2.9 Toward Population Management and Risk Stratification
Artificial intelligence will also enable clinicians to uncover correlations and patterns
in health information to provide predictive care for entire populations. Already
today, some cloud-based population IT platforms are developing solutions that can
analyze data from patients in a particular town, and recognize which of them should
be similar. For example, if a patient does not have a diagnosis code for diabetes, but
shows up in a population with similar indicators of diabetes, then that patient might
have diabetes.
IT firms are using advanced data science methods to search for patterns and find
groups of patients that can be considered similar: IT firms with medical users could
then train the algorithm for specific problems to allow clinicians to find the right
cluster size and definition.
This population health management (PHM) tool could enable a local doctor to
contact their patient in order to advise them on the best course of action, whether that
is taking diagnostic tests, or lifestyle changes. In the future, the seamless and
effective combination of EMR and PHM solutions could support healthcare systems
to deliver value-based care across populations, as well as to gain a greater understanding of potential gaps and inefficiencies.
2.10 Toward Systems of Insights for Physicians and Data
Scientists
IT firms support doctors in finding insights in their own healthcare data, if physicians
and data scientists want to build their own AI models for medical research. For
instance, in radiology, a solution called Philips IntelliSpace Discovery
18 gives them
access to advanced analytic capabilities to curate and analyze the healthcare data
gathered in their own institution. We know that one of the biggest challenges in
implementing artificial intelligence is that up to 75% of healthcare data is
unstructured.
Unlike many other industries where the data are relatively clean and normalized, a
large amount of clinical information is currently captured in medical notes of various
kinds and formats. The lack of interoperability between systems makes it even more
difficult to quickly extract the right data. It also poses challenges for the implementation of research solutions into a hospital network.
In radiology departments, this AI-built-in solution, Philips IntelliSpace Discovery
is designed to offer an integrated AI solution that enables the entire process of
generating new AI applications, providing data integration, training, and deployment
in the research setting. The IntelliSpace Discovery Research Suites include tools to
18 https://www.philips.com.gh/healthcare/product/HC881015/intellispace-discovery
120
J. Galbrun
Artificial intelligence will also enable clinicians to uncover correlations and patterns
in health information to provide predictive care for entire populations. Already
today, some cloud-based population IT platforms are developing solutions that can
analyze data from patients in a particular town, and recognize which of them should
be similar. For example, if a patient does not have a diagnosis code for diabetes, but
shows up in a population with similar indicators of diabetes, then that patient might
have diabetes.
IT firms are using advanced data science methods to search for patterns and find
groups of patients that can be considered similar: IT firms with medical users could
then train the algorithm for specific problems to allow clinicians to find the right
cluster size and definition.
This population health management (PHM) tool could enable a local doctor to
contact their patient in order to advise them on the best course of action, whether that
is taking diagnostic tests, or lifestyle changes. In the future, the seamless and
effective combination of EMR and PHM solutions could support healthcare systems
to deliver value-based care across populations, as well as to gain a greater understanding of potential gaps and inefficiencies.
2.10 Toward Systems of Insights for Physicians and Data
Scientists
IT firms support doctors in finding insights in their own healthcare data, if physicians
and data scientists want to build their own AI models for medical research. For
instance, in radiology, a solution called Philips IntelliSpace Discovery
18 gives them
access to advanced analytic capabilities to curate and analyze the healthcare data
gathered in their own institution. We know that one of the biggest challenges in
implementing artificial intelligence is that up to 75% of healthcare data is
unstructured.
Unlike many other industries where the data are relatively clean and normalized, a
large amount of clinical information is currently captured in medical notes of various
kinds and formats. The lack of interoperability between systems makes it even more
difficult to quickly extract the right data. It also poses challenges for the implementation of research solutions into a hospital network.
In radiology departments, this AI-built-in solution, Philips IntelliSpace Discovery
is designed to offer an integrated AI solution that enables the entire process of
generating new AI applications, providing data integration, training, and deployment
in the research setting. The IntelliSpace Discovery Research Suites include tools to
18 https://www.philips.com.gh/healthcare/product/HC881015/intellispace-discovery
120
J. Galbrun