This Week in Physical Medicine & Rehabilitation — Aug 13, 2026

Welcome to This Week in Physical Medicine and Rehabilitation. This week we're covering 10 notable papers spanning prediction and measurement in neurorehabilitation, shoulder diagnosis and injection practice, and the economics and add-ons of rehabilitation care. Let's dive in.

We start with prognostication, because so much of inpatient rehabilitation planning hinges on it. In the Archives of Physical Medicine and Rehabilitation, Cogan and colleagues took a national sample of nearly forty thousand adults admitted to United States inpatient rehabilitation facilities with stroke in fiscal year 2023 and asked whether the data we are already legally required to collect — the Inpatient Rehabilitation Facility Patient Assessment Instrument — can predict discharge function [1]. Using random forest regression with twenty-nine predictors, and then distilling those forests into visual summary trees a clinician can actually read, the models explained roughly six-tenths of the variation in Section GG self-care and mobility at discharge. Unsurprisingly, admission function was the single strongest predictor, but the notable signal was bladder and bowel incontinence, which predicted both self-care and mobility outcomes. The practical message is that no new data collection is needed to generate defensible early estimates for discharge planning conversations, and that continence status deserves more weight in your prognostic thinking than it usually gets. A parallel prediction paper in Spinal Cord addresses a much more specific question with real day-to-day consequences: who will achieve independent clean intermittent catheterization after motor-complete spinal cord injury [2]. In 194 patients from a Japanese specialist centre, all classified as American Spinal Injury Association Impairment Scale A or B at discharge, forty-seven percent achieved independent catheterization. Upper extremity motor score discriminated this outcome with excellent accuracy, and importantly the thresholds differed sharply by sex and age — roughly nineteen for men under sixty-five, twenty-eight for older men, twenty-four for younger women, and much higher, around fifty, for women aged sixty-five and above. Scores measured as early as four weeks post-injury performed just as well, which means you can set realistic bladder management expectations, and target hand and arm training accordingly, within the first month rather than at discharge. The single-centre retrospective design and the fact that these are study-specific thresholds mean they should be treated as a guide rather than a rule.

Staying with measurement, a longitudinal study of 286 patients with subacute stroke, also in Archives, tackled a problem that quietly undermines a lot of our outcome reporting: which minimal important change value for the Berg Balance Scale should we actually believe [3]. Comparing Berg change between thirty and sixty days after stroke against improvement in ambulation category and functional independence, the authors found that thresholds of two and five points behaved acceptably as validated change scores, while the frequently cited nine-point threshold did not. In patients who still needed walking assistance, five points held up; a four-point threshold in independent walkers did not. So when you document that a patient has meaningfully improved in balance, five points on the Berg is the defensible number in a subacute inpatient population.

Turning to the shoulder, three papers this week converge on how we diagnose and inject it. In the American Journal of Physical Medicine and Rehabilitation, Qin and colleagues pooled twenty-three studies and 739 patients receiving subacromial corticosteroid injection for impingement syndrome [4]. Pain improved by about three points on a ten-centimetre visual analogue scale overall, with substantial functional gains, but meta-regression found no relationship whatsoever between corticosteroid dose and either pain or function. Doses below forty milligrams of triamcinolone equivalent produced essentially the same result as forty milligrams or more. This is indirect comparison rather than head-to-head randomisation, so it is not proof of equivalence, but it makes a reasonable case for a therapeutic ceiling and for defaulting to lower doses, which matters given the cumulative cartilage, tendon, and glycaemic costs of steroid. Diagnosis of the same region gets a reality check in Archives, where Green and colleagues prospectively examined seventy-one adults before shoulder surgery, comparing clinical tests and high-resolution ultrasound against blinded intraoperative findings for long head of biceps pathology [9]. The orthopaedic special tests — Speed's, O'Brien's, Uppercut, Yergason's, and groove palpation — performed poorly on their own. Qualitative ultrasound was reasonably specific for full-thickness tears, at eighty-six percent, but only moderately sensitive at sixty-three percent, and it was frankly weak for detecting dislocation or subluxation, missing most cases. The best performance came from combining ultrasound findings with the clinical tests and patient age. Quantitative measures were cut-off dependent, with a maximal tendon surface area above twenty-one square millimetres giving the best balance. The takeaway is to stop treating a negative special test or a normal-looking tendon on ultrasound as ruling out biceps pathology, and to interpret imaging within the full clinical picture. The third shoulder paper, in PM and R, is arguably the most striking [5]. Among 114 infants with birth brachial plexus palsy and persistent neurological deficit, universal ultrasound screening at three months found posterior shoulder subluxation in eighty-three percent, with fifty-eight percent in the severe range by alpha angle. Prior literature had reported anywhere from seven to fifty-six percent. Critically, physical examination did not identify these children: three-quarters of infants with entirely normal passive external rotation had subluxation on imaging, and forty-four percent of those with severe subluxation had normal passive external rotation. Very few infants who were normal at three months developed subluxation by six months, suggesting the process is largely established early. If you follow these infants, exam-based screening alone is not sufficient, and early ultrasound at around three months deserves consideration — though this is retrospective single-centre data and it remains unproven that earlier detection changes joint outcomes.

Our final theme is what we add to rehabilitation and what it costs. In Spinal Cord, a Swiss retrospective analysis of fifty-three patients with chronic spinal cord injury and denervated gluteal muscles who required surgical pressure injury treatment found median total costs of roughly 143,000 Swiss Francs, driven largely by indirect operating costs, with a median hospital stay of eighty-nine days [6]. Projected costs for twenty years of preventive electrical stimulation came to about 12,600 Swiss Francs. That is a modelled projection rather than a trial of prevention, so it demonstrates cost asymmetry rather than proven efficacy — but the asymmetry is large enough to justify pushing for reimbursement discussions. On the therapy side, an Archives observational study of 1,619 inpatient rehabilitation patients used inverse probability of treatment weighting to compare those who had at least one session involving a trained facility dog against those who did not [7]. Canine-assisted intervention was associated with about 2.4 additional points of composite Section GG improvement, a result that just crossed statistical significance, with self-care and mobility domains trending in the same direction but not significant individually. There was a dose-response trend with cumulative exposure. This is non-randomised and vulnerable to selection of healthier, more engaged patients, so treat it as hypothesis-generating. Two systematic reviews round out the theme. In the Journal of Rehabilitation Medicine, nine randomised trials totalling 627 patients showed heat therapy for hand osteoarthritis produced moderate short-term pain reduction and improved grip strength, with multimodal approaches like paraffin and mud outperforming plain heated gloves, and longer, more frequent sessions predicting better grip outcomes [8]. And in Archives, a meta-analysis of twenty-one exercise trials in older adults with Alzheimer's disease found benefit across four of five cognitive domains plus activities of daily living, emotional well-being, and quality of life — though not caregiver burden — with the most effective dose being sessions under sixty minutes at least three times weekly rather than fewer, harder bouts [10].

If you only have time for one paper this week, make it the birth brachial plexus ultrasound screening study in PM and R [5]. It shows that the physical examination we rely on misses most cases of a structural complication that is far more common than we believed, and that alone should change how you follow these infants.

Here are the key takeaways from this week in Physical Medicine and Rehabilitation. First, mandated Medicare assessment data alone can predict discharge self-care and mobility after stroke reasonably well, and incontinence is a stronger prognostic marker than most of us treat it as. Second, upper extremity motor score at four weeks can tell you early who will manage independent catheterization after motor-complete spinal cord injury, with thresholds that differ by sex and age. Third, five points is the defensible minimal important change on the Berg Balance Scale in subacute stroke — not nine. Fourth, higher corticosteroid doses buy you nothing extra in subacromial injection for impingement, so default lower. Fifth, in infants with birth brachial plexus palsy, normal passive external rotation does not exclude posterior shoulder subluxation; consider ultrasound screening at three months. And finally, frequent moderate exercise, under an hour and at least three times weekly, supports cognition and function in Alzheimer's disease, while multimodal heat therapy offers short-term relief in hand osteoarthritis.

That's your roundup for This Week in Physical Medicine and Rehabilitation. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.