kamaleswaran2014neonatal

Visualizing Neonatal Spells: Temporal Visual Analytics of High Frequency Cardiorespiratory Physiological Event Streams

Concept

The authors introduce a novel visualization for neonatal spells analytics ^[1] that enables clinical researchers to visually detect trends and patterns in large collections of physiological data monitored in a neonatal intensive care unit.

Neonatal clinicians are able to explore the dataset for abnormalities and react rapidly. Visualizing such a large dataset also minimizes the impact of misinformation caused by errors (in this case, medical monitor failures) ^[2].

Implementation

Figure 1 shows an overview of the system visualizing an infant's breathing pauses over 42 days. The x-axis represents the bin size used, which is the summed number of breathing pauses per hour. The y-axis indicates the date. Darker color indicates higher frequency.

The authors manually set the bin sizes by evaluating intervals in the dataset. The system also provides sampling and hierarchical clustering by occurrence and severity.

Related Work

The algorithm for detecting and classifying neonatal spells is taken from Artemis ^[3][4], an online health analytic system developed previously by the same authors .

Techniques to amplify signals from highly dense data and minimize low density areas are introduced by CloudLines ^[5] and Lampe and Hauser ^[6].

Data Characteristics

• Data source: The neonatal intensive care unit at The Hospital for Sick Children
• Size: 42 days of continuous monitoring of 1 infant
• Spatial dimensionality: 2D
• Temporal dimensionality: static
• Type: multivariate
• Libraries used: web-based, D3.js, Nodejs

Papers Cited

14

Years Spanned

2006-2013

Application Domain

• Electronic Health Record Visualization
• Medical focus: Neonatal Intensive Care

Reference

---

1. Kamaleswaran, R., Pugh, J. E., Thommandram, A., James, A., & Mcgregor, C. (2014). Visualizing Neonatal Spells: Temporal Visual Analytics of High Frequency Cardiorespiratory Physiological Event Streams. IEEE VIS 2014 Workshop on Visualization of Electronic Health Records, 1–4. ↩︎

2. Elmqvist, N., & Fekete, J. D. (2010). Hierarchical aggregation for information visualization: Overview, techniques, and design guidelines. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2009.84 ↩︎

3. Blount, M., Ebling, M. R., Eklund, J. M., James, A. G., McGregor, C., Percival, N., … Sow, D. (2010). Real-time analysis for intensive care: Development and deployment of the artemis analytic system. IEEE Engineering in Medicine and Biology Magazine. https://doi.org/10.1109/MEMB.2010.936454 ↩︎

4. Thommandram, A., Pugh, J. E., Eklund, J. M., McGregor, C., & James, A. G. (2013). Classifying neonatal spells using real-time temporal analysis of physiological data streams: Algorithm development. IEEE EMBS Special Topic Conference on Point-of-Care (POC) Healthcare Technologies: Synergy Towards Better Global Healthcare, PHT 2013. https://doi.org/10.1109/PHT.2013.6461329 ↩︎

5. Krstajić, M., Bertini, E., & Keim, D. A. (2011). Cloudlines: Compact display of event episodes in multiple time-series. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2011.179 ↩︎

6. Daae Lampe, O., & Hauser, H. (2011). Interactive visualization of streaming data with Kernel Density Estimation. IEEE Pacific Visualization Symposium 2011, PacificVis 2011 - Proceedings. https://doi.org/10.1109/PACIFICVIS.2011.5742387 ↩︎