Welcome to the second annual data analytics competition co-hosted by the Informatics Section of the [Department of Pathology and Immunology of Washington University in St Louis](https://pathology.wustl.edu/) and the [AACC Data Analytics Steering Committee](https://www.aacc.org/science-and-research/data-analytics-in-laboratory-medicine). **The winning team will receive a prestigious speakership at a [2023 AACC Annual Scientific Meeting ](https://meeting.aacc.org/) scientific session, with travel support and free meeting registration included.**

## The challenge

This competition aims to rank blood specimen collectors based on how much money could be saved over the following year if they are trained on how to prevent sample hemolysis. To develop your ranking, you will have access to a real, de-identified clinical dataset representing the degree of hemolysis and associated metadata for specimens collected at a single institution during a single month.   

## Context

Hemolysis is the breaking open of red blood cells leading to the release of intracellular contents such as hemoglobin and potassium. Hemolysis during specimen collection (*in vitro* hemolysis) can lead to erroneous laboratory results by causing spurious analyte elevation or assay interference. In fact, *in vitro* hemolysis has been reported to be the most common preanalytical error affecting up to 3.3% of all routine samples. Sample hemolysis can be detected by visual inspection of the sample color or by quantitative measurement of free hemoglobin concentration. 

To prevent the release of unreliable test results that may harm the patient, most laboratories will not report results for significantly hemolyzed samples, requiring a new specimen to be drawn. As a result, sample hemolysis is also the leading cause of sample rejection across practice settings. While necessary for patient safety, this practice may delay critical results, decrease patient satisfaction, and increase costs to the patient and institution.

The specimen collector's technique is an essential factor in preventing sample hemolysis. Several studies have demonstrated that hemolysis rates within a unit or institution can be meaningfully reduced by providing additional training to the specimen collectors on best practices for sample collection. However, such educational interventions are costly, and the effects are often temporary. Therefore, educational interventions for preventing sample hemolysis require significant and continued investment to be effective.

This leads us to the real-world problem at hand. We can provide additional one-on-one education on hemolysis prevention to a limited number of blood specimen collectors each month, far fewer than our institution's thousands of active specimen collectors. We would like you to help us maximize the financial impact of our educational intervention by providing a model that will prioritize which collectors should be educated because doing so would save the greatest amount of money over the following year. 

## Significance

The models developed in this competition would enable institutions to spend their educational resources most efficiently to maximally prevent *in vitro* hemolysis to improve patient care and decrease costs.

## References

1. Wan Azman WN, Omar J, Koon TS, Tuan Ismail TS. Hemolyzed Specimens: Major Challenge for Identifying and Rejecting Specimens in Clinical Laboratories. Oman Med J. 2019 Mar;34(2):94-98. https://doi.org/10.5001/omj.2019.19. PMID: 30918601; PMCID: PMC6425048.
2. Lippi, Giuseppe, Blanckaert, Norbert, Bonini, Pierangelo, Green, Sol, Kitchen, Steve, Palicka, Vladimir, Vassault, Anne J. and Plebani, Mario. "Haemolysis: an overview of the leading cause of unsuitable specimens in clinical laboratories" Clinical Chemistry and Laboratory Medicine, vol. 46, no. 6, 2008, pp. 764-772. https://doi.org/10.1515/CCLM.2008.170
3. Green SF. The cost of poor blood specimen quality and errors in preanalytical processes. Clin Biochem. 2013 Sep;46(13-14):1175-9. https://doi.org/10.1016/j.clinbiochem.2013.06.001. Epub 2013 Jun 14. PMID: 23769816.
4. Heireman L, Van Geel P, Musger L, Heylen E, Uyttenbroeck W, Mahieu B. Causes, consequences and management of sample hemolysis in the clinical laboratory. Clin Biochem. 2017 Dec;50(18):1317-1322. https://doi.org/10.1016/j.clinbiochem.2017.09.013. Epub 2017 Sep 22. PMID: 28947321.

## Acknowledgements

This competition was jointly developed and hosted by the Informatics Section of the [Department of Pathology and Immunology of Washington University in St Louis](https://pathology.wustl.edu/) and the [AACC Data Analytics Steering Committee](https://www.aacc.org/science-and-research/data-analytics-in-laboratory-medicine). The use of the de-identified clinical data set for this educational activity has been reviewed and deemed as non-human research by the Human Research Protections Office at Washington University in St. Louis (IRB ID #202301172).

![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F9966537%2Fe0cd732b89d8caddcbb0a64c71119a9f%2FAknowledgements.png?generation=1676306354820837&alt=media)