Recruitment feasibility for deep brain stimulation compared to transcranial magnetic stimulation treatment trials for methamphetamine use disorder

Abstract

Purpose: Retrospective comparison of recruitment strategies for deep brain stimulation (DBS) to transcranial magnetic stimulation (TMS) for treatment-refractory methamphetamine use disorder. Methods: Recruitment occurred over 27 months for the DBS trial and over 12 months for the TMS trial. Data were collected from the initial screening survey through baseline assessment. Recruitment rates from the screening survey to phone screen and from phone screen to baseline assessment were compared across the 2 trials using chi square tests. Reasons for exclusions were tabulated. Results: Of 2892 screening survey respondents for the DBS trial, 167 were phone screened, 9 gave informed consent, and 2 completed the surgery. Of 255 screening survey respondents for the TMS trial (conservatively including 103 individuals who initiated the survey but did not answer any questions), 44 were phone screened, 16 gave consent and 6 received TMS treatments. There were higher rates of progression from screening survey to phone screen for TMS versus DBS (17% vs. 6%, respectively; p < 0.001) and from phone screen to baseline assessment (36% vs. 5%, respectively; p < 0.001). Conclusions: Recruitment for a DBS trial of methamphetamine use disorder proved difficult. Fewer individuals with methamphetamine use disorder both qualified for and were interested in pursuing an invasive neuromodulation treatment approach, as compared to a non-invasive approach, for their disorder. Understanding reasons for recruitment failure is important for determining feasibility of future studies.

Highlights

• •Few individuals qualified for and were interested in this invasive approach.
• •DBS study recruitment progression rates were significantly lower than a TMS study.
• •Barriers to recruitment are reviewed to inform future studies.

1. Introduction

In 2024 about 1.6 million Americans ages 12 years and older met criteria for a methamphetamine use disorder (MUD; (SAMHSA, 2025)). MUD is associated with medical morbidities (Weisbrod et al., 2025) and increased mortality (Han et al., 2021; Stockings et al., 2019) and the available treatments mainly utilize psychosocial interventions (AshaRani et al., 2020); currently there are no FDA-approved medications treatments (Chan et al., 2019; Khalili et al., 2025).

Recent advances allow targeted modulation of specific brain regions using a variety of techniques (Mahoney III et al., 2020). Deep brain stimulation (DBS) which allows long-term stimulation of subcortical structures, was approved in the United States almost 30 years ago for movement disorders (Deng et al., 2021) and currently is being investigated for various substance use disorders (SUD) (Mahoney III et al., 2020; Zammit Dimech et al., 2024). Transcranial magnetic stimulation (TMS) has been studied for MUD specifically (Chang et al., 2022). Both DBS and TMS are experimental (without established efficacy) for the treatment of MUD but address an unmet need. Although DBS is widely performed for movement disorders, the feasibility of DBS for SUD is unknown. A primary endpoint for our DBS study included demonstrating the feasibility of adequate recruitment (n = 5 over 2 years). Here, we report on results of recruitment efforts for a trial of DBS neuromodulation of the nucleus accumbens (Sakai et al., 2024) and compare these results to a non-invasive trial using transcranial magnetic stimulation (TMS) to the left dorsolateral prefrontal cortex for individuals with MUD.

2. Methods

We reviewed screening survey logs from two IRB-approved trials—one for DBS (3447 responses) and the other for TMS (255 responses). For both trials, recruitment utilized online advertising/social media campaigns, print flyers, and recruitment via current patients within the university treatment system.

2.1. DBS trial recruitment [NCT05558358]

Between 12-05-2022 and 03-27-2025, we received 3447 responses to two screening surveys (one abbreviated covering 7 inclusion and exclusion criteria (IEC) followed by a longer survey covering all IEC; see supplemental Table 1 for IEC). Some respondents started a survey on more than one occasion. By matching respondent name, along with one additional identifier (phone number, email, or date of birth) we removed duplicates which resulted in 2892 unique respondents (note - we also removed duplicates where the name was not identical, e.g., a nickname or a hyphenated last name, but the additional identifier matched). We calculated rates of item endorsement for each survey, totaled the number of phone screens, consent sessions, and baseline assessments.

2.2. TMS trial recruitment [NCT06320366]

Using the first year of recruitment for a non-invasive neuromodulation study of TMS for MUD (07-25-2024 to 07-21-2025), we compared the first 255 screening survey respondents from that study to the DBS study. The IEC are overlapping but have key differences highlighted in Supplemental Table 1. Although the timing of recruitment overlapped through March 27, 2025 (we received 83 TMS survey responses during that time), our team prioritized offering DBS to eligible respondents (and dedicated a much larger advertising budget to DBS recruitment) through March of 2025. We acknowledge that prioritizing DBS to eligible participants may have skewed our results but if such an effect occurred, it would have been to favor DBS recruitment. While our TMS screening survey was opened 255 times, only 152 individuals answered at least one screening question. Here, we conservatively count the number of respondents as 255. We totaled the number of phone screens, consent sessions, and baseline assessments. Importantly, we did not offer respondents an “either or” choice between the DBS and TMS trials at the time of survey, outreach, or consent.

2.3. Analyses

We compared the number of survey respondents phone screened (yes/no) for the two trials using the chi square test. Amongst those phone-screened, we compared the number of respondents who consented to participation and started their baseline assessment using the chi square test.

3. Results

3.1. Review of DBS recruitment

Of the 2892 respondents, 1038 (35.9%) passed the abbreviated survey screen (see Supplemental Table 2 below for the rates of exclusion by responses). Our study staff reached out to those individuals to phone screen, provided additional information about the study, and offered to complete the more extensive survey together. Per protocol, staff were to attempt to reach each respondent on 6 occasions using 3 methods (i.e., email, text, and phone call). Individuals who outreached the study team by other means (e.g., emailing or calling the lab directly in response to a news story, referral, etc.) were contacted in the same manner. Most individuals were unresponsive to research coordinator outreach and did not complete the longer survey, and 4 were in the process of being contacted (6 contacts were not yet completed) when the study recruitment ended. The longer survey in Research Electronic Data Capture (REDCap; (Harris et al., 2009)), which was sometimes done while on the phone with study staff, was started 577 times. We phone-screened 167 individuals and 42 were found to be eligible and remained interested in participating (upon hearing more details about the invasive nature of the study or about the requirement for residential care, many individuals were uninterested; some during the phone screen disclosed an exclusionary concern). Thirty-seven scheduled a consent session, 22 attended a consent session, and 9 consented and started a baseline assessment. Seven of these 9 were excluded (2 for current major depression, 2 for inadequately controlled diabetes, 1 for sedative hypnotic use disorder, 1 for opioid use disorder, and 1 for moderate MUD). Two participants met the study inclusion criteria and none of the exclusion criteria following baseline assessment, subsequently completing the surgery.

3.2. Review of TMS trial recruitment and between-trial comparisons

After reviewing 255 survey responses, we phone screened 44 individuals, 25 were eligible and interested in participating following phone screening, and 16 consented and at least started their in-person baseline assessment (10 were eligible to participate following baseline assessment and 6 received TMS treatments). Fig. 1 compares the progression from survey response to phone screen (χ² = 49.4; df = 1; p < 0.001) and from phone screen to baseline assessment (χ² = 32.0; df = 1; p < 0.001). To visualize the differences, percent of sample is presented in Fig. 1; counts were utilized for statistical tests (Supplemental Table 3). (See Table 1.)

Fig. 1.

Comparing Recruitment for the DBS and TMS studies.

Table 1.

Number (percent) across various stages of the recruitment process for both the deep brain surgery (DBS) and transcranial magnetic stimulation (TMS) trials.

	At least started a screening survey; n	Phone screened; n (% of those who at least started a screening survey)	Eligible and interested in proceeding following phone screen; n (% of those who at least started a screening survey)	Consented and at least started baseline assessment; n (% of those who at least started a screening survey)	Eligible after screening; n (% of those who at least started a screening survey)
DBS Trial	2892	167 (5.8)	42 (1.5)	9 (0.3)	2 (0.07)
TMS Trial	255	44 (17.3)	25 (9.8)	16 (6.3)	10 (3.9)

4. Discussion

Few individuals who contacted our study team qualified for and remained interested in participating in the DBS trial. The DBS trial had significantly lower rates of successful phone screens and progression to consent/baseline assessment, in comparison to a TMS trial run by the same study personnel.

4.1. Trial differences

Although both trials have many overlapping IEC, the DBS trial has some key differences that likely impacted study participation. First, the DBS trial required prior multiple treatment failures (i.e., to use an invasive approach we sought to recruit those who are treatment refractory) and unfortunately, many of our survey respondents with severe MUD denied or reported inadequate prior SUD treatment. To assure abstinence around the time of surgery, we required patients in the DBS protocol to be admitted for 1 week prior to surgery for detox and then return to residential treatment following surgery. Potential participants who were most highly functioning (those employed, etc.) often noted that they could not commit to residential care (e.g., worries about losing employment). Future DBS studies with this population might consider alternative approaches, such as a stepped-care model (Uchtenhagen, 2020). Many individuals who were interested in the DBS trial had medical exclusions (e.g., pulmonary hypertension, Raynaud's phenomenon, HIV, Factor V Leiden mutation, etc.) and some were considered to be at elevated risk for surgical morbidity as a result (Behrouz et al., 2016; Weisbrod et al., 2025); many of these concerns do not represent a contraindication for TMS (Kim & Paik, 2021; Rossi et al., 2021). Although not systematically recorded, many individuals, or their families, raised concerns about the invasive nature of the procedure and, despite many instances of making desperate pleas for help in our screening instruments, were not interested in pursuing this approach. One family, who had had a member undergo DBS for Parkinson's did not have such concerns and viewed the approach as standard as opposed to experimental. One of our neurosurgery team members [PDG] also noted that Parkinson's patients often have years to consider whether they would pursue DBS (as medication trials are continued but begin to fail) and may even attend support groups to help in considering whether to pursue DBS. Although the DBS trial had much longer follow-up compared to TMS (1 year vs. 12 weeks), some participants raised questions about end-of-study procedures, including about costs related to battery changes or device removal. After these initial questions, we offered participants enrollment in a 5-year open-label study and agreed to pay up to $25,000 for device removal or first battery change if insurance declined.

4.2. Comparison with other DBS addiction trials

We completed about 1 surgery per year in the DBS MUD trial. Three individuals who completed baseline evaluations were good surgical candidates and interested in pursuing this approach but were excluded for (1) meeting only moderate MUD, (2) meeting criteria for a sedative hypnotic use disorder, or (3) recent re-initiation of opioid use (meeting opioid use disorder criteria). With our small numbers, these 3 additional individuals could potentially have more than doubled the number of surgeries completed in our trial. Thus, the stringency of exclusions impacted feasibility. It is possible that recruitment for other SUDs may proceed at a faster pace. For example, patients on medication assisted treatment who are continuing to use opioids, may be open to novel intervention approaches. Consistent with this, Rezai and colleagues completed DBS surgeries for 4 patients with opioid use disorder over a 2-year period – though it is important to note they had fewer exclusions for co-occurring SUDs (Rezai et al., 2023). One hurdle faced in our trial was the exclusion of co-morbid opioid use disorder at a time of increasing fentanyl contamination of the methamphetamine supply in the Mountain region (Huhn et al., 2024). Davidson recruiting for an open-label pilot for alcohol use disorder, reported a pace of about 3  surgeries per year in their trial (Davidson et al., 2022), but Bach in their blinded sham-controlled 3-site study of alcohol use disorder completed 12 surgeries over 3 years (about 1.33 surgeries/site/year (Bach et al., 2023)). That study apparently sought to recruit 30 participants (Zammit Dimech et al., 2024). Thus, we estimate the range of recruitment pace for recent DBS trials for addiction in North America and Europe to range from about 1–3 surgeries/site/year.

4.3. TMS results compared with the prior literature

Recent work examining non-invasive neuromodulation trials for addictive disorders has supported a non-participation rate (amongst those determined to be eligible, the percent who were not randomized) of about 56% (Sevda et al., 2025). In our trial, many individuals had some level of screening (n = 255), but amongst those assessed at baseline to be eligible to participate (n = 10), six were randomized.

4.4. Conclusions

Our results raise concerns about the scalability of DBS as a neuromodulatory treatment for MUD. Barriers to enrollment faced by our team, which might be avoided in future trials, included requirement for treatment-refractoriness (when treatments available are of near moderate effects sizes and when many patients have not had prior treatment), exclusions for most co-morbid SUDs, considerations about time required in residential care and requirement of severe MUD. Although the efficacy of neuromodulation in addiction treatment remains uncertain, DBS has advantages over other neuromodulation techniques including long-term stimulation, in-office adjustment of stimulation parameters, and local field potential recording for biomarker identification. DBS may provide important insights into neural markers which can guide adaptive approaches to stimulation (Qiu et al., 2024). Lastly, while our experience suggests that an adequately powered clinical trial of DBS for severe MUD is unlikely given stringent IEC, prior studies have also suggested that such work has positively impacted the modest number of individuals (Zammit Dimech et al., 2024) who have undergone this procedure worldwide for SUD, though these studies have generally used open label designs and included small numbers of participants (Mahoney III et al., 2020).

CRediT authorship contribution statement

Joseph Sakai: Writing – original draft, Funding acquisition, Formal analysis, Data curation, Conceptualization. Jody Tanabe: Writing – review & editing, Writing – original draft, Funding acquisition, Data curation, Conceptualization. Kristen Raymond: Writing – review & editing, Data curation. Pamela David Gerecht: Writing – review & editing, Data curation. Hannah Gebhardt: Writing – review & editing, Data curation. Melina Kilen: Writing – review & editing, Data curation. John A. Thompson: Writing – review & editing, Data curation. Drew S. Kern: Writing – review & editing. Steve Ojemann: Writing – review & editing. Marilyn Coors: Writing – review & editing. Daniel Kramer: Writing – review & editing. Susan K. Mikulich-Gilbertson: Writing – review & editing, Formal analysis. Katrina Foster: Writing – review & editing, Conceptualization. Aviva Abosch: Writing – review & editing, Data curation, Conceptualization.

Funding

Data presented in this paper were supported by grants from the National Institute on Drug Abuse (UG3DA054746, K24DA058882) and the Hewit and Kane Family Foundations. The project received additional support from NIH/NCATS Colorado CTSA Grant Number UM1 TR004399. Contents are the authors' sole responsibility and do not necessarily represent official NIH views.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Corresponding author will upload here: https://declarations.elsevier.com/

Dr. Thompson has received support from Medtronic (unrelated to this study). Kern – Received honorarium for consulting/advisement from Abbott, AbbVie Pharmaceutics, Alpha Omega Engineering, Boston Scientific, Insightec, Medtronic; received grants from Medtronic. The University of Colorado School of Medicine, Departments of Neurology and Neurosurgery receives educational grants for fellowship training from Abbott, AbbVie Pharmaceutics, Boston Scientific, Insightec, & Medtronic. Ojemann Received honorarium for consulting from Alpha Omega Engineering and Grants for Fellowship support from Medtronic and Boston Scientific. Dr. Abosch receives investigator-initiated research support and consulting support from Medtronic that are unrelated to this study. Dr. Sakai received travel support from Altimmune as part of his participation in the RECLAIM trial. He has received grants from the National Institute on Drug Abuse (NIDA), the National Institute of Mental Health, the Hewit Foundation and the Kane Family Foundation. Dr. Abosch has received grant funding from NIDA, receives grant funding from the National Institute of Neurologic Disorders and Stroke (NINDS), investigator-initiated research funding from industry (Medtronic), and consulting income from industry (Medtronic). all other authors declares none.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.