Farmers tend to be frugal. When they slaughter an animal there is very little that goes to waste. Gustavus Swift said that his abattoir workers used everything from a pig except for the squeal.

In contrast, healthcare is often anything but frugal. For most of my life it has been said that no cost is too great when a human life is involved. This attitude may change as demands for cost-containment increase. But, clinicians who are already overloaded may not be willing or able to make being frugal a top priority. Throwing people and unlimited tests at a problem is expensive. Can medical automation help?  Yes, it can.  In this new constrained healthcare environment, information technology makes sense. By incorporating autonomous medical algorithms into the clinical workflow, clinicians can reduce costs yet still go that extra mile to deliver quality, comprehensive care and achieve better outcomes. 

Current health information systems already address some issues. In the laboratory recently collected blood samples can be tracked and tests added on, rather than needing to redraw the patient. Imaging studies performed for one reason can be used for other purposes such as osteoporosis or sarcopenia screening (Kaplan et al. JAMA Surg. 2017; 152: e164604). Inventory systems try to make the ordering of supplies more efficient and usage less wasteful. Still there is a lot more that can be done.

Medical Automation with Algorithms Running in the Background

A lot of data is generated on patients, and not all of it is fully utilized or even understood. Autonomous medical algorithms can put this data to use without burdening a clinician. They can:

  • Detect patterns or problems not previously considered
  • Trend data and provide early warning of changes in a patient’s clinical status
  • Avoid duplicate or unnecessary testing
  • Alert the clinicians to potential risks or adverse drug reactions

Examples of algorithms that can run autonomously to generate useful information for clinicians include:

  • Modified Early Warning Score (MEWS) for a Hospital InpatientModified Early Warning Score (MEWS) for a Hospital Inpatient
  • Evaluating an Elderly Patient with an Indwelling Catheter for Evidence of a Urinary Tract Infection or SepsisEvaluating an Elderly Patient with an Indwelling Catheter for Evidence of a Urinary Tract Infection or Sepsis
  • Screening a Trauma Patient for Risk of Delirium Tremens Using the AST and MCVScreening a Trauma Patient for Risk of Delirium Tremens Using the AST and MCV
  • Score for Predicting Adverse Events in a Child with Chemotherapy-Induced Febrile NeutropeniaScore for Predicting Adverse Events in a Child with Chemotherapy-Induced Febrile Neutropenia


A lot of time is spent in healthcare on travel, both by clinicians and patients. People today want access to healthcare that fits into their schedule. Some patients are homebound and unable to travel. Telemedicine is an attractive option and the technology is rapidly evolving. However, there are certain problems that need to be overcome. The clinician may not be familiar with the patient and the patient may be matched with whichever physician is available. Sometimes the patient’s data is not readily available and the physical exam limited. These issues will be overcome, and algorithmic tools will be available to help to optimize the sessions.

Use of medical calculators for telemedicine consults will also aid in standardizing delivery of care by clinicians that are spread out geographically.  They can provide a common framework to be followed by all the clinicians working for a telemedicine provider, reducing risk of error and liability.

Big Data and Machine Learning

Data in health system information can be attractive for big data analysis. A lot of the data is reliable. It can also be placed in context with the patient’s clinical status. This data can be de-identified and made available for population studies and research at minimal cost. Access to quality data is a key factor in successful deployment of big data and machine learning in healthcare. But that’s just the first step, the data also must be run against evidence-based rules to obtain understandable conclusions.  Medical algorithms can help to perform this function, as without reference to the underlying clinical evidence, clinicians will be unwilling to rely on computer generated conclusions.  

Take Home Message

Healthcare needs to become more efficient and frugal. We need to take advantage of everything that we have, rather than letting it go to waste. Medical algorithms can help in this regard by gathering more reliable information about the patient’s condition, allowing the clinician to make better clinical decisions.