Sacroiliac Joint (SIJ) Dysfunction / Pelvic Girdle Pain

The naming, first, because it decides what the numbers mean. SIJ pain is a response to an anaesthetic block. SIJ dysfunction is an aberrant position or movement — and it has no reference standard at all, which is why "validity of the tests for this disorder is unknown" [1]. Every accuracy number on this page belongs to PAIN. None of it belongs to DYSFUNCTION. Type "SIJ dysfunction" into PubMed and you will surface treatment trials that never defined the diagnosis; search sacroiliac joint pain, sacroiliac provocation and pelvic girdle pain instead.

Then split the branch, in the first thirty seconds of the subjective.

Who: Someone with unilateral buttock or posterior pelvic pain — postpartum, after a fall onto the buttock, or insidious.
What: Paramidline, below L5, unilateral, ± groin, ± posterior thigh. Not midline. Nothing below the knee, no dermatomal band, no neurological deficit.
How it behaves: Aggs: standing, walking, stairs, single-leg stance, rolling in bed, sit-to-stand. Eases: rest, external compression.

The illness script in one line:

A joint that hurts when it is loaded asymmetrically, in someone who will tell you it feels "stuck" or "out" — so the pain sits off the midline below L5, external compression helps, and the only way anyone has ever confirmed it is by injecting it.

The one thing to take from this page:

The cluster crosses the joint off. It does not write it in. Three or more positive pain provocation tests moves you from 20% to about 35% certainty. All of them negative moves you to about 92% certainty the SIJ is not the source. Use it in the direction that works.

[2][3][4][5][6][7][8][9][10][11][12][13][14][15][16]

No Australian guideline exists for this presentation. There is no Australian guideline for sacroiliac joint pain; for pregnancy-related pelvic girdle pain, the Australian Pregnancy Care Guidelines carry a weak recommendation that pregnancy-specific exercises, physiotherapy, acupuncture or a support garment may provide some pain relief. Anchored to the European guidelines for the diagnosis and treatment of pelvic girdle pain [9] for the conservative and pregnancy pathway, and to the multispecialty consensus practice guidelines on sacroiliac joint complex pain [15] for the block and interventional half. The European guideline is eighteen years old, consensus-based, and nothing has replaced it — a live search on 2 Sep 2026 returned no later physiotherapy guideline in this pattern.

Australia's only lumbopelvic standard is the ACSQHC Low Back Pain Clinical Care Standard (2022) [17]. It was read live on 2 Sep 2026: eight quality statements covering assessment and the focused neurological examination, imaging reserved for suspected serious pathology, education, self-management and physical activity, physical and psychological interventions, medicines, and review and referral. A string search of that text returns zero occurrences of "sacroiliac", "SIJ" and "pelvic girdle". It is the right safety anchor and it is not a pattern anchor. Say both halves.

And the other half, which is this pattern's own. The absence of an Australian guideline is not an absence of Australian evidence. Schwarzer, Aprill and Bogduk built the first prevalence figure at the University of Newcastle. Laslett published the reasoning-plus-tests validity study in the Australian Journal of Physiotherapy. Hancock, Maher and Han built both cross-structure systematic reviews at the University of Sydney. We wrote a large share of this literature and then never wrote the standard.

Patterns > Labels. "SIJ dysfunction" names a position nobody can measure. Unilateral, paramidline, below L5, worse on single-leg loading, eased by compression, with the lumbar spine cleared and the inflammatory questions asked names the problem you can work with on Monday. Check it with your educator.

Five stages, in order. And the honest opening, because it is a finding rather than a failure: in the most direct test of this question — 85 patients, 12 tests chosen by a multidisciplinary expert panel, two examiners, intra-articular block with ≥90% relief as the criterion standard — "no historical feature, none of the 12 sacroiliac joint tests, and no ensemble of these 12 tests demonstrated worthwhile diagnostic value" [18].

So the subjective here does what a subjective is actually for. It splits the branch, generates the hypothesis, sets your dosage and rules out the dangerous neighbours. One historical feature has a number attached and you will meet it in the body chart.

MAIN PROBLEM

"It's right here, one side, in my backside. Started after I had the baby / after I came down the stairs badly. Standing in the queue kills it and I can't roll over in bed. It feels like it's out of place."

Two things to catch in that sentence. One side, one region, below the belt line — and "out of place", which is a patient's report of what it feels like, not a finding. Write it down as their words. You'll have been taught that the joint does not go out — control over correction. That one holds up, and it is the first thing you will say back to them.

Ask the branch question now: "Has this been there since a pregnancy?" Everything downstream turns on the answer.

BODY CHART

• P1: unilateral, paramidline, below L5, over the buttock or posterior pelvis. P2 commonly into the groin or the posterior thigh. Mark P+ / P++ where they show you the worst of it, and mark the side.
• Ask them to point with one finger. Provocative injection into the SIJ produced buttock hypoaesthesia running about 10 cm caudally and 3 cm laterally from the PSIS, matching the area of maximal pain — in 10 asymptomatic volunteers [2]. That is a ten-person map. It generates a hypothesis. It confirms nothing.
• Then do the one thing on this page that has a likelihood ratio attached: establish midline versus paramidline explicitly. Absence of midline low back pain: +LR 2.41 (95% CI 1.89–3.07), −LR 0.35 (0.12–1.01) [13]. Midline pain was reported by 95.8% of patients with internal disc disruption versus 12.9% of those with SIJ pain (χ²=101.4, P<.0001) [19]. Midline points away from this pattern and towards the disc.
• Groin pain earns its own mark. In the original Australian prevalence study it was the only referral pattern associated with a positive block [3].
• Nothing below the knee. No dermatomal band. If there is, you are on the lumbar radicular pain page.

BEHAVIOUR

• Aggs: standing, walking, stairs, single-leg stance, rolling over in bed, sit-to-stand, getting out of the car.
  Eases: rest, and external compression — which is the pattern's own SE line and the reason a belt is a reasonable trial rather than a gimmick.
• 24-hr pattern: standing, walking and stairs build it; the transitions — rolling over in bed, sit-to-stand, out of the car — are in the Aggs; rest and external compression ease it.
• Night: rolling over in bed is in the Aggs. Pain at night is a §4 flag, and morning stiffness over 30 minutes in a young adult sends you to §4.
• Severity: VAS 0–10 at best / at worst / right now.
• Coding: usually In or Is — position- and load-driven, settling when unloaded — moving to P when irritable. Morning stiffness <30 min = M; >30 min = In, and >30 min in a young adult sends you straight to section 4.
• Irritability decides your dosage, and it has three parts. Resting symptoms + the vigour of the activity that provokes it + how long it takes to settle. In pregnancy and early postpartum, irritability is usually the limiting factor, not range.
• Asymmetrical, painful functional movements — sit-to-stand, stairs, single-leg tasks — are worth recording because you can re-test them. They have never been measured against any reference standard in this pattern. Use them as asterisk signs, not as evidence.

HISTORY

• Onset: during or after a pregnancy, a fall onto the buttock, a fall onto one leg, or insidious. Date it and date what it has done since.
• Read the prevalence literature knowing where it came from. 13/43 (30%) of consecutive patients with chronic LBP maximal below L5-S1 obtained gratifying relief from a single fluoroscopically guided block (Schwarzer et al. 1995). In a pre-selected cohort taken to double blocks, only 10/54 (18.5%) survived the confirmatory injection [4]. Pooled, the range is 10% to 62% depending on setting, most studies near ~25%, with an uncontrolled-block false-positive rate of ~20% [10]. The 2025 consensus lands on 15–30% of patients with axial pain predominantly below L5 [15]. In the tertiary accuracy studies the median proportion testing positive is 53% — because those patients were referred for the block (Han et al. 2023).
  Read that spread the way you read the facet literature: the higher figures come from populations selected to have the condition. A student's placement caseload sits at the low end.
• Pregnancy is a different scale entirely. About 45% of all pregnant women and 25% postpartum report pelvic girdle pain and/or pregnancy-related low back pain; serious pain in ~25% during pregnancy, severe disability in ~8%, still serious postpartum in ~7%. Of all patients, about half have PGP, a third have pregnancy-related LBP and a sixth have both — and PGP and LBP are distinct entities, not one thing [6]. Point prevalence of PGP in pregnancy is ~20% [9].
• Ask what makes it centralise, because it changes the answer. Adding a McKenzie evaluation to the provocation tests took the clinical examination to sensitivity 91% (95% CI 62–98), specificity 83% (68–96), +LR 6.97 (2.70–20.27) against double blocks [20]. Specificity of ≥3 positive provocation tests rises from 78% to 87% in patients whose symptoms cannot be made to centralise [1]. If you can centralise their pain, stop testing the SIJ and go to non-specific low back pain.
• Previous episodes, previous imaging, what they have been told, and what helped last time. If they have been told a joint is "out", find out who told them and what they now believe.

SPECIAL / SCREENING QUESTIONS

Every patient. Every time. On this page the screening questions are the highest-yield part of the subjective, because the miss here is a disease, not a structure.

• Inflammatory back pain / axial spondyloarthritis. Age at onset <40, insidious onset, night pain, better with exercise, not better with rest, morning stiffness >30 min (coding In), good response to NSAIDs. Then the company axSpA keeps: uveitis, inflammatory bowel disease, psoriasis, enthesitis, dactylitis, family history. Section 4.
• Fracture. Training spike, running volume, RED-S or low energy availability, amenorrhoea, osteoporosis, long-term corticosteroids, prior pelvic radiotherapy. Pain that does not ease with relative rest.
• Infection. Fever, rigors, night sweats, feeling unwell, recent instrumentation, immunosuppression, IV drug use, unremitting non-mechanical pain.
• Cauda equina. Bilateral leg symptoms, saddle anaesthesia, bladder or bowel change, sexual dysfunction. Buttock pain is not a reason to skip the question. Cauda equina syndrome.
• The ordinary serious ones: unexplained weight loss, cancer history, night pain that is progressive and constant.

Rank the mechanical patterns — one discriminating feature each, and that is all you need to move between them. Then rule out the dangerous ones in section 4, which comes before any of this in practice.

Two more that sit beside the ranking rather than inside it.

• Gluteal tendinopathy / GTPS is lateral — over the greater trochanter, worse lying on that side. Lateral, not posterior. Gluteal tendinopathy.
• Pubic symphysis pain is central pubic pain on walking, rolling in bed and stairs, and it commonly coexists. It matters prognostically: in a 1,789-woman pregnancy cohort, none of the women classified as symphysiolysis still had pain at 6 months, versus 21% of those with pelvic girdle syndrome — all three pelvic joints — still in pain at two years [5].

And the branch, which is not a competitor. Pregnancy-related pelvic girdle pain is a separate pattern with its own prevalence, its own guideline, its own test, its own trial and its own prognosis [6][9][21][22][11]. Do not merge the two on a page or in your reasoning.

Clear the lumbar spine first, and write down that you did. Across 43 RCTs of conservative management for SIJ pain, only 2 (4.7%) performed a reasonable thoracolumbar screen before making the diagnosis, 9 (20.9%) partially, and 32 (74.4%) did none at all [16]. More than 95% of the trials that generated this pattern's treatment evidence skipped the step you are about to do. That is not a formality — it is the reason section 3 exists.