Psychological Tests Before Pain-Relief Implants Fail to Predict Two-Year Success

For millions of people living with chronic pain that resists medication and surgery, spinal cord stimulation offers a tantalizing promise: a small implanted device that delivers gentle electrical pulses to the spinal cord, scrambling pain signals before they reach the brain. Yet the technology does not work equally well for everyone, and for decades clinicians have debated whether a patient’s psychological profile before implantation can forecast who will benefit. A new retrospective study from a Spanish hospital team now adds an unusually rigorous data point to that debate, tracking forty consecutive patients for two full years after their devices were switched on.

The research, led by Esther Gómez Botella of the Hospital General Universitario Dr. Balmis in Alicante and collaborators at the Alicante Institute for Health and Biomedical Research and Miguel Hernández University, was published in BMC Psychology. What makes the study distinctive is its selection policy: unlike many programs that exclude patients deemed psychologically unsuitable, the Alicante team excluded no one on psychological grounds. Every patient who proceeded to implantation between 2012 and 2022 had completed a standardized psychological assessment beforehand, allowing the researchers to ask, without the distortion of selection bias, whether baseline personality and symptom profiles actually shaped outcomes.

The tool they used was the Millon Clinical Multiaxial Inventory, third edition, a widely administered questionnaire that scores patients on twenty-four clinical scales spanning anxiety, depression, somatoform tendencies, personality patterns, and severe psychopathology. Each patient’s profile was recorded before surgery. The team then measured outcomes at one month, six months, one year, and two years after implantation, using instruments that are standard in modern pain research: a visual analogue scale for pain both at rest and during movement, the EuroQol five-dimension questionnaire for quality of life, a visual analogue scale for perceived overall health, and the Pittsburgh Sleep Quality Index for sleep.

The headline finding is emphatically positive. Before implantation, patients rated their resting pain at an average of 8.56 on a ten-point scale, and their movement-related pain at 9.39, scores that describe near-maximal, disabling discomfort. Two years later, those averages had fallen to 4.22 and 5.24 respectively, reductions of roughly 51 and 44 percent. More than half of the patients maintained at least a fifty percent reduction in pain at the two-year mark, a threshold clinicians regard as a meaningful clinical response rather than mere statistical noise.

The benefits extended well beyond raw pain scores. The average quality-of-life index on the EuroQol instrument climbed from a dismal 0.18 to 0.60, a shift that in health-economic terms corresponds to a dramatic recovery in everyday functioning. Patients’ self-rated overall health rose from 31.4 to 54.9 on a hundred-point scale, and sleep quality improved substantially, with Pittsburgh index scores falling from 15.5 to 10.1 as insomnia and fragmented rest eased. All of these changes were statistically robust, with p-values below 0.001, and, critically, they persisted across every follow-up visit over the entire two-year window, arguing against the gradual fading of benefit that has plagued some neurostimulation cohorts.

Against that encouraging backdrop, the psychological question yielded a deliberately cautious answer. The researchers used linear mixed-effects models, a statistical framework that can track repeated measurements within each individual while handling missing follow-up data gracefully, to test each of the twenty-four MCMI-III scales as a potential predictor of each of the five outcomes. Because twenty-four scales times five outcomes produces a hundred and twenty statistical comparisons, the team applied false-discovery-rate correction, a modern multiple-testing safeguard that filters out the spurious associations that inevitably emerge when so many hypotheses are examined simultaneously.

Several scales did show nominal, or uncorrected, associations with outcomes. The most consistent patterns pointed to schizoid personality traits predicting poorer quality of life and dysthymia, a chronic form of low-grade depression, predicting poorer perceived health. But none of these associations survived false-discovery-rate correction, and most became weaker still when the researchers adjusted for the Desirability validity scale, an MCMI-III measure that flags patients who answer in a socially flattering rather than candid manner. Notably, the cohort showed no clinically meaningful elevations on the severe personality and psychotic-spectrum scales, meaning the analysis could not test whether such severe profiles might matter.

The authors are careful about what their data do and do not demonstrate. Their findings, they conclude, neither support nor exclude a prognostic role for pre-implantation psychological screening; the results are best treated as hypothesis-generating material for prospective research rather than as a basis for deciding who deserves an implant. That distinction matters, because in practice psychological screening has sometimes functioned as a gatekeeping instrument, with patients denied neuromodulation on the basis of unvalidated cut-offs. This study’s no-exclusion design offers a rare window into what happens when that gate is left open: outcomes were strong across the board.

The study’s technical strengths include its adherence to the STROBE reporting guidelines for observational research, its use of IMMPACT-aligned outcome measures recommended by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials, and sensitivity analyses using bootstrap confidence intervals to check that the results did not hinge on assumptions about statistical distributions. Its funding came from the Instituto de Investigación Sanitaria y Biomédica de Alicante, which played no role in the design, analysis, or decision to publish, and the authors declare no competing interests, an important transparency note in a field where device manufacturers are heavily involved.

The limitations are equally clear-eyed. Forty patients from a single center is a modest sample, and retrospective designs cannot control every confounding variable the way a randomized trial can. The two-year follow-up, though longer than most, still falls short of the decade-long horizon on which implanted devices operate. Nonetheless, the practical message for patients and clinicians is striking: spinal cord stimulation delivered substantial, durable relief across pain, mood-linked health perception, quality of life, and sleep, and the psychological questionnaires collected before surgery proved far better at describing patients than at predicting their futures. For now, the authors suggest, psychological assessment belongs alongside implantation as a support tool, not in front of it as a filter, and the definitive test of psychological prognostication awaits larger, prospective, deliberately designed studies.

Subject of Research: Psychological predictors of two-year outcomes after spinal cord stimulation for chronic pain

Article Title: Pre-implantation psychological assessment and two-year outcomes of spinal cord stimulation for chronic pain: a retrospective observational study

Article References: Gómez Botella, E., Gómez Salinas, L., Sastre Peris, Y., Bernabéu Juan, P., Pérez-Jover, V., & van-der Hofstadt Román, C. J. (2026). Pre-implantation psychological assessment and two-year outcomes of spinal cord stimulation for chronic pain: a retrospective observational study. BMC Psychology. https://doi.org/10.1186/s40359-026-05596-x

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05596-x

Keywords: chronic pain, spinal cord stimulation, MCMI-III, psychological assessment, quality of life, sleep quality, neuromodulation, EQ-5D-5L, prognosis, pain management, retrospective study, Pre-implantation