Interoceptive bodily awareness in patients seeking pain relief with osteopathic manipulative treatment: an observational cohort pilot study

Context: Interoceptive bodily awareness (IBA) is one's attentional focus on and relationship with comfortable and uncom-fortable (e.g., pain) internal body sensations. Integrating IBA into research on osteopathic manipulative treatment (OMT) is growing, both as an outcome and predictor of treatment outcomes; however, it has yet to be studied in a clinical setting. Objectives: We aimed to conduct a pilot study to measure IBA, with the Multidimensional Assessment of Interocep-tive Awareness (MAIA), in patients seeking OMT for pain, and to test if OMT exposure may be associated with higher IBA as measured by the MAIA. The primary outcome was the change in MAIA scores, and the secondary outcomes were reduction in pain intensity, reduction in pain interference, and increase in participants ’ perception of change post-OMT. Methods: A convenience sample was recruited from individuals presenting for OMT appointments at a College of Osteopathic Medicine OMT teaching clinic. Participants were recruited into our single-arm observational cohort study (n=36), and categorized into one of two groups, OMT-naïve (n=19) or OMT-experienced (n=17), based on prior exposure to OMT. We measured MAIA scores and clinical pain-related outcomespriorto,immediatelyafter,andat1and3weeksafter a usual-care OMT session in the clinic. Covariates including experience with mind – body activities, non-OMT body work, and physical and emotional trauma were also collected to explore potential relationships. We utilized t tests to compare MAIA scores and pain outcomes between groups and across time points. Stepwise regression models were utilized to explore potential relationships with covariates. Results: The OMT-experienced group scored higher on the MAIA scales “ Not-worrying ” (p=0.002) and “ Trusting ” (p=0.028) at baseline. There were no signi ﬁ cant changes in the MAIA scores before and after the single OMT session. Analysis of secondary outcomes revealed that all pain outcomes signi ﬁ cantly decreased post-OMT (p<0.05), with the largest relative improvements in the acute pain and OMT-naïve subgroups, with diminishing e ﬀ ects over time. Conclusions: Assessing IBA with MAIA in a clinical OMT setting is feasible. There were signi ﬁ cant positive correlations between OMT exposure and two of the eight MAIA scales. Future studies are justi ﬁ ed to further explore this relationship.

Central to osteopathic philosophy and treatment [1,2], and akin to the biopsychosocial model in pain management, is the recognition of the reciprocal relationship between the "biological, psychological, and social factors in the understanding of illness, treatment, and the maintenance of health" [3].Currently, there is a paucity of research on how the therapeutic effects of OMT may be acting "top-down" through cognitive and psychological states in addition to the "bottom-up" action of physiologic and mechanical changes on the tissue level [4].As our understanding of pain neurophysiology has evolved to recognize the key role of these "top-down" mechanisms in pain pathogenesis, integrating these perspectives into OMT research and practice presents novel investigative questions.
Pain is an experienced physical sensation with strong emotional elements and is viewed as a combination of somatosensory and interoceptive processes [5].Interoception has been defined as a sense of the internal physiological condition of the body [6], a key element of bodily awareness [7].Neuroscience has revealed a neuroanatomically, functionally, and phylogenetically distinct region of the brain, the anterior insular cortex (AIC), which integrates bottom-up interoceptive afferent signals with exteroception, cognition, and top-down regulatory functions [8].Interoception appears to be integral to the body's ability to maintain homeostasis [6,9].
Interoception has been objectively assessed as interoceptive accuracy (e.g., an individual's accuracy in heart rate perception), accuracy-related confidence ('metacognition' of one's accuracy) [30], and as self-reported interoceptive sensibility [9].Another term for self-reported interoceptive sensibility is interoceptive awareness, which includes the quality of the individual's attentional, emotional, and regulatory components of bodily awareness [9,30].Assessing bodily awareness by interoceptive accuracy alone does not distinguish between adaptive and maladaptive bodily awareness [28,31].
Interventions focused on training direct even-handed attention toward subtle bodily sensations and emotional regulation have been shown to be beneficial for multiple chronic diseases including chronic pain [7,12,[32][33][34][35][36][37].Furthermore, integrating mind-body approaches into pain management have been shown to be superior to usual care, associated with improved pain and lower opioid use [38].Given their clinically meaningful benefits and cost-effectiveness, mind-body approaches to pain management have been classified as high-value interventions [32].
It has been argued that interventions that target IBA may be therapeutic by interceding with the central neurophysiology of chronic pain and by "restoring a person's sense of presence and agency in the world" [39], the latter being a key aspect of an internal health locus of control, which has been positively associated with improved treatment response in chronic low back pain [40,41].D'Alessandro et al. [42] proposed a "reading system able to interpret musculoskeletal disorders from an [integrated] perspective, where the properties of the nervous system are embraced into a more holistic and functional-related context" with a central role for interoception and sensitization, which may be targeted by osteopathic treatment.This reflects a growing interest in integrating interoception into research on OMT, both as outcome and predictor of treatment outcomes.A study looking at the immediate effects of an OMT mobilization technique applied to the temporomandibular joint found treatment associated with increases in interoceptive accuracy and interoceptive accuracy correlated with posttreatment range of motion [43].Another study looking at myofascial release found that baseline interoceptive sensitivity was associated positively with posttreatment range of motion and negatively with posttreatment pain thresholds [44].A study of high-velocity low-amplitude (HVLA) thrust techniques applied to the thoracolumbar junction showed no posttreatment effects on interoceptive accuracy but an inverse correlation of baseline interoceptive accuracy with posttreatment range of motion [45].A randomized controlled trial (RCT) of OMT for chronic low back pain utilized the heartbeat detection task in the functional magnetic resonance imaging (fMRI) scanner and focused on the insula, the brain's hub for interoceptive processing.It found that the OMT group improved in interoceptive accuracy [46].Of note, the OMT intervention in the fMRI study, was more akin to a clinical OMT encounter (a 30 min session, including a structural assessment and treatments selected by the practitioner from a pool of indirect techniques) than the previously mentioned studies that entailed a single technique applied to a single region.
The Multidimensional Assessment of Interoceptive Awareness (MAIA) was developed as a self-report measure to assess IBA, with the particular goal of differentiating between adaptive and maladaptive forms [47].The MAIA has been applied in multiple studies exploring the relationship between IBA and painful conditions [12,13,22,48].To our knowledge, no studies of manual therapy and/or OMT have assessed IBA with the MAIA in a clinical setting.
The aim of this study was to pilot a protocol for assessing bodily awareness utilizing the updated version, MAIA-2, in patients seeking OMT for the treatment of pain in a busy clinic setting.We also aimed to explore the relationship between exposure to OMT, bodily awareness, pain outcomes, and potential covariates in our sample to guide hypothesis development for future studies.We (1) hypothesized that patients with higher prior exposure to OMT would score higher on the MAIA-2 (implying higher levels of IBA) at baseline, and we (2) wanted to explore whether MAIA scores increase post-OMT.

Ethical approval and population
Our study was approved by the Touro University California College of Osteopathic Medicine (TUCCOM) Institutional Review Board (IRB #FWA00009823).A convenience sample was recruited at the university's OMT Teaching Clinic, where OMT is provided free-of-charge to the campus community by predoctoral fellows in Osteopathic Manipulative Medicine (OMM) and precepted by osteopathic physician faculty boardcertified in neuromuscular medicine/osteopathic manipulative medicine (NMM/OMM).Between January and April 2019, study staff screened potential participants on arrival to the clinic, and informed consent was obtained utilizing paper forms.Participants were compensated with $10 gift cards after completing surveys at each time point.Funding was provided by TUCCOM.

Study design
This single-arm observational cohort pilot study recruited participants into two groups, OMT-naïve and OMT-experienced.In prior studies of mind-body interventions that showed an association with changes in MAIA scores, intervention doses were much higher (1-3 sessions/week of contemplative training for 8 weeks or longer [31]) than our single OMT session in this pilot study.Thus, expecting that we may not observe changes in MAIA scores after a single OMT session, we a priori split our sample according to previsit exposure levels to determine potential cumulative changes in MAIA scores associated with 'a higher dose' of total OMT exposure.
The primary outcome in this study was IBA assessed by the MAIA-2.Secondary outcomes were pain intensity, pain interference, and participant's perception of change post-OMT.Although mindful attentional states have been shown to be associated with clinical outcomes such as pain intensity [13], pain prognosis [49], and depression [33], prior studies did not determine a clinically meaningful change in IBA, or effect size.For this pilot study, we aimed to enroll 30 to 40 participants.
Inclusion criteria were: 18 years of age or older, seeking OMT for the relief of pain, qualifying as either "OMT-naive" (no OMT sessions in the past 3 months, <3 total lifetime sessions) or "OMT-experienced" (at least 1 OMT session in the past 3 months, >5 total lifetime sessions), willingness to stay after the appointment to complete the post-OMT survey, and willingness to respond to follow-up email surveys within 3 days of receiving.Exclusion criteria were: unable to read and write in English, and having already participated in this OMT study.
Patients presenting to the OMT clinic were approached by the study staff seated at the front desk, which was located down the hall from the treatment rooms.The front-desk study staff screened and enrolled interested and eligible participants.OMT practitioners (fellows and faculty preceptors) remained in their treatment rooms between patients to avoid seeing whether their next scheduled patient was performing enrollment paperwork.All fellows were aware that the study was occurring (however, they were not aware of the objectives or hypotheses) to adhere to the blinding procedures of remaining in their rooms, and all patients received OMT as usual care regardless of study enrollment status.Patients were instructed not to disclose their study participation status to the OMT practitioners.If OMT practitioners became unblinded to their patient's study participation status, this participant was flagged so that their data could be excluded from analysis.

Measures
Descriptive data were collected at baseline via questionnaire.All questionnaires were collected securely through Research Electronic Data Capture (REDCap).Questionnaires were administered at four time points: immediately before and after the OMT session on study laptops in the clinic, and at 1-and 3-weeks follow-up via secure email.
MAIA-2 was administered at each time point.The MAIA-2 is a 37-item questionnaire that assesses IBA across the following eight distinct but related domains: noticing, not distracting, not worrying, attention regulation, emotional awareness, self-regulation, body listening, and trusting.Higher scores on the MAIA-2 questionnaire imply higher levels of IBA.The MAIA has shown acceptable internal consistency reliability (alpha range 0.66-0.82),good discrimination between known groups, and good sensitivity to change in longitudinal studies [50].
Pain was assessed at each time point with the Numeric Rating Scale (0=no pain; 10=worst imaginable pain), and the impact of pain on daily functional status was assessed with the four-item Patient-Reported Outcomes Measurement Information System (PROMIS)-SF4a Pain Interference Scale.The Global Rating of Change (GRC) scale was assessed at each follow-up time point.This scale provides an opportunity for participants to integrate their impression of improvement [51].The Numerical Rating Scale (NRS), PROMIS, and GRC have been recommended as core outcome measures in chronic pain trials and have been validated, having shown acceptable internal consistency and reliability [52,53].
We chose to capture prior experience with mind-body activities that have an association with MAIA scores [48].Given the known interrelated neurophysiology of touch and interoception [54,55], we also captured experience with the non-OMT body work.All experience questions were answered on a 0-10 scale (0=no experience, 10=highly experienced).Given the relationship between traumatic experiences, interoception, and body-based trauma therapies [56], we collected data on the extent and timing of exposure to both physical and emotional trauma.Exposure to trauma questions were answered on a 0-10 scale (0=none, 10=many times and/or a major event).

Analyses
Our research questions were: are there differences in baseline levels of IBA as measured by the MAIA-2 based on historical exposure to OMT? Are there differences in IBA as measured by the MAIA-2 before and after a single OMT appointment?Are changes in clinical outcomes (pain interference, pain intensity, patient's global impression of change) associated with changes in IBA?In addition, in exploratory analyses, we examined whether potential covariates (exposure to mind-body practices, exposure to body work, exposure to trauma, pain interference, pain intensity, and patient's global impression of change) play a role in associations between OMT and IBA.We utilized Stata Statistical Software Version 15.1 (StataCorp LLC, College Station, TX).
For scoring the MAIA-2 scales, we utilized the instructions as laid forth in the MAIA-2 questionnaire [47].If a scale had less than 50 % of item responses, we did not utilize the score for that scale.
To investigate whether there were differences in baseline levels of IBA between OMT-naive and OMT-experienced participants, we conducted t tests comparing average baseline MAIA-2 scores.We utilized stepwise regression models to further examine the potential association between prior exposure to OMT and MAIA-2.For variables that t tests had revealed a significant between-group difference (p<0.05),we added potential covariates to stepwise regression models one at a time, utilizing p=0.08 as the cutoff for inclusion in the final model and reported results with a beta coefficient and 95 % confidence interval (CI).We further investigated the potential association between MAIA-2 scales and exposure to OMT for the entire pooled sample with exposure to OMT as a continuous independent variable.Changes in the MAIA-2 scores pre-post OMT session were compared between subgroups by t tests.
We utilized t tests to examine pre-post changes in clinical outcomes (pain intensity, pain interference, patient's global impression of change).We conducted clinical outcome tests for the entire sample and separately for OMT-naive vs. OMT-experienced and acute vs. chronic pain.

Recruitment, enrollment, and retention
Sixty-two consecutive potential participants were approached.Eight declined due to time constraints, thus 54 were assessed for eligibility.Among those 54 individuals, 12 were ineligible, 42 were eligible and enrolled, 36 of whom were included in the analysis.Reasons for exclusion from the analysis were: (1) the treatment team became unblinded to patients' study enrollment status prior to or during the OMT session (n=5); and (2) a participant had shoulder surgery during the study period and thus was excluded on the basis that their pain scale was likely an outlier (n=1).We had 100 % retention of participants through the follow-up period, as shown in Figure 1 participant study flow diagram.
Covariates (mind-body activities, non-OMT bodywork, physical and emotional trauma assessed on 0-10 scales) were similar between OMT-naive vs. OMT-experienced, and between participants with acute vs. chronic pain.Therefore, we did not adjust analyses by group for covariates and present our participant characteristics for the sample as a whole.Across the entire sample, mean prior exposure to mind-body activities and non-OMT bodywork scored 5.97 (±2.64) and 4.89 (±2.57), respectively.For prior trauma, mean scores were 4.00 (±3.37; mean age of occurrence at 17.7 [±9.73] years) for physical and 5.17

Interoceptive bodily awareness
At baseline, the OMT-experienced group scored higher than the OMT-naive group on MAIA-2 scales "Not-Worrying" (3.11 vs. 2.41, p=0.002) and "Trusting" (4.02 vs. 3.31, p=0.028) (Figure 2a).Linear regression analysis of these MAIA scales by prior exposure to OMT as a continuous variable showed that 9 % of the variance of "Not-Worrying" (β=0.05;95 % CI, −0.02 to 0.08) and 21 % of the variance for "Trusting" (β=0.04;95 % CI, −0.01 to 0.09) was explained by the number of prior OMT sessions (Table 3).Associations between prior exposure to OMT by group and baseline MAIA scores were not modified by covariates (mind-body activities, non-OMT bodywork, physical and emotional trauma) in logistic regressions, and therefore analyses were not adjusted by covariates.MAIA scores did not significantly differ between the acute and chronic pain subgroups; however, patients with acute pain trended toward higher scores on the MAIA scale "Emotional Awareness" compared with chronic pain (3.92 vs. 3.48, p=0.078) (Table 2).There were no significant differences in MAIA scores at baseline compared to after the single OMT session or at any of the follow-up time points.
Our first hypothesis was accepted partially in that the OMT-experience group scored higher on two of the eight scales.Our second hypothesis was not accepted in this study because we did not observe changes in MAIA scores following the single OMT session.

Pain outcomes
In analyzing our sample as a whole, participants reported significant decreases in all pain outcomes: both current pain and average pain intensity over the previous 7 days, pain interference, and patient impression of change at all followup time points compared to pretreatment, with diminishing effects over time (Table 4, Figure 3).
Subgroup analysis (OMT-naive vs. OMT-experienced groups, acute vs. chronic pain groups) revealed that OMT-naive participants reported the greatest decrease in pain interference, compared to all other subgroups.Participants with acute pain experienced the greatest decrease in pain intensity compared to all other subgroups (Table 5).
Linear regression models for baseline MAIA scores and pain outcomes revealed that "Emotional Awareness" and "Body Listening" baseline scores were significantly related to change in pain interference at 1 week (β=−0.42;95 % CI, −0.74 to −0.10, and β=−0.25;95 % CI, −0.44 to −0.05, respectively), and "Noticing" and "Trusting" baseline scores were related to the change in participants' impression of change at 3 weeks (β=0.10;95 % CI, −0.11 to 2.00, and β=0.85; 95 % CI, 0.08 to 1.61, respectively) (Table 6).While there was no significant association between baseline MAIA scores and changes in pain intensity, we observed trends for both "Not Worrying" and "Emotional Awareness" baseline scores to be associated with changes in current pain at 1 week (p=0.065and 0.056, respectively).

Discussion
We found significantly higher MAIA scores for the "Not-Worrying" and "Trusting" aspects of IBA in our OMT-experienced group compared to OMT-naïve.Given that this baseline data analysis is cross-sectional, we are unable to infer whether this is a characteristic of individuals who more consistently seek OMT, or whether repeated exposure to OMT may contribute to higher scores in these scales.However, the emergence of these two scales in our data are particularly interesting in the context of osteopathy.
The "Not-Worrying" scale measures a person's tendency to not experience emotional distress or worry with sensations of pain.It is associated with an awareness of how emotions affect behavior beyond the ability to name the Baseline data and sample characteristics presented for the entire study sample.Covariates (mind-body activities, non-OMT bodywork, physical and emotional trauma assessed on - scales) were similar between OMT-naive vs. OMT-experienced groups, and between participants with acute vs. chronic pain.OMT, osteopathic manipulative treatment; SD, standard deviation.a OMT session: > min in duration, performed by an osteopathic manipulative medicine (OMM) University Fellow, OMM faculty, or a community doctor of osteopathic medicine (DO).b Type of touch in OMT: =gentle, =mix of gentle and forceful, =forceful.c Exposure to mind-body activities and non-OMT body work: =no experience, =highly experienced.d Exposure to physical and emotional trauma: =none, =many times and/or a major event.The bold numbers for n and % are the totals for the acute and chronic pain groups.emotions associated with sensations.These characteristics may translate into an individual facing less difficulty in engaging in goal-directed behavior and fewer difficulties with impulse control [47].Thus, individuals who score higher on this scale may be more likely to seek OMT as a result of higher levels of goal-directed behavior of seeking relief and treatment.The "Not-Worrying" scale is the scale most strongly (negatively) correlated with other validated measures of anxiety [47].Thus, cumulative prior OMT exposure could also drive the higher scores on this scale, which would be in line with findings from prior studies linking spinal manipulation and reductions in anxiety [57].
Additionally, multiple studies have found increased levels of endocannabinoids after OMT, which have a direct anxiolytic effect [58].
The "Trusting" scale measures a person's tendency to experience their body as safe and trustworthy, and it is distinct from simply the inverse of anxiety.Trusting one's bodily sensations, or perceiving them as helpful for decisionmaking, has been shown to be an important component of chronic pain management [59] and the sense of self [8].One of the most enduring quotes by Andrew Taylor Still is: "To find health should be the object of the physician.Anyone can find disease" [60].OMT is an inherently therapeutic medical intervention; more than an outside force directed at pathology, it aims to facilitate a body's self-regulating and selfhealing mechanisms.Thus, central to the philosophy of the osteopathic physician is trusting the intelligence of the body.
MAIA scores may potentially guide clinicians toward incorporating mind-body approaches into the treatment (via cueing their patient's attention toward their physiologic and emotional experience) and to recommending mind-body approaches to therapy, training, and practices outside of the clinical encounter.For instance, if a patient scored highly on the "Noticing" subscale (awareness of uncomfortable, comfortable, and neutral body sensations), but low on the "Trusting" subscale (experience of one's body as safe and trustworthy), a clinician may explore the emotional and historical aspects of trust as they relate to the patient's symptoms and her body in general, and recommend mind-body practices that focus on building trust.Although we did not observe significant post-OMT changes in MAIA scores, these may become apparent with more OMT sessions.OMT trials that have led to significant and lasting decreases in chronic low back pain involved eight sessions [61].Similarly, studies that have shown significant changes in MAIA scores in response to a mind-body intervention involved longer multi-session interventions [31].A longitudinal trial with more OMT sessions would be necessary to capture changes in IBA as measured by MAIA.
The more robust treatment response in the acute pain group compared to the chronic pain group were as expected.Chronic pain has been described as "stuck" in terms of "the multiple influences on the persistence of pain and pain behavior, and their stubborn resistance to therapeutic intervention" [62].

Limitations
This study was a small pilot study and was not powered for statistical significance, thus our statistical analyses may only suggest possible associations between data points.Because the study was conducted without a control group, causal inferences are not permitted.Our convenience sampling method limits the generalizability of our study.Additionally, given that there were differences in the MAIA scores between OMT exposure groups at baseline, this limits the power of comparing post-outcomes between the groups.In future studies with larger cohorts examining the relationship between OMT and MAIA, we recommend adjusting for prior OMT exposure.
In prior controlled OMT studies that found a positive correlation between OMT and interoception, participants were  unable to distinguish the group to which they had been assigned.This suggested that the findings were not due to a placebo effect [42,46].Because it is impossible for practitioners to be blinded to the treatment in procedural studies, including OMT and surgery, and given that baseline MAIA scores may predict treatment responses [63], it is important to have at least a single-blind placebo control group to reduce bias.Given that our sample's baseline pain levels were relatively low, we cannot infer that our findings would be consistent among individuals with higher baseline pain levels.
In this study, we did not track the physical examination findings of participants nor the specific techniques utilized in the OMT session, and it is known that physiologic responses to OMT and spinal manipulation are approach-specific and vary based on the technique administered, anatomic location of delivery, physical complaint, physical findings, presence of comorbidities, and the clinician's and/or patient's body type.
While conducting an observational study on usual care, OMT is limited in its reproducibility due to the inherent variation between patient presentation as well as the practitionerdependent approach to OMT, gathering this level of detail in future larger studies would enable more in-depth analysis of the potential relationships between OMT and IBA.Additionally, future studies should capture more comprehensive details of the patient's chief complaint, comorbidities, and pain history (including onset, duration, location, characteristic, aggravating factors, relieving factors, timing, surgical history, interventional pain procedures, and the past or current use of any opioid or analgesic medications).

Conclusions
Our findings justify further research on the relationship between OMT and IBA in patients seeking pain relief with OMT.Osteopathic assessments often include a subjective history for psychological health and physical and emotional trauma.
Integrating the MAIA-2 may add a helpful and validated dimension andif appropriatehas the potential to guide the clinician toward incorporating mind-body approaches into the treatment.Future research to guide the integration of the MAIA-2 into patient-clinician interactions and whether this translates into meaningful outcomes are needed.

Figure 2 :
Figure 2: (A) Mean baseline MAIA scores (with standard deviation) for scales with a significant difference between OMT experience groups (experienced vs. naïve; t tests).(B) Association of baseline MAIA scores (for scales with significant difference between experience groups) with the number of previous OMT sessions (as continuous variable from 0-20 sessions, >20 sessions=21 on scale).Pearson correlation coefficients: not-worrying: r=0.46; trusting: r=0.29.

Figure 3 :
Figure 3: Pain outcomes for the entire sample.

Table  :
Baseline data and sample characteristics.

Table  :
Association of baseline MAIA scores with the number of prior OMT sessions (linear regression models).

Table  :
Pain outcomes (entire sample).tests for change in pain outcomes from baseline to after osteopathic manipulative treatment (OMT) visit and and -week follow-up.p values: a <.; t b <.; c <.; d <..SD, standard deviation.

Table  :
Change in pain outcomes by subgroup.

Table  :
Association of baseline MAIA scores with impression of change and pain interference (linear regression models).p≤..MAIA, Multidimensional Assessment of Interoceptive Awareness; β, beta coefficient; CI, confidence interval.R  , explained variance.