Highlights
In the NeTKoH stepped-wedge cluster randomized clinical trial, neurological teleconsultations embedded in rural general practice did not increase the proportion of patients managed exclusively by general practitioners.
Compared with standard care, teleconsultations reduced referrals to neurologists but increased referrals to other specialists and hospitals, consistent with a triage-enhancing rather than care-substituting effect.
Patient-reported quality of life and health status did not differ significantly between groups over 3 months, and adherence to recommendations was lower during the intervention phase.
The findings are clinically important for health systems serving structurally underserved regions: teleneurology may improve the direction of care pathways even when it does not immediately reduce specialist-sector utilization.
Background
Access to neurological care is a persistent problem in rural and underserved regions. Neurological symptoms such as new headache syndromes, dizziness, sensory disturbances, tremor, gait disorder, suspected seizure, cognitive complaints, or focal deficits frequently present first in primary care, yet the availability of office-based neurologists is often limited. Delayed specialty access may prolong diagnostic uncertainty, defer time-sensitive workup, and increase potentially avoidable emergency or hospital use. At the same time, many neurological presentations seen in general practice are nonspecific, self-limited, or better explained by non-neurological disease, making triage difficult.
Telemedicine has been proposed as a practical mechanism to support primary care clinicians in regions with specialist shortages. In neurology, telemedicine is already well established for specific use cases such as acute stroke systems of care, where telestroke networks have improved access to expert assessment and reperfusion decision-making. Whether this model can be broadened to routine ambulatory neurology in primary care is less clear. The relevant policy question is not simply whether video consultation is feasible, but whether it changes patient flow, reduces unnecessary referrals, preserves outcomes, and strengthens care in routine practice.
The NeTKoH study addressed this question in Western Pomerania, a rural region in northeast Germany, by creating a telemedicine network linking general practitioners, patients, and neurologists. Its premise was clinically intuitive: if GPs can obtain rapid specialist input at the point of care, more patients with neurological symptoms might be safely managed in general practice, reducing the need for face-to-face secondary care. The trial’s actual findings were more nuanced and arguably more informative for implementation science.
Study Design and Methods
Trial design
NeTKoH, short for Neurological Teleconsultation With General Practitioners to Strengthen Specialist Care in Western Pomerania, was a stepped-wedge cluster randomized clinical trial conducted from January 1, 2021, through July 31, 2025. In a stepped-wedge design, participating practices begin in the control condition and then cross over to the intervention at randomized time points until all clusters receive the intervention. This design is often used when an intervention is expected to be beneficial, implementation must be phased, or logistics preclude simultaneous rollout.
For service-delivery interventions in real-world primary care, the stepped-wedge approach has practical advantages, but it also introduces analytic and interpretive challenges. Secular trends, changing clinician behavior over time, and selection effects around the timing of crossover can influence results. The authors appropriately acknowledged that lack of blinding of GPs to randomization could have introduced selection bias.
Setting and participants
The trial included 41 participating GP practices in northeast Germany. Adult patients aged 18 years or older were eligible if they presented with symptoms for which a neurological consultation was considered necessary by the GP. Patients were continuously approached for inclusion. Follow-up duration was 3 months, which is reasonable for assessing short-term care pathways and initial management decisions, though it may be too short to detect downstream effects on function, diagnosis, or health care utilization for chronic neurological conditions.
Between October 15, 2021, and October 25, 2024, 986 patients were enrolled; 3 were excluded, leaving 983 patients in the analysis. Of these, 605 patients, or 61.5%, were female, and the median age was 55 years, with a range of 18 to 90 years. The intervention phase included 517 patients and the control phase 469 patients.
Intervention and comparator
In the intervention condition, GP practices were equipped with a telemedicine system enabling a real-time video consultation involving the GP and patient on one side and a neurologist on the other. The neurologist provided immediate assessment and recommendations regarding further diagnostics, need for specialist referral, hospital evaluation, or ongoing management in primary care.
The comparator was standard care, in which the GP managed the patient according to usual practice, including referral to neurology or other services when indicated.
Outcomes
The primary outcome was the proportion of patients who continued to be managed solely within general practice. This is a pragmatic endpoint aligned with the intervention’s original policy rationale: if teleneurology substitutes for in-person specialty care, then more patients should remain under GP-led management.
Secondary observations included referral patterns, patient-reported quality of life and health status, and adherence to recommendations. These outcomes are particularly important because a failure to increase exclusive GP management does not necessarily mean failure of the intervention; it may instead indicate that teleconsultation identified unmet needs requiring escalation.
Key Results
Primary outcome: fewer patients remained solely in general practice
The central finding was directionally opposite to the original hypothesis. Teleconsultations resulted in fewer patients being managed solely within general practice than standard care: 38.3% versus 50.7%, with an adjusted odds ratio of 0.58 and a 95% confidence interval of 0.38 to 0.88. Because the confidence interval does not cross 1.0, this result is statistically significant.
At first glance, this could be interpreted as a negative trial. If the intended objective was to increase GP-only management, the intervention did not succeed. However, endpoint interpretation matters. Keeping more patients in primary care is only desirable if it is clinically appropriate. An intervention that uncovers red flags, clarifies alternate etiologies, or appropriately escalates care could worsen the primary metric while improving real-world quality of care.
Referral patterns: less neurology, more other specialty and hospital referral
Referral patterns strongly suggest a triage effect. Compared with standard care, teleconsultations decreased referrals to neurologists from 41.4% to 36.4%. At the same time, referrals to other specialists increased from 4.7% to 11.9%, and referrals to hospitals increased from 2.6% to 12.8%.
This pattern is clinically plausible. Patients presenting with “neurological” symptoms in primary care often have conditions whose optimal next step is not an outpatient neurologist. Dizziness may reflect vestibular, cardiovascular, or internal medicine causes. Sensory symptoms may raise concern for compressive, rheumatologic, metabolic, or psychiatric disease. Acute or concerning focal deficits, severe progressive symptoms, or possible central nervous system emergencies may warrant hospital evaluation rather than ambulatory specialty follow-up. In this context, teleconsulting neurologists may have improved sorting of patients into the most appropriate destination.
The reduction in direct neurology referrals also suggests that teleneurology did not simply add another layer before the same referral. Instead, it may have refined decision-making. This distinction is important for workforce planning in regions where neurologist capacity is limited.
Patient-reported outcomes
No significant differences were observed in quality of life or health status between teleconsultation and standard care. This neutral result should be interpreted cautiously. First, the follow-up period was only 3 months. Second, the population likely included a heterogeneous set of symptom complexes with varying prognoses and expected trajectories. Third, process improvements in triage do not always translate quickly into measurable patient-reported outcome gains, particularly if baseline symptom burden is modest or outcome instruments are not disease-specific.
Still, the absence of an obvious patient-reported benefit means that implementation arguments for broad teleneurology deployment should not rely on short-term quality-of-life improvement alone. The stronger case may be around access, equity, timeliness, and more appropriate routing of care.
Adherence to recommendations
Adherence to recommendations was lower in the intervention phase. The abstract does not provide granular reasons, but this finding deserves attention because care models succeed only if recommendations are actionable and accepted by both clinicians and patients. Lower adherence could reflect several real-world barriers: limited appointment availability in referral sectors, patient reluctance to travel or be hospitalized, mismatch between tele-neurologist advice and GP or patient preferences, or workflow friction in implementing recommendations generated during a video consultation.
From a health services standpoint, this may be one of the most actionable results in the trial. Teleconsultation can improve diagnostic direction, but operational gains will be blunted if downstream recommendations are difficult to execute.
Follow-up care status
At follow-up, more than half of patients in either group were still managed solely in general practice. This suggests that, despite early differences in referral behavior, a substantial proportion of presentations remained amenable to GP-led longitudinal management. It also indicates that the intervention changed initial pathways more than the eventual broad location of care for all patients.
Clinical Interpretation
The trial’s main contribution is conceptual: teleneurology in primary care may function more effectively as a triage and escalation tool than as a substitution model intended to keep patients out of secondary care. This distinction matters for clinicians, health system leaders, and policymakers.
For GPs, rapid access to neurological expertise appears useful not because it necessarily reassures enough patients to remain in primary care, but because it can identify which patients should go elsewhere, including urgently. For neurologists, the model may reduce some low-yield face-to-face referrals while redirecting patients to more appropriate services. For health systems, especially in rural settings, the value proposition may be improved allocation of limited specialist resources rather than crude referral reduction.
These findings also highlight a recurring issue in digital health evaluation: the “success” of a telemedicine service depends heavily on the chosen endpoint. If the endpoint is decreased specialist use, then the intervention underperformed. If the endpoint is more appropriate specialist and hospital use, then the results are more encouraging. In symptom-based fields such as neurology, where diagnostic uncertainty is common, an intervention that improves discrimination may appropriately increase escalation in a subset of patients.
Strengths of the Trial
The study has several notable strengths. It was randomized at the cluster level, minimizing contamination within practices and reflecting real-world implementation. The stepped-wedge design allowed all participating practices eventually to receive the intervention, a practical advantage for service interventions. The study was conducted in a structurally underserved rural region, where findings are highly relevant to current access challenges. The sample size was substantial for a pragmatic telemedicine trial, with 983 analyzable patients across 41 practices.
Another strength is the focus on routine primary care rather than a narrowly defined neurological disease. Many access problems arise at the symptom-to-diagnosis interface; studying teleneurology at this point of first contact is clinically meaningful and operationally relevant.
Limitations and Methodological Considerations
The authors rightly noted that GPs could not be blinded to randomization status, creating potential selection effects. In a stepped-wedge trial, clinicians aware that teleconsultation is available may alter which patients they enroll or how readily they judge neurological consultation necessary. This can distort comparisons between control and intervention phases.
The primary endpoint itself is debatable. Exclusive management within general practice is a utilization metric, not a direct patient-centered outcome. It assumes that remaining in primary care is preferable, which is only true when clinically justified. Given the observed increase in hospital and non-neurology referrals, the intervention may have exposed under-recognized complexity rather than failed.
The 3-month follow-up may have been too short to detect effects on definitive diagnosis, symptom resolution, emergency presentations, costs, or longer-term patient experience. The abstract also does not specify the case mix in detail, limiting assessment of which symptom categories benefited most. Heterogeneity is likely substantial; teleneurology may be more valuable for some presentations than others.
Adherence being lower in the intervention phase complicates interpretation of outcome neutrality. If recommendations are not followed, then the trial evaluates both teleconsultation quality and implementation barriers simultaneously. Future studies should separate these components more explicitly.
Finally, generalizability outside rural Germany should be considered carefully. Local referral structures, neurologist availability, hospital proximity, reimbursement, digital infrastructure, and GP scope of practice all influence the effect of teleconsultation.
Implications for Practice and Policy
For clinicians working in underserved areas, this trial supports tele-neurological consultation as a decision-support mechanism, especially for uncertain presentations in which the key question is not diagnosis alone but level and destination of care. Practices considering implementation should set expectations accordingly: the likely early benefit is better triage, not necessarily fewer referrals overall.
For health systems, the increase in hospital referrals should not automatically be viewed as undesirable. If teleconsultation identifies patients needing urgent assessment who would otherwise have waited for outpatient review, this could represent improved safety. However, implementation should be coupled with pathway design, including clear criteria for hospital escalation, referral access to non-neurology specialties, and feedback loops between sectors.
For policymakers, outcome selection in telemedicine programs deserves careful thought. Programs designed for rural equity should measure timeliness, appropriateness of referral destination, diagnostic accuracy, avoidable delays, emergency events, and patient travel burden, not only whether specialty utilization declines.
What Future Research Should Address
Future trials should stratify by presenting syndrome to identify where GP-embedded teleneurology has the highest clinical yield. Likely candidates include headache, dizziness, sensory complaints, movement symptoms, transient focal deficits, and seizure-like events, each of which has different triage requirements.
Longer follow-up is needed to determine whether improved initial routing translates into better diagnostic timelines, reduced preventable admissions, lower emergency use, or improved functional outcomes. Economic evaluation is also essential. A service that increases some referrals may still be cost-effective if it prevents inappropriate specialist visits, reduces delay-related morbidity, or decreases travel in rural areas.
Implementation science should examine why adherence to recommendations was lower during the intervention phase. Understanding barriers at the patient, GP, specialist, and system levels could substantially increase the impact of such programs.
Conclusion
The NeTKoH stepped-wedge cluster randomized clinical trial provides an important corrective to simplified expectations about telemedicine in ambulatory neurology. In rural general practice, teleneurology did not increase the proportion of patients managed solely by GPs; in fact, it reduced that proportion. Yet the intervention also reduced referrals to neurologists while increasing referrals to other specialists and hospitals, suggesting that its main value may lie in sharpening triage rather than replacing secondary care.
For clinicians and health systems, the message is practical: teleneurology may be most useful when the problem is uncertainty about where a patient should go next. In underserved regions, that function alone may be clinically meaningful. The next generation of studies should evaluate not just whether teleconsultation reduces referrals, but whether it sends the right patient to the right place at the right time.
Funding and Trial Registration
Trial registration: German Clinical Trials Register, DRKS00024492.
The abstract provided does not specify the funding source. Readers should consult the full JAMA Neurology article for detailed funding, sponsor, and conflict-of-interest disclosures.
References
1. Kiel S, Wainwright KL, Angermaier A, Chenot JF, Wischmann HA, Schulz RS, Filser PJ, Flöel A, Kurth T, von Podewils F. Neurological Teleconsultations in General Practice: A Stepped-Wedge Cluster Randomized Clinical Trial. JAMA Neurology. Published online May 26, 2026. PMID: 42189536. https://pubmed.ncbi.nlm.nih.gov/42189536/
2. Demaerschalk BM, Miley ML, Kiernan TEJ, et al. Stroke telemedicine. Mayo Clinic Proceedings. 2009;84(1):53-64. PMID: 19121255.
3. Wechsler LR, Demaerschalk BM, Schwamm LH, et al. Telemedicine Quality and Outcomes in Stroke: A Scientific Statement for Healthcare Professionals From the American Heart Association/American Stroke Association. Stroke. 2017;48(1):e3-e25. PMID: 27980189.
4. Hussey MA, Hughes JP. Design and analysis of stepped wedge cluster randomized trials. Contemporary Clinical Trials. 2007;28(2):182-191. PMID: 16829207.
