Standardized Neurotrauma Registries as the Highest Priority for Advancing Global Neurosurgery: A Narrative Review

Traumatic brain injury (TBI) and spinal cord injury (SCI) from neurotrauma remain a leading cause of global mortality and disability, with the greatest burden falling on low and middle-income countries (LMICs). In 2021, there were an estimated 20.8 million head injuries globally, of which 56.6% were moderate to severe TBI [1]. Also, more than 15 million people around the world are living with SCI, with most of them due to traumatic events such as falls, road traffic accidents, and violence [2]. The lack of data limits policymakers and stakeholders from efficiently prioritising neurosurgical services within national surgical, obstetric, and anaesthesia plans (NSOAPs).

LMICs have the highest burden of neurotrauma, with nearly 90% of injury-related deaths [3]. Nevertheless, reliable epidemiological data to characterise the incidence and quantify the extent of neurotrauma are missing, as well as to ascertain neurotrauma outcomes over time in many of these countries. This lack of information inhibits effective policy development and resource distribution [4].

Factors such as a high prevalence of road traffic accidents (RTAs), violence, and poor adherence to and enforcement of injury prevention measures exacerbate this burden, but many low- and middle-income countries (LMICs) do not have available national-level data with which to describe the incidence of neurotrauma or the course of outcomes over time [2,4].

Registries, organised approaches for the collection, storage, and examination of patient-level information, are important in generating valid estimates of incidence, identifying delays in care, assessing treatment responses, and informing resource allocation. Nevertheless, the majority of current neurotrauma surveillance is either partial, non-standardised, or non-existent in LMICs, thereby undermining comparability and severely constraining policy-relevant decision-making [4,5,6].

Recent scoping and systematic reviews highlight significant variation in data dictionaries amongst LMIC trauma registries, with many registries lacking high-priority injury data items, including mechanism, injury severity scores (ISS), prehospital time, and functional outcomes [4]. Furthermore, investigations into barriers to neurotrauma care show that delays in presentation and delivery of care, because of geographic, infrastructural, personnel, and financial impediments, are associated with worse outcomes but remain under-researched at many sites [5,7]. In view of the devastating impact of neurotrauma and the wide surveillance lacuna, there is a pressing need for research to standardise neurotrauma registries. These registries may be the most high-yield intervention in global neurosurgery: they can yield the essential data that undergirds planning, monitoring, and performing quality improvement for neurosurgical care across the globe [4,6].

Definition of trauma versus neurotrauma registries

A trauma registry collects information on all significant injuries to body regions and can be configured to include neurotrauma-specific data fields. In contrast, a neurotrauma-specific registry is limited to intracranial and spinal injuries, providing a neurosurgical perspective on the case and the neurological outcome. Including a neurotraumatology module in a preexisting trauma registry fosters compatibility, exploits the already existent foundation, and minimizes replications. Separate neurotrauma registries, which provide greater clinical detail, are likely to be more resource-intensive and may be less sustainable. A modular approach, in which a core trauma data set can be extended to include a neurotrauma subset, is recommended to provide a compromise between clinical detail and interoperability.

This narrative review, therefore, aims to evaluate the current status of neurological trauma data collection in LMICs, catalogue existing registry models, and discuss some of the methodological and implementation challenges encountered. Moreover, it also aims to propose a minimum neurotrauma dataset for surveillance, which can be customised across diverse settings, as well as develop a research agenda to guide further registry development and integration into health systems.

This was a narrative review synthesizing evidence on neurotrauma surveillance and registries within global neurosurgery. We performed a search on PubMed and Google Scholar from inception to September 2025 for terms related to neurotrauma, registries, and LMICs. Grey literature was included from the WHO, WFNS, and the Global Surgery Foundation, and reference lists were screened. Eligible studies included peer-reviewed literature, reviews, policy documents, and reports from registries, addressing any aspect of the burden of neurotrauma, registry models, data quality, outcomes, cost, or the process of establishing a registry in an LMIC. Studies focusing on high income only and without global relevance were excluded. Language was restricted to English (excluded due to lack of translation resources). Data were extracted on study objectives, populations, registry features, outcomes, and findings. The findings were synthesized into five themes: burden and unmet need, registry design and implementation, outcome and quality metrics, economic and policy considerations, and research gaps. Meta-analysis was not possible due to heterogeneity; the results were narratively synthesized.

A systematic search was performed in PubMed and Google Scholar from inception to 30 September 2025. Representative search terms combined key concepts for neurotrauma, registries, and global health settings using Boolean operators:

(“neurotrauma” OR “head injury” OR “traumatic brain injury” OR “spinal cord injury”)

AND (“registry” OR “database” OR “surveillance” OR “trauma system”)

AND (“low- and middle-income countries” OR “LMIC” OR “developing countries” OR “global surgery”)

All the retrieved records were screened by title and abstract for their relevance, and the potentially relevant articles were retrieved for the full-text review. Discrepancies were resolved by discussion.

There were 616 records identified, 48 were full-text reviewed, and 12 met the inclusion criteria for the final narrative synthesis (Table 1)

Table 1 highlights the summary of the search and selection process.

A narrative review methodology was chosen in place of a scoping review because the primary focus was to provide an interpretive synthesis of concepts, frameworks, and implementation experiences, rather than a systematic mapping of all available evidence. This approach allows for the combination of diverse types of data from diverse data sources, such as policy documents, organizational reports, and registry narratives, that are not amenable to a formal quantitative synthesis.

Quality assessment

For transparency and rigour in the interpretation of the included evidence, all studies were appraised retrospectively using a well-established structured instrument. The Critical Appraisal Skills Programme (CASP) checklists were used for different study designs (CASP Cohort Study checklist for observational or pilot studies; CASP Qualitative checklist for qualitative studies; an adapted JBI checklist was applied for scoping and review articles). Two independent reviewers evaluated each study for key domains: the extent to which the study design was appropriate, the representativeness of the sample, the validity of the measurement of the outcome, the completeness of the data/follow-up, the risk of bias, and the transparency of the analysis or costing methods, if applicable. Discrepancies were discussed and adjudicated. An overall quality rating of High, Moderate, or Low was also assigned to each study, and this was used to inform the evidence strength labels that are reported alongside each recommendation. This method allows for a systematic, replicable evaluation of methodological quality, even when conducted retrospectively on studies in a narrative synthesis.

Table 2 highlights the structured quality appraisal using CASP/JBI principles and provides transparent justification for each source.

Currency and living evidence statement

Searches were up to date until September 2025. Given the rapidly changing environment for global neurotrauma registries, the dataset and synthesis should be considered a living evidence resource and may be subject to updates (e.g., annually) as new registry data and implementation reports become available.

Synthesis of the literature and recent program reports yielded four consistent, high-priority findings regarding neurotrauma registries in LMICs: (a) the burden is large but inadequately captured; (b) registry coverage and design is fragile; (c) the data content is heterogeneous and frequently incomplete; (d) barriers to implementation and sustainability are predictable and reiterative; and (e) there are significant gaps in outcomes, follow-up, and health-economic data that undermine the value for policy. Each finding is summarized below, along with its practical implications for registry design.

Burden is large but incompletely captured

Global estimates confirm that TBI and SCI remain leading causes of death and disability, with low and middle-income countries having the largest share of severe events and road-traffic-related etiology. Population-based models predict tens of millions of new TBIs incidents per year and a large years-lived-with-disability burden, highlighting urgent surveillance needs to define unmet surgical demands. Yet, primary sources of surveillance in LMICs are fragmented: national case ascertainment is far from complete, and numerous injuries never make it to hospitals captured by registries; this results in a systematic underestimation of the actual burden and impedes service planning [1,8].

Registry coverage & design is fragile

In a scoping review of national trauma registries within LMICs, 15 countries with 16 national registries with neurotrauma-specific elements were highlighted. The study reviewed and compared the data dictionaries of these trauma registries against the WHO’s Minimum Dataset for Injury (MDI), and identified core elements such as Glasgow Coma Scale (GCS), mechanism of injury, anatomic site, and injury severity scores, included in most registries, but crucial variables such as helmet use and alcohol intoxication were among the least represented. Concordance varied: Cameroon, Colombia, Iran, Myanmar, and Thailand had higher similarity with MDI elements; many others captured fewer elements [4].

Another registry, the Latin America & Caribbean Neurotrauma Registry (LATINO-TBI), highlighted the feasibility of multicenter registry data aggregation in LMIC regions. This registry records key epidemiological and clinical data and has facilitated preliminary inter-institutional benchmarking. However, LATINO-TBI suffers from significant incompleteness regarding items such as prehospital care and discharge outcome, essential for a complete epidemiologic and system evaluation [9].

Coverage, therefore, is patchy: although many core data elements are frequently present, there is significant variation in the definition and collection of “minimum neurotrauma data.” This inconsistency is also reflected in the number of registries per region, with documented efforts more numerous in Latin America and Southeast Asia, compared to Sub-Saharan Africa and certain areas of South Asia that remain underrepresented [4].

Table 2 presents a pragmatic minimum neurotrauma dataset for LMIC registries.

Implementation barriers & facilitators

Multiple studies report that barriers to registry establishment and implementation in LMICs are significant and multifactorial. In Colombia, a qualitative evaluation of a traumatic brain injury registry project highlighted incomplete clinical data capture, suboptimal availability of technological (IT) resources, time constraints due to high clinical workload, and difficulty in obtaining ethical approval as key issues [9]. Local champions and administrative buy-in were some of the strongest facilitators; embedding data collection within clinical workflow contributed to reducing data burden [10].

General review of trauma registries implementation (2017–2023) in LMICs was consistent with these findings: poor funding, uncoordinated efforts among administrative stakeholders, infrastructural barriers (i.e., computers and stable internet access), limited human resources for data entry and validation, and challenges in data analysis and reporting were recurring themes. Interviews with stakeholders stressed that, in addition to technical challenges, low priority is often assigned to registry maintenance after initial funding runs out, challenging sustainability [11].

Facilitators included: existence of a motivated local leader; simplicity of the data collection tool;  use of mobile or digital platforms; and connecting registry data to positive returns such as better patient care, institutional benchmarking, and research production [10].

Table 3 summarizes common implementation barriers with pragmatic mitigations and exemplar evidence

Outcome measurement & long-term follow-up

Most LMIC registries consistently record in-hospital mortality and basic disposition, but considerably fewer record validated functional outcomes (e.g., Extended Glasgow Outcome Scale), cognitive status, return-to-work, or rehabilitation engagement at defined post-discharge intervals. Loss to follow-up is prevalent due to patient mobility, expense, and limited clinical capacity. As a result, registries infrequently produce the disability, dependency, and economic burden measures that inform cost-effectiveness and budget impact analyses. Novel and remote follow-up approaches (SMS/phone, community health worker follow-up) are promising but have not been broadly validated across diverse LMIC contexts [12,13].

Economic, administrative, and policy integration

Direct cost, out-of-pocket expense, and indicators of catastrophic financial effect on the household are rarely collected by registries; likewise, health system resources (intensive care unit (ICU) days, imaging counts, implant/consumable use) are inconsistently recorded. This lack of reliable information makes it exceedingly difficult to generate the cost-per-DALY or catastrophic expenditure evidence that Ministries of Health prioritize neurosurgical investments within NSOAPs or benefit packages. Including a small, standardized set of economic items (hospital bills, length of stay, ICU use, selected household cost items) in registry minimum datasets would result in a significant increase in policy traction [8,14].

Priority registry design features

From the evidence above, pragmatic design features that maximize yield in LMICs are:

• Adopt a short, standardized minimum dataset based on WHO/NIH CDE principles; prioritize timestamp and a single validated 30-/90-day functional outcome. [15,16]
• Ensure data collection is possible at the bedside (paper forms that can be easily transcribed into digital format) and train a small cadre of dedicated data clerks rather than depending entirely on clinicians. [17]
• Design for sustainability from the outset: transition roadmap from seed/donor funding to institutional/MoH support, regular reporting that illustrates value, and integration into national reporting systems as much as possible. [14]
• Add a brief economic module to predict direct hospital costs, and a household cost proxy for early policy-relevant analyses. [8]

Principal findings

This review discovered that the establishment of neurotrauma registries in LMICs is increasingly feasible but remains heterogeneous in content and lacks sustainable outcomes, especially with respect to long-term patient outcomes, economic data, governance, and operational failures. There is a need for and possibility of harmonizing a short, pragmatic minimum dataset focusing on timestamps, with a simple validated functional outcome at 30/90 days, and a simple economic module at a basic level. Remote follow-up methods (phone/SMS) combined with modest investments in dedicated data personnel materially improve outcome capture, while early engagement with national health authorities increases the likelihood of sustainability and policy translation. These crucial insights point to a single, high-yield next step: a collaborative pilot execution of a standardized registry package across representative LMIC settings with an implementation evaluation embedded in the design [18–20].

Interpretation and relation to existing literature.

The results are consistent with global injury surveillance guidance emphasizing core, actionable fields to maximize usefulness and feasibility in resource-limited settings and feasibility in low-resource settings [15]. Data on global burden highlight the magnitude and disparity of neurotrauma: injuries represent a significant and increasing proportion of death and disability among young adults, especially in LMICs, reiterating the need for dependable surveillance to inform planning and advocacy [21]. The literature on tele-follow-up/mobile health provides pragmatic solutions for the longstanding chronic-outcome data gap: phone, SMS, and smartphone approaches have been validated within multiple trauma cohorts and can be modified to collect GOS/E-like end points along with rudimentary quality-of-life measures for situations in which in-person follow-up is not feasible [19,20]. The economics of injury do reveal significant out-of-pocket and catastrophic costs in LMIC contexts, illustrating the policy traction that even a modest registry economic module could bring to bear if connected to national financing discussions [22].

Implementation and sustainability pathway

A pragmatic three-year stepwise plan is proposed. Pilot implementation and proof of concept across one to three sites, finalizing dataset, training staff, collecting 6–12 months of data, and embedding costing or outcome validation in Year 1; scale-up and formalization to 10–20 sites, securing memorandum of understanding (MoUs), developing a regional training-of-trainers model, and initiating Ministry of Health engagement in Year 2. MoU development) and building a regional cadre of trainers (regional training-of-trainers (ToT)) as well as engaging with the relevant Ministry of Health, by Year 3: with the registry into local (or Ministry) streams of funding through integration into institutional budgets and the establishment of a governance board with long-term data-use agreements. A simple three-year progression can be expressed thus:
Year 1: pilot and/or proof of concept:

• Apply at 1-3 sites; finalize dataset; train staff
• Gather 6–12 months of pilot data
• Perform embedded costing and/or outcome validation substudies
• Initial Ministry of Health (MoH) consultation

Year 2: expansion and formalization:

• Expand to 10-20 sites
• Sign MoUs and regional training-of-trainers (ToT) models
• Design digital and IT support systems
• Start formal coordination with the MoH reporting pathways

Year 3: transition to sustainability:

• Incorporate registry costs into institution or MoH budgets
• Establish a national governance board and long-term data-use agreements
• Institutionalize stewardship, access, and quality assurance mechanisms

From Uganda, implementation research demonstrates that the introduction of a structured trauma registry significantly increases documentation completeness and can form the basis for institutional and national dissemination and adoption, providing proof of concept for a staged pilot-to-scale implementation pathway [23].

The per-site indicative resource requirements are a data clerk (0.5-1 FTE), tablet or device (one-off, around $150-300), SIM or data stipend ($5-15 per month), and training and oversight ($2,000-5,000 per annum). When precise local cost data are not available, an embedded costing study is recommended during the pilot phase.

Strengths of the evidence base and of this review.

This data consists of multicentre registry reports and recent service evaluations and implementation studies that illustrate real-world limitations and pragmatic mitigations; this results in recommendations that are grounded in operational realities rather than being purely theoretical. The review integrates both technical (data elements, outcome measures) and programmatic (staffing, information technology, governance, ethical considerations) aspects and serves as a unified road map for pilots and policy makers.

Limitations and sources of bias.

Available registry reports and implementation studies are biased towards certain regions (e.g., Latin America, a portion of Southeast Asia) and tertiary hospitals, which results in an overestimation of feasibility and underrepresentation of rural/peripheral environments. Also, included studies were restricted to English-language and may exclude important regional reports published in Spanish, Portuguese, French, or other languages. There is a potential for publication bias of successful pilots; negative or unsuccessful attempts at registries are much less frequently reported, thereby limiting lessons on modes of failure. Data on outcomes and costs are limited, restricting quantitative synthesis; variability in terminology and definitions used across studies further limits comparability. Also, a technology-dependent approach assumes sufficient mobile penetration and network connectivity, which is not universally present across LMIC contexts.

Out of the 12 registry reports, implementation studies and syntheses that we identified in our targeted search, 4 (33.3%) were based in Latin America (including multi-institution LATINO-TBI and country-level work in Colombia), 3 (25.0%) were Africa- based (Ghana, South Africa, Nigeria prototype), 1 (8.3%) was from South Asia (India), 0 (0%) were from the Middle East & North Africa (MENA) region, 0–1 (0– 8.3%) applied to Southeast Asia in primary registry reports located by our search, and 2 (16.7%) included multi-region or global syntheses and policy/ tool kit documents.

This overrepresentation of certain regions may impart an optimistic bias regarding feasibility and success in those better-reported regions (particularly Latin America), and downplay challenges, barriers, and contextual limitations in those less reported, namely Africa (beyond a small number of centers) and South Asia /MENA, where primary registry reports and pilot data are few and far between.

Table 4 summarizes identified neurotrauma registries, pilot projects, and implementation-focused reports included in the review.

Future pilot work should explicitly target underrepresented regions, particularly Africa and South Asia, to enhance global relevance and equitable evidence generation.

Ethics, governance, and data sharing

Essential governance provisions should include the appointment of a data steward who is responsible for quality assurance, access control, and compliance, and provide clear documentation of data flows, encryption techniques, and storage locations. There must be an application of standard de-identification protocols, which include removal of direct identifiers (names, full date of birth, address), use of pseudonymized registry IDs, restriction of dates to month and year or use injury-day offsets, transfer via a secure, encrypted channel, and well-defined data-retention periods.

In the case of minimal risk observational registries, a consent waiver/opt-out model may be appropriate, subject to local IRB approval. Explicit consent needs to be sought for follow-up contact (phone/SMS), linkage with identifiable hospital or administrative records, and any cross-border data transfer if identifiers are retained. A minimum data-sharing agreement (MDSA) should specify purpose and permitted analyses, allowed use, dataset details, and de-identification measures, security and access controls, publication and authorship rights, duration, renewal, and dispute mediation. Institutional sign-off is required from both sides for cross-border provision of data with local ethics clearance. Template IRB language and a model MDSA should be attached to enable swift approvals.

Experience from regional neurotrauma registries and WHO injury-data standards highlights the need for formalized governance, anonymous data formats, and a well-defined data sharing agreement, reinforcing the recommendation of stewardship and ethical oversight [6,9].

Outcome measure selection

We suggest a pragmatic 30-day functional outcome measure using the 3-level simplified Glasgow Outcome Scale (GOS) (good outcome, moderate disability, severe disability/death), to best maintain discriminative validity while evaluating feasibility. The Glasgow Outcome Scale – Extended (GOSE) allows better discrimination but entails training and extended follow-up, making it potentially impractical for many developing world settings. The 3-level scale favors the completeness and reliability of follow-up information.

30-day outcome assessment algorithm: determine patient status via telephone or clinic visit at 30 (±7) days postinjury: (1) Favorable – patient can care for self (GOSE 5–8); (2) Moderate disability – at home but requires some assistance (GOSE 3–4); (3) Severe disability or death – dependent or dead (GOSE 1–2). If validation data are not available, a nested substudy of the 3-level classification against the full GOSE in a 10–20% sample to provide an estimate of misclassification is advisable.

Simplified outcome measures are recommended because time, personnel, and follow-up capacity are significant limitations in LMIC environments. A brief scale reduces the burden on the interviewer, limits training requirements, and enhances completeness of follow-ups, especially when follow-up is done by phone or SMS. Trade-offs include: precision: simplified instruments (such as 3-level GOS) lose some discriminative granularity compared with full GOSE; feasibility: the simplified instruments substantially decrease missingness and interviewer error, and this may result in the production of usable data with ultimately higher quality usability despite lower resolution; validation need: where possible, a nested comparative substudy (at 10–20% of cases) can assess the extent of misclassification and corroborate suitability in the local setting.

Christie et al. conducted a pilot study of telephone follow-up after injury in Cameroon and demonstrated that brief follow-up data could be collected by phone (83% provided numbers; 48% completed follow-up) and that brief triage-style assessments identified unmet needs and deaths that would not be identified otherwise. These operational results form the basis for the recommendation to favour pragmatic short functional end points (greater completeness at the cost of granularity) [24]. In Cameroon, telephone-based follow-up tools that are short and pragmatic have been shown to produce substantially higher completion rates compared to more comprehensive functional evaluations, providing strong evidence to recommend simple, minimally burdensome outcome measures as the priority [25].

Economic module

A simple economic module should allow hospital and household cost elements to be captured in simple bands for routine data collection. The following hospital items are the minimum: admission billing band (0=free/subsidized; 1=low; 2=medium; 3=high), length of stay (days), ICU days, major procedures (yes/no), and imaging (CT yes/no). The following items are at least required from a household perspective: out-of-pocket payments (bands: 0, <$50, $50–199, $200–999, ≥$1000), lost income (none, <1 month, ≥1 month), and whether assets were sold or borrowed (yes/no).

As part of the analysis, the cost per patient can be calculated using site-specific unit costs estimated by applying local billing bands to approximate unit costs (e.g., median bill per band times number of days/procedures). Then, a proportion experiencing catastrophic expenditure can also be calculated: out-of-pocket not less than one month of household income, or ≥$200. In the absence of local cost data, a simple concurrent cost sub-study is recommended to determine local-specific conversion factors.

Evidence from various LMIC contexts reports significant levels of out-of-pocket and catastrophic household spending following injury, which indicates both the feasibility and policy relevance for registries to capture a limited standardized set of economic indicators for reporting [22,26].

To promote easy adoption, the economic module can consist of a concise list of these core cost elements: hospital-level cost items: admission billing band (subsidized/low/medium/high). length of stay and ICU days, major procedures (yes/no), CT imaging (yes/no); household-level cost items: out-of-pocket expenditures (simple bands), lost income (none, <1 month, ≥1 month, sale of assets or borrowing (yes/no). These relate to common economic indicators by allowing the calculation of patient-level costs by multiplying the length of stay and procedures by local unit costs or billing band adjustment factors. Also, catastrophic expenditure can be defined as out-of-pocket payments ≥1 month of household income or ≥$200 (or a locally validated threshold). Indicators of financial distress (such as asset sales and borrowing) are well-established and commonly used in health economics as early warnings that households are at risk of being unable to withstand healthcare shocks.

Unanswered questions and next research steps.

Some key gaps in knowledge warrant empirical investigation: (1) Which minimal outcome measure provides the best tradeoff between discriminative validity and feasibility across a variety of LMIC contexts? (2) What is the most cost-effective follow-up approach (phone versus community health worker visits) in each setting? (3) To what extent can registries be financially sustained at the institutional level without continued donor funding, and which models of financing (e.g., line-item hospital budgets, MoH integration, donor-to-government transit) are most successful in this regard? These questions necessitate pragmatic, multi-centre implementation trials and economic evaluations, integrated within registry pilots.

Uniform, sustainable neurotrauma registries that capture a small core dataset, incorporate remote outcome follow-up, and have modest economic indicators are the best investment the global neurosurgery community can make today. Piloting across representative LMIC sites with an embedded implementation evaluation and an explicit sustainability pathway will generate burden, outcome, and cost evidence required to inform workforce planning, financing, and system reforms that mitigate mortality and disability from neurotrauma.