Predictive model is accurate, may improve lumbar epidural steroid injection efficacy

In the current climate of bundled payments and value-based health care reform, investigators concluded from a study they conducted that creating an accurate, predictive model to improve efficacy of lumbar epidural steroid injections for lumbar degenerative disease was important.

Medicine is a science of uncertainty and an art of probability, no more so than in the field of lumbar degenerative disease, Ahilan Sivaganesan, MD, said at American Association of Neurological Surgeons Annual Meeting, where he discussed a model he and colleagues designed to predict the clinical effectiveness of lumbar epidural steroid injections (ESI).

“This is a first attempt at a decision support tool for lumbar ESI. It is centered around patient-reported outcomes as opposed to radiographic or other intermediate variables. It is statistically valid and accurate. I think, over time, models, such as this, can help in patient selection; also in having real, intelligent conversations with a patient to give them personalized expectations,” Sivaganesan said.

Ahilan Sivaganesan

The investigators studied 225 patients who underwent lumbar ESI for degenerative pathology during a period of 2 years. They used baseline data and patient reported outcomes at the 3-month follow-up.

Using a previously reported minimum clinically important difference (MCID) for Oswestry Disability Index (ODI) of 7.1%, patients were divided into those who reached the MCID after the injection and those who did not reach that MCID.

The investigators found variables in the patients, based on previous clinical intuition and research, which directly correlated to the subjects either reaching or not reaching the MCID for ODI after undergoing the injection.

“A severe central stenosis, a transforaminal or interlaminar route, a higher baseline for ODI and a diagnosis of disc herniation increases the odds of a patient achieving the MCID for ODI. A symptomatology greater than 12 months, prior surgery and higher baseline anxiety scores all contributed to decreased odds of achieving that MCID for ODI,” Sivaganesan said.

Using the variables and 80% of the patients from the study, the investigators developed a predictive model to determine the effectiveness of lumbar ESI. To test the model in the “real world,” the investigators used the 20% of patients not included in the development of the model to predict if they would reach the MCID for ODI 3 months after treatment.

The area under the receiver-operating characteristic curve for the model, based on the 20% of patients not used to build the model itself, was 72%.

“[This] does not blow things out of the water, but is quite robust. With further improvements, we can increase the performance of the model even more,” Sivaganesan said.

According to the investigators, a patient with higher baseline ODI, who used narcotics for a short period of time, has no prior surgeries and little anxiety and received a transforaminal or interlaminar injection has a 96% chance of reaching the MCID for ODI at 3 months follow-up. Conversely, a patient with spondylolisthesis who has a longer symptom duration, a prior surgery and higher anxiety at baseline has a 17% chance of reaching the MCID for ODI at the same follow-up. – by Robert Linnehan

Reference:
Chotai S, et al. Paper #623. Presented at: American Association of Neurological Surgeons Annual Meeting; May 2-6, 2015; Washington, D.C.

For more information:
Ahilan Sivaganesan, MD, can be reached at Vanderbilt University Medical Center, Department of Neurological Surgery, 1211 Medical Center Dr., Nashville, TN 37232; email: ahilan.sivaganesan@vanderbilt.edu.

Disclosure: Sivaganesan reports no relevant financial disclosures.