Back Pain Doctor

Sacroiliac Joint Rehabilitation: A Staged Programme

Rehabilitation is the only treatment in this section with durable randomised evidence behind it, and it is also the one most often delivered in a form that does not work. The difference between the trial that produced large two-year effects and the trials that produced nothing is not the exercises. It is the supervision, the individualisation and the sequence.

The design principle that separates the trials that worked from the ones that did not

Two randomised trials of stabilising exercise for postpartum pelvic girdle pain reached opposite conclusions. Mens' trial of diagonal trunk muscle exercises found nothing. Stuge's trial found large effects that were still present two years later. Stuge subsequently published a direct comparison of the two designs, and the differences are the whole lesson.

The trial that found nothing The trial that worked
Muscles targetedGlobal muscles onlyLocal muscles first, global muscles added gradually
DeliveryInstructed by videotapeSupervised and corrected in person
IndividualisationNot individualisedChoice of exercise, order and dosage individualised
Pain during exerciseNot managedExplicitly avoided
DurationShorterTwenty weeks

What that means for you

  • A printed sheet of pelvic exercises done unsupervised is closer to the trial that failed than the one that worked.
  • Sequence matters: deep local control is established before global loading is added, not alongside it.
  • Exercises that hurt are modified, not pushed through. Pain avoidance was a design feature of the successful trial, not a concession.
  • The successful programme ran for twenty weeks. Anyone promising resolution in three sessions is not describing this.

The honest caveat, stated up front

The evidence for exercise in this specific condition is better than for anything else here, but it is not strong in absolute terms. A 2025 narrative review of motor control exercise for pelvic girdle pain found six eligible studies, of which five reported no statistically significant change in short-term pain, and concluded that the exercises used across the literature were generally unspecific and not designed according to motor control theory.

The 2025 consensus guidelines make the same point differently: the evidence base for non-interventional therapy in sacroiliac joint complex pain is indirect, extrapolated mostly from general low back pain studies.

So the position taken here is: this is the best-supported treatment available, it is worth doing properly, and it is not guaranteed to work. Both halves of that are true and neither is left out of the conversation.

The programme

What follows is a template built on the design features of the trials above. It is a starting structure, not a prescription — exercise selection, dosage and rate of progression are individualised at assessment, and the phase criteria matter more than the exercise names. Nothing here replaces assessment by a clinician who has examined you.

Approximately weeks 0-4

Phase 1 — Settle and re-establish control

Reduce the repeated provocation that is keeping symptoms alive, and re-establish deep trunk and pelvic floor control without loading the painful pattern.

  • Load triage, not rest Daily

    Identify the two or three specific tasks that reliably flare it — commonly stairs, prolonged standing, single-leg dressing, rolling in bed — and modify them temporarily. Walking and general activity continue. Bed rest is not part of this.

  • Transversus abdominis and pelvic floor co-activation 10 holds of 10 seconds, 2-3 times daily

    Taught in supine, progressed to sitting and standing. Gentle, low-effort contraction — this is a timing and awareness task, not a strength task. Breathing continues throughout.

  • Supine bridging, double leg 2 x 10, daily, pain-free range only

    Establishes gluteal recruitment without single-leg load. Stop short of any symptom reproduction.

  • Supported rolling technique Practised each time you turn in bed

    Log-rolling with knees together and a pre-set trunk contraction. Often the single highest-value change in the first fortnight, because night pain drives much of the distress.

  • Hip mobility within tolerance Daily, 30-60 seconds per position

    Hip flexor and hip rotation range, because restricted hip movement pushes load into the pelvis. Not aggressive stretching of the posterior pelvis.

  • Pelvic belt, if it helps During provocative activity only

    A non-rigid belt has randomised evidence for short-term symptom control in pregnancy-related pelvic girdle pain. Used to permit activity, not to replace it, and weaned as capacity improves.

Before moving on

  • Night pain and rolling pain reduced.
  • Able to hold the deep co-activation in sitting and standing without breath-holding.
  • Double-leg bridging pain-free.
  • Daily walking tolerance stable or improving for at least a week.

Approximately weeks 4-10

Phase 2 — Build capacity

Add global muscle work to the local control that is now established, and begin restoring strength through the hips and trunk.

  • Single-leg bridging 3 x 8 each side, progressing to 3 x 12

    The first genuine asymmetrical load. Introduced only once double-leg bridging is comfortable. Pelvis stays level; if it drops, regress.

  • Side-lying hip abduction, progressing to side plank on knees 3 x 10 each side; side plank 3 x 20-30 seconds

    Gluteus medius capacity, which is central to controlling the pelvis in single-leg stance.

  • Bird-dog 3 x 8 each side, held 5 seconds

    Trunk control with contralateral limb load — the pattern that most closely mimics gait demand. Quality over range.

  • Sit-to-stand and split squat 3 x 10, progressing to added load

    Restores the leg strength that lets the pelvis be loaded without compensation. Split squat introduced once sit-to-stand is comfortable.

  • Hip hinge pattern, unloaded then loaded 3 x 10, load added by tolerance

    Deadlift pattern taught without load first. This is the movement most people avoid and most need.

  • Graded walking programme Increase weekly distance by no more than 10-20 per cent

    Progressive, measured, and not reset by every minor flare.

Before moving on

  • Single-leg bridge with a level pelvis, 12 repetitions each side.
  • Side plank on knees held 30 seconds each side without symptom reproduction.
  • Stairs manageable without a marked increase in pain.
  • No more than mild soreness settling within 24 hours after sessions.

Approximately weeks 10-16

Phase 3 — Single-leg load and force transfer

Restore tolerance of the loads that actually provoke this problem — single-leg stance, step-through, and asymmetrical trunk load.

  • Step-ups and step-downs 3 x 8-10 each side, increasing step height

    The most direct rehearsal of stair load. Controlled descent is the harder and more important half.

  • Single-leg deadlift 3 x 8 each side, load added by tolerance

    Combines single-leg stance with a hip hinge and a trunk demand. Close to the mechanics of the active straight leg raise.

  • Loaded carries — suitcase and farmer's 3 x 20-30 metres each side

    Asymmetrical carries load force closure directly and are well tolerated by most people at this stage.

  • Full side plank 3 x 30-45 seconds each side

    Progression from the knee version once that is comfortable.

  • Split squat and lunge with load 3 x 8-10 each side

    Progressive external load. This is where capacity is genuinely built.

  • Belt weaned Reduce to provocative tasks only, then discontinue

    If a belt was used in Phase 1, this is where it comes out. Continued reliance at this stage is a signal to re-examine rather than to keep the belt.

Before moving on

  • Single-leg stance held 30 seconds with a level pelvis.
  • Step-downs from a standard stair height, controlled, both sides.
  • Loaded carries without symptom reproduction.
  • Return to the previously avoided daily tasks.

Approximately weeks 16-20 and beyond

Phase 4 — Return to full loading

Rebuild the specific capacity your work, training or sport demands, and establish what you keep doing afterwards.

  • Progressive resistance training 2-3 sessions weekly, continuing indefinitely

    Squat, hinge, single-leg and carry patterns loaded progressively. This is the maintenance dose, not a temporary phase.

  • Impact reintroduction, where relevant Graded, starting well below previous volume

    For runners: short intervals of running within walking, increasing total weekly volume by no more than 10 per cent. See the running injury guidance.

  • Task-specific rehearsal Built into training

    Lifting technique for manual work, rotation and change of direction for field sport, prolonged standing tolerance for those who need it.

  • A maintenance minimum Two sessions weekly

    The realistic floor that holds the gains. Most recurrences follow a period of doing nothing plus a load spike, and both halves of that are avoidable.

Before moving on

  • Full return to the activities that matter to you.
  • A training habit you will actually keep.
  • A written plan for what to do at the first sign of a flare, so it does not become a relapse.

Flares — what to do and what not to do

A flare is not a re-injury

  • Symptom increase after a load increase is expected and is not damage. The rule of thumb used here is that soreness settling within 24 hours is acceptable; soreness that is worse the next morning means the last progression was too big.
  • Drop back one phase for a week rather than stopping. Complete rest reliably makes the next attempt harder.
  • Do not restart the programme from the beginning. You keep the capacity you built.
  • Reassess if a flare lasts more than two weeks, if the pain pattern changes, or if any of the red flag features on the diagnosis page appear.

Where procedures fit around this

The sequence used here is deliberate. Rehabilitation comes first, and is delivered properly, because it is the only element with durable randomised support and because a meaningful proportion of people improve without any procedure at all.

Where a procedure is considered, it is considered as something that may create a window in which the loading work becomes possible — not as an alternative to it. The original Sydney dextrose protocol required a three-month stability programme before any injection, and that principle remains sound.

The specific evidence for each option is set out separately: PRP, prolotherapy, and shockwave and EMTT.

Related on this site

Common questions

How long does sacroiliac rehabilitation take?
The trial that produced durable results ran for twenty weeks, with effects still measurable two years later. That is the realistic frame for a persistent problem. Some people improve considerably sooner; the point of quoting the trial duration is that programmes abandoned at three or four weeks are not the thing that was tested.
Can I do these exercises from a sheet at home?
Partly, but the evidence suggests that is the weaker version. The randomised trial that found no benefit delivered exercises by videotape without individualisation; the trial that found large benefits supervised and corrected them, individualised the selection and dosage, and avoided provoking pain. A home programme is worth doing, but it should be one that has been set and adjusted for you.
Should exercises hurt?
No. Pain avoidance was an explicit design feature of the successful trial, not an optional extra. Mild soreness that settles within 24 hours is acceptable; pain during the exercise means it should be modified or regressed.
Do I need to keep exercising once the pain settles?
Yes, at a maintenance dose. Two sessions a week of progressive loading is the realistic floor. Most recurrences follow a period of doing nothing combined with a sudden increase in load.
Is core strengthening the answer?
Not by itself, and not in the way it is usually described. The trial evidence supports establishing deep local control first and then progressively adding global strength and single-leg loading — a sequence, not a set of abdominal exercises. A 2025 review found that most motor control exercise programmes used in this population were unspecific and produced no significant short-term pain change.
Can I exercise while pregnant with pelvic girdle pain?
Generally yes, with the programme adapted to the stage of pregnancy, and this should be coordinated with your maternity care. A randomised trial in pregnancy found a non-rigid lumbopelvic belt gave better short-term pain and disability outcomes than home exercise alone, so a belt is a reasonable adjunct during this period. Any new pain with fever or neurological symptoms needs prompt assessment rather than an exercise adjustment.

Evidence reviewed

Every clinical statement on this page traces to one of the sources below, including the ones that point the other way. Where a study is small, unblinded or authored by people with a commercial interest in the result, that is noted alongside it rather than left out.

Last reviewed 2026-09-04. This is general information about a condition and its treatments, not personal medical advice, and it is no substitute for assessment by a clinician who has examined you.

More in this section

The joint and its ligaments

Why the ligaments behind the joint matter as much as the joint itself — and why that changes where a needle goes.

Lumbopelvic instability

A useful clinical model that has been badly oversold. Here is the part that holds up.

Getting the diagnosis right

No single test proves it. Here is what each one is actually worth.

Protocols for clinicians

Written for referrers and physiotherapists: the parameters, the outcome measures and the effect sizes, without the patient framing.

PRP at the sacroiliac joint

Two randomised trials, opposite conclusions, and the better-designed one is the negative one.

Prolotherapy at the sacroiliac joint

The longest trial record at this joint — and a real-world cohort where fewer than a quarter of patients improved.

Shockwave and EMTT

Two small randomised trials, one sham-controlled, and a device with no trials at all.

Referral, ablation and surgery

The procedures not done here, what their evidence actually shows, and when a referral is the right answer.

Ready for a clearer plan for your back or musculoskeletal pain?

Book an assessment with Dr Joshua Hatch.

Your assessment focuses on understanding the likely source of your pain and the most appropriate non-surgical options for your diagnosis, with the aim of reducing pain and improving function.

Book an appointment
Book an appointment with Back Pain Doctor
Book now Call