even though the two voices spoke the same words; two weeks after participants were first enrolled, they reported that they liked the sessions with Laurence’s voice more than they liked sessions with Emily’s voice (average ratings were 7.56 and 5.56, respectively). Moreover, the average pre- to post-session reduction in pain intensity for Laurence’s recordings was0.71; the average reduction associated with sessions with Emily’s voice was just 0.08 at the 2-week assessment point. The relative ratings and impact of Laurence’s vs. Emily’s voice was consistent with the feedback we received from an outside consultant for the project, who told us that the “… female voice is presenting a very polished sounding speech, but the speech pattern doesn't communicate clearly because it seems to lack an understanding of what is being said.”

In addition, at this same point in time (i.e., two weeks after participants were first enrolled) we found that the rate that Study 1 participants chose a 5-min session (52%) was much higher than the rates they chose sessions with a 10-min (14%) or 20-min (35%) duration. However, the 5-min sessions resulted in markedly less pain reduction (mean reduction = 0.31 on the 0–10 scale) than the 10-min (mean = 0.83 reduction) or 20-min (mean = 1.56 reduction) sessions. This combination of (1) low rates of satisfaction with and pain reduction in response to Emily’s voice, (2) a very high frequency of choosing the 5-min duration sessions, and (3) a small impact of the 5-min sessions on current pain, led us to speculate that the version of the application being evaluated in Study 1 might not have the same benefits as the original prototype, unless we made some changes.

We decided to modify the application in two ways based on these findings. First, we added an additional female voice option (“Carla”), anticipating that participants would be more satisfied with and respond more to this female voice option than Emily's voice. After introducing Carla’s voice as an option, we found that participants were about as (highly) satisfied with her voice as they were with Laurence’s voice.

Second, we altered the application's onboarding information for new participants to emphasize listening to the longer sessions (i.e., 10 minutes or 20 minutes) instead of the 5-min sessions, at least at first, as they were learning how to use therapeutic hypnosis. The new onboarding information told participants that at first the most pain reduction/benefit would likely come from listening to the longer sessions, that the 5-minute sessions were meant to be used when participants are more familiar with self-hypnosis, and that they can use the 5-minute sessions to cue (or recall) positive responses that they experienced from listening to the longer sessions. Because these changes were made after the studies had begun, and because Study 2 began after Study 1, more Study 2 participants than Study 1 participants (1) had access to Carla's voice and (2) heard this onboarding information.

Procedures

Consent and enrollment

Study 1 (Chronic Low Back Pain).

We recruited the Study 1 participants using Research Match, a national research participant registry managed by the National Institute of Health. We used a multi-step process to identify eligible participants and recruit a racially varied sample. First, we set the Research Match location filters to be anywhere in the United States, health condition filters to '“Chronic Low Back Pain,” demographic filters to any sex/gender, and race/ethnicity filters to in include individuals who identified as American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, Mixed race, or Hispanic/Latino/Latina, until we reached our ≥ 50% recruitment goal of participants of individuals who identified as being in these groups. We over-enrolled individuals who identified as being in these groups in order to facilitate generalization of the findings to as many individuals as possible. After screening and enrolling participants from this first recruitment wave, we evaluated how close we were to our recruitment goals for sex and race identity. After we achieved our goal of enrolling 25 participants who identified as American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, Mixed race, or Hispanic/Latino/Latina, we adjusted the Research Match filters to include individuals who identified as White and non-Hispanic/non-Latino/Latina until we reached our recruitment goal of N = 50. Potentially eligible participants received a study email invitation from Research Match with a link to our online REDCap screening form. The online screening form provided a UW Institutional Review Board (IRB) approved study summary and collected participants' contact information and screening question responses.

The Study 1 screening form assessed for all study inclusion and exclusion criteria except for exclusion criteria (1), (2), and (3), which were assessed via telephone interview later. The inclusion criteria included: (1) 18 years old or older; (2) self-reported chronic low back pain (i.e., pain in the low back that has been ongoing for 3 months or more; and occurs on at least half the days in the past 3 months); (3) average pain intensity in the past week of ≥ 4 on a 0–10 scale, where 0 = “No Pain” and 10 = “Pain as bad as you can imagine”; (4) reads, speaks, and understands English; (5) regular access to the internet per self-report; and (6) access to an internet enabled device. The study exclusion criteria included: (1) history or current diagnosis of primary psychotic or major thought disorder (i.e. Schizophrenia, Psychosis, Bipolar Disorder) with current active symptoms such as active delusional or psychotic thinking (self-report) within the past 5 years per self-report; (2) psychiatric hospitalization within the past 6 months per self-report; (3) psychiatric or behavioral conditions, including active suicidal ideation or intent to harm oneself or others, in which symptoms were unstable or severe within the past 6 months as disclosed during self-report screening; (4) active malignancy (e.g., cancer not in remission) or actively in cancer treatment per self-report; (5) planned surgery in the next six months that is related to the back pain per self-report; and (6) currently receiving or have received hypnosis treatment for any pain condition per self-report.

Potential participants who met the inclusion and exclusion criteria based on the REDCap screening form were then contacted by study staff via telephone within 2 days of staff receiving a REDCap notification of a completed online screening form. During the phone call, study staff administered questions to assess the potential participant’s history or current diagnosis of primary psychotic or major thought disorder, whether they had been hospitalized for a psychiatric treatment within the past 6 months, and the presence of active suicidal ideation or intent to harm oneself or others. Participants that endorsed a history of being diagnosed with schizophrenia, psychosis, or bipolar disorder but did not have any active symptoms were still eligible. Participants that endorsed active suicidal ideation or any of nine follow up questions regarding symptoms that could interfere with participation (e.g., delusional thinking) were referred to the study PI (MPJ), who is a licensed clinical psychologist, for a final eligibility decision (and potential referral to clinical resources). However, none of the study participants who were screened and were otherwise eligible endorsed active suicidal ideation or delusional symptoms.

Study 2 (any type of chronic pain).

The recruitment process for Study 2 was nearly identical to Study 1. The only difference was that instead of requiring that the participant have chronic low back pain, the Study 2 participants could endorse having any type of chronic pain.

Informed Consent.

For both Study 1 and Study 2, participants provided verbal informed consent over the telephone with study staff. The UW IRB approved a waiver of written informed consent for this study.

Statistical Analyses

Descriptive Analyses of application Use

We computed the total number and percent of sessions of different durations (i.e., 5, 10, or 20 minutes) were listened to, and the number and percent that each voice (i.e., Emily, Laurence, or Carla) was listened to across the entire study period for both groups.

Feasibility Analyses

To evaluate the feasibility of the application, and for both studies, we first computed the mean, standard deviation, median and range for the number of times participants in the Access conditions listened to sessions as well as the number and percent of the participants who listened to at least four sessions during the first four weeks of access (i.e., per protocol). To evaluate satisfaction with the application, we computed the average satisfaction rating for those participants in the Access groups who listened to at least one session during the first four weeks of access, as well as the number and percent of the responses to each of the satisfaction response options. To evaluate interest in continued access to the application, we computed the average of the interest ratings of the participants in the Access groups who used the application per protocol (i.e., at least four times from baseline to the 4-week assessment point), as well as the number and percent of the responses to each of the interest question response options. To understand the participants' willingness to pay for continued access to the application, we computed the number and percent of responses to each of the payment question response options for the Access group participants in each study who had used the application per protocol. Given the inherent differences between the two studies, Fisher’s exact tests were used to compare satisfaction and interest for continued access between the two studies. To assess the usability of the application, we computed the average SUS for participants in the Access groups of each study who had used the application at least once.

Clinical Effects Analyses

To estimate the possible effects of the application on clinical outcomes, we conducted two types of analyses. First, to estimate the immediate effects of the application's sessions that target pain intensity, we computed the ranges and average pre- to post-session changes in current pain intensity ratings for each participant individually (including participants in the Access condition who had access to 8 full weeks from baseline to the 8-week assessment point, and participants in the waitlist condition who had access for 4 weeks from the end of week 4 to the end of week 8 assessment point) and computed the median, minimum, and maximum of these values. Median statistics were computed as the measure of central tendency given that the distributions of these changes were anticipated to be positively skewed (i.e., we anticipated that most participants would access the application four times during every 4-week period of access, but a fewer number might access the application many times). Of interest was the general consistency of responses to the application's sessions both within and between individuals. These same analyses were conducted to assess the effects of the application on positive mood. We also conducted linear mixed-effects models (one for each study) for the pre- to post-session changes in pain intensity and positive mood, including a random intercept for each participant to account for clustering, to estimate the effect size and significance levels of these changes. Given the possibility that participants who found the application to be particularly effective might choose to listen to more sessions than those who found the application to be less effective, as a sensitivity analysis, we assessed if there was evidence of informed cluster sizes by testing if the effectiveness of the application on pain intensity and positive mood is associated with the number of times sessions were listened to. This allowed us to determine if the magnitude of the pre- to post-session improvements in pain intensity and positive mood were associated with frequency of listening (i.e., if those who reported more benefits tended to listen more often, which could bias the estimates of average benefit).

Next, we calculated the mean and standard deviation for each of the five continuous outcomes (i.e., average pain intensity, worst pain intensity, pain interference, anxiety, sleep disturbance) at baseline, 4-weeks, and 8-weeks stratified by treatment condition (Access vs. Waitlist group). Standardized effects were calculated to determine the within group change between baseline and both 4- and 8-weeks as well as the between group difference in these changes. We computed 95% confidence intervals were provided for the between group effect sizes for the continuous variables to determine the range plausible effect sizes in the population for these variables. Finally, we conducted Mann Whiteney-U Tests to compare the change in opioid use between the two conditions for each study at each time point.

Results

Frequency of application Use

The frequency of use of the application as a function of voice and duration for the Study 1 and Study 2 participants in both the Access and Waitlist groups is presented in Table 2. Although Study 1 participants were listening to the 5-minute sessions at a > 50% rate compared to the longer-duration sessions early in the study, as can be seen in Table 2, by the 4-week assessment, on average, the Study 1 participants had listened to 5-minute sessions only 22% of the time. Thus, they appeared to respond to the recommendations made by the research staff to listen mostly to the longer sessions. Also, this rate was only slightly larger than that of the Study 2 participants, who listened to the 5-minute sessions 16% of the time. This is likely because more Study 2 participants received the staff recommendations to listen to the longer sessions and Study 2 participants received these recommendations before being given access to the application.

Table 2.

Total counts and proportion of Rose sessions listened to during the first 4 weeks of each study by the Access group participants, stratified by Voice and Duration.

Study Voice 5 minutes 10 minutes 20 minutes Total
1 Carla 10 (2%) 37 (10%) 97 (27%) 144 (39%)
Emily 4 (1%) 1 (<1%) 36 (10%) 41 (11%)
Laurence 67 (18%) 26 (7%) 87 (23%) 180 (49%)
Total 81 (22%) 64 (18%) 220 (60%) 365 (100%)
2 Carla 12 (2%) 16 (3%) 53 (11%) 81 (17%)
Emily 10 (2%) 7 (1%) 85 (17%) 102 (21%)
Laurence 57 (12%) 111 (23%) 136 (23%) 304 (62%)
Total 79 (16%) 134 (28%) 274 (56%) 487 (100%)
Open in a new tab

At the 4 week assessment, Study 1 and Study 2 participants had similar rates of listening to the different session durations, but there was a substantial difference between Study 1 and 2 in the total number of sessions listened to. Table 2 shows that 26 Study 1 Access group participants listened to a total of 365 sessions during the first four weeks of participation, while 25 Study 2 participants listened to > 100 more sessions (487) over this same period. The between-group differences were even more pronounced for the 20-minute sessions: as a group, Study 2 participants listened to the 10- and 20-minute sessions 408 times. Study 1 participants listened to these longer-duration sessions just 284 times ‒ a difference of 124 (or 30%). Consistent with these findings, during the four-week study periods, the Study 1 Access group participants listened to a mean of 14.0 sessions (SD = 17.7) and a median of 6 sessions (minimum = 0, maximum = 72). The Study 2 Access group listened to a mean of 18.7 sessions (SD = 12.6) and median of 14 sessions (minimum = 4, maximum =54).

Feasibility Results

Use Feasibility Criterion

Fifteen (58%) of the Study 1 participants in the Access arm listened to application sessions per protocol (i.e., ≥ 4 times during the 4-week period). All the Study 2 participants in the Access arm (100%) listened to the application sessions per protocol. Thus, application use feasibility criteria (i.e., ≥ 50% using the application per protocol) were met for both studies.

Satisfaction Feasibility Criterion

Among those in the Access groups who accessed the application at least once, 63% of the Study 1 participants were either somewhat satisfied (38%) or very satisfied (25%) with the application; 84% of the Study 2 participants were either somewhat satisfied (28%) or very satisfied (56%) with the application. Thus, the satisfaction feasibility criterion (i.e., ≥ 50% of participants reporting that they are somewhat satisfied or very satisfied with the application) was met. The between-study differences in satisfaction rates were not statistically significant (Χ2(1) = 1.9, p = .09), although the medium effect size for the between-group difference (Cohen's h = 0.48), suggests a trend for Study 2 participants to be more satisfied.

Interest in Continued Access Feasibility Criterion

With respect to interest in having continued access to the application if it were available, 79% of the Study 1 participants expressed interest (29% somewhat, 25% very, and 25% extremely interested), and 76% of the Study 2 participants expressed interest (4% somewhat, 24% very, and 48% extremely interested). Thus, the feasibility criterion with respect to interest in continued access was met for both studies. The between-study differences in rates of interest were not statistically significant (Χ2(1) < −.1, p > .99) and the effect size for this difference was less than small (Cohen's h = 0.07).

Willingness to Pay for Continued Access

With respect to willingness to pay for access to the application if it were available as a product, at the four-week follow-up, 11 of the 17 (65%) Study 2 participants expressed a willingness to pay for access to the application if it were available as a product. Thus, this criterion was met for Study 2. Because the question assessing this criterion was incorrectly worded for almost all the Study 1 participants, we are unable to determine if the feasibility criterion regarding willingness to pay for access to the application was met for Study 1.

Usability Criterion

The average System Usability Scores for the Study 1 and Study 2 participants assigned to the Access condition and who used the application at least once were 81.04 and 86.80, respectively. Thus, the feasibility criterion (i.e., mean SUS score ≥ 68) was met for both studies.

Effects on Clinical Outcomes

Immediate Effects on Pain Intensity and Positive Mood

Across the entire 8-week study period and collapsed across both the Access and Waitlist groups, the median immediate effects of listening to an application session was a 0.71-point decrease in the 0 – 10 NRS (Range of decrease, −0.44 to 2.70) in the Study 1 participants and much larger decrease if 1.62 (range of decrease −0.02 to 4.67) in the Study 2 participants. The median patient-average improvement in positive mood (“inspiration”) was 0.50 points (on the 0–10 NRS, range, −1.00 to 3.00) for the Study 1 participants and over twice this (1.33; range, −2.0 to 4.75) for the Study 2 participants. After accounting for clustering by participants, immediate reduction in pain intensity was 0.83 (SE = 0.12; t = −6.7; p < .01) and 1.55 (SE = 0.15; t = −10.0; p < .01) points in Study 1 and Study 2, respectively, after using the application. Improvements in positive mood were 0.58, (SE = 0.18; t =3.2; p < .01) and 1.67 (SE = 0.23; t = 7.4; p < .01) points higher for the Study 1 and Study 2 participants, from pre- to post-session. Using a bootstrap approach (Nevalainen, Oja, & Datta, 2017), no evidence for informed cluster sizes were present for either the pain (Study 1: p = .74; Study 2: p = .12) or mood (Study 1: p = .64; Study 2: p = .82) measures.

Average Pain Intensity, Worst Pain Intensity, Sleep Disturbance, Pain Interference, and Anxiety

The means and SDs and associated effect sizes of the five continuous secondary outcome variables at each assessment point for Study 1 and Study 2 as a function of treatment condition are presented in Table 3. Given that this was a pilot feasibility study and was not powered to detect statistically significant effects, the focus of these analyses was on the effect size differences between the Access and Waitlist conditions, especially at the primary endpoint (Week 4). For the Study 1 participants, after controlling for the baseline outcome score, the 4-week (primary endpoint) effect size (Cohen's d) differences between the Access and Waitlist conditions were 0.11, 0.43, 0.38, −0.18, and −0.18 (i.e., weak to medium effects, favoring the Waitlist participants for four of the five domains) for average pain intensity, worst pain intensity, sleep disturbance, pain interference, and anxiety, with negative values indicating a greater reduction (or smaller increase) among the Access condition. A different pattern of findings emerged for Study 2, with the differences between the Access and Waitlist conditions being −0.15, −0.40, −0.14, −0.15, and −0.18 (again, weak to medium effects), all favoring the Access group participants.

Table 3.

Means and standard deviations of the secondary clinical outcomes at baseline, 4 weeks, and 8 weeks after study enrollment, and standardized effect sizes for the change in outcome from baseline.

Study 1 Study 2
Mean (Standard deviation) Standardized effect size Mean (Standard deviation) Standardized effect size
Variable Baseline 4 weeks 8 weeks B/L to 4weeks B/L to8 weeks Baseline 4 weeks 8 weeks B/L to 4 weeks B/L to 8 weeks
Average pain intensity (0–10 NRS)
 Waitlist group, N 5.71 (1.88) 5.42 (1.79) 4.96 (1.72) −0.16 −0.4 5.92 (1.83) 5.17 (1.49) 3.46 (2.08) −0.36 −1.3
 Access group, N 5.81 (1.44) 5.69 (1.74) 5.38 (2.12) −0.08 −0.29 5.83 (1.70) 5.04 (1.51) 4.46 (1.84) −0.52 −0.86
 Effect size between treatment groups (95% CI) 0.11 (−0.46, 0.68) 0.12 (−0.45, 0.70) −0.15 (−0.72,0.43) 0.34 (−0.23, 0.92)
Worst pain intensity (0–10 NRS)
 Waitlist group, N 7.63 (1.38) 6.92 (1.47) 6.91 (1.81) −0.51 −0.52 7.29 (0.95) 6.63 (1.35) 4.96 (2.39) −0.7 −2.44
 Access group, N 7.23 (1.34) 7.19 (1.60) 6.69 (1.91) −0.03 −0.4 7.35 (1.02) 6.16 (1.57) 5.73 (2.39) −1.17 −1.59
 Effect size between treatment groups (95% CI) 0.43 (−0.14, 1.01) 0.07 (−0.51, 0.65) −0.40 (−0.98, 0.18) 0.28 (−0.29, 0.85)
Sleep Disturbance (PROMIS)
 Waitlist group, N 58.5 (8.7) 57.3 (9.1) 53.5 (10.8) −0.13 −0.57 57.2 (9.5) 54.0 (10.2) 50.0 (12.6) −0.44 −0.97
 Access group, N 57.3 (7.9) 58.4 (9.7) 57.1 (10.0) 0.14 −0.03 55.2 (7.4) 41.7 (10.7) 50.5 (10.6) −0.48 −0.64
 Effect size between treatment groups (95% CI) 0.38 (−0.19, 0.95) 0.70 (0.11, 1.30) −0.14 (−0.72, 0.44) 0.24 (−0.33, 0.81)
Pain Interference (PROMIS)
 Waitlist group, N 64.4 (7.2) 64.9 (7.2) 61.7 (7.6) 0.07 −0.37 61.0 (4.59) 61.0 (5.9) 56.1 (9.1) 0 −1.08
 Access group, N 64.3 (6.7) 64.4 (6.4) 64.0 (9.0) 0.01 −0.05 60.5 (5.3) 59.4 (5.3) 56.6 (8.7) −0.21 −0.74
 Effect size between treatment groups (95% CI) −0.18 (−0.75, 0.40) 0.35 (−0.25, 0.94) −0.15 (−0.73, 0.44) 0.13 (−0.45, 0.71)
Anxiety (PROMIS)
 Waitlist group, N 55.5 (10.9) 59.2 (13.7) 57.0 (12.4) 0.34 0.14 54.0 (7.4) 54.8 (9.6) 54.1 (10.2) 0.11 0.01
 Access group, N 58.0 (10.5) 60.4 (11.3) 58.2 (11.3) 0.23 0.02 57.8 (4.8) 57.4 (6.7) 55.4 (8.0) −0.08 −0.5
 Effect size between treatment groups (95% CI) −0.18 (−0.75, 0.39) −0.20 (−0.78, 0.37) −0.18 (−0.5, 0.40) −0.28 (−0.85, 0.29)
Open in a new tab

Eight weeks after randomization, after participants in both conditions had the opportunity to access the application, the average decreases (improvements) in average and worst pain intensity were 0.59 and 0.63 points (on the 0 to 10 scale) for the Study 1 participants and 1.92 and 1.98 for the Study 2 participants. Very small improvements were observed for sleep disturbance and pain interference in the Study 1 participants (PROMIS score decreases were 2.6 and 1.5 for these outcomes), but larger improvements were observed for these outcomes in the Study 2 participants: decreases of 6.0 and 4.5, which represent about 50% of the standard deviation of these scales. The average decrease in 6 points observed in the PROMIS Sleep Disturbance T score for the Study 2 participants is at the low end of the range and just below the median amount that is considered clinically meaningful for this measure (i.e., 6 – 12, median 8; Schultz et al., 2023). The decrease of 5 points observed in the PROMIS Pain Interference T scores for the Study 2 participants is in the middle of the ranges considered to be clinically meaningful for this measure (i.e., 3.5 – 5.5) in one study (Amtmann et al., 2016) and 4–6 in another (Yost et al., 2011). Negligible changes in PROMIS Anxiety scores were observed in the participants from either study (i.e., a very small increase in the baseline to 8-week anxiety score in the Study 1 participants of 0.9 points and a very small decrease in the anxiety score in the Study 2 of 1.2 points).

Opioid Use

As noted previously, opioid use was more frequent at baseline in the Access condition (10 participants) than in the Waitlist condition (2 participants) for the Study 1 participants, and remained so at the primary endpoint (Week 4; 10 vs 3) and at Week 8 (8 vs 2). Opioid use was infrequent and similar at baseline between the two conditions in Study 2 (Access condition: 1 participant, Waitlist condition: 2 participants) and remained constant across all time points. Consistent with these statistics, the test for between-group differences in change in opioid medication use were not statistically significant at Week 4 (Study 1: W = 288, p = .66, Study 2: W = 299, p > .99) nor Week 8 (Study 1: W = 265.5, p = 0.35, Study 2: W = 287.5, p > .99).

Discussion

The key findings from the two feasibility studies described in this paper were: (1) the interactive media application evaluated was found to be feasible to use across all a priori feasibility criteria; (2) access to and use of the application was associated with significant improvements in some, but not all, pain and function domains; (3) the beneficial effects of the application appeared to build over time; and (4) different versions of the application appeared to result in different outcomes. These findings have important implications for determining if additional development and testing of the application, specifically, is warranted, as well as for how digital applications for providing therapeutic hypnosis can and should be designed.

Feasibility

The application met all five of the feasibility criteria that were evaluated. Specifically, when offered, the application was accessed and used by 58% of the Study 1 participants and 100% of the Study 2 participants four times or more during the first four weeks of potential access; the feasibility criterion for this domain was ≥ 50%. In addition, 63% and 84% of the Study 1 and Study 2 participants who accessed the application at least once were either somewhat or very satisfied with it; the feasibility criterion for this domain was ≥ 50%. Consistent with this, 79% and 76% of the Study 1 and Study 2 participants expressed an interest in continued access to the application after the study, and 65% of the Study 2 participants endorsed a willingness to pay for this access (willingness to pay for access could not be reliably assessed in the Study 1 participants). The cutoffs for these feasibility criterion variables were ≥ 50% and ≥ 40%, respectively. Finally, the average System Usability Scores for the application were 81.04 (grade “A”; Lewis & Sauro, 2018) and 86.80 (grade “A+”; Lewis & Sauro, 2018)) for the Study 1 and Study 2 participants, respectively. These were not only above the standard 68-feasibility cutoff, but are both the industry goal for application usability (i.e., 80; Lewis & Sauro, 2018). In sum, the findings provide strong support for the feasibility of the application, indicating that additional development and testing are warranted.

Changes in Pain and Function Associated with Application Use

The application had both immediate and longer-lasting benefits on some, but not all, pain and function domains assessed. Given that the content of the sessions offered by the application was designed to reduce pain intensity and improve positive mood (here operationalized as increases in ratings of “inspiration”), application use would be expected to have immediate effects on these variables. The findings from both studies are consistent with this expectation, with average pain intensity reductions of 0.71 and 1.62 points, and average increases in inspiration ratings of 0.50 and 1.30 points on 0–10 scales for Study 1 and Study 2 participants, respectively. The immediate reductions in pain intensity reported by the participants in Study 2, but not Study 1, are comparable to those reported in response to clinical hypnosis (i.e., reduction of 1.31 points on a 0–10 scale) and a mindfulness mediation exercise (i.e., reduction of 1.18 points on a 0–10 scale) provided by trained clinicians (Garland et al., 2017). The immediate improvements in positive mood cannot be compared to those that are the result of in-person treatments sessions, as prior research has not been conducted to examine these effects, to our knowledge. In any case, the immediate effects on pain intensity and positive mood reported by the Study 2 participants support the potential of the application for providing immediate relief in situations when such relief might be sought; for example, during pain flare-ups. Research to evaluate this potential benefit is warranted.

Use of the application was also associated with robust improvements in some, but not all, measures of day-to-day (i.e., average) pain and function domains. The improvement in average pain intensity reported by both the Access and Waitlist participants eight weeks after randomization were 2.37 and 1.46 for the Study 1 and Study 2 participants, respectively. These large effects are considerably greater than those associated with opioid use in chronic pain populations (i.e., average improvement of 0.69 on 0–10 scales [Busse et al., 2018]). It is important to keep in mind, however, that the current feasibility study was not powered to detect significant treatment effects. The low sample size means that the outcome estimates are less reliable, and could be larger or smaller than the estimates that would result from a fully-powered clinical trial with a larger sample size. At most, the findings support the potential for the interactive media application having substantial effect on average pain intensity that might (or might not) exceed those produced by opioids. A fully-powered study is needed to have reliable estimates of these effects.

With respect to the design of a fully-powered study, our findings help identify the secondary outcome domains that are most likely to be impacted by the application, and that therefore should be assessed in such a study. Specifically, we found large and clinically meaningful benefits for the secondary outcome measures assessing pain at its worst in the past 7 days, sleep disturbance, and pain interference, but virtually no change in anxiety or opioid use for either Study 1 or Study 2. The lack of an impact on anxiety is not due to floor effects, as the participants in both studies had average anxiety scores that were larger than half a standard deviation above normative values (i.e., T scores > 55). One possible explanation for the differences in outcome between the different outcome domains is that the content of the therapeutic hypnosis sessions in the application specifically target “inspiration,” pain intensity/comfort, and sleep quality, all of which are hypothesized to be directly related to pain. (Barbieri et al., 2024; Chang et al., 2022; de Vries et al., 2018; Jensen, 2011). None of the application sessions target anxiety reduction or opioid use directly. To the extent that an improvement (decrease) in general anxiety or decrease in opioid use are goals of individuals with chronic pain, the current findings suggest that the application might need to be adapted to include one or more sessions that target these outcome domains. In the meantime, a fully-powered clinical trial evaluating the application as it is should ideally include measures of worst pain intensity, pain interference, and sleep disturbance as primary or secondary outcomes, depending on the population being studied.

The Beneficial Effects of the Application may Build Over Time

A clear pattern of continued improvement over time emerged in the Access participants across eight weeks of access for average pain intensity, worst pain intensity, and pain interference. For each of these three outcome domains, while improvements were observed at 4 weeks, substantial additional improvements occurred over the subsequent 4 weeks of access. We also observed a pattern of continued improvement across several outcome domains in the original pilot study evaluating the application, when examining changes from baseline to 2 weeks and then from 2 weeks to 4 weeks (Jensen et al., 2025). Overall, while there is evidence that the application has immediate effects on an individual's current experience (e.g., of pain intensity or mood), the beneficial effects of the application on an individual's day-to-day pain and function may require some time. For some outcome domains, in order to obtain maximum benefit, it might take as long as 8 weeks or even longer for the full benefits to be observed. Future research should extend the follow-up period, if possible, to determine the time needed to obtain full benefit from the application.

Different Versions of the Application May Result in Different Outcomes

Early monitoring of the effects of the application on the feasibility and clinical outcomes indicated that the female voice offered in the version of the application used when Study 1 began resulted in lower satisfaction ratings as well as less immediate improvement in the pain intensity ratings than the male voice. Moreover, we observed a very high rate of use of the 5-minute session duration option, while at the same time, the immediate beneficial effects of the 5-minute sessions were much lower than those of the 10- and 20-minute sessions. We therefore adapted the application to add a second female voice and provide additional onboarding instructions recommending greater use of the longer duration sessions, at least until participants had more experience and practice with self-hypnosis, at which point the shorter sessions might then be more effective. Because these changes were made a few weeks after the start of Study 1 and soon after the start of Study 2, the Study 1 participants had much less access to the updated version of the application than the Study 2 participants. This difference may explain, at least in part, the higher rates of application use (33% greater) by the Study 2 participants, the higher rates of satisfaction reported by the Study 2 participants, and the overall larger benefits on the clinical outcomes reported by the Study 2 participants. Overall, these findings support the premise that how the therapeutic content is offered by a digital therapeutic (i.e., the specific voices that provide the content and the instructions regarding how to best use the content) matters.

Study Limitations

The study has several limitations that should be considered when interpreting the findings. First, as alluded to previously, the two study samples varied in two important ways ‒ pain diagnosis (i.e., low back pain versus mixed chronic pain conditions) and relative access to different versions of the application (i.e., mostly having access to the original version vs. mostly having access to the adapted version). Thus, we cannot be certain which of these two factors (or both) might have contributed to differences in outcomes between the two studies. Further research is needed to examine the potential effects of pain type and the specific components of the application (e.g., voice listened to, session duration, specific hypnotic suggestions provided) on response to therapeutic hypnosis. Second, the findings from these feasibility studies should not be viewed as providing definitive answers regarding efficacy. Additional research is needed with larger sample sizes to draw conclusions regarding the efficacy the application evaluated here. Third, we did not assess hypnotizability in either study, even though many experts recommend such assessment when feasible (e.g., Kekecs et al., 2024). We did not ask participants to undergo hypnotizability testing due to limited resources, the need to balance assessment burden with motivation to participate in the study, and because identifying predictors of treatment response was not a study aim. Nevertheless, future research to determine the role that hypnotizability has as a predictor of responses to Rose would be useful.

Summary and Conclusions

Despite the studies’ limitations, the findings indicate that the application is feasible and that more work to develop and evaluate the application is warranted. The findings also provide important information that would be useful in the design of a fully-powered clinical trial to evaluate the efficacy of the application. For example, the findings support the use of immediate improvements in pain intensity as an appropriate primary outcome, and immediate improvements in positive mood, as well as improvements in day-to-day average pain intensity, worst pain intensity, pain interference, and sleep disturbance as secondary outcomes. To the extent that reductions in general anxiety or opioid use are treatment goals, the findings suggest that new content may be needed that target improvements in these outcomes. The findings also provide support for the importance of ongoing research to evaluate the efficacy of applications like the one evaluated here to ensure that a rigorously-derived and evolving evidence-base will be used to maximize their benefits.

Acknowledgement:

We would like to acknowledge Thomas S. Marlowe, MD, for his valuable consultation on the quality of the different voice options offered by the web application studied in this paper.

Funding:

The research reported here was supported by the National Center for Complimentary and Integrative Health of the National Institutes of Health (award number R41AT011996). The REDCap instance used in this study is supported by the Institute of Translational Health Sciences, which is funded by the National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1TR002319. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

Conflict of interest: Mark P. Jensen is the author or editor of seven books and facilitates workshops related to the topic of this paper. He receives royalties for the sale of the books and receives financial remuneration for facilitating workshops. M. Elena Mendoza is the co-author of a book related to the topic of this paper and receives royalties for the sale of this book. Laurence I. Sugarman is the author or editor of four books related to the topic of this paper and facilitates workshops related to the topic of this paper. He receives royalties for the sale of these books and receives financial remuneration for facilitating workshops. Mark P. Jensen, Owen Gottlieb, and Laurence I. Sugarman also own equity in a company (HypnoScientific, Inc.) that is developing hypnosis content for use in software applications designed to provide symptom relief. None of the other authors have any conflicts of interest to declare.

Data availability statement:

The data that were collected for the findings presented here were generated as part of an STTR-funded project and are subject to specific access restrictions. The raw data are potentially available from the corresponding author on reasonable request, and subject to approval and compliance with data use agreements.

References

  1. Adachi T, Fujino H, Nakae A, Mashimo T, & Sasaki J (2014). A meta-analysis of hypnosis for chronic pain problems: a comparison between hypnosis, standard care, and other psychological interventions. International Journal of Clinical and Experimental Hypnosis, 62(1), 1–28. doi: 10.1080/00207144.2013.841471 [DOI] [PubMed] [Google Scholar]
  2. Amtmann D, Kim J, Chung H, Askew RL, Park R, & Cook KF (2016). Minimally important differences for Patient Reported Outcomes Measurement Information System pain interference for individuals with back pain. Journal of Pain Research, 9, 251–255. doi: 10.2147/JPR.S93391 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Banger A, Kortum PT, & Miller JT (2008). An empirical evaluationof the System Usability Scale. International Journal of Human-Computer Interaction, 24(6), 574–594. [Google Scholar]
  4. Barbieri A, Veneruso M, Chiarella L, Nobili L, & Cordani R (2024). Hypersomnia and cataplexy in Moebius syndrome. Sleep Medicine, 124, 421–423. doi: 10.1016/j.sleep.2024.10.012 [DOI] [PubMed] [Google Scholar]
  5. Bartels SL, Pelika A, Taygar AS, & Wicksell RK (2025). Digital approaches to chronic pain: A brief meta-review of eHealth interventions – Current evidence and future directions. Current Opinion in Psychology, 62,101976. 10.1016/j.copsyc.2024.101976 [DOI] [PubMed] [Google Scholar]
  6. Bhandari RP, Harrison LE, Richardson PA, Goya Arce AB, You DS, Rajagopalan A, Birnie KA, & Sil S (2022). Clinical utility of CAT administered PROMIS Measures to track change for pediatric chronic pain. Journal of Pain, 23(1), 55–64. doi: 10.1016/j.jpain.2021.06.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Breivik H, Collett B, Ventafridda V, Cohen R, & Gallacher D (2006). Survey of chronic pain in Europe: Prevalence, impact on daily life, and treatment. European Journal of Pain, 10(4), 287–333. doi: 10.1016/j.ejpain.2005.06.009 [DOI] [PubMed] [Google Scholar]
  8. Busse JW, Wang L, Kamaleldin M, Craigie S, Riva JJ, Montoya L, Mulla SM, Lopes LC, Vogel N, Chen E, Kirmayr K, De Oliveira K, Olivieri L, Kaushal A, Chaparro LE, Oyberman I, Agarwal A, Couban R, Tsoi L,. .. Guyatt GH (2018). Opioids for chronic noncancer pain: A systematic review and meta-analysis. JAMA, 320(23), 2448–2460. doi: 10.1001/jama.2018.18472 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Youngblut JM, & Casper GR (1993). Single-item indicators in nursing research. Research in Nursing & Health, 16(6), 459–465. doi: 10.1002/nur.4770160610 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chang JR, Fu SN, Li X, Li SX, Wang X, Zhou Z, Sabina M Pinto SM, Samartzis D, Karppinen J, & Wong AY (2022). The differential effects of sleep deprivation on pain perception in individuals with or without chronic pain: A systematic review and meta-analysis. Sleep Medicine Reviews, 66, 101695. doi: 10.1016/j.smrv.2022.101695 [DOI] [PubMed] [Google Scholar]
  11. Chimenti RL, Rakel BA, Dailey DL, Vance CGT, Zimmerman MB, Geasland KM, Williams JM, Crofford L,C, & Sluka KA, (2021). Test-retest reliability and responsiveness of PROMIS Sleep Short Forms within an RCT in women with fibromyalgia. Frontiers in Pain Research, 2, 682072. doi: 10.3389/fpain.2021.682072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. de Vries GE, Wijkstra PJ, Houwerzijl EJ, Kerstjens HAM, & Hoekema A (2018). Cardiovascular effects of oral appliance therapy in obstructive sleep apnea: A systematic review and meta-analysis. Sleep Medicine Reviews, 40, 55–68. doi: 10.1016/j.smrv.2017.10.004 [DOI] [PubMed] [Google Scholar]
  13. Dworkin RH, Turk DC, Farrar JT, Haythornthwaite JA, Jensen MP, Katz NP, Kerns RD, Stucki G, Allen RR, Bellamy N, Carr DB, Chandler J, Cowan P, Dionne R, Galer BS, Hertz S, Jadad AR, Kramer LD, Manning DC,. .. IMMPACT. (2005). Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain, 113(1–2), 9–19. doi: 10.1016/j.pain.2004.09.012 [DOI] [PubMed] [Google Scholar]
  14. Ellis JD, Dunn KE, & Huhn AS (2024). Harm reduction for opioid use disorder: Strategies and outcome metrics. American Journal of Psychiatry, 181(5), 372–380. doi: 10.1176/appi.ajp.20230918 [DOI] [PubMed] [Google Scholar]
  15. Garland EL, Baker AK, Larsen P, Riquino MR, Priddy SE, Thomas E, Hanley AH, Galbraith P, Wanner N, & Nakamura Y (2017). Randomized Controlled Trial of Brief Mindfulness Training and Hypnotic Suggestion for Acute Pain Relief in the Hospital Setting. Journal of General Internal Medicine, 32(10), 1106–1113. doi: 10.1007/s11606-017-4116-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hoy D, Bain C, Williams G, March L, Brooks P, Blyth F, Woolf A, Vos T, & Buchbinder R (2012). A systematic review of the global prevalence of low back pain. Arthritis & Rheumatology, 64(6), 2028–2037. doi: 10.1002/art.34347 [DOI] [PubMed] [Google Scholar]
  17. Jensen MP (2011). Psychosocial approaches to pain management: an organizational framework. Pain, 152(4), 717–725. doi: 10.1016/j.pain.2010.09.002 [DOI] [PubMed] [Google Scholar]
  18. Jensen MP (2018). Measurement of pain. In Fishman SM, Ballantyne JC, & Rathmall JP (Eds.), Bonica's management of pain. Media, PA: Williams & Wilkins. [Google Scholar]
  19. Jensen MP, Chan JF, Gottlieb O, Sugarman LI, & Stensland E (2025). Use and benefits of a digital therapeutic for chronic pain management: A pilot study. Psychology of Consciousness: Theory, Research, and Practice , Advance online publication. doi: 10.1037/cns0000425 [DOI] [Google Scholar]
  20. Kekecs Z, Moss D, Whorwell PJ, Varga K, Terhune DB, Shenefelt PD, Palsson OS, De Benedittis G, & Elkins G (2024). Best practice recommendations for conducting and reporting controlled trials in clinical hypnosis research. Journal of Evidence-Based Integrative Medicine, 29, 2515690X241274538. doi: 10.1177/2515690X241274538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Lewis JR (2018). The System Usability Scale: Past, present, and future. International Journal of Human-Computer Interaction, 34, 577–590. doi: 10.1080/10447318.2018.1455307 [DOI] [Google Scholar]
  22. Lewis JR, & Sauro J (2018). Item benchmarks for the System Usability Scale. Journal of Usability Studies, 13(3), 158–167. [Google Scholar]
  23. MacLean RR, Shor R, Reilly ED, Reuman L, Solar C, Halat AM, & Higgins DM (2024). Engagement in digital self-management interventions for chronic pain: A systematic review. Clinical Journal of Pain, 41, e1289. doi: 10.1097/AJP.0000000000001289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Manhapra A, MacLean RR, Rosenheck R, & Becker WC (2023). Are opioids effective analgesics and is physiological opioid dependence benign? Revising current assumptions to effectively manage long-term opioid therapy and its deprescribing. British Journal of Clinical Pharamcology. doi: 10.1111/bcp.15972 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. McCauley JD, Thelen MH, Frank RG, Willard RR, & Callen KE (1983). Hypnosis compared to relaxation in the outpatient management of chronic low back pain. Archives of Physical Medicine and Rehabilitation, 64(11), 548–552. [PubMed] [Google Scholar]
  26. Nahin RL (2015). Estimates of pain prevalence and severity in adults: United States, 2012. Journal of Pain, 16(8), 769–780. doi: 10.1016/j.jpain.2015.05.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Nevalainen J, Oja H, & Datta S (2017). Tests for informative cluster size using a novel balanced bootstrap scheme. Statistics in Medicine, 36(16), 2630–2640. doi: 10.1002/sim.7288 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. PROMIS. (2024, 2024). The NIH Patient-Reported Outcomes Measurement Information System. Retrieved from http://www.healthmeasures.net/explore-measurement-systems/promis.
  29. Qureshi AR, Patel M, Neumark S, Wang L, Couban RJ, Sadeghirad B, Bengizi A, & Busse JW (2023). Prevalence of chronic non-cancer pain among military veterans: A systematic review and meta-analysis of observational studies. BMJ Military Health. doi: 10.1136/military-2023-002554 [DOI] [PubMed] [Google Scholar]
  30. Rizzo RRN, Medeiros FC, Pires LG, Pimenta RM, McAuley JH, Jensen MP, & Costa LOP (2018). Hypnosis enhances the effects of pain education in patients with chronic nonspecific low back pain: A randomized controlled trial. Journal of Pain, 19(10), 1103 e1101–1103 e1109. doi: 10.1016/j.jpain.2018.03.013 [DOI] [PubMed] [Google Scholar]
  31. Schultz NM, Morga A, Siddiqui E, & Rhoten SE (2023). Psychometric evaluation of the PROMIS SD-SF-8b instrument in individuals experiencing vasomotor symptoms due to menopause. Health and Quality of Life Outcomes, 21(1), 126. Examining the Concurrent and Predictive Validity of Single Items in Ecological Momentary Assessments [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Song J, Howe E, Oltmanns JR, & Fisher AJ (2022). Examining the concurrent and predictive validity of single items in ecological Momentary assessments. Assessment, 30(5), 662–1671. doi: 10.1186/s12955-023-02206-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Svendsen MJ, Wood KW, Kyle K, Cooper K, Nørregaard CD, Sandal LF, Stoch MJ, Mair FS, & Nicholl BI (2020). Barriers and facilitators to patient uptake and itilisation of digital internveitions for the self-management of low back pain: A system review of qualitative studies. BMJ Open, 10, e038800. doi: 10.1136/bmjopen-2020-038800 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Tan G, Fukui T, Jensen MP, Thornby J, & Waldman KL (2010). Hypnosis treatment for chronic low back pain. International Journal of Clinical and Experimental Hypnosis, 58(1), 53–68. doi: 10.1080/00207140903310824 [DOI] [PubMed] [Google Scholar]
  35. Tan G, Rintala DH, Jensen MP, Fukui T, Smith D, & Williams W (2015). A randomized controlled trial of hypnosis compared with biofeedback for adults with chronic low back pain. European Journal of Pain, 19(2), 271–280. doi: 10.1002/ejp.545 [DOI] [PubMed] [Google Scholar]
  36. Verster JC, Mulder KE, Verheul MC, van Oostrom EC, Hendriksen PA, Scholey A, & Garssen J (2024). Test-retest reliability of single-item assessments of immune fitness, mood, and quality of life. Heliyon, 9(4), e15280. doi: 10.1016/j.heliyon.2023.e15280 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Weatherly S, McKenna T, Wahba S, Friedman A, Goltry W, Wahid T, Abourahma H, Lee K, Rehman A, Odeh A, & Costin J (2024). Effectiveness of digital health interventions (DHI) in chronic pain management: A scoping review of current evidence and emerging trends. Cureus, 16(10), e72562. doi: 10.7759/curesus.72562 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Yost KJ, Eton DT, Garcia SF, & Cella D (2011). Minimally important differences were estimated for six Patient-Reported Outcomes Measurement Information System-Cancer scales in advanced-stage cancer patients. Journal of Clinical Epidemiology, 64(5), 507–516. doi: 10.1016/j.jclinepi.2010.11.018 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that were collected for the findings presented here were generated as part of an STTR-funded project and are subject to specific access restrictions. The raw data are potentially available from the corresponding author on reasonable request, and subject to approval and compliance with data use agreements.

RESOURCES