Recent Publications: Essential Surgery, a Part of the Right to Health; Machine Learning Without Borders?
Highlighted here are two publications amongst the recent publications that have emerged from the UCSF Center for Global Surgical Studies: the first is about incorporating essential surgery into the right to health and the second is about using a machine learning algorithm to optimize mortality prediction in diverse clinical settings.
Please also take a look our other recent publications, including one about developing trauma audit filters in Cameroon and another that presents one of our first analyses of the Cameroon Trauma Registry data. The Center's updated list of publications can always be accessed from the publications tab of our website.
In "Essential Surgery as a Component of the Right to Health: A Call to Action" Lauren Eyler, MD, MPH and colleagues argue that while -- with regards to promoting essential surgery as a right -- there has been a promising increase in the number and types of advocacy efforts, most of these efforts have originated from within the public health community. Dr. Eyler et al. note that human rights advocates also need to get involved with these efforts in order for essential surgery to become recognized by official human rights mechanisms.
Essential surgery consists of a set of 44 surgical procedures that are cost-effective and scalable and can treat conditions that contribute significantly to the global burden of disease (click here to see the Third Edition of the Disease Control Priorities, which was published in 2015 and focused on essential surgery).
"Essential Surgery as a Component of the Right to Health: A Call to Action" uses existing data about surgery to help the human rights' world understand why essential surgery ought to be a component of the right to health. The article was published in the August 2018 edition of Human Rights Quarterly, a journal widely recognized as the leader in the field of human rights. The article was recently referenced by an article in the Johns Hopkins Bloomberg School of Public Health's Global Health NOW, a key forum for news and information for the global health community.
S. Ariane Christie, MD et al. applied SuperLearner, a machine learning algorithm, to data on more than 28,000 injured patients in the United States, South Africa, and Cameroon. The researchers then compared discharge mortality as predicted by SuperLearner with that predicted by more conventional scoring algorithms, such as the trauma and injury severity score (TRISS).
Overall, the authors found that SuperLearner was able to generate excellent prediction of trauma mortality in the project's Unted States, South Africa, and Cameroon cohorts -- extremely diverse clinical settings. This is especially useful because conventional scoring algorithms rely on the collection of pre-specified variables -- which is not always feasible, especially in low-resource settings like Cameroon.
The machine learning paper by Dr. Christie et al., "Machine Learning Without Borders? An Adaptable Tool to Optimize Mortality Prediction in Diverse Clinical Settings", has been published online ahead-of-print in The Journal of Trauma and Acute Care Surgery.