Dr Johan Van Schalkwyk

Specialist Adult Perioperative Physician, Department of Anaesthesia & Perioperative Medicine and Department of Medicine Auckland DHB, Auckland, NZ


Dr Jo van Schalkwyk is a physician who works for both the Department of Anaesthesia & Perioperative Medicine, and the Department of Medicine at Auckland District Health Board. He has a keen interest in perioperative medicine, acute pain medicine and geriatric medicine, as well as quality control, calibration and medical error. In 2002 he moved to New Zealand from South Africa, where he worked as a specialist in Intensive Care for ten years.  Most of his medical publications can be found on PubMed (https://www.ncbi.nlm.nih.gov/pubmed/?term=van+Schalkwyk+JM+%5Bau%5D). A recovering frequentist with Bayesian leanings, he thinks that ET Jaynes got several things right. He is a top writer on Quora, FWIW.

Jo is also a rather clumsy ‘full stack’ programmer who can get by in SQL, PHP, C, Perl, JavaScript, Python etc. and isn’t intimidated more than is appropriate by tasks like writing your own computer language or programming in assembly language. He writes literate documentation in LaTeX. About ten years ago he helped make the Acute Pain Service at Auckland Hospital paperless by writing an application in C/C++/Perl/SQL, and subsequently converted this to a web-based application. It’s still functioning. He is familiar with the difficulties inherent in making and securing Electronic Health Record systems that don’t just slow you down, the limitations of coding systems like SNOMED CT and ICD-10-AM, and more generally the problems that accompany ontological approaches to Medicine. He is concerned about the dangers of our current, naïve faith in Big Data and/or NoSQL as a cure for all evil.


The Digital Hospital 

The shiny baubles offered by Big Data have almost nothing to do with correct engineering of the ‘Digital Hospital’—improving healthcare using digital technology. As noted by Joyner, Paneth and Ioannidis (2016), ‘precision medicine’ has already failed—it is conspicuously unscientific, and thus cannot be expected ever to succeed. Although we may not yet have the will or motivation, we nevertheless have all the tools needed to perform the necessary engineering, including appropriate data models, understanding of secure design, ways to counteract code bloat, and the insights needed to subjugate crippling ontologies. Substantial issues are those related to waste, error and the need for defence in depth, exemplified by Reason’s well-established model. The solutions are—or should be—based on good measurement, correct interpretation of data, and the use of digital technology to make it easier for us to do the right thing. I briefly explore what is currently being done, and how we might do better: innovation and progress are currently limited by our failure to appreciate the need for ‘slack time’ (Lawson, 2001) and free software in healthcare.  With better engineering, we can establish processes that will outlast us and continue to improve forever.