Weekly MSK Literature Review

Monday: Prognostic Factors for Shoulder Pain Recovery in Physical Therapy

What predicts which shoulder pain patients will recover -and which will not?

• 60% of patients recovered at 26 weeks (65% in the working population), with SPADI scores dropping from a median of 49.5 to 16.9 and pain scores from 6.0 to 2.0

• Shorter symptom duration was the strongest independent predictor of recovery, reinforcing the value of early access to care

• Lower baseline disability (SPADI) at intake predicted better recovery across all models

• Absence of anxiety or depression (EQ-5D) was independently associated with recovery in the total population

• Higher working alliance scores predicted recovery, meaning the quality of the therapeutic relationship had measurable prognostic weight

• Employment status mattered: having a paid job increased recovery likelihood in the total population, though not in the working-population subgroup alone

• The final model showed modest discrimination (AUC 0.67), and the authors caution against clinical implementation without external validation

Why it matters: Psychosocial factors, specifically mood and therapeutic alliance, are not soft variables -they are independent predictors of shoulder recovery in outpatient PT. Clinicians who screen for anxiety and depression at intake and actively cultivate working alliance are doing prognostic work, not just rapport building.

Karel, Y. H. J. M., Verhagen, A. P., Thoomes-de Graaf, M., Duijn, E., van den Borne, M. P. J., Beumer, A., Ottenheijm, R. P. G., Dinant, G.-J. J., Koes, B. W., & Scholten-Peeters, G. G. M. (2017). Development of a prognostic model for patients with shoulder complaints in physical therapist practice. Physical Therapy, 97(1), 71–80.

Tuesday: Early Physical Therapy in Rotator Cuff Tears -Faster Recovery, Not Different Recovery

Does starting PT early in a rotator cuff tear change long-term outcomes, or just the speed of getting there?

• Patients who completed PT within the first 3 months had SPADI scores 10.96 points lower than those who did not (p = .02), with the gap widest at 3 months (-13.03 points)

• The between-group difference was not statistically significant at 6, 12, or 18 months, suggesting early PT accelerates recovery without altering the long-term trajectory

• Outcomes improved linearly up to approximately 16 PT sessions, after which gains plateaued or slightly declined -a clear dose-response threshold

• Clinical meaningfulness is uncertain: whether the 10.96-point SPADI difference clears the MCID depends on which threshold (8, 10, or 13.2 points) is applied

• Mental health (MHI-5) was not a significant predictor of outcomes in this cohort, contrasting with findings in adjacent literature

Why it matters: The 16-session threshold is the most actionable finding here. It gives clinicians and payers a defensible, evidence-informed ceiling for conservative rotator cuff management -and raises a pointed question about what is driving continued utilization beyond that point.

Dickinson, R. N., Ayers, G. D., Archer, K. R., Fan, R., Page, C., Higgins, L. D., Kuhn, J. E., Baumgarten, K. M., Matzkin, E., & Jain, N. B. (2019). Physical therapy versus natural history in outcomes of rotator cuff tears: The Rotator Cuff Outcomes Workgroup (ROW) cohort study. Journal of Shoulder and Elbow Surgery, 28(5), 833–838. https://doi.org/10.1016/j.jse.2018.10.001

Wednesday: Early Functional Change Predicts LBP Outcomes -PROs Are Doing the Work

Can the first three visits tell you whether a low back pain patient is on track?

• Across 6,523 patients treated in 26 outpatient PT clinics, 42.7% achieved at least 30% disability improvement by visit 3; 49% reached that threshold by visit 6

• Baseline MDQ score (visit 1) was the strongest single predictor of improvement by visit 3

• The combination of visit 1 and visit 3 MDQ scores dominated all other predictors -including age, sex, payer type, symptom duration, and pain location -for predicting improvement by visit 6

• Models using only MDQ1 and MDQ3 achieved excellent discriminative accuracy (AUC 0.84 and 0.85), meaning early PRO data alone can reliably identify who is and is not improving

• Demographic and payer variables contributed minimally, suggesting that outcome trajectory is driven by the patient's clinical response, not their background characteristics

Why it matters: If a two-point PRO dataset (visit 1 and visit 3) can predict visit 6 outcomes with AUC above 0.84, there is no clinical or administrative justification for waiting until episode completion to assess trajectory. EMR platforms that surface early PRO trends in real time are not a convenience feature -they are a decision-support infrastructure.

Brennan, G. P., Snow, G., Minick, K. I., & Stevans, J. M. (2023). Significant clinical improvement was predicted in a cohort of patients with low back pain early in the care process. Physical Therapy, 103(9), pzad082.

Thursday: Patient Expectations Predict PT Outcomes -But No One Is Measuring Them the Same Way

How much do patients' beliefs about their own recovery actually shape the result?

• Across 21 studies and 4,879 participants, higher baseline recovery expectations were positively associated with better outcomes in 47 of 48 analyses -a near-uniform directional effect

• The vote-counting score of 10.65 out of a possible ±21 confirms a clear positive relationship: patients who expect to improve tend to improve more

• Evidence spanned low back pain (7 studies), neck pain (6), shoulder pain (3), and knee pain (3), suggesting the expectation-outcome link is not condition-specific

• Measurement was highly heterogeneous: 13 studies used investigator-created Likert questions, with no universally accepted instrument across the field

• 18 of 21 included studies scored at least 10 out of 12 on CASP quality appraisal, indicating the finding is supported by generally sound methodology

Why it matters: The consistency of the expectation-outcome association across conditions and study designs means this is a real signal, not noise. The barrier to acting on it is not evidence -it is the absence of a standardized measurement tool. Clinics that add even a single validated expectation item to intake are capturing prognostic data that most practices are leaving on the table.

Wassinger, C. A., Edwards, D. C., Bourassa, M., Reagan, D., Weyant, E. C., & Walden, R. R. (2022). The role of patient recovery expectations in the outcomes of physical therapist intervention: A systematic review. Physical Therapy, 102(4), pzac008.

Friday: Psychological Distress Trajectory Predicts Long-Term Pain Outcomes

Can a four-week change in psychological distress tell you who will still be in pain at 12 months?

• Using the OSPRO cohort (n = 279), three predictive models were built to estimate the risk of failing to achieve 50% pain reduction at 12 months: a longitudinal model (AUC 0.79), an early-change model incorporating 4-week distress shift (AUC 0.71), and a baseline-only model (AUC 0.68)

• Baseline psychological distress (OSPRO-YF) and its change by week 4 were among the strongest predictors of 12-month pain outcomes across all models

• Anatomical region independently predicted risk: shoulder and low back pain carried higher likelihood of poor long-term pain reduction than other regions

• History of previous pain episodes was protective, suggesting experienced patients may engage with treatment differently

• The early-change and baseline-only models were selected for the P3 tool specifically because they are clinically feasible -capturing useful predictive signal without requiring extended follow-up data to run

• The P3 tool enables risk stratification at intake and risk updating at four weeks, supporting modification of care pathways before an episode completes

Why it matters: A patient whose psychological distress is not improving in the first four weeks is signaling treatment risk in real time. The P3 tool converts that signal into an actionable risk estimate -giving clinicians a structured basis for escalating care, adding behavioral health resources, or adjusting the plan before 12 months of poor outcomes accumulate.

Horn, M. E., George, S. Z., Li, C., Luo, S., & Lentz, T. A. (2021). Derivation of a risk assessment tool for prediction of long-term pain intensity reduction after physical therapy. Journal of Pain Research, 14, 1515–1524.

This Week at a Glance

Early intervention is a prognostic variable, not just a scheduling preference. Karel et al. found shorter symptom duration to be the strongest predictor of shoulder recovery; Dickinson et al. showed early PT produces the largest functional gains in rotator cuff tears, even when long-term outcomes converge. Getting patients into care quickly changes their clinical trajectory.

PROs measured early are more predictive than demographic or structural variables. Brennan et al. demonstrated that visit 1 and visit 3 MDQ scores alone outperform age, sex, payer, and symptom duration in predicting LBP outcomes. Karel et al. showed similar primacy of baseline SPADI. The patient's own functional report is the sharpest prognostic instrument available.

Psychological factors appear in every paper this week, across every condition. Karel et al. found anxiety and depression independently predicted shoulder recovery. Horn et al. identified psychological distress trajectory as a key predictor of 12-month pain outcomes. Wassinger et al. showed recovery expectations predict results across MSK conditions. These are not adjunct considerations -they are core clinical signals.

Dose and trajectory matter as much as treatment type. Dickinson et al. identified a 16-session threshold beyond which PT gains in rotator cuff tears plateau. Brennan et al. showed that early trajectory (visit 3 function) predicts final outcomes better than any single baseline variable. Both findings point toward the same operational principle: monitor response, not just delivery.

Measurement standardization is the field's most persistent gap. Wassinger et al. found 13 different expectation instruments across 21 studies, preventing meta-analysis. Karel et al. noted external validation is needed before their prognostic model is clinically deployable. Horn et al. offer the P3 tool as a step toward standardized, feasible risk stratification. Progress is being made, but heterogeneity continues to slow translation.

What This Means in Practice

Clinicians: Assess psychological status and recovery expectations at intake -not as screening extras, but as prognostic variables with direct bearing on treatment planning and patient communication.

EMR and technology developers: The Brennan et al. AUC findings make a direct case for real-time PRO trending built into the clinical workflow. Visit 3 function predicting visit 6 outcomes is only actionable if the platform surfaces that signal before visit 4.

Payers and utilization managers: The 16-session dose threshold from Dickinson et al. is the kind of evidence-grounded benchmark that belongs in benefit design conversations. Extended utilization beyond that point in rotator cuff management warrants scrutiny.

Value-based care strategists: Multiple papers this week demonstrate that early trajectory predicts long-term outcomes with high accuracy. Risk stratification tools like the P3 are the infrastructure for proactive, differentiated care pathways -not retrospective quality reporting.

Researchers and educators: The expectation measurement problem identified by Wassinger et al. is an open methodological problem with high clinical stakes. A validated, brief, condition-agnostic expectation instrument would unlock a body of comparative research that is currently fragmented.

Bottom line: Across shoulder, rotator cuff, and low back pain, the evidence this week converges on a single conclusion -psychological state, patient expectations, and early functional response are more predictive of outcomes than diagnosis, demographics, or treatment type alone. The clinics and systems that build infrastructure around those signals will outperform those that do not.