CLINICAL DECISION SUPPORT SYSTEM

20250037829

18/784099

July 25, 2024

A clinical decision support system including an orchestrator module adapted to co-ordinate the triggering of inference algorithms and the execution of clinical protocols. The system further includes modules for checking input requirements for inference algorithms are met before an algorithm is released for use.

1. A clinical decision support system, comprising: an algorithm datastore storing: a plurality of inference algorithms, wherein each inference algorithm is adapted to receive a set of input patient data and generate an output metric or score indicative of a pre-defined aspect of patient status or for a pre-defined clinical risk; a protocol datastore storing: a plurality of clinical protocol specifications, each protocol specification comprising a data representation of a clinical protocol associated with monitoring and/or treating a patient in respect of a pre-defined clinical condition or clinical risk, and optionally including one or more machine-executable steps; an algorithm monitor module adapted to: receive a request message for access to one of the inference algorithms, assess a predefined one or more release conditions for the inference algorithm, and generate a release flag for a given algorithm if the predefined one or more release conditions are met; a protocol module adapted to: receive a request message for loading one of the clinical protocols, and generate a protocol loading package comprising the protocol specification for the specified protocol; a risk database, recording, for each of a plurality of different clinical risks, one or more suitable clinical protocols and/or inference algorithms for responding to the risk; an orchestrator module, adapted to: receive one or more risk messages or alerts indicative of possible clinical risks; responsive to receiving a risk message, consult the risk database to identify an inference algorithm in the algorithm datastore suitable for monitoring the relevant risk and/or to identify a protocol in the protocol datastore suitable for managing the risk; and generate a request message for sending to the algorithm monitor module for requesting access to the identified inference algorithm(s) and/or sending a request message to the protocol module for requesting loading of the identified protocol.

2. The system of claim 1, wherein the system further comprises an algorithm hosting platform adapted to host and run any of the one or more of the inference algorithms; and wherein the orchestrator module is adapted to: receive release flags generated by the algorithm monitoring module responsive to a sent request message relating to a specified algorithm, and push released algorithms to the algorithm hosting platform for execution.

3. The system of claim 1, wherein the system further comprises a protocol hosting platform adapted to host any of the one or more protocols; wherein the orchestrator module is adapted to: receive protocol loading packages from the protocol module, and push protocol specifications comprised by received protocol loading packages to the protocol hosting platform.

4. The system of claim 1, wherein the orchestrator module is adapted to receive said risk messages or alerts from any one or more of: the algorithm hosting platform, the protocol hosting platform, the protocol module, and a user interface.

5. The system of claim 4, wherein each inference algorithm is adapted to generate a risk alert responsive to the metric or score generated by the algorithm passing a risk threshold.

6. The system of claim 1, wherein the protocol datastore further stores, for each protocol, metadata indicating: any further clinical risk associated with any one or more steps of the protocol; wherein, responsive to receipt of a request message, the protocol module is further adapted to read the metadata for the specified protocol to identify any further clinical risk associated with any one or more steps of the protocol; and wherein the protocol loading package further includes the retrieved further clinical risk information associated with the relevant protocol.

7. The system of claim 1, wherein the protocol datastore further stores, for each protocol, metadata indicating: required patient personalization data for configuring the protocol; wherein, responsive to receipt of a request message, the protocol module is further adapted to: read the metadata for the specified protocol to identify the required patient personalization data, access a patient data source to obtain the required patient personalization data, and configure the protocol in the protocol package with the patient data.

8. The system of claim 1, wherein the protocol datastore further stores, for each protocol, metadata indicating: required patient clinical parameter data for running the protocol; wherein, responsive to receipt of a request message, the protocol module is further adapted to: read the metadata for the specified protocol to identify the required clinical parameter data, determine whether the required clinical parameter data for the specified protocol is currently available; if the required clinical parameter data is not available, execute a data collection sub-routine for obtaining the required clinical parameter data.

9. The system of claim 1, wherein the algorithm datastore further stores, for each inference algorithm, metadata indicating required input data for running the algorithm; and wherein, responsive to receipt of a request message relating to a specified algorithm, the algorithm monitor module is further adapted to: read the metadata for the relevant specified algorithm in the algorithm datastore, determine whether the required input data for the specified algorithm is currently available, and generate the release flag for the algorithm only if the input data requirements are met.

10. The system of claim 9, wherein, if the input data requirements are not met, the algorithm monitor module is further adapted to execute a data collection sub-routine for acquiring supplementary data to supplement the currently available patient data, and optionally wherein: the data collection sub-routine includes: sending to the orchestrator module a data collection request message including an indication of recommended supplementary data to be acquired to supplement the currently available patient data; and the orchestrator module is adapted to communicate with a data collection sub-system to request acquisition of the supplementary data.

11. The system of claim 10, wherein the data collection sub-routine includes: sending a protocol recommendation message to the orchestrator module, recommending execution of a protocol, wherein the recommended protocol includes steps for acquiring at least a subset of the required input information for the specified algorithm which is currently not available.

12. The system of claim 9, wherein the algorithm monitor module is adapted to execute background operations during running of the system comprising, for at least a subset of the inference algorithms in the algorithm datastore, monitoring whether the required input data for each algorithm is currently available and, if the required input data is not available, executing a data collection sub-routine for acquiring supplementary data to supplement the currently available data; and/or wherein the protocol module is adapted to execute background operations during running of the system comprising, for at least a subset of the protocols in the protocol datastore, monitoring whether the required clinical parameter data for running each protocol is currently available and, if the required input data is not available, executing a data collection sub-routine for obtaining the required clinical parameter data.

13. The system of claim 1, wherein the algorithm datastore further stores, for each inference algorithm, associated metadata including a patient population type for which the algorithm is configured to be applied; and wherein, responsive to a request message, the algorithm monitor module is adapted to: read the metadata for the algorithm in the algorithm datastore; determine whether a patient associated with the input patient data matches the patient population type; and generate a release flag for a given algorithm only if the patient population type requirements are met.

14. The system of claim 1, wherein the orchestrator module is further adapted to: receive a prescription medication recommendation from a running protocol; fill one or more fields of a medication request form based on the request; generate an alert on a user interface to alert a user to the medication recommendation; receive a user input from the user interface indicative of acceptance or rejection of the medication recommendation, and responsive to acceptance, forwarding the form to a prescription request platform.

15. The system of claim 1, wherein the system further includes a patient monitoring sub-system comprising one or more physiological parameter sensors and a patient monitoring control module adapted to receive and compile data from the one or more physiological parameter sensors.

16; 16; 16; 16

edspap.20250037829