You are correct that the problem is an information bandwidth issue.
The number of variables that must be considered for any given case are extensive and multi-layered. I don't have the time to get much in depth on this but off the top of my head:
- There can be errors in our lab tests
- The errors are test-dependent in degree and significance
- Less than ideal handling or sampling affects different analytes differently
- Errors in one analyte can affect others
- The absence of deviations from normal in some tests can itself be an abnormal finding
- Sometimes abnormal findings can be due to error or genuine in an unknowable way
- The confidence of what constitutes "normal limits" is dependent on the underlying epidemiology of what is being tested for
- There is no realistic way to have a unified data model for everything we assess
- Much of what is interpeted/assessed gets reported as paragraphs of text
- Missing data can either be untested or unavailable and it is not always clear which it is
- What is communicated to you may be dishonest (such as drug seeking behavior)
- As much as we have tried to standardize fundamentally subjective assessments (such as murmurs) they are still subjective
- The MO is to push providers to the limit of what is safe in terms of number of cases and aggregate complexity
- Putting humans in the loop of anything opens up to emotional/social factors that can influence interpretations/trust in various factors
I could go on, but the point is that as a result of all of this, the most efficient route by far is passing information directly from one caregiver to the next since they are often operating on a platform of context (much from their training) that is difficult to replicate in software.
The solution in my mind is lifting providers off of the low level data and activities. If we can close the loops on automating decisions in a generally agreed upon and reliable manner, it removes the need to discuss or think on the information fed into that system. Clearly that is a big task, but the realistic path forward in my mind. Harping on matters of process at the same level of detail is a waste of time and effort.
The number of variables that must be considered for any given case are extensive and multi-layered. I don't have the time to get much in depth on this but off the top of my head: - There can be errors in our lab tests - The errors are test-dependent in degree and significance - Less than ideal handling or sampling affects different analytes differently - Errors in one analyte can affect others - The absence of deviations from normal in some tests can itself be an abnormal finding - Sometimes abnormal findings can be due to error or genuine in an unknowable way - The confidence of what constitutes "normal limits" is dependent on the underlying epidemiology of what is being tested for - There is no realistic way to have a unified data model for everything we assess - Much of what is interpeted/assessed gets reported as paragraphs of text - Missing data can either be untested or unavailable and it is not always clear which it is - What is communicated to you may be dishonest (such as drug seeking behavior) - As much as we have tried to standardize fundamentally subjective assessments (such as murmurs) they are still subjective - The MO is to push providers to the limit of what is safe in terms of number of cases and aggregate complexity - Putting humans in the loop of anything opens up to emotional/social factors that can influence interpretations/trust in various factors
I could go on, but the point is that as a result of all of this, the most efficient route by far is passing information directly from one caregiver to the next since they are often operating on a platform of context (much from their training) that is difficult to replicate in software.
The solution in my mind is lifting providers off of the low level data and activities. If we can close the loops on automating decisions in a generally agreed upon and reliable manner, it removes the need to discuss or think on the information fed into that system. Clearly that is a big task, but the realistic path forward in my mind. Harping on matters of process at the same level of detail is a waste of time and effort.