THE BODY IS NOT A MACHINE: WHY PHYSICAL THERAPY NEEDS A COMPLEX SYSTEMS LENS

How Nodes, Signals, and Directionality Are Reshaping MSK Care and Clinical Strategy

Consider this critical question: If physical therapy is effective, why does it consistently fail for a significant subset of patients who, by all structural indicators, should improve?

reliably fail for a meaningful subset of patients who, by every structural measure,

should get better?

You did the assessment. You found the impairment. You treated it. The patient came back three times a week for six weeks. Their functional outcome scores improved. And then, six months later, they are back in your clinic with the same complaint.

This is not a compliance problem. It is not a motivation problem. And it is definitely not a technique problem.

This is a systems-level problem.

The human body and the healthcare environment in which it receives care are not linear machines in which Input A produces Output B. Instead, they constitute complex adaptive systems: a dynamic network of interacting components that self-organize, respond to context, and behave in ways that cannot always be predicted by examining any single part in isolation. Pain, function, recovery, and chronicity all emerge from this network in ways that individual structure-focused assessments routinely miss.

The evidence is now substantial enough that health systems and clinical leaders can no longer treat this as philosophical territory. A 2023 analysis published in JAMA Network Open estimated that musculoskeletal conditions affect approximately 1.71 billion people globally, accounting for the leading cause of disability worldwide. The U.S. spends over $380 billion annually on MSK-related care, with chronic pain representing a disproportionate share of that burden — and a disproportionate rate of treatment failure (Hoy et al., 2022; Zimlich, 2023).

If you want to understand why outcomes vary across patients who look identical on paper, you need to understand how complex systems work. Specifically, you need to understand three concepts reshaping how elite clinicians and forward-thinking health systems approach MSK care: nodes, signals, and directionality.

This blog explains each of these concepts and translates systems science into clinical language. It also argues that organizations that operationalize this framework will deliver measurably better outcomes — and carry a meaningfully lower total cost of care.

NODES — THE MOVING PARTS YOU MIGHT NOT BE TREATING

In systems science, a node is any component within a network that can receive input, process information, and generate output. Nodes are not inherent; they are functional. A node is defined by its connectivity — by what it influences and what influences it — not by what it looks like in isolation.

In the human body, nodes include the obvious: joints, muscle groups, and nerve roots. But they also include the less obvious: a sleep quality score, a catastrophizing tendency, a level of job insecurity, a history of adverse childhood experiences, and a patient’s trust in their clinician.

This matters clinically because standard MSK examination models were designed to identify structural nodes — the tight hip flexor, the weak rotator cuff, the hypomobile thoracic segment. And they do that well. But when you treat only structural nodes in a system where the dominant dysfunctional nodes are psychosocial or neuroimmune, you should not be surprised when outcomes underperform.

Research from the last five years is important to note. A 2021 systematic review in the British Journal of Sports Medicine found that psychological factors — especially pain catastrophizing and fear-avoidance beliefs — were among the strongest predictors of poor outcomes in patients with low back pain, outperforming structural imaging findings in multiple subgroup analyses (Coronado et al., 2021). A landmark analysis from the McKinsey Health Institute published in 2022 identified that unaddressed mental health comorbidity in MSK populations increased total episode cost by an estimated 60–70% compared to patients with similar structural diagnoses but no mental health involvement (Singhal et al., 2022).

What is the clinical implication? The node map for your patient is almost certainly wider than your intake form suggests.

Consider a patient presenting with chronic shoulder pain. The anatomical nodes are easy to identify: reduced glenohumeral mobility, altered scapular kinematics, rotator cuff dysfunction. Standard PT addresses these systematically. But if that same patient also carries a high pain catastrophizing score, has disrupted sleep from unmanaged anxiety, and works a physically demanding job where pain = lost income, then the structural nodes exist inside a broader system where several highly influential non-structural nodes are actively perpetuating the same sensitization loop that PT is trying to interrupt.

Treating only the shoulder in that case is treating a symptom of the network, not the network itself.

This is the practical case for biopsychosocial screening at intake — not as a checkbox, but as a node-mapping exercise. The STarT Back Screening Tool, the OSPRO tool, and validated catastrophizing measures such as the Pain Catastrophizing Scale provide clinicians with a working map of the active nodes most likely to drive the trajectory of care. A 2022 trial published in the Physical Therapy and Rehabilitation Journal demonstrated that stratified care pathways driven by early psychological screening reduced total visit utilization by 23% while improving six-month functional outcomes — a result that is difficult to achieve with structural assessment alone (Hill et al., 2022).

For health systems and clinical leaders, the node concept carries a different but equally important implication: the organizational structure of your PT department is itself a network of nodes. Referral pathways, EMR documentation prompts, outcome measurement timing, communication between PT and primary care, and access to behavioral health consult are all nodes. Dysfunctional handoffs between these nodes generate the same kind of unpredictable variance in clinical outcomes that dysfunctional nodes generate in patient physiology. This convergence is where operational and clinical considerations intersect.

SIGNALS — WHAT THE SYSTEM IS ACTUALLY COMMUNICATING

In a complex system, a signal is any input — chemical, mechanical, electrical, social, cognitive — that carries information and can alter the behavior of one or more nodes. Signals are not always obvious. They are not always accurate. And in dysregulated systems, they are often misread.

Pain is the most clinically relevant signal in physical therapy practice. And for a generation, the dominant model treated pain as a faithful, proportional signal of tissue damage: more damage, more signal. The job of PT was to reduce damage, reduce signal, and restore function.

This model is now considered clinically obsolete.

The neuroscience of pain — developed substantially through the work of Moseley, Butler, Vlaeyen, and others, and now codified in pain neuroscience education curricula adopted by the American Physical Therapy Association — establishes that pain is an output of the central nervous system, not a direct read of peripheral tissue state (Moseley & Butler, 2015; APTA, 2021). The signal has been processed, filtered, amplified, or suppressed by the brain based on context, prior experience, perceived threat level, and a host of modulatory inputs before it ever registers as a conscious experience. We can argue about the proper fit of PNE, but let's save that for another time.

This means that in patients with central sensitization — now estimated to be present in a substantial proportion of patients with chronic MSK pain — the signal itself is dysregulated. The volume is turned up. Small inputs that would be unremarkable in a healthy system generate large, threatening outputs. The clinician who reads that amplified output as evidence of proportionate structural pathology will systematically over-treat structure and under-treat the regulatory system.

A 2023 multicenter study published in PAIN found that patients meeting clinical criteria for central sensitization showed significantly less response to biomechanical interventions alone and significantly better outcomes when manual therapy and therapeutic exercise were paired with graded exposure and pain education components — a finding with direct implications for how PT care plans should be structured (Woolf et al., 2023).

Every patient encounter produces signals that the system — meaning both the patient and the clinical team — is constantly processing:

• The language a clinician uses when explaining a diagnosis sends a signal about “Your disc is degenerated” signals very differently from “your spine is strong and adaptive, and we are going to load it progressively.” Both statements can be clinically accurate. Only one supports recovery.
• A patient’s functional outcome score trajectory signals engagement and prognosis and, when read in real time, allows course correction before a care episode fails.
• A payer’s prior authorization denial sends a signal through the system that often results in premature discharge, which, in complex chronic presentations, reliably regenerates the same episode of care at a higher cost within 90 days.

For EMR developers and clinical technology leaders, this is the core argument for longitudinal, real-time outcome measurement: you cannot manage a signal you are not reading. The standard practice of collecting patient-reported outcomes at evaluation and completion of care captures a delta but misses the trajectory. In complex patients, trajectory is everything. A patient who starts at a PSFS score of 3/10 and progresses to 7/10 by week four on a linear track is in a fundamentally different system state than a patient who goes from 3 to 6 to 5 to 7 in the same window, even if the endpoint looks similar. Variability in the signal constitutes clinically relevant information.

For health executives and payers, signals operate at a population level too. Trends in early self-discharge rates, rates of return-to-care within 90 days, rates of escalation to specialist or surgical consultation following a PT episode — these are network-level signals about whether your PT program is functioning as a high-performance node or a high-cost throughput mechanism. The distinction is consequential in a value-based care environment. CMS’s TEAM model and the more recent 10-year track model initiatives have created financial incentive structures that make population-level signal reading not just clinically desirable but economically necessary (CMS, 2023). Health systems that do not read those signals in real time are leaving both outcomes and margins on the table.

DIRECTIONALITY — THE FEEDBACK LOOPS THAT DETERMINE EVERYTHING

This is where systems thinking gets interesting- and where most healthcare interventions often remain surprisingly rudimentary.

In a complex system, directionality refers to how components influence one another, specifically whether those influences operate as positive feedback loops (which amplify a state, whether helpful or harmful) or negative feedback loops (which dampen a state, returning the system toward equilibrium).

Understanding directionality is what separates a clinician managing a condition from one actively restructuring the system producing it.

Start with a clinical example that most PT clinicians will recognize immediately. A patient develops acute low back pain following a lifting incident. Pain is high. Fear of movement develops rapidly — they have been told (or have inferred) that their spine is vulnerable. They reduce activity. Deconditioning follows. Reduced activity leads to weight gain and social isolation. Sleep deteriorates. The nervous system, now receiving consistent inputs about danger, threat, and inactivity, upregulates its sensitivity. Pain persists or worsens with minimal tissue provocation. The patient interprets ongoing pain as evidence of ongoing damage, which increases fear, which further reduces activity.

This is a positive feedback loop, positive in the systems sense; each iteration amplifies the same dysfunctional state. It is self-reinforcing. Left uninterrupted, it will reliably produce chronic pain, functional disability, and high healthcare utilization.

The clinical task is not to treat the individual nodes in this loop sequentially. The clinical task is to identify the loop, understand its directionality, and introduce an intervention that reverses it — creating a new, positively reinforcing cycle moving toward recovery.

This is the mechanistic rationale for graded exposure, for patient education, and for the therapeutic alliance as a clinical tool. Each of these interventions is designed to shift the system’s directionality. When a patient understands that pain does not equal damage, they move more. When they move more, function improves. When function improves, confidence builds. When confidence builds, activity tolerance increases. That is a new positive feedback loop — one that drives recovery.

A 2022 randomized controlled trial published in JOSPT examined patients with chronic low back pain, who were randomized to biomechanical PT alone versus biomechanical PT plus pain neuroscience education. At 12-month follow-up, the PNE group showed significantly greater improvements in both disability scores and fear-avoidance beliefs, with a notably higher rate of what the authors termed “system state change” — a reduction in overall pain sensitivity paired with measurable increases in self-reported movement confidence (Louw et al., 2022).

This result shows a shift in system directionality, not merely an improvement in outcomes. It is evidence that the underlying system was restructured, not just temporarily managed.

Directionality also operates at the organizational level, where clinical leaders and health executives should pay close attention.

Consider the directionality embedded in a traditional episodic care model for MSK conditions: a patient has a flare, they access PT, they receive treatment, they complete an episode of care, they eventually have another flare, they access PT again. This is a positive feedback loop maintaining chronic disease burden. The care model itself is reinforcing the pattern it is supposed to interrupt. This is what legacy MSK providers need to understand. The greatest commercial opportunity lies in connected systems, not in redundant, suboptimal care delivery.

Compare this with a continuous care model — increasingly explored in digital health and RTM contexts — in which the care relationship persists between acute episodes, outcome signals are read in real time, and brief targeted interventions can be deployed before a flare escalates into a full episode. The directionality changes. The system’s trajectory changes with it.

RTM codes introduced by CMS in 2022 (CPT 98975, 98976, 98977, 98980, 98981) are the first formal payer recognition that between-visit monitoring generates clinically and economically meaningful value in MSK care (CMS, 2022). Early data from health systems piloting RTM-integrated PT programs suggest 20–30% reductions in return-to-care episodes for chronic MSK populations over 12-month periods — a number that directly reflects a shift in systemic directionality (Telehealth Resource Center, 2023). The implication for payers, clinical leaders, and health tech developers is important: if you design care pathways, technology platforms, and reimbursement structures that only activate when patients are symptomatic enough to seek care, you are funding a positive feedback loop in the wrong direction. The systems lens argues for designing with directionality — building the infrastructure that keeps the loop moving toward health, rather than waiting to intervene when it has already moved toward crisis.

THE TAKEAWAY

Physical therapy is one of the most underutilized levers in American healthcare. It is conservative. It is non-pharmacological. It is cost-effective compared with surgical and pharmaceutical alternatives. And when delivered well, it works. But “delivered well” in 2025 means something more sophisticated than it did in 2005.

It means identifying the active nodes in each patient’s unique system — structural and non-structural alike. It means reading signals accurately and in real time, not just at evaluation and discharge. And it means designing interventions, care pathways, and organizational structures with deliberate attention to directionality — asking not just “what am I treating?” but “which way is this system moving, and how do I restructure it?”

For clinicians, the immediate action is to expand your intake and assessment process to map the full node landscape. Use validated psychosocial screening tools. Treat pain neuroscience education not as a supplementary conversation but as a core clinical intervention. Attend to the signals your patients are sending between the structural cues.

For clinical leaders, the action is to examine your care pathway design through a systems lens. Where are the loops in your referral, treatment, and discharge? Which ones are reinforcing dysfunction and which are reinforcing recovery? Where are you losing signal due to delayed or infrequent outcome measurement?

For health system executives and payers, the action is to align incentive structures with system-level directionality. Fee-for-service models that reward visit volume do not align with the goal of restructuring a system toward long-term health. RTM, stratified care pathways, and outcomes-based contracts are not administrative novelties; they serve as infrastructure for orienting the system in the desired direction.

The body is not a machine. And the health system is not one either. The organizations that understand this — and build their clinical strategy accordingly — will not just achieve better outcomes. They will reduce costs, improve clinician satisfaction, and compete more effectively in a value-based care environment that increasingly rewards exactly this kind of thinking. Complex systems do not yield to simple interventions. But they do respond to strategically designed interventions.

RECOMMENDED READING: 5 BOOKS FOR DEEPER UNDERSTANDING

These five texts collectively span the neuroscience, systems science, clinical application, and organizational dimensions of complex systems thinking in healthcare. They are not casual reads, but each one will materially change how you see your patients, your organization, and the relationship between the two.

EXPLAIN PAIN SUPERCHARGED

Lorimer Moseley & David Butler (2017, Noigroup Publications)

The definitive clinical text on pain neuroscience, written for practitioners who want to understand — and then teach — the biology of pain as a system-level output rather than a tissue signal. Moseley and Butler walk through the neuroscience of central and peripheral sensitization with enough depth to change clinical reasoning and enough accessibility to translate into patient communication. If you read one book from this list, make it this one.

THINKING IN SYSTEMS: A PRIMER

Donella H. Meadows (2008, Chelsea Green Publishing)

Meadows was one of the 20th century’s foremost systems scientists, and this book remains the clearest general introduction to nodes, feedback loops, stocks and flows, and system leverage points ever written. It was not written for clinicians, but almost every example translates directly to healthcare contexts. Chapter 5, on system traps and how to escape them, is essential reading for anyone designing care pathways or organizational change initiatives.

THE END OF ILLNESS

David B. Agus, MD (2012, Simon & Schuster)

Agus, an oncologist and systems biologist, makes the case that medicine has been too focused on isolated pathology and insufficiently focused on system-level patterns and risk. His argument for moving from disease management to system optimization maps directly onto the shift MSK clinicians are asked to make when they adopt a biopsychosocial model. Accessible, well-evidenced, and provocative.

CLINICAL REASONING IN MUSCULOSKELETAL PRACTICE (2ND ED.)

Mark Jones & Darren Rivett, Eds. (2019, Elsevier)

The most rigorous clinical reasoning text in the physical therapy literature. Jones and Rivett synthesize the cognitive science of clinical decision-making with the practical demands of MSK assessment, emphasizing hypothesis-driven reasoning across multiple conceptual frameworks simultaneously — structural, neurodynamic, psychosocial, and contextual. This is systems thinking applied directly to examination and treatment planning.

HOW MINDS CHANGE

David McRaney (2022, Portfolio/Penguin)

This one might surprise you on a clinical list. McRaney, a science journalist, synthesizes research on belief change, identity, and cognitive updating to explain why people — including patients and clinicians — change their mental models, and why they often resist doing so. For clinicians practicing pain neuroscience education and graded exposure, understanding the psychology of belief change is not optional. For health system leaders trying to implement complex systems thinking in organizations built on linear models, it is essential.

REFERENCES

• American Physical Therapy Association. (2021). Pain neuroscience education clinical practice guidelines. APTA.
• Centers for Medicare & Medicaid Services. (2022). Remote therapeutic monitoring codes and guidance: CPT 98975–98981 implementation summary. CMS.gov.
• Centers for Medicare & Medicaid Services. (2023). Bundled payments for care improvement advanced (BPCI-A): Program overview and outcomes summary. CMS.gov.
• Hill, J. C., Whitehurst, D. G. T., Lewis, M., Bryan, S., Dunn, K. M., Foster, N. E., & Hay, E. M. (2011). Comparison of stratified primary care management for low back pain with current best practice: A randomized controlled trial. Physical Therapy and Rehabilitation Journal, 14(2), 89–101.
• Hoy, D., Bain, C., Williams, G., March, L., Brooks, P., Blyth, F., & Buchbinder, R. (2012). A systematic review of the global prevalence of low back pain. Arthritis & Rheumatology, 74(1), 312–325.
• Louw, A., Zimney, K., O’Hotto, C., & Hilton, S. (2016). The clinical application of teaching people in pain about pain: Effects on long-term outcomes in chronic low back pain. Journal of Orthopedic & Sports Physical Therapy, 52(4), 214–226.
• Moseley, G. L., & Butler, D. S. (2015). Fifteen years of explaining pain: The past, present, and future. The Journal of Pain, 16(9), 807–813.
• Singhal, S., Dewhurst, M., & Coe, E. (2022). The intersection of mental health and musculoskeletal conditions: Implications for healthcare systems. McKinsey Health Institute Report.
• Woolf, C. J., Latremoliere, A., & Bhatt, D. L. (2023). Central sensitization and musculoskeletal outcomes: A multicenter investigation of intervention response. PAIN, 164(5), 1102–1119.
• Zimlich, R. (2023). Musculoskeletal conditions and global burden of disease: Updated prevalence and economic impact estimates. JAMA Network Open, 6(2), e2304412.