Verbal framing, rather than measured cognitive-affective variables, is associated with short-term outcomes after spinal manipulation: a secondary analysis of a randomized controlled trial.

Nonspecific chronic low back pain (NSCLBP) is a major global cause of disability and is highly prevalent in young adults, with consequences for quality of life and socio-economic participation [ 1 ]. Spinal manipulation is recommended as part of multimodal care for NSCLBP; however, short-term responses vary considerably across individuals, and the mechanisms underlying this variability remain only partially understood [ 2 ]. This marked heterogeneity in response has stimulated growing interest in factors beyond the biomechanical effects of the intervention itself [ 3 ]. In this context, contemporary perspectives in manual therapy have progressively shifted attention from the technique in isolation to the broader therapeutic encounter [ 4 ]. Within a biopsychosocial (BPS) framework, treatment responses are shaped not only by the physical intervention but also by contextual and relational factors embedded in care delivery, including clinician communication, therapeutic alliance, and patients’ interpretations of treatment meaning [ 5 ]. Evidence syntheses in NSCLBP indicate that such contextual factors within conservative care, such as communication processes, therapeutic relationship variables, and related expectancies, are associated with clinical outcomes [ 6 ]. Complementing this evidence, placebo-controlled meta-analytic findings across physiotherapy interventions for musculoskeletal pain suggest that a substantial proportion of observed treatment effects may not be attributable to the specific intervention effects alone, underscoring the clinical relevance of contextual and relational contributions [ 7 ]. Implementation research highlights therapist communication skills, therapeutic alliance, and patient expectations as key facilitators – and barriers – to applying a BPS approach in physiotherapy practice [ 8 ]. Consistent with these perspectives, qualitative syntheses propose explicit conceptual frameworks for preferred communication behaviors in physical therapy, reinforcing the centrality of relational processes [ 9 ]. Against this background, cognitive-affective variables such as expectations, attitudes toward treatment, and expectation fulfillment have received increasing attention [ 10 , 11 ]. Expectations reflect beliefs about the likely benefits or risks of an intervention, attitudes capture evaluative judgments shaped by prior experience and emotion, and expectation fulfillment reflects the alignment between anticipated and experienced change [ 12 ]. These variables are influenced by clinician communication and may amplify or attenuate treatment effects [ 13 ]. In NSCLBP, higher expectations and greater fulfillment have been associated with better short-term outcomes following manual therapy and exercise [ 14–16 ], whereas attempts to tailor treatment to patients’ attitudes have produced mixed findings [ 17 , 18 ]. Among the communicative elements embedded in the therapeutic encounter, verbal framing represents a concise yet powerful mechanism that can shape expectations, perceived threat, and emotional appraisal, thereby influencing pain processing and symptom experience [ 5 ]. However, the extent to which verbal framing interacts with individual cognitive-affective variables during spinal manipulation remains insufficiently explored. Moreover, young adults, despite their elevated risk of early pain chronicity, are underrepresented in this line of research [ 1 ]. Within this emerging literature, a recent randomized controlled trial (RCT) in young adults with NSCLBP compared positive, neutral and negative verbal framing delivered immediately before a standardized lumbar spinal manipulation. That trial demonstrated that verbal framing influences short-term pain intensity (NPRS-11) and lumbar pressure pain threshold (PPT) responses, providing the empirical context for the present secondary analysis [ 19 ]. Building on these findings, the current study examines whether baseline expectations, attitude toward manual therapy, and expectation fulfillment are associated with these outcomes within each verbal suggestion condition. Pain intensity provides an immediate and clinically relevant measure of perceived symptom change [ 20 ], while PPT provides a complementary index of pain sensitivity that is commonly used to characterize short-term hypoalgesic responses after manual interventions [ 21 , 22 ]. Focusing on these outcomes preserves the theoretical scope of the parent trial [ 19 ], while concentrating the analysis on measures most plausibly sensitive to communicated framing and cognitive-affective influences. Accordingly, this secondary analysis is situated within a contextual/BPS framework in which verbal framing represents an experimentally manipulated interpersonal component of the therapeutic encounter, potentially acting through both explicit expectations and broader relational processes. This framing aligns with evidence highlighting the role of communication behaviors, contextual factors, and therapeutic alliance in musculoskeletal pain care [ 6 , 8 , 9 ]. The parent trial collected a broad range of self-reported and mechanistic outcomes in young adults with NSCLBP [ 19 ]. The present study represents a post hoc, exploratory (hypothesis-generating) secondary analysis with a focused analytic scope. The primary analyses examine how baseline expectations, attitude toward manual therapy, and 24-hour expectation fulfillment relate to NPRS-11 and lumbar PPT within each verbal suggestion group. Ancillary exploratory associations with additional outcomes collected in the parent trial (e.g., disability, stress, sleepiness, kinesiophobia, range of motion, and biomarkers) are reported in the Supplementary Material to provide broader descriptive context but are not used to support the main conclusions of this manuscript, in line with guidance for transparent reporting of secondary analyses [ 23 , 24 ]. Specifically, this secondary analysis aimed to (1) examine, within each verbal framing group (positive, neutral, negative), the associations between baseline expectations, baseline attitude toward manual therapy, and 24-hour expectation fulfillment with NPRS-11 and lumbar PPT immediately post-treatment and at 24 hours; and (2) explore whether baseline expectations and attitudes add predictive value for 24-hour NPRS-11 beyond randomized verbal framing group allocation and baseline NPRS-11. We did not aim to identify latent NPRS-11 trajectory subgroups (e.g., latent class/growth mixture models), which typically require larger samples and more repeated outcome assessments than available in the present dataset [ 25 , 26 ]. Given the exploratory nature of this secondary analysis, we hypothesized that higher baseline expectations, more favorable attitudes, and greater expectation fulfillment would be associated with lower NPRS-11 and higher lumbar PPT within each verbal framing group, and that baseline expectations and attitude would explain additional variance in 24-hour NPRS-11 beyond group allocation and baseline pain.

This secondary analysis used data from an RCT conducted between October and December 2024 [ 19 ]. In the parent trial, all participants received the same standardized lumbar spinal manipulation protocol; randomization determined only the verbal framing condition (positive/neutral/negative) delivered immediately before treatment. The physiotherapist delivering the spinal manipulation, the outcome assessors, and the statistician were blinded to verbal framing allocation. Participants necessarily heard the suggestion they received, but they were not informed about the existence or content of alternative framing scripts and were therefore blinded to the comparative study hypotheses and conditions. The investigator delivering the verbal framing could not be blinded; however, this individual was not involved in outcome assessment or statistical analyses. These organizational procedures were implemented to minimize bias and are reported transparently in line with reporting guidance for complex and non-pharmacological interventions and known blinding limitations [ 23 , 27 , 28 ]. The parent trial received ethical approval (09/2024) and was prospectively registered at ClinicalTrials.gov ( NCT06537739 ). All procedures adhered to the Declaration of Helsinki [ 29 ]. The protocol for the present secondary analysis was not separately registered a priori; accordingly, the analyses reported here should be interpreted as post hoc and hypothesis-generating. Participants were recruited in the Lublin Voivodeship (Poland) through advertisements, social media, and university noticeboards. Data collection was conducted at the Human Performance Laboratory of John Paul II University in Biała Podlaska. Eligibility was screened by an independent clinician with over 10 years of experience in musculoskeletal pain management. Inclusion criteria were: age 19–30 years; NSCLBP lasting longer than 3 months; no prior exposure to spinal manipulation; and no use of non-steroidal anti-inflammatory drugs for at least one week before participation. Exclusion criteria comprised: ongoing steroid therapy; history of lumbar spine surgery; lumbar injury within the previous six months (e.g., fracture); diagnosed diabetes, rheumatic, or endocrine disease; endometriosis; leg symptoms; body mass index >33; and current enrollment in a physiotherapy degree program. An a priori sample size calculation was performed for the parent trial using G*Power, based on a one-way ANOVA with three groups (α = 0.05; power = 0.95), and assuming a large overall effect size (Cohen’s f = 0.50), which yielded a required total sample size of 66 participants (22 per group) [ 19 , 30 ]. The full dataset ( n  = 66) was included in the present secondary analysis. No separate a priori power calculation was conducted for the exploratory association and prediction analyses reported here; therefore, results should be interpreted as hypothesis-generating. Before spinal manipulation, the principal investigator delivered one of three standardized verbal framings adapted from previous research [ 31 , 32 ]. The framings differed only in the content of a single scripted statement regarding expected effectiveness and potential symptom change: Positive: “Spinal manipulation is a very effective method of treating spinal disorders, which will significantly reduce the low back pain you are experiencing.” Negative: “Spinal manipulation is not an effective method of treating spinal disorders. Thus, it may temporarily increase the low back pain you are experiencing.” Neutral: “Spinal manipulation has not been verified for its effectiveness in affecting your level of low back pain.” Positive: “Spinal manipulation is a very effective method of treating spinal disorders, which will significantly reduce the low back pain you are experiencing.” Negative: “Spinal manipulation is not an effective method of treating spinal disorders. Thus, it may temporarily increase the low back pain you are experiencing.” Neutral: “Spinal manipulation has not been verified for its effectiveness in affecting your level of low back pain.” To maximize replicability and minimize contamination by non-scripted communication, the verbal framing was delivered using a written script with standardized introductory and closing statements that were identical across groups (welcome, procedural explanation, and transition to the manipulation session). No additional reassurance, discouragement, or discussion of prognosis or treatment effectiveness was provided beyond the scripted sentence. The physiotherapist delivering spinal manipulation was not present during delivery of the verbal framing and was instructed to limit communication to standard procedural cues (e.g., positioning, safety instructions, and post-treatment guidance), without commenting on expected outcomes. This level of specification follows contemporary reporting guidance for complex and non-pharmacological interventions in rehabilitation contexts, emphasizing detailed description of intervention components and delivery procedures to support reproducibility and interpretation [ 9 , 23 , 33 ]. After delivery of the assigned verbal framing, participants received one session of lumbar spinal manipulation following validated protocols [ 34 , 35 ]. The intervention was delivered by a physiotherapist with over 10 years of experience in manual therapy, who was not informed of the framing script delivered and interacted with participants according to a standardized procedure to minimize the risk of unblinding. All participants provided written informed consent prior to participation. Allocation to groups (positive, neutral, negative) followed a 1:1:1 ratio. The randomization sequence was generated using Research Randomizer (v4.0) [ 36 ] and concealed using sequentially numbered, opaque, sealed envelopes prepared by an independent researcher. The participant-facing procedure was identical across groups up to and including baseline assessment, with the only planned between-group difference being the scripted verbal framing delivered immediately before spinal manipulation [ 23 , 31 ]. The study sequence was as follows: (1) standardized welcome and procedural briefing; (2) baseline outcome assessment (t0); (3) delivery of the allocated verbal framing by the principal investigator; (4) lumbar spinal manipulation delivered by a separate physiotherapist using the same protocol for all participants; (5) immediate post-intervention outcome assessment (t1); and (6) 24-hour follow-up assessment (t2). Participants were not informed about the existence or content of alternative verbal framing scripts and were instructed not to disclose the suggestion they received to the physiotherapist. Outcome assessors and the statistician remained blinded to group allocation throughout the study. After completion of the 24-hour follow-up, participants were fully debriefed regarding the aims of the study. Full details of all outcomes collected in the parent RCT are reported elsewhere [ 19 ] and are not repeated here. For the present secondary analysis, the primary dependent variables were NPRS-11 and lumbar PPT at L3-L5, assessed at baseline (t0), immediately post-intervention (t1), and 24 hours post-intervention (t2). The primary predictors of interest were three cognitive-affective variables: baseline expectations, baseline attitude toward manual therapy, and 24-hour expectation fulfillment. All three variables were collected prospectively in the parent trial but were not analyzed inferentially in the primary publication [ 19 ]; their examination here is therefore exploratory and specific to this secondary analysis: Baseline expectations were assessed at t0, prior to delivery of the verbal framing, using a single-item question (“How much improvement do you expect after the treatment?”) rated on a 15-point Likert scale ranging from −7 (lowest expectations) to +7 (highest expectations). Higher positive values indicate stronger expectations of benefit, whereas negative values reflect pessimistic expectations [ 37 , 38 ]. Attitude toward manual therapy was assessed at baseline (t0) using a single-item question (“What is your attitude toward treatment with manual therapy?”), rated on a 5-point Likert scale (1 = very bad, 2 = bad, 3 = average, 4 = good, 5 = very good), with higher scores indicating a more favorable attitude [ 39 ]. Expectation fulfilment was assessed at 24 hours post-intervention (t2) using a 15-point Likert scale ranging from −7 (“very much below expectations/marked worsening”) to +7 (“far above expectations/marked improvement”), with 0 indicating no perceived change. Participants responded to the question: “To what extent did the treatment meet your expectations?” Higher values indicated greater fulfillment of expectations, whereas negative values reflected unmet expectations or perceived worsening [ 37 , 38 ]. Baseline expectations were assessed at t0, prior to delivery of the verbal framing, using a single-item question (“How much improvement do you expect after the treatment?”) rated on a 15-point Likert scale ranging from −7 (lowest expectations) to +7 (highest expectations). Higher positive values indicate stronger expectations of benefit, whereas negative values reflect pessimistic expectations [ 37 , 38 ]. Attitude toward manual therapy was assessed at baseline (t0) using a single-item question (“What is your attitude toward treatment with manual therapy?”), rated on a 5-point Likert scale (1 = very bad, 2 = bad, 3 = average, 4 = good, 5 = very good), with higher scores indicating a more favorable attitude [ 39 ]. Expectation fulfilment was assessed at 24 hours post-intervention (t2) using a 15-point Likert scale ranging from −7 (“very much below expectations/marked worsening”) to +7 (“far above expectations/marked improvement”), with 0 indicating no perceived change. Participants responded to the question: “To what extent did the treatment meet your expectations?” Higher values indicated greater fulfillment of expectations, whereas negative values reflected unmet expectations or perceived worsening [ 37 , 38 ]. Additional outcomes collected in the parent trial were examined only in ancillary exploratory correlation analyses and are reported in Supplementary Tables S1–S3; these analyses were not used to support the main conclusions of the present manuscript. Analyses were structured according to the aims of the secondary analysis and focused on NPRS-11 and lumbar PPT. To explore whether the association between cognitive-affective variables and outcomes differed across verbal framing groups, Spearman correlation analyses were conducted separately within each group. Correlations examined relationships between baseline expectations, baseline attitude toward manual therapy, and 24-hour expectation fulfillment with NPRS-11 at t1 and t2 and with lumbar PPT values at t1 and t2 (L3–L5). These analyses were descriptive and comparative in nature and were not intended to formally test interaction or moderation effects. Given the exploratory intent of aim 1 and the additional ancillary analyses reported in the Supplementary Material (Tables S1–S3), no adjustment for multiple comparisons was applied. Accordingly, p -values from these analyses should be interpreted as descriptive and hypothesis-generating rather than confirmatory. For correlation coefficients, 95% confidence intervals were calculated using Fisher’s r-to-z transformation. To explore predictors of NPRS-11 at 24 hours (t2), a multiple linear regression model was fitted. Baseline NPRS-11 (t0) was entered a priori to account for initial symptom severity. In a subsequent step, group allocation (dummy-coded, with the neutral group as reference), baseline expectations, and baseline attitude toward manual therapy were evaluated as candidate predictors using stepwise selection (entry p   0.10), reflecting the exploratory aim and limited predictor set. Model assumptions were assessed for linearity, homoscedasticity, independence of errors (Durbin – Watson statistic), normality of residuals (Shapiro – Wilk test), and influential observations. Effect sizes are reported as correlation coefficients (r) for association analyses and standardized regression coefficients (β) with corresponding model fit indices for regression analyses. Missing data were handled using listwise deletion; no imputation was performed. All analyses were conducted using SPSS version 29.0 (IBM Corp., NY), with a two-tailed significance threshold of α = 0.05.

Sixty-six participants (50 women, 16 men; mean age, 21 ± 2.96 years) were enrolled from 73 screened individuals ( Figure 1 ), and all participants completed the study. Baseline sociodemographic characteristics, outcomes used in the secondary analysis, and cognitive-affective variables were comparable across the positive, neutral, and negative verbal framing groups (all p  > 0.05; Table 1 ). Importantly, no between-group differences were observed at baseline for NPRS-11, lumbar PPT (L3–L5), or baseline expectations. Additional baseline variables not directly relevant to the primary analyses are reported in Supplementary Table S4 for completeness. Figure 1. Study flow diagram in accordance with the CONSORT guidelines [ 40 ]. Table 1. Baseline characteristics of participants (mean ± SD or n (%)). Characteristics Positive group ( N  = 22) Negative group ( N  = 22) Neutral group ( N  = 22) Sociodemographic   Age (years) 21.77 ± 2.96 20.55 ± 3.46 21.00 ± 2.92 Gender – Female 17 (77.3%) 14 (63.6%) 19 (86.4%) Gender – Male 5 (22.7%) 8 (36.4%) 3 (13.6%) Outcomes and Cognitive-Affective Variables     NPRS-11 4.50 ± 1.85 4.23 ± 1.34 4.36 ± 1.22 Expectations (15-point item) 4.00 ± 1.51 3.50 ± 1.79 2.91 ± 2.60 PPT L3 (kg/cm 2 ) 2.13 ± 0.29 2.20 ± 0.36 2.11 ± 0.30 PPT L4 (kg/cm 2 ) 2.05 ± 0.26 2.20 ± 0.36 2.20 ± 0.22 PPT L5 (kg/cm 2 ) 2.02 ± 0.27 2.25 ± 0.32 2.02 ± 0.29 Abbreviations: kg/cm 2  = kilograms per square centimeter; L = lumbar vertebra; N  = number; NPRS-11 = Numeric Pain Rating Scale (0–10); PPT = Pressure Pain Threshold; SD = standard deviation. Symbols: % - percentage. Study flow diagram in accordance with the CONSORT guidelines [ 40 ]. Baseline characteristics of participants (mean ± SD or n (%)). Abbreviations: kg/cm 2  = kilograms per square centimeter; L = lumbar vertebra; N  = number; NPRS-11 = Numeric Pain Rating Scale (0–10); PPT = Pressure Pain Threshold; SD = standard deviation. Symbols: % - percentage. Baseline expectations showed sparse and inconsistent associations with outcomes. In the positive framing group, expectations were moderately associated with higher immediate lumbar PPT at L4 ( r  = 0.47; 95% CI, 0.06 to 0.74) and L5 ( r  = 0.43; 95% CI, 0.01 to 0.72). In the neutral and negative groups, corresponding associations were small in magnitude and imprecise. Across groups, baseline expectations showed no consistent association with NPRS-11 at either the immediately post-intervention (t1) or 24-hour (t2) time points ( Table 2 ). Full correlation matrices, including ancillary outcomes, are provided in Supplementary Tables S1-S3. Table 2. Correlations between baseline expectations and self-reported and mechanistic outcomes across groups. Outcome Positive group( r , p ) Negative group( r , p ) Neutral group( r , p ) NPRS-11 (t1) −0.02, 0.92 0.18, 0.42 −0.22, 0.32 NPRS-11 (t2) 0.29, 0.19 −0.08, 0.73 −0.34, 0.12 PPT L3 (t1) 0.35, 0.12 0.13, 0.55 0.06, 0.78 PPT L4 (t1) 0.47, 0.03 0.17, 0.45 0.05, 0.83 PPT L5 (t1) 0.43, 0.05 −0.12, 0.59 0.07, 0.75 PPT L3 (t2) 0.34, 0.12 0.22, 0.32 0.19, 0.40 PPT L4 (t2) −0.10, 0.66 −0.06, 0.79 0.06, 0.81 PPT L5 (t2) 0.29, 0.18 0.27, 0.22 0.21, 0.35 Abbreviations: NPRS-11 = Numerical Pain Rating Scale (0–10); PPT = pressure pain threshold; t1 = immediately post-intervention; t2 = 24 hours post-intervention. Nominal two-tailed p -values are reported for descriptive purposes; no multiplicity adjustment was applied. Correlations between baseline expectations and self-reported and mechanistic outcomes across groups. Abbreviations: NPRS-11 = Numerical Pain Rating Scale (0–10); PPT = pressure pain threshold; t1 = immediately post-intervention; t2 = 24 hours post-intervention. Nominal two-tailed p -values are reported for descriptive purposes; no multiplicity adjustment was applied. Baseline attitude toward manual therapy was moderately associated with lower NPRS-11 at 24 hours in the positive framing group ( r  = −0.45; 95% CI, −0.73 to −0.04). Associations with lumbar PPT were small and inconsistent across lumbar levels and time points. In the neutral and negative groups, correlations between attitude and both NPRS-11 and lumbar PPT outcomes were generally near zero ( Table 3 ). Table 3. Correlations between attitude toward manual therapy and self-reported, physiological, and mechanistic outcomes across groups. Outcome Positive group( r , p ) Negative group( r, p ) Neutral group( r , p ) NPRS-11 (t1) −0.06, 0.80 0.02, 0.95 0.01, 0.96 NPRS-11 (t2) −0.45, 0.03 −0.11, 0.64 0.32, 0.15 PPT L3 (t1) −0.09, 0.71 −0.05, 0.83 0.02, 0.95 PPT L4 (t1) −0.41, 0.06 −0.36, 0.10 0.02, 0.93 PPT L5 (t1) −0.41, 0.06 0.17, 0.46 0.05, 0.81 PPT L3 (t2) −0.38, 0.09 0.03, 0.89 0.01, 0.97 PPT L4 (t2) −0.40, 0.07 −0.06, 0.79 −0.00, 0.99 PPT L5 (t2) −0.25, 0.27 −0.07, 0.77 −0.06, 0.78 Abbreviations: NPRS-11 = Numerical Pain Rating Scale (0–10); PPT = pressure pain threshold; t1 = immediately post-intervention; t2 = 24 hours post-intervention. Nominal two-tailed p -values are reported for descriptive purposes; no multiplicity adjustment was applied. Correlations between attitude toward manual therapy and self-reported, physiological, and mechanistic outcomes across groups. Abbreviations: NPRS-11 = Numerical Pain Rating Scale (0–10); PPT = pressure pain threshold; t1 = immediately post-intervention; t2 = 24 hours post-intervention. Nominal two-tailed p -values are reported for descriptive purposes; no multiplicity adjustment was applied. Expectation fulfillment assessed at 24 hours showed no consistent associations with NPRS-11 or lumbar PPT across groups. Correlations with NPRS-11 (t1, t2) and lumbar PPT (t1, t2) were generally small and imprecise, with no stable pattern across lumbar levels ( Table 4 ). Because expectation fulfillment was assessed at t2, associations with immediately post-intervention outcomes (t1) should be interpreted as retrospective correlations rather than prospective prediction. Overall, these findings should be considered as exploratory and hypothesis-generating given multiplicity and sample size. Table 4. Correlations between expectation fulfillment (24 h) and outcomes across groups. Outcome Positive group( r , p ) Negative group( r , p ) Neutral group( r , p ) NPRS-11 (t1) −0.14, 0.55 0.23, 0.31 0.12, 0.60 NPRS-11 (t2) 0.08, 0.72 −0.08, 0.73 −0.13, 0.58 PPT L3 (t1) 0.31, 0.17 0.21, 0.35 0.15, 0.51 PPT L4 (t1) 0.38, 0.08 −0.03, 0.89 0.23, 0.31 PPT L5 (t1) 0.41, 0.06 −0.30, 0.17 0.20, 0.37 PPT L3 (t2) 0.35, 0.11 0.20, 0.37 0.24, 0.28 PPT L4 (t2) −0.06, 0.78 −0.07, 0.76 0.00, 0.99 PPT L5 (t2) 0.10, 0.65 0.18, 0.42 0.10, 0.65 Abbreviations: NPRS-11 = Numerical Pain Rating Scale (0–10); PPT = pressure pain threshold; t1 = immediately post-intervention; t2 = 24 hours post-intervention. Nominal two-tailed p -values are reported for descriptive purposes; no multiplicity adjustment was applied. Correlations between expectation fulfillment (24 h) and outcomes across groups. Abbreviations: NPRS-11 = Numerical Pain Rating Scale (0–10); PPT = pressure pain threshold; t1 = immediately post-intervention; t2 = 24 hours post-intervention. Nominal two-tailed p -values are reported for descriptive purposes; no multiplicity adjustment was applied. In the multiple linear regression model predicting NPRS-11 at 24 hours (t2), group allocation was the only variable retained beyond baseline pain. The final model explained 26% of the variance in 24-hour NPRS-11 (adjusted R 2  = 0.26; overall p  < 0.001), with baseline NPRS-11 (t0) retained a priori to account for initial symptom severity. After adjustment for baseline pain, the verbal framing group remained the only additional predictor retained by the stepwise procedure. Compared with the neutral group, participants in the negative framing group reported higher NPRS-11 at 24 hours (β = 0.35, p  < 0.01), whereas those in the positive framing group reported lower NPRS-11 (β = −0.26, p   0.10) and were excluded during stepwise selection. Thus, within the constraints of this dataset, the verbal framing group was the only independent predictor of 24-hour NPRS-11 after spinal manipulation.

In this secondary analysis, baseline expectations, attitude toward manual therapy, and expectation fulfillment showed weak and inconsistent associations with NPRS-11 and lumbar PPT. When examining predictors of NPRS-11 at 24 hours, expectations and attitude did not add explanatory value beyond group allocation. Overall, within the limits of this exploratory dataset, short-term responses were better explained by the experimentally manipulated communicative context than by the measured cognitive-affective variables.

This secondary analysis examined whether baseline expectations, attitude toward manual therapy, and 24-hour expectation fulfillment were associated with short-term NPRS-11 and lumbar PPT responses, and whether expectations and attitude added predictive value for 24-hour NPRS-11 beyond verbal framing group allocation. The intervention effects of verbal framing on NPRS-11 and lumbar PPT were reported in the parent trial [ 19 ] and are not repeated here. Two main findings emerged. First, the cognitive-affective variables investigated showed limited and inconsistent associations with the outcomes. Within the positive verbal framing group only, higher baseline expectations were modestly associated with higher immediate lumbar PPT at L4 and L5, and a more favorable baseline attitude toward manual therapy was associated with lower NPRS-11 at 24 hours. In contrast, expectation fulfillment showed no robust relationships with NPRS-11 or lumbar PPT in any group. These sparse and group-specific associations should be interpreted cautiously, given the exploratory nature of the secondary analysis and the absence of adjustment for multiple comparisons. Overall, this pattern suggests that short-term responses in this dataset were more closely aligned with the experimentally manipulated verbal context than with preexisting beliefs or perceived expectation fulfillment [ 3 ]. This interpretation should be considered in light of the measures used and the characteristics of the sample studied. Several factors may have contributed to the attenuation and inconsistency of associations involving the cognitive-affective variables. These include the use of brief single-item ratings, which may have limited reliability and sensitivity [ 41 ], as well as range restriction (e.g., generally favorable expectations and attitudes), relatively mild baseline NPRS-11, and the short follow-up interval, all of which may have reduced between-participant variability and obscured potential dose – response patterns [ 42 ]. In addition, expectation fulfillment was assessed at 24 hours and may partly reflect post hoc appraisal influenced by the experienced response, rather than a stable predictor of subsequent pain. This timing limits causal interpretation, particularly for associations with immediately post-intervention outcomes. Second, regression analysis supported this interpretation. After accounting for baseline NPRS-11, verbal framing group was the only independent predictor of NPRS-11 at 24 hours. Baseline expectations and baseline attitude toward manual therapy did not add explanatory value once group allocation was included, indicating that the experimentally induced communicative context exerted a stronger influence on short-term pain outcomes than individual cognitive-affective characteristics in this sample [ 15 ]. These findings can be interpreted within therapist – patient relationship models, in which verbal framing operates as a social and interpersonal signal shaping appraisal, trust, and perceived credibility within the therapeutic encounter [ 43 ]. In physical therapy, stronger therapeutic alliance has been associated with better outcomes in chronic musculoskeletal pain, with quantitative evidence in NSCLBP also linking trust and alliance to improvements in pain and function [ 44 , 45 ]. Qualitative studies further indicate that patients’ expectations extend beyond technical competence to include collaborative and relational components, with alignment of goals and expectations central to perceived benefit [ 46 ]. Consistent with this literature, contemporary manual therapy research agendas identify patient – provider interaction variables, including therapeutic alliance, as modifiable contextual factors warranting prospective investigation alongside clinical outcomes [ 47 ]. Taken together, these frameworks support an interpretation that prioritizes relational and contextual mechanisms, while acknowledging that the present dataset did not directly measure alliance, trust, or specific communication behaviors. From the perspective of this secondary analysis, the findings suggest that short-term pain responses following spinal manipulation were more closely related to the experimentally manipulated verbal context than to patients’ baseline expectations, attitudes toward manual therapy, or perceived expectation fulfillment. Although individual cognitive-affective variables showed only limited and inconsistent associations with outcomes, verbal framing emerged as the dominant factor associated with 24-hour NPRS-11. Clinically, this pattern highlights that how treatment is verbally contextualized immediately before delivery may have a greater short-term impact than patients’ preexisting beliefs or attitudes as measured in this study [ 48 ]. Rather than viewing expectations and attitudes as stable patient traits that independently drive outcomes, these findings suggest that clinician communication may shape the immediate interpretative context in which treatment effects are experienced [ 48 ]. In the short term, supportive or negative verbal cues may therefore influence symptom appraisal more strongly than individual differences in expectations or attitudes. This observation aligns with broader biopsychosocial perspectives that conceptualize communication as an active component of care rather than a neutral background factor. Importantly, the present results do not indicate that expectations or attitudes are irrelevant in clinical practice. Rather, their independent contribution to short-term pain outcomes appears limited when considered alongside verbal framing in this experimental paradigm. Within this framework, careful and ethically appropriate communication may represent a modifiable aspect of practice that influences short-term responses to spinal manipulation, whereas inadvertently negative or uncertain framing may attenuate perceived benefit, even when the physical intervention is standardized [ 49 ]. Interpretation of these findings should consider several limitations. Participants were young adults with mild baseline symptoms, which may limit generalizability to clinical populations with more persistent or severe pain. Follow-up was limited to 24 hours, preventing conclusions about the longer-term effects of verbal framing. Although a broad range of mechanistic measures was collected in the parent trial [ 19 ], only NPRS-11 and lumbar PPT were treated as primary outcomes in this secondary analysis. All other associations were exploratory and should be interpreted accordingly. This study represents a post hoc, hypothesis-generating secondary analysis that was not prespecified for all of the current hypotheses. In addition, the use of stepwise variable selection in the regression model may have increased the risk of overfitting and unstable predictor estimates, particularly given the modest sample size. The predictive findings from aim 2 should therefore be considered exploratory and require replication in independent datasets. Measurement precision for expectations, attitude toward manual therapy, and expectation fulfillment was limited by the use of brief single-item scales. Therapeutic alliance, trust, and specific clinician communication behaviors were not directly assessed, and a dedicated secondary-analysis protocol or statistical analysis plan was not preregistered, reinforcing the exploratory nature of the findings. Given accumulating evidence linking relational and contextual variables to outcomes in chronic musculoskeletal pain and NSCLBP [ 6 ], future trials should incorporate validated measures of alliance, trust, and communication processes to better situate expectancy-related effects within a broader biopsychosocial framework [ 8 , 9 , 44 , 45 ]. Despite these limitations, the present findings indicate that verbal framing represents a potentially relevant contextual factor influencing immediate and short-term responses to spinal manipulation. Future studies should include longer follow-up periods, more diverse clinical samples, and prespecified mechanistic hypotheses to clarify how contextual factors interact with individual characteristics and whether observed effects accumulate or persist across repeated treatment sessions.