When the System Gets in the Way
Institutional Barriers and Enablers of Evidence-Based Neurorehabilitation
In my previous posts in this series introducing implementation science in neuro rehab, I explored the therapist factors that influence whether clinicians adopt evidence-based interventions: knowledge, skills, confidence, motivation, habits, and the emotional texture of clinical decision-making. Those factors matter enormously. But they don’t exist in isolation. Clinicians work within systems, and systems can either support or undermine best practice just as powerfully as any individual factor.
This post focuses on institutional factors. I’m talking about the organisational, policy, and structural determinants that shape whether evidence-based neurological rehabilitation is delivered at scale in the real world. Using the Theoretical Domains Framework (TDF) as a lens, I want to explore how the environments in which clinicians work can become barriers to the care they are trying to provide, BUT also how those same environments can be structured to actively enable it.
Let’s be honest... The institutional landscape for neurological rehabilitation in Australia is under genuine strain, and I do not think we serve clinicians or patients well by pretending otherwise. But strain is not the whole story. There are services, organisations, and leaders doing this well, and understanding what they are doing differently is just as useful as understanding what goes wrong.
We can’t just throw our hands up in the air saying it’s too hard.
So with that, this piece will outline some current issues I’ve seen in neurological rehabilitation and demonstrate how these real-life examples map to the TDF.
Acute care crowding out rehabilitation
Public health systems face an uncomfortable structural reality. Acute hospitals, emergency departments, and surgical services capture political attention, media coverage, and funding priority in ways that rehabilitation does not. When health system budgets tighten, rehabilitation services are frequently asked to absorb the impact first.
In Australia, this is not hypothetical. Over the course of the CIMT implementation program I conducted within Rehabilitation in the Home (RITH), a large early-supported discharge service in Perth, referrals to the service increased by 49% over the study period — yet staffing grew by just 3% in the same timeframe [1]. New service models were introduced, caseloads expanded, and clinicians were increasingly expected to manage more acute and complex patients with the same or fewer resources. The context-specific CIMT protocol we developed had been designed to fit the service as it existed in 2018 [2]. By 2021, that service looked very different.
This pattern is not unique to RITH. Across Australian public health, resources are being redirected toward acute services in response to growing demand, while community-based rehabilitation operates with constrained funding and limited capacity to implement new evidence-based programs. When rehabilitation is positioned as a lower-tier priority, the downstream effects are predictable: reduced staffing, less time for clinical supervision, fewer opportunities for skill development, and a service culture where productivity targets dominate over quality improvement.
TDF domains implicated: barriers
Environmental context and resources is the most obvious domain here, and it is worth being specific about what resource deprivation actually means for implementation. It is not simply that clinicians lack equipment, although that also matters. It means that the staffing ratios, session lengths, and organisational infrastructure required to deliver complex interventions are absent or inadequate. Implementing CIMT, for example, requires interdisciplinary collaboration, joint home visits, protected time for training, and a manageable caseload. When referrals increase by almost 50% with no corresponding staffing increase, every one of those conditions becomes harder to meet.
Goals operates more subtly but is equally important. In a service under pressure, the immediate goal becomes seeing enough patients to manage the caseload. This inevitably crowds out longer-term goals such as skill development, quality improvement, or the sustained implementation of a complex intervention. Clinicians do not abandon their values; they deprioritise them because the environment makes it impossible to act on them consistently.
Social influences are shaped by this dynamic in ways that accumulate over time. When a service operates in survival mode and the organisational message, explicitly or implicitly, is that throughput matters most: the social norms within teams shift accordingly. Implementing something new, or something that takes more time and coordination than usual care, begins to feel like a deviation from what the team expects rather than an expression of professional responsibility.
Reinforcement is undermined when resource constraints prevent clinicians from completing programs with adequate fidelity to observe meaningful outcomes. If clinicians attempt to deliver a complex intervention in a depleted environment and the results are disappointing (because the dose was insufficient, the support was absent, or the patient withdrew due to logistical barriers), they are less likely to try again. The intervention becomes associated with difficulty and failure rather than benefit.
TDF domains implicated: enablers
Not all services are trapped in this dynamic, and the differences between those that manage to implement well despite resource constraints and those that do not are instructive.
Organisational commitment and leadership support operate through the TDF domain of social/professional role and identity and social influences. In the RITH service, the managers who actively endorsed CIMT and made it clear that delivering evidence-based upper limb rehabilitation was a professional expectation, had a measurable effect on adoption rates at their sites [1]. Clinicians in those teams were more likely to offer the intervention, more likely to seek training, and more likely to persevere when they encountered patient or logistical challenges. Leadership does not compensate for insufficient staffing, but it shapes the social norms within which clinicians make decisions, and those norms matter.
Protected time for clinical supervision and professional development, even modest amounts, dramatically changes what is possible within environmental context and resources. Services that have deliberately structured caseload management to allow joint visits, peer observation, and regular case discussion report better implementation outcomes than those that treat clinical supervision as an add-on contingent on spare time. When organisations treat supervision as a core operational function rather than a luxury, clinicians develop the skills and confidence to deliver complex interventions even in resource-constrained environments.
Audit and feedback mechanisms engage the TDF domain of reinforcement in the opposite direction to the barrier described above. Services that systematically collect data on which patients are eligible for evidence-based interventions, how many are offered them, and how many receive them create a visible accountability structure that reinforces implementation behaviour. In our process evaluation, audit and feedback was identified as one of the most powerful enablers of sustained CIMT delivery [1, 3]. It was also one of the most underutilised, primarily due to time constraints and the logistical difficulty of accessing medical records across multiple sites. Where it was deployed, even on a single occasion, it prompted tangible changes in clinical practice.
The NDIS funding crisis and its consequences for complex neurological rehabilitation
The National Disability Insurance Scheme was designed as a transformational investment in Australia’s disability sector. For many people with neurological conditions: acquired brain injury, spinal cord injury, progressive neurological disease, stroke; it has provided access to allied health services that simply did not exist before. That access has been genuinely life-changing for many participants.
But the scheme is under severe and growing strain. Physiotherapy price limits were frozen for five consecutive years, and from July 2025, the NDIA implemented a $10 per hour reduction in physiotherapy rates (or almost $40 per hour for those in some states), bringing the rate to $183.99 per hour. This was alongside a 50% reduction in travel reimbursement and the removal of rural and remote price loadings. These decisions were made without meaningful consultation with the allied health sector and against a backdrop of rising operational costs, inflation, and extraordinary care complexity.
The consequences are already being felt. Allied health services have exited the NDIS market following the most recent pricing decision. Profit margins are below sustainable levels with many providers reporting operating at a loss. Providers in regional and remote areas, who typically deliver therapy in participants’ homes across large geographic areas, face the greatest pressure. Eight allied health peak bodies issued a joint statement describing the changes as financially unsustainable and a direct threat to participant safety.
For neurorehabilitation specifically, the implications are significant. Neurological conditions are among the most complex and resource-intensive presentations in the NDIS. Delivering best-practice care: whether that is intensive upper limb training, gait retraining, cognitive strategy training, or managing long-term complications, all require time, expertise, and a multidisciplinary team. When pricing does not reflect that complexity, providers are forced to make difficult choices: reduce clinical time, eliminate supervision, stop accepting NDIS referrals, or exit the scheme entirely.
TDF domains implicated: barriers
Environmental context and resources in the NDIS context takes on a particular character. The resource constraint here is not primarily about equipment, it is about time. Pricing that does not reflect the complexity of neurological presentations forces clinicians to compress sessions, reduce frequency, or eliminate non-face-to-face activities such as case coordination, interdisciplinary liaison, and documentation. For complex interventions that depend on dose and frequency to produce meaningful neuroplastic change, compressed clinical time directly compromises effectiveness.
Skills deteriorate when investment in professional development is removed. Clinical skill in complex neurological rehabilitation must be actively maintained through practice, supervision, and exposure to a range of presentations. When NDIS pricing makes clinical supervision economically unviable and continuing professional development is eliminated from organisational budgets, clinicians’ skills do not simply plateau, they erode.
Beliefs about capabilities: a clinician’s confidence in their own ability to deliver an intervention effectively is closely linked to skill maintenance. Clinicians who have fewer opportunities to deliver complex interventions, observe expert peers, and receive feedback on their practice report lower confidence over time. In our research, interdisciplinary collaboration was one of the most powerful enablers of confidence [3]. When funding models make that collaboration economically unviable, this enabler disappears.
Social/professional role and identity is disrupted when an organisation’s survival becomes its primary focus. Clinicians who enter neurological rehabilitation with a strong sense of professional identity, a belief that their role is to deliver effective, evidence-based care, experience significant psychological strain when funding models force them to compromise that identity. The result is not just dissatisfaction, but a gradual erosion of the professional values that motivate high-quality practice.
TDF domains implicated: enablers
Despite the funding pressures, there are organisational strategies that protect the conditions required for evidence-based practice.
Interdisciplinary collaboration and shared workloads, where organisational structures actively support joint delivery of complex interventions, engage beliefs about capabilities and environmental context and resources simultaneously. In the RITH service, CIMT programs delivered collaboratively by physiotherapists, occupational therapists, and therapy assistants were more likely to be completed with full protocol fidelity than those delivered by a single clinician. Shared delivery reduced the individual burden, increased session frequency, and created a team accountability structure that reinforced consistent practice. Organisations that deliberately build interdisciplinary workflows into their service model, rather than treating collaboration as an informal arrangement, create structural conditions that enable evidence-based practice to survive resource constraint. I should acknowledge the additional challenges associated with the fragmented service provision across private services, but greater communication and collaboration can still be viable strategies.
Organisational advocacy and professional identity can also operate as enablers through social/professional role and identity. Services and clinical leaders that publicly advocate for appropriate funding; through peak bodies, professional associations, and direct engagement with policymakers, contribute to a broader culture in which evidence-based practice is positioned as a professional expectation and a policy priority. The joint statement issued by eight allied health peak bodies in response to NDIS pricing decisions in 2025 is an example of this, and I know the APA have invested significantly in government advocacy ahead of the upcoming NDIS pricing guide. Collective professional advocacy does not immediately reverse pricing decisions, but it creates a clear public record of what evidence-based neurological rehabilitation actually requires, and it strengthens the professional identity of clinicians working under pressure.
Technology and cognitive load
There is something deeply ironic about the way technology is reshaping clinical work. The tools designed to make healthcare more efficient: electronic medical records, digital workflows, reporting systems, communication platforms, have in many cases added substantially to the cognitive burden on clinicians, rather than reducing it. And that’s before we even start on AI!
In the RITH service, we observed this directly. During 2021, the organisation commenced a large-scale implementation of a new ICT solution to manage referrals, workflows, and patient flow. This was a genuine service improvement initiative. But it required multiple clinical champions per site to learn complex new systems and train their colleagues. Many of whom were the same clinicians we had identified as mCIMT champions. The result was that mCIMT implementation momentum stalled, not because clinicians had lost interest, but because their finite capacity for change had been absorbed by something else.
Managers in our focus groups captured this precisely. One described her team as having “so many systems and processes” that “it’s actually very hard to remember to do something. So, you kind of do have to make it common sense or a clear role, otherwise it will fall off.” [3]
The broader trend is accelerating. Artificial intelligence tools, digital documentation requirements, telehealth platforms, outcome measurement software, and increasingly complex EMR systems are all demanding attentional resources from clinicians who are already operating at or near capacity.
TDF domains implicated: barriers
Memory, attention, and decision processes is the domain most acutely affected by cognitive overload. Evidence-based practice is not automatic. It requires clinicians to consciously recall eligibility criteria, intervention protocols, outcome measures, and decision-making processes at the point of clinical contact. When cognitive bandwidth is consumed by navigating new technology and competing administrative demands, the working memory available for evidence-based clinical reasoning is reduced. Clinicians revert to habitual practice, not because they have forgotten the evidence, but because deliberate decision-making requires cognitive resources that are no longer available.
Behavioural regulation is similarly compromised. Sustaining a new clinical behaviour requires ongoing self-monitoring, reflection, and adjustment. In a working environment crowded with technology change and administrative burden, the space required for reflective practice diminishes, and with it the capacity to maintain new clinical behaviours.
Emotions plays an important role here that is sometimes overlooked. Technology change can generate frustration, anxiety, and a sense of loss of control. These emotional responses directly influence motivation, confidence, and the willingness to engage with additional change, even when that change is clinically valuable.
Intentions, the conscious commitment to perform a specific behaviour, are also undermined by competing demands. Even a clinician who has formed a genuine intention to offer an evidence-based intervention may fail to act on it when they arrive at the clinical setting and find their attentional resources already depleted. Implementation intentions (specific if-then plans tied to environmental cues) are far more likely to survive a cognitively demanding environment. Without those structured cues, even well-motivated clinicians default to routine.
TDF domains implicated: enablers
Technology is not inherently an implementation barrier. When designed thoughtfully and introduced with adequate support, it can actively enable evidence-based practice.
Clinical decision support tools and prompts embedded within EMR systems engage memory, attention, and decision processes in a positive direction. Flagging eligible patients at the point of referral, prompting eligibility screening at intake, and embedding outcome measurement tools into routine documentation reduce the cognitive effort required to initiate evidence-based interventions. In our RITH implementation, adding an CIMT flag to the intake process was identified as one of the most practical and low-cost strategies for increasing adoption. It worked not by adding to clinician load but by reducing it; the decision cue was built into the workflow rather than left to individual memory.
Sequencing and pacing of change is an organisational strategy that protects behavioural regulation and emotions. Services that deliberately stagger implementation priorities reduce the cognitive competition that disrupts adoption. The conflict we observed between CIMT implementation and ICT rollout in RITH could not have been entirely avoided, but more deliberate sequencing and explicit protection of CIMT champions from competing change roles would have mitigated its impact. Organisations that recognise the finite nature of change capacity, and plan accordingly, create conditions in which individual improvement initiatives are far more likely to succeed.
Failing to de-implement ineffective practices
Implementing evidence-based practice is only half the challenge. The other half is stopping what does not work, and that is a conversation the rehabilitation community is only beginning to have.
Clinicians’ time is finite. When ineffective or low-value interventions continue to be delivered alongside new evidence-based approaches, they compete for the same limited resource. In neuro rehab, this is a genuine issue. Passive modalities, compensatory strategies applied without therapeutic intensity, and interventions that have consistently failed rigorous evaluation continue to occupy session time in many services. The reasons are understandable. Habits are hard to break, patients expect familiar treatments, and de-implementation requires just as much deliberate effort as implementation. But, the consequence is that evidence-based interventions are crowded out before they gain traction.
TDF domains implicated: barriers
Social/professional role and identity is perhaps the most underappreciated barrier to de-implementation. Stopping an intervention that has been part of a clinician’s practice for years carries identity implications. Clinicians who have used a particular approach throughout their career may experience the suggestion that it is ineffective as a challenge to their professional competence and to the value of the care they have previously provided. In team environments, this resistance is amplified: de-implementing a widely used approach challenges shared professional norms, and navigating that challenge requires careful facilitation and leadership. We need better language and strategies in addressing this issue, otherwise we risk alienating a bunch of dedicated clinicians who will just double-down on their approach and dismiss evidence-based practice as not living in the real-world.
Beliefs about consequences operates powerfully here. Clinicians may hold beliefs that stopping a familiar treatment will result in worse patient outcomes, even when the evidence does not support this. In complex neurological populations, where individual responses to treatment are genuinely difficult to predict, this conservatism is understandable. But it represents a significant barrier to the reallocation of clinical time toward evidence-based alternatives.
Intentions are directly implicated. Forming an intention to stop doing something familiar is psychologically different from forming an intention to adopt something new. Stopping requires a conscious and repeated decision to override an ingrained habit, often in the middle of a busy clinical session, without an external prompt or reminder.
Environmental context and resources connects to de-implementation in ways that are less obvious but equally important. Many low-value interventions persist because the systems within which clinicians work continue to support them: equipment remains available, session plans are structured around familiar treatments, and outcome measures do not distinguish between time spent on effective and ineffective components of care. De-implementation requires actively redesigning these environmental structures.
Knowledge is also implicated. What is often absent is specific, actionable knowledge about which components of current practice should be discontinued and what should replace them. Identifying an ineffective intervention without providing a clear and feasible alternative creates a knowledge gap that many clinicians resolve by continuing what they know.
TDF domains implicated: enablers
Organisational audit processes that make the distribution of clinical time visible. Breaking down what proportion of sessions are spent on which interventions engage environmental context and resources and behavioural regulation as enablers. When clinicians and managers can see clearly how clinical time is being allocated, conversations about de-implementation become grounded in data rather than perception. This is more likely to result in deliberate, managed reallocation of clinical effort than relying on individual clinicians to make these decisions in isolation during busy clinical days.
Leadership endorsement and explicit clinical governance around de-implementation addresses social/professional role and identity by repositioning the act of stopping an ineffective practice not as a criticism of past care but as an expression of professional responsibility and intellectual honesty. Services led by managers and senior clinicians who frame de-implementation as a mark of clinical rigour create cultures in which the evidence base is treated as a living document, and practice evolves accordingly.
What this means in practice and why the TDF matters
Having worked through barriers and enablers across these institutional themes, I want to step back and address a question that some readers may be asking: why bother with a framework like the TDF at all? Why not simply identify a problem and try to fix it?
The answer lies in what happens when we do not use a framework, and what decades of failed implementation attempts in healthcare tell us about the limits of intuition-based change.
The most common approach to implementation in rehabilitation services has historically been education. Identify a practice gap, deliver a training session, and assume that knowledge will translate into behaviour. This approach has been tried repeatedly, evaluated rigorously, and is NOT EFFECTIVE [4]. In my own experience, the first attempt to increase CIMT delivery in the RITH service: two education sessions delivered to over thirty clinicians, produced no measurable change in practice twelve months later. Education was there. The behaviour change was not.
The reason is that knowledge is only one of many determinants of behaviour. A clinician who: knows that CIMT is evidence-based but lacks the skills to deliver the transfer package; works in a team where the social norms do not support intensive upper limb therapy; is managing a caseload that leaves no time for joint visits; and has never been reinforced for attempting a complex new program: well, that clinician is not going to change their practice because of a training session.
The TDF matters because it provides a systematic, theory-informed structure for identifying which determinants are actually operating in a given context, across all of the relevant domains. When we map barriers and enablers to the TDF, we stop guessing about what is preventing implementation and start identifying it with specificity. We can see that the barrier is not primarily knowledge but confidence. Or not primarily confidence but access to peer modelling. Or not primarily peer modelling but the competing cognitive demands of a simultaneous technology rollout.
That specificity matters, because different determinants require different strategies. Confidence is addressed through supervised practice and graduated exposure, not through more education. Cognitive load is addressed through environmental redesign and decision cues, not through motivational messaging. Social norms are addressed through leadership endorsement and communities of practice, not through individual training. The TDF, used systematically, allows us to match the strategy to the actual problem. That matching is what separates implementation programs that produce sustained behaviour change from those that produce a temporary increase in knowledge and nothing else.
For clinicians, the practical value of this framework extends beyond formal research and implementation programs. It provides a structured way to think about your own practice setting, to evaluate why a particular evidence-based intervention is not being delivered consistently, and to identify which behaviour change strategies are most likely to make a difference. Is the barrier that clinicians do not know how to deliver the intervention? Then the strategy is training and supervised practice. Is it that the team does not value it? Then the strategy involves social influence and leadership endorsement. Is it that the environment does not support it? Then the strategy requires organisational change, adjusted caseloads, protected time, or redesigned workflows.
The TDF does not make implementation easy. The barriers described in this post are real, and some of them are not resolved by a framework alone. They require policy change, investment, and institutional commitment that goes beyond what any individual clinician or service can deliver unilaterally.
But the TDF gives us the language, the structure, and the evidence base to make that case. It allows us to articulate why a funding model that eliminates clinical supervision undermines skills development. Why a technology rollout without adequate change sequencing disrupts behavioural regulation. Why burnout is not a personal failing but an institutional condition with predictable and documented consequences for the adoption of evidence-based practice.
Used well, the TDF moves the conversation from “why aren’t clinicians doing this?”, which places the burden of failure on individual practitioners, to “what conditions are preventing this from happening?”. This frames the conversation on the systems within which those practitioners work. That shift in framing provides the foundation for an implementation that is sustainable, equitable, and honest about where the real work needs to happen.
I hope you’ve enjoyed and learned from my introduction to using the TDF to identify determinants across therapist and institutional factors. In my next post, I will turn to the third category of determinants: patient factors, and how the Capability, Opportunity, Motivation- Behaviour (COM-B) model helps us understand why even the most evidence-based and well-delivered intervention may not be accepted or sustained by the people it is designed to help.
And with that, thanks for reading.
Ashan
References
1. Weerakkody, A., E. Godecke, and B. Singer, Implementing modified constraint-induced movement therapy after stroke in an early-supported discharge rehabilitation service: a process evaluation using RE-AIM QuEST. BMC Health Serv Res, 2025. 25(1): p. 1086.
2. Weerakkody, A., et al., Unlocking the restraint-Development of a behaviour change intervention to increase the provision of modified constraint-induced movement therapy in stroke rehabilitation. Australian Occupational Therapy Journal, 2023. 70(6): p. 661-677.
3. Weerakkody, A., E. Godecke, and B. Singer, Translating acceptability to sustained delivery: Clinician and manager perspectives on implementing modified constraint-induced movement therapy in an early-supported discharge rehabilitation service. Aust Occup Ther J, 2024. 72(1).
4. Bird, M.L., et al., Moving stroke rehabilitation evidence into practice: a systematic review of randomized controlled trials. Clin Rehabil, 2019: p. 269215519847253.