> For the complete documentation index, see [llms.txt](https://docs.bitdoctor.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.bitdoctor.ai/clinical-measurement-reports/irregular-heartbeat.md).

# Irregular Heartbeat

## Description 

Irregular heartbeats (IHB) are those that occur outside the user’s normal heart rhythm (e.g., a premature ventricular contraction). BitDoctor Technology can be used to detect and count the number of IHB the user experiences within a given measurement period.  A bpm of 60-100 is considered healthy/normal.

## Method 

The ability to detect at least one irregular heartbeat via BitDoctor Technology was assessed against the presence of irregular heartbeats as determined by electrocardiograms (ECG) annotated by experienced cardiology nurses. 

## Participants 

Subjects were recruited from several cardiology clinics. These patients had been referred for ECGs and a large proportion had irregular heartbeats. 

## Additional Data Collection Setup

Same as the [setup as HR](/clinical-measurement-reports/heart-rate.md#additional-data-collection-setup).

## Additional Data Collection 

Each 30-minute recording was then sub-divided into 1-minute samples, resulting in 12,000 1- minute samples. Some of these were excluded due to movement or poor signal quality, leaving 24 10,997 samples for analysis. The electrocardiograph for each sample was then annotated by a cardiology nurse to identify segments with IHB. 

## Performance of IHB detection algorithm 

BitDoctor signal was extracted from each video and the IHB detection algorithm was employed to identify IHB within the signal. This algorithm employs machine-learning to identify IHB. 

The performance of this algorithm was assessed by recall, precision and f1 score. Recall is the percentage of cases where the BitDoctor-based algorithm detected an IHB when there was indeed an IHB (according to ECG). It was calculated as True Positives / (True Positives + False Negatives). Precision is the percentage of cases where an IHB actually occurred, out of all cases where the BitDoctor-based algorithm claimed an IHB had occurred. It was calculated as True Positives / (True Positives + False Positives). F1 score is the ‘harmonized mean’ of these measures, calculated as 2 \* (precision \* recall) / (precision + recall). 

## Results 

Recall was 96%, precision was 82% and the F1 score was 89% in terms of detecting at least one IHB.