The study, published in Frontiers in Stroke, offers the first real-world snapshot of quality monitoring practices across telestroke networks worldwide. An international team led by Ayush Agarwal of the All India Institute of Medical Sciences in New Delhi and Christine Tunkl of University Hospital Heidelberg combined two complementary data sources: a global survey of telestroke network leaders and a systematic review of the published literature. Of 254 networks identified worldwide, 88 responded to the survey, a response rate of 34 percent, with 71 percent of responses coming from networks in high-income countries. The systematic review analyzed 92 studies describing 64 acute telestroke networks across 17 countries, 92 percent of which were based in high-income settings.

The headline finding is deceptively simple: quality monitoring is far from universal. Only 64 of the 88 surveyed networks, or 72 percent, reported performing any quality monitoring at all, meaning roughly one in four networks operates without systematically measuring its own performance. The picture was brighter in the published literature, where every one of the 64 networks described in studies reported some quality metrics, but the authors caution that this likely reflects reporting choices rather than genuine differences in practice. Networks that fail to publish their data may still be monitoring it, and networks that responded to the survey may not have captured everything they do.

To make sense of the data, the researchers applied a framework first described by the health care quality pioneer Avedis Donabedian in 1988, which divides quality assessment into three interconnected constructs: structure, process, and outcomes. Structure refers to the settings where care occurs, such as standard operating procedures, the mode of teleconsultation, and the methods used to transfer brain images. Process refers to the mechanics of delivering care, including door-to-needle time, door-to-CT time, and thrombolysis rates. Outcomes refer to the effects of care, such as stroke severity scores after treatment, mortality, functional recovery, and safety measures like symptomatic intracranial hemorrhage.

The structural picture was relatively encouraging. In the survey, 87 percent of networks reported operating under standard operating procedures, 87 percent used real-time video consultation, and 92 percent documented their teleconsultations formally. For image transfer, 73 percent relied on Picture Archiving and Communication Systems, the standard hospital infrastructure for storing and sharing medical images, while 26 percent used specialized applications and 9 percent resorted to consumer messaging platforms such as WhatsApp or Viber. In the published literature, 95 percent of networks described real-time video examination, and the reliance on PACS was lower at 26 percent, though more than half of the published studies simply did not specify how images were transferred.

Process metrics told a more uneven story. Door-to-treatment time was the most frequently documented indicator, reported by 95 percent of survey networks and 87 percent of those in the literature. Thrombolysis rate, the proportion of eligible patients who receive clot-dissolving drugs, was the single most common performance indicator across both data sources, monitored by 93 percent of survey networks and 87 percent of published networks. But other recommended process measures lagged far behind. Door-to-consultation time was tracked by only 4 percent of survey networks, and inter-hospital transfer rates, which reveal how often telestroke consultations lead to patient transfers for advanced treatment, appeared in 50 percent of published studies but were not offered as a survey option and were reported by none of the responding networks.

The outcome metrics showed the starkest gaps, and these are the measures that matter most to patients. In-hospital mortality was documented by 78 percent of survey networks but only 31 percent of published networks. NIHSS scores at discharge, a standard measure of neurological deficit, were recorded by 65 percent of survey networks but a mere 1 percent of published networks. NIHSS at 24 hours was almost entirely absent from both sources, appearing in just 1 percent of survey networks and 4 percent of published ones. Functional outcome measured with the modified Rankin Scale at three or six months, the gold standard for judging whether stroke treatment restores independence, was reported by 50 percent of survey networks and 35 percent of published networks. Secondary prevention measures were tracked by 34 percent of survey networks and appeared in none of the published studies.

The authors argue that this lopsided focus on speed over results is not simply negligence but reflects systemic constraints. Responsibility for data monitoring is often shared ambiguously between the telestroke provider and the recipient hospital, and shared data infrastructures that would allow standardized collection are frequently lacking. The patient pathway in telestroke typically spans multiple hospitals, as patients are transferred for higher-level care, fragmenting the record and complicating the attribution of outcomes. And without clear incentives or regulatory requirements, data monitoring is easily deprioritized as time-consuming and resource-intensive. The result is a field that has matured technically while remaining surprisingly immature in how it measures its own success.

There is evidence that monitoring pays off. A telestroke program in Catalonia, Spain, that tracked baseline and 24-hour NIHSS scores, thrombolysis rates, mortality, symptomatic hemorrhage, and three-month functional outcomes achieved a progressive drop in door-to-needle time between 2013 and 2015, an improvement the program’s researchers attributed directly to quality monitoring. A report from Michigan stroke hospitals between 2022 and 2023 found that telestroke evaluation increased the odds of thrombolysis by 61 percent among eligible patients but was associated with 44 percent lower odds of achieving door-to-needle times within 60 minutes, a gap that quality data helped identify and that innovations such as activating the teleconsultation before neuroimaging and storing the telestroke cart in the CT suite helped close. Registry data from Germany showed that telestroke implementation was associated with an absolute decrease of 14.4 percent in interhospital transfers across all stroke center levels, generating substantial cost savings.

To close the gap, the research team proposes a minimum dataset for acute telestroke services, organized along the Donabedian framework. On structure, it includes the teleconsultation mode and technical or service failures. On process, it includes door-to-needle time, intravenous thrombolysis rate, and the total number of stroke teleconsultations. On outcomes, it includes in-hospital mortality and symptomatic intracranial hemorrhage. The authors also recommend formal agreements that clearly define monitoring responsibilities between providers and recipient hospitals, regular joint data review, and eventually a national or global certification program with quality monitoring as a key requirement. Until the global stroke community agrees on a standardized core set of indicators, they warn, comparing outcomes across networks and driving meaningful improvement will remain an aspiration rather than a practice.

Subject of Research: Quality monitoring practices and recommended performance indicators in telestroke networks worldwide

Article Title: Quality monitoring in telestroke networks: real-world insights from survey response data and literature review

Article References: Quality monitoring in telestroke networks: real-world insights from survey response data and literature review. (n.d.). https://doi.org/10.3389/fstro.2026.1879025

Image Credits: AI Generated

DOI: 10.3389/fstro.2026.1879025

Keywords: telestroke, stroke care, quality monitoring, telemedicine, thrombolysis, quality indicators, NIHSS, door-to-needle time, patient outcomes, health care quality, stroke networks, Donabedian model