Response to Letter to the Editor for “Nonoperative Management of the Adult Spinal Deformity Patient: A Biopsychosocial Intervention versus Care-as-usual - A Comparative Treatment-Based Cohort Study with One-Year Follow-Up”

We thank the authors for their interest in our study and for their thoughtful methodological comments. We appreciate the opportunity to clarify several aspects of the study design and interpretation. We fully agree that rigorous discussion of methodology is essential, particularly in an area where scientific evidence regarding nonoperative treatment of adult spinal deformity (ASD) remains limited.

The concern regarding incomplete participant flow in Figure 1 appears to arise from a misinterpretation of the figure. The comment states that 2308 patients were excluded because scoliosis type was unspecified. However, Figure 1 reports that 2368 patients were excluded at this step. The difference is explained by the figure resolution, where the number “6” may have been interpreted as a “0”. When the correct number is used, all patients are accounted for and no unexplained attrition is present. We appreciate the opportunity to clarify this issue and acknowledge that a higher-resolution figure may have prevented this misunderstanding.

The authors further raise concerns regarding the construction of the cohorts and the possibility of confounding arising from baseline differences. We agree that the observational design and use of historical controls do not permit causal inference, which was not the intention of the study. The primary aim was exploratory: to evaluate whether ASD patients participating in a biopsychosocial intervention demonstrated different treatment outcomes over one year compared to ASD patients receiving usual nonoperative care, thereby contributing to the ongoing international discussion regarding nonoperative ASD management. While the limitations associated with this design were discussed in the manuscript, we acknowledge that the exploratory nature of the study and the resulting restrictions on causal interpretation could have been stated more prominently. We appreciate the opportunity to clarify this important aspect of the study’s objectives and interpretation.

The biopsychosocial cohort and care-as-usual cohort were selected from existing clinical populations because the current evidence base does not yet provide a well-defined, evidence-based nonoperative treatment pathway for ASD that could serve as the basis for a randomized comparison. Consequently, observational comparisons using well described clinical cohorts remain an important step in the development of evidence in this field. We chose to study patients as they present in routine practice, acknowledging the substantial heterogeneity that characterizes the ASD population. ¹ While this approach inevitably results in baseline differences between cohorts, it also enhances the clinical relevance and external validity of the findings. More restrictive matching procedures could have produced more homogeneous samples, but at the cost of substantially smaller cohorts and findings that may have been less representative of the patients encountered in daily clinical care.

We agree with the authors that the calculation of change scores does not eliminate confounding and cannot by itself establish a treatment effect. We therefore wish to emphasize that change scores were not intended as a method for causal adjustment. Rather, they were used to describe differences in outcome trajectories between cohorts that had similar diagnostic eligibility criteria and comparable baseline functioning as measured by the Oswestry Disability Index (ODI). The study should therefore be interpreted as providing evidence of an association between participation in the biopsychosocial program and greater improvement over one year, rather than proof of a causal treatment effect.

At the same time, we believe it would be an oversimplification to attribute the observed differences solely to baseline imbalances. Several potentially influential baseline characteristics differed between cohorts. For this reason, we explored the relationship between these characteristics and change in functioning. Although such analyses cannot exclude residual confounding, they demonstrated only limited associations with the primary outcome and therefore provided no indication that any single measured characteristic was a dominant explanation for the observed treatment differences. We agree, however, that absence of strong correlations should not be interpreted as proof of exchangeability and that unmeasured confounding cannot be excluded.

Importantly, we believe the value of observational research lies not only in establishing causality but also in identifying clinically relevant signals that warrant further investigation. The size and consistency of improvement observed in the biopsychosocial cohort, in both primary and all secondary outcome measures, suggest that this treatment approach deserves continued scientific attention. This conclusion is consistent with recent international efforts to further define and develop nonoperative management strategies for adults with spinal deformity. ²

Finally, we appreciate the authors’ comments regarding statistical methodology. We agree that future studies should employ more advanced methods for baseline adjustment, including ANCOVA and other approaches designed to account for baseline differences. More importantly, prospective comparative studies will be required to strengthen causal inference and further clarify the effectiveness of biopsychosocial interventions in ASD.

In conclusion, we agree that our findings should not be interpreted as establishing causality. However, we maintain that the study provides meaningful exploratory evidence that participation in a biopsychosocial intervention is associated with greater improvements in functioning than those observed in a care-as-usual cohort. Together with the growing recognition of the importance of nonoperative care in ASD, these findings support continued investigation of biopsychosocial treatment approaches through more rigorous prospective research.

ORCID iDs

Evelien H. W. Hoevenaars https://orcid.org/0000-0001-7858-8834

Ferran Pellisé https://orcid.org/0000-0002-0644-7757

Javier Pizones https://orcid.org/0000-0001-7394-9637