Correction: contactless radar-based heart rate Estimation in palliative care– a feasibility study and possible use in symptom management

Stefan G. Grießhammer, Anke Malessa, Hui Lu, Julia Yip, J Leuschner, Florian Christgau, Nils C. Albrecht, Marie Oesten, Thanh Tran, Robert Richer, Maria Heckel, Bjoern M. Eskofier, Alexander Koelpin, Tobias Steigleder, Christoph OstgatheBMC 2025 · DOI · OpenAlex

Introduction

In the evolving landscape of medical technology, the transition from invasive to non-invasive monitoring represents a significant leap in patient comfort and care quality. This is particularly true in palliative care, where the focus shifts from curative treatment to symptom management and the preservation of dignity. A recent study published in BMC Palliative Care (2024) titled "Contactless radar-based heart rate estimation in palliative care – a feasibility study and possible use in symptom management" explored the potential of radar technology to meet these needs. However, a subsequent formal correction was issued to address a minor arithmetic discrepancy in the reported accuracy of the system. This article summarizes the study's findings and the specific statistical adjustments made to ensure the integrity of the clinical data.

The Challenge of Monitoring in Palliative Care

Traditional heart rate monitoring relies on contact-based sensors, such as electrocardiogram (ECG) electrodes or pulse oximetry clips. While effective in acute care settings, these methods present unique challenges for patients at the end of life. Physical sensors can cause skin irritation, restrict movement, and serve as a constant, visible reminder of medical intervention, which may be distressing for both patients and their families.

Contactless radar-based systems offer a promising alternative. By utilizing low-power electromagnetic waves to detect the subtle mechanical movements of the chest wall caused by cardiac activity, these systems can estimate heart rate without any physical contact. The feasibility of such technology depends on its ability to provide accurate data that matches the precision of gold-standard reference devices.

Methodological Validation and the Correction

The original research involved a comprehensive evaluation of a radar-based system against a ground-truth reference across a large dataset of cardiac activity. The researchers focused on "HR difference"—the variance between the radar's estimated heart rate and the actual heart rate.

To validate the system, two primary error margins were established:

  1. A tight window of +/- 5 beats per minute (bpm): This represents a high level of clinical precision.
  2. A standard deviation window of +/- 8.2 bpm: This threshold was defined as ±1.96 times the standard deviation of the mean difference, a common statistical benchmark for agreement between two measurement methods.

The correction notice specifically addresses the calculation of the percentage of heartbeats that fell within the +/- 8.2 bpm window. Upon re-examination of the raw data, it was determined that the previously reported figure of 97.25% required a minor downward adjustment to reflect the mathematical reality of the sample size.

Updated Results and Data Integrity

The study analyzed a total of 57,048 heartbeats collected from patients in a palliative care environment. This large sample size provides a robust foundation for assessing the system's performance. The corrected findings are as follows:

  • Within +/- 5 bpm: 53,583 out of 57,048 beats were accurate, resulting in a success rate of 93.93%.
  • Within +/- 8.2 bpm: 55,439 out of 57,048 beats were accurate. The corrected percentage is 97.18% (adjusted from the original 97.25%).

While a change of 0.07% may seem marginal in a practical clinical setting, the correction is vital for the scientific record. In medical research, absolute precision is necessary for meta-analyses, where data from multiple studies are pooled to determine the overall efficacy of a technology. Furthermore, for engineers developing safety-critical medical devices, every decimal point contributes to the validation of the underlying algorithms.

Insights and Future Directions

Despite the slight adjustment, the core conclusion of the study remains unchanged: contactless radar-based heart rate estimation is highly feasible for palliative care. Achieving over 97% accuracy within a narrow error margin suggests that the technology is nearing the level of reliability required for routine clinical use.

The proactive nature of this correction reflects a commitment to transparency and data integrity, which is especially important when researching vulnerable populations. Moving forward, the focus of this research will likely shift from basic feasibility to practical integration. Future work may include integrating these systems into real-time symptom management dashboards, testing robustness against patient movement, and evaluating the psychological impact on families when traditional "wired" monitoring is removed.

Conclusion

The correction to the 2024 study on radar-based heart rate estimation serves as a reminder of the rigorous standards governing medical literature. By ensuring that the statistical evidence is reported with total accuracy, the researchers have strengthened the foundation upon which future contactless monitoring technologies will be built. The high level of precision demonstrated—97.18% within the standard deviation window—confirms that radar technology is a viable and compassionate tool for the future of palliative care.

Summary updated

BibTeX
@inproceedings{griehammer2025correction,
  author = {Stefan G. Grießhammer and Anke Malessa and Hui Lu and Julia Yip and J Leuschner and Florian Christgau and Nils C. Albrecht and Marie Oesten and Thanh Tran and Robert Richer and Maria Heckel and Bjoern M. Eskofier and Alexander Koelpin and Tobias Steigleder and Christoph Ostgathe and Bjoern Eskofier},
  title = {{Correction: contactless radar-based heart rate Estimation in palliative care– a feasibility study and possible use in symptom management}},
  booktitle = {BMC},
  year = {2025},
  doi = {10.1186/s12904-025-01741-2},
}

Loading related content...

Figure