The short- and long-term effect of osteopathic manipulative treatment on pain, and psychosocial factors in adults with chronic low back pain

Data source

A total of 101 patients were consented over 6 months (September 2021 to March 2022) and followed for 1 year from date of consent (September 2022 to March 2023).

Patients were identified from the AthenaHealthNet Electronic Health Records (EHR) of new patients seeking CLBP treatment and recruited by convenience at their first visit to the MSUCOM OMM clinic. Once the provider qualified the patient, they were consented by the Research Coordinator (i.e., not the practitioner) to avoid bias. The paper consent form was reviewed with and provided to the patient, and any questions regarding the consent and research were answered. Patients were not compensated for their participation.

The inclusion criteria consisted of ambulatory patients 18–85 years old with CLBP lasting longer than 3 months. The exclusion criteria included patients with spinal or neurological disorders, spinal fractures within the past year, severe osteoporosis, scoliosis, or osteoarthritis in the spine or hip joints. The patients must not have a history or diagnosis of retroperitoneal tumors or lymphadenopathy, or a diagnosis of lumbar spine radiculopathy verified by imaging, negative electromyography/nerve conduction velocity (EMG/NCV) testing as well as matching symptoms. A patient was also rejected if he or she had a spinal surgery completed within the past year. Rhizotomies, epidural injections, and spinal cord stimulator implants were acceptable if completed at least 3 months prior to the beginning of the study and if they were not the apparent source of increased pain following the procedure as identified by the patient. Scoliosis was accepted if it was considered mild, defined as cases with a Cobb angle less than 20°. Generally speaking, patients meeting these criteria are termed to have “nonspecific” CLBP.

Outcome measures

The instruments to measure pain, function, and biopsychosocial factors were the National Institutes of Health (NIH) Minimum Dataset for Chronic Low Back Pain (NMD) [10]. The dataset evolved from the NIH Task Force on CLBP and is a 40-item checkbox questionnaire developed to increase the use of standardized measurements and a definition for CLBP (Supplementary Material). The research task force (RTF) impact score was calculated from the NMD. The RTF score was tabulated from adding up the numbers assigned to answers from selected questions, including the answers from a visual analog scale, pain interference on day-to-day activities, and responses on the ability to do day-to-day activities. The score determines the level of impact that CLBP has on daily function, with a range between 8 (lowest possible pain score) and 50 (highest possible pain score). The biopsychosocial score (pain sensitivity questionnaire [PSQ]) was calculated from the NMD set including scores from psychological questions and sleep disturbance. The score determines the level of impact on psychological factors like depression, feeling of worthlessness, anxiety, and sleep disturbance, with a range from 10 (minimal psychological affects) to 44 (most psychological effects). The survey also has patients record whether they have ever utilized any other treatments such as exercise therapy, injections, opioids, or psychological counseling.

The pain instrument was given in person by the research coordinator at the first patient visit after initial treatment (baseline) and sent electronically via research electronic data capture (REDcap) at 2, 4, and 8 weeks and at 6 and 12 months after initial treatment. This sampling frequency was selected to ensure that changes in both the short term and long term are documented. Study data were collected and managed utilizing REDCap electronic data capture tools hosted at Michigan State University [11, 12]. REDCap is a secure, web-based software platform designed to support data capture for research studies, providing: (1) an intuitive interface for validated data capture; (2) audit trails for tracking data manipulation and export procedures; (3) automated export procedures for seamless data downloads to common statistical packages; and (4) procedures for data integration and interoperability with external sources.

Description of treatment

Treatment protocols for this study were not standardized, and each patient was treated as the practitioner saw fit based on the principles found in the Foundations of Osteopathic Medicine, 4th Edition [13]. All providers participating in this study were AOA Board (ABONMM)-certified practitioners in Neuromusculoskeletal Medicine/OMM. Their practice experience (since completion of residency) ranges from 4 to 28 years, with an average of 12 years. Although each practitioner may have approached the patient in a slightly different manner from the other practitioners, they always followed the established osteopathic practices and principles of the profession [13]. There are also two residents who participated in this study; as they treated each patient, the residents were supervised by one or more of the providers.

Power analysis

A power analysis was conducted in G*power version 3.1 software [14] to observe a drop in pain utilizing a paired t test for a Cohen’s d z-score effect size of 0.322 at 80 % power and 5 % level of significance, and it required a sample size of 61 total patients.

Statistical analysis

To determine whether patients receiving OMT experienced improvements in pain symptoms and biopsychosocial factors over time, a repeated-measures multivariate longitudinal regression model [15] was utilized to model the association between pain symptoms and biopsychosocial factors over time.

Time trends of pain symptoms and biopsychosocial factors were analyzed as overall mean and by decile of the level at baseline. Pairwise differences or measurements were compared between time points and baseline to identify the time when there was a significant improvement. Mixed-effects models were utilized to test the association of time (from baseline to 12 months) with pain scores and biopsychosocial factors. Comparisons were estimated for comparing 4 weeks and 6 months to baseline and 6–12 months to baseline.

RTF pain impact scores

As displayed in Table 1A, the baseline RTF impact score assessment had an estimated marginal mean impact score of 24.7. Within the initial 2 weeks, this score decreased to 22.8. At the 4-week mark, the RTF score was 21.5. At 8 weeks, the score was 21.3. At 6 months, the RTF impact score was 18.9, and at the 12-month mark, it was 18.7 (Figure 2).

Figure 2:

Time trends of mean RTF impact pain scores (orange) and mean biopsychosocial scores (PSQ, blue). A significant decrease occurs at 6 months vs. baseline in both the pain score and biopsychosocial score.

Comparisons were made between time points (Table 1B). The difference between 4 weeks and baseline was statistically significant (3.23, p=0.001), while there was no additional change between 6 months and 12 months (0.21, p=0.850).

Biopsychological scores (PSQ)

As presented in Table 2, the baseline assessment of the PSQ scores had an estimated marginal mean of 20.3. Within the initial 2 weeks, the score was 19.7. The 4-week assessment was 19.3. At 8 weeks, the score was 19.6. At the 6-month mark, the PSQ score was 16.9, and at 12 months, it was 15.9 (Figure 2).

Comparisons were made between time points (Table 2B). The difference between 4 weeks and baseline was not statistically significant (1.03, p=0.141), while there was a statistically significant reduction (3.44, p<0.001) between 6 months and baseline. The additional change between 12 months and 6 months was not statistically significant (0.93, p=0.281).

Limitations

The dropout rate over the year of the trial could have resulted in possibly the loss of higher-quality participants, incomplete results, and an impaired interpretation of the data resulting in data loss and poor data quality. When asked why the patient did not complete a timeline survey, many patients stated that they felt better or were no longer attending the OMM clinic, so they did not know they needed to continue with the surveys. Other reasons were due to psychosocial issues such as depression or anxiety. A better explanation at the onset of the trial, along with continued communication about response importance, may have allowed for better retention. A financial incentive may have also helped them to realize how their participation benefits them. The survey allowed for self-reporting of other treatments but only if the participant has ever had a different treatment modality that was not specific to the year of OMT. A future study would need to include other treatment modalities such as exercise therapy, analgesic usage, injections, or psychological counseling sought during the year of the study, in order to better understand whether the OMT was indeed effective. The survey data was self-reported, with no system to validate self-reported outcomes. Participants could exaggerate or underreport their pain levels, or not recall the levels as they were in the previous time points, and thus report them in a different manner.