One-sided, just below the belt line, and you can cover it with a fingertip. Nothing is out of place — but the joint has stopped tolerating being loaded on one leg, and that is a trainable problem.
Medically reviewed by Dr. Carson Bailey, PT, DPT · September 2026
The sacroiliac joint is the weight-bearing junction between the base of the spine and the pelvis, and it is thought to be the source of roughly 15% to 30% of mechanical low back pain.[1] It is also the joint most often misdescribed: it is the largest joint in the body by surface area and moves only about 2 to 4 mm in any direction, so it does not slip out of place and cannot be put back in.[1] The tests clinicians use to identify it are better at ruling it out than ruling it in — a positive cluster leaves about 35% certainty, a negative one about 92%.[2] RISE Physical Therapy treats SI joint pain one-on-one in 45-minute sessions in Boise and Nampa, Idaho, and builds the plan around single-leg tolerance rather than realignment.
SI joint pain is pain coming from the sacroiliac joint — the pair of joints where the sacrum, the triangular bone at the base of the spine, meets the ilium on each side of the pelvis. Their job is structural rather than mobile: everything your upper body weighs has to cross them to reach your legs. The joint surface covers around 17.5 cm², making it the largest joint in the body, and the movement available to it is roughly 2 to 4 mm in any direction.[1]
The symptom pattern is unusually specific for a back complaint. People with SI joint pain typically point to one spot with one finger — the dimple at the back of the pelvis, an inch or two below the belt line, on one side. From there it spreads downward into the buttock and the back of the thigh. Of the documented referral patterns, the posterior or lateral thigh is involved in about 50% of cases, pain travels below the knee in roughly 28%, and reaches the foot in around 14%.[1] That last group is the reason SI joint pain is so often mistaken for a disc problem.
What provokes it is asymmetry — loading one leg at a time. The complaints we hear most are stairs, stepping out of a car, rolling over in bed, standing on one leg to put trousers on, and sitting on a hard surface or with a wallet under one buttock. Unlike most spinal pain, it usually has a start date: pain from this joint usually follows an inciting event rather than creeping in.[1] If you can name the fall, the pregnancy or the collision, that history is doing real diagnostic work.
SI joint pain has a short and fairly consistent list of causes: trauma, pregnancy, repetitive one-sided loading, and previous lumbar fusion surgery.[1] It tends to arrive in two age groups for different reasons — younger adults after an injury or a pregnancy, and older adults as the joint degenerates.[1]
Trauma is the cleanest version: a fall directly onto one buttock, a missed step off a kerb, or a motor vehicle collision where one leg was braced against the floor. Repetitive loading is the version athletes and tradespeople get, where months of running on a cambered road, kicking off the same leg, or carrying a load on one hip gradually outstrip what the joint tolerates. Lumbar fusion matters because it removes motion from the segments above and the pelvis absorbs the difference — which is why new SI pain after spine surgery is a recognised pattern rather than a coincidence.
And a category that is not mechanical at all: inflammatory sacroiliitis. In axial spondyloarthritis the SI joints are inflamed by a systemic disease process, and the giveaway is that the symptoms behave backwards — worse with rest, better with movement, stiff for more than half an hour in the morning, often waking the person in the second half of the night, and usually starting before age 45. That pattern is a rheumatology referral, not a course of exercise, and separating it from mechanical SI pain is part of what a first evaluation is for.
SI joint pain is diagnosed with a cluster of pain provocation tests, and the honest answer is that those tests are much better at ruling the joint out than ruling it in. The standard approach is to stress the joint in five or six different directions and treat three or more positive tests as meaningful. A 2021 systematic review with meta-analysis pooled the studies and put numbers on what that buys you: a positive likelihood ratio of 2.13 and a negative likelihood ratio of 0.33, which at a 20% background prevalence translates to about 35% certainty after a positive cluster and about 92% certainty after a negative one — with the certainty of the underlying evidence graded very low.[2]
Read that carefully, because it changes what an honest clinician can promise you. If the tests are negative, we can be reasonably confident your problem is somewhere else and go looking. If they are positive, the SI joint has become the leading suspect and nothing more. Image-guided injection remains the reference standard for actually confirming the joint as the pain source,[1] and no scan can make the diagnosis on its own — plenty of degenerate-looking SI joints belong to people with no pain at all. So the first block of treatment is best understood as a test of the hypothesis, which is why it should come with an agreed date for deciding whether it worked.
Which brings us to the claim almost everyone with this diagnosis has heard: that their SI joint is "out" and needs putting back. The anatomy does not allow it. A joint with 17.5 cm² of interlocking surface, bound by some of the strongest ligaments in the body, with 2 to 4 mm of total available motion, is not subluxing and relocating between appointments.[1] Manipulation and mobilisation of the region genuinely do reduce pain, sometimes within a single session — that effect is real and it is worth using. What it is not is evidence that something was displaced. The distinction matters because the realignment story quietly teaches people that their pelvis is fragile and needs regular correcting, and that belief costs more over a decade than the technique ever gains.
Yes — pregnancy is one of the most common settings for SI joint pain, and the numbers are larger than most people expect. Around 45% of pregnant women report pelvic girdle pain, low back pain or both; roughly 25% describe the pain as serious, and about 8% report severe disability from it.[3] After delivery, approximately 25% still have complaints and around 7% have problems serious enough to need intervention.[3]
The reassuring part of that data is the trajectory: about 93% of postpartum pelvic girdle pain resolves on its own within three months.[3] The part worth acting on is the remainder — the 7% whose pain does not settle in that window are, in the authors' own framing, at considerable risk of prolonged serious pain.[3] That is the single most useful thing to know if you are three months postpartum and still hurting on stairs: waiting it out has already been tried, and the group you are now in is the one where early treatment has the most to offer. If pelvic floor symptoms sit alongside the joint pain, our pelvic floor therapy and postpartum recovery pages cover that side of it.
The research on physical therapy for SI joint pain is consistently positive and consistently small. Active treatment beats doing nothing in every trial that has tested it; almost every one of those trials is limited enough that the honest conclusion is directional rather than precise.
Nine studies of manipulation, exercise and kinesio taping for SI joint dysfunction. All three reduced pain and disability, manipulation fastest — but the included studies were rated low to average quality, with small samples and limited blinding.
View study →120 adults aged 30–60 randomised to motor control exercise, balance training, both, or usual care for 12 weeks. The combined group reached a median pain score of 2.8/10 versus 6.5 for usual care, and disability of 15% versus 37%.
View study →The 2017 review is the closest thing the field has to a summary, and its verdict was that manipulation, exercise and taping all help, with manipulation producing the quickest change in pain.[4] Its authors were candid about what sat underneath that: nine studies ranging from single cases to 65 participants, most without blinding, several without proper randomisation, and inconsistent follow-up.[4] It is a reasonable basis for choosing a direction of treatment and a poor basis for predicting your result.
The 2023 trial is the better-built one, and it is the reason our programmes lean the way they do. Adults aged 30 to 60 were randomised across four arms for 12 weeks of twice-weekly supervised sessions. The group that got motor control exercise and balance training finished with a median pain score of 2.8 out of 10 against 6.5 for usual care, an Oswestry Disability Index of 15% against 37%, and quality-of-life scores that held at the 24-week follow-up — outperforming either component on its own.[5] The authors name their own limits plainly: a narrow age band of 30 to 60, and no follow-up beyond 24 weeks, so nothing is known about whether the gains hold at two years.[5]
What we take from all of it is modest and useful. Combining hands-on treatment with progressive loading looks better than either alone; the loading that matters is single-leg and balance-based rather than generic core work; and anyone quoting you a percentage success rate for an SI joint protocol is quoting a number that does not exist in this literature.
Injections and fusion belong after a genuine course of conservative care has been tried and measured, not instead of one. The usual escalation runs from exercise, manual therapy, anti-inflammatories and a pelvic belt, through image-guided steroid injection and radiofrequency denervation of the nerves supplying the joint, to surgical fusion.[1] An image-guided injection does double duty — it is both a treatment and the reference standard for confirming the joint is genuinely the source.[1]
Fusion has the strongest surgical evidence and it needs reading with its caveats attached. A randomised trial comparing minimally invasive SI joint fusion against conservative management reported substantially better one-year results for surgery — low back pain scores of 14.0 versus 41.6 and disability scores of 8.7 versus 25.0 — with benefits persisting at two years.[6] The caveats are not small: outcome assessment was not blinded and the primary measures were patient-reported, in a trial where everyone knew which group they were in; participants could cross over from conservative care to surgery at six months, and those who did improved at similar rates.[6] Older surgical experience was considerably less happy — one study cited in the standard clinical reference on this condition reported that 82% of patients were unsatisfied after surgery.[1]
None of that makes surgery wrong for the right person. It makes the sequence matter. Because the diagnosis itself is only about 35% certain on clinical testing,[2] operating on an unconfirmed hypothesis is a real risk, and conservative treatment that fails is itself informative — it is part of how the case for anything more invasive gets built honestly.
A handful of patterns turn pelvic pain from a physical therapy problem into a medical one. Go to an emergency department the same day for:
And see a physician before starting therapy for:
Screening for every one of these is a fixed part of the pelvic and spine evaluations we run in Boise and Nampa. It is why an appointment you booked about stairs includes questions about your sleep, your morning stiffness and your general health.
SI joint care at RISE is built around one question: how much load can you take through one leg before it hurts? That is the demand the joint is failing, it is measurable, and it responds to training. We are equally explicit about what we are not doing — we are not realigning anything, and we are not signing you up for maintenance adjustments. Getting that right takes a full 45 minutes one-on-one with the same Doctor of Physical Therapy every session, because the diagnosis is provisional and the plan has to keep being checked against how you actually respond.
Provocation cluster, lumbar and hip screening, red-flag and inflammatory questions — so you leave knowing what we think it is, how confident that is, and what would change our mind.
Manual therapy and manipulation, a pelvic belt if it helps immediately, and short-term changes to the asymmetric loads that keep re-provoking the joint through the day.
Progressive hip and gluteal strengthening with motor control and balance work, taken far enough that stairs, cars and bad nights stop being events worth planning around.
CONDITION
CONDITION
CONDITION
CONDITIONOne-on-one care for SI joint pain with a Doctor of Physical Therapy in Boise and Nampa.
Carson is a co-founder of RISE Physical Therapy and treats orthopedic, sports, and spine conditions with a hands-on, movement-first approach. He earned his Doctorate of Physical Therapy from Pacific University in Oregon. Meet the full team →