JMIR Publications

ABSTRACT

Background:

Stroke has a significant global impact, with 12.2 million incidents in 2019 and second leading cause of death worldwide. In Finland, stroke prevalence is 1.5% within population. Cerebrovascular diseases (CVDs), including strokes and transient ischemic attacks (TIAs), cause long-term disabilities and economic burdens. Strokes can be ischemic, caused by blood clots, or hemorrhagic, caused by bleeding in the brain. Immediate diagnosis and treatment are crucial. Proper management of TIAs is also essential due to the high risk of subsequent strokes. Currently wearable sensors, AI-based prediction, and automatic video analysis are not used in cerebrovascular disease diagnosis and recovery estimation.

Objective:

The aim of this study was to describe a study design and study cohort characteristics of the Stroke-Data study conducted at two university hospitals in Finland. The study aimed to collect data from healthy controls and patients with TIA or stroke to investigate the possibilities of sensors, artificial intelligence and video analysis in cerebrovascular disease diagnosis and recovery estimation.

Methods:

The Stroke-Data study was a prospective national multi-center study conducted at Oulu University Hospital (OUH) and Kuopio University Hospital (KUH) from October 2021 to December 2022. The study collected multi-modal data on stroke and transient ischemic attack (TIA) patients, as well as healthy controls. Data collection was performed by study nurses and a research assistant. Inclusion criteria for patients included being over 18 years old and diagnosed with a cerebrovascular accident or TIA within the last three days. Healthy control subjects were over 18, without chronic diseases or long-term medications. Data collection included electroencephalogram (EEG); electrocardiogram (ECG); near-infrared spectroscopy (NIRS); accelerometers to analyze gait and balance; retinal fundus images; video recordings from neurological examination to analyze the face, body movements, and speech; clinical assessments; questionnaires and electronic health records.

Results:

In total we recruited 263 participants were recruited, of whom 123 were controls (59 [IQR: 45–72] years], 68% female), 31 were patients with TIA (71 [IQR: 60–80] years, 26% female) and 103 were patients with stroke (70 [IQR: 59–76] years, 41% female). We excluded sSix participants were excluded (five with other conditions than TIA or stroke and one control with previous CVD). The results of this study will be published and presented in peer-reviewed journals and international conferences.

Conclusions:

We successfully conducted a multicenter data collection study on healthy controls and stroke and TIA patients with multimodal data collection methods. The protocol is unique as similar multimodal data collection has not yet been performed.