When Should You Refer for an Epidural Steroid Injection? A Colleague's Perspective

One of the most common referral questions I get from primary care and physical therapy colleagues is some version of: “Is this patient a candidate for an epidural steroid injection, or should we keep trying conservative care first?” I want to lay out how I approach that question, so referrals arrive at the right time — not too early, not too late.

Who tends to benefit

The patients I see the most benefit in are those with radicular pain — pain, numbness, or tingling radiating down an arm or leg, typically from a compressed or irritated nerve root in the cervical or lumbar spine. Epidural steroid injections are used less often, and are generally less effective, for patients whose pain stays localized to the neck or back without a radicular component. If your patient’s presentation is primarily axial pain, that’s useful information for me before I see them — it changes the conversation we’ll have.

A 2020 comprehensive systematic review of lumbar transforaminal steroid injections for radicular pain found that 74% of patients with disc herniations achieved 50% or greater pain-reduction at three months. View the study.

What the procedure involves

Different physicians have different approaches to how they perform spinal injections. I want my patients to be as comfortable as possible while also making sure to accurately target the impacted nerve. I use fluoroscopic guidance to place a small needle adjacent to the affected nerve root, confirm needle position and flow with contrast, and then inject a corticosteroid along with a small volume of local anesthetic (typically lidocaine).

One clinical detail worth flagging for co-management: because the anesthetic component takes effect immediately, a patient’s short-term response in the recovery area can be a useful data point — if their radicular pain resolves quickly, even briefly, that’s consistent with the injected level being a relevant pain generator. I mention this not as a formal diagnostic protocol, but as something referring physicians and PTs sometimes find helpful context when patients call afterward describing a “it worked, then came back” pattern in the first few hours.

Setting expectations before the referral

It helps if patients arrive already understanding that:

• The anesthetic effect is fast but short-lived (a couple of hours); it is not the treatment effect.
• The corticosteroid itself takes days up to about a week to reach full effect.
• Duration of relief is genuinely variable and depends heavily on the underlying pathology: a large disc herniation that resolves on its own can yield durable relief, while foraminal or central stenosis tends to be less predictable and sometimes responds poorly.

Framing it this way up front reduces the “it didn’t work” calls that come in around hour three, when it’s simply the anesthetic wearing off before the steroid has taken effect.

Screening before you refer

A few factors are worth checking before sending a referral, since they affect candidacy or timing:

• Uncontrolled diabetes — glycemic control matters, since corticosteroids can raise blood glucose temporarily.
• Active infection, local or systemic.
• Anticoagulation status and any other bleeding risk relevant to a spinal procedure.

None of these are automatic disqualifiers, but they are worth flagging in the referral to help me plan the visit and, when needed, coordinate with you on medical management and correct timing.

Frequency and the co-management piece

I typically limit injections to a maximum of four per year per patient, even when they’re providing relief, given the cumulative effects repeated corticosteroid dosing can have on bone density and on blood glucose and blood pressure. For patients you’re co-managing — particularly those with diabetes or osteoporosis risk — that’s useful context to have on your side of the chart as well.

I also see this treatment as a bridge, not a replacement for physical therapy. The relief window it creates is often what makes it possible for a patient to actually participate in PT and make progress, which is why I’d rather coordinate timing with your team than treat the injection as a stand-alone intervention.

My approach to candidacy

I don’t recommend an epidural steroid injection unless I genuinely expect it to give a patient meaningful relief. If a patient arrives having been told broadly that “these just mask pain,” I’ll walk them through an individualized consult of their imaging and symptoms rather than applying a blanket rule in either direction. My goal is that every referral you send gets an honest answer — including when the honest answer is that this isn’t the right tool for that patient.

If you have a patient you’re considering referring, or want to discuss a case directly, I’m glad to connect — details below.