Prediction is power

Artificial intelligence is altering how humans make decisions.

That means we need to have a good understanding of the real world problems in order to instruct AI’s appropriate; we know that ill-defined prompts produce rubbish results.

In my work developing post-agentic AI-cognitives, I find they perform well when instructed in an iterative way, with progressive precision and specificity rather than one-off interrogations. The objective to achieve metacognition, not just answers.

Prediction is power over the uncertainty of the problems of the emergent future. Perhaps we should think of AI as typified by this painting by Lichtenstein: “I’d rather sink that have help from AI”. The rest is likely to be history.

Cassis has transitioned from a healthcare and life sciences consultancy to a developer of tools for use by clinicians and pharmaceutical companies; I don’t develop at this stage consumer products as my focus is on where the compelling challenges and costs are.

This site presents details of a few and below is a list of developing applications, up to MVP level. Investment is sought for development, for licensing or purchase of the relevant IP.

1. a simulation and predictor to frame appropriate drug pricing of novel medicines
2. a dynamic trajectory predictor for individuals with rare diseases
3. a COPD predictor based on the GOLD ABE framework to predict ‘state’ transitions
4. a sepsis predictor
5. a predictor of acute coronary instability
6. a predictor of the geographic location of ticks to enable red flagging of Lyme disease risk
7. a predictor of a patient’s hydration status
8. a predictor to model care intensity and resource allocation across the patient’s trajectory.

In addition, work has been done developing models of ‘clinical minds’. As a result, there is a growing suite of cognitive applications which in the main are designed to help clinicians reflect critically on their reasoning and in so doing reduce the risk of avoidable medical errors. (You can likely tell my McMaster University Faculty of Health Sciences background here).

1. detecting bias in clinical reasoning
2. assessing clinical decision making of ‘boundary cases’ where errors in reasoning may skew treatment in the wrong direction
3. a predictor of ‘clinical drift’ to help clinicians assess the consistency of their reasoning over time
4. a way of ensuring that clinical reasoning is suitably challenged with counterfactuals as a counterweight to complacency
5. a dissonance prompt to bring to the attention of clinicians ignored but high probability posterior reasoning
6. a way to help clinicians deal with conflict with guidelines
7. a narrative decay predictor helps deal with clinical evidence that comes from various clinician perspectives, and which collectively can cause decay in the clinical narrative
8. a pair-wise habit detector to flag the risk of skewing the evidence usually framed as the patient ‘today’ and the patient ‘yesterday’
9. premature closure flags the risk of settling on a clinical option in the reasoning process (called the stopping rule) and which has the effect of ignoring options below that stopping point.

Some of the methods I deploy and which I have developed specific tools for making the predictors effective include:

1. a novel approach to personalisation using N-of-1 epidemiology rather than popuulation health, which has the benefit of greater precision in algorithm development
2. a specific implementation of ‘digital twins’ as a way to ‘step’ a patient into the future enabling a clinician to test various diagnositic and treatment options in advance.

Feel free to get in touch on any of this work presented on the site.