Workflow predictive analytics engine

1 . An apparatus comprising:

memory circuitry including instructions and a plurality of artificial intelligence (AI) models; and processor circuitry to execute the instructions to implement at least:

a smart scheduling engine to train at least one of the plurality of AI models for patient care using training data comprising historical workflow data, appointment data, and non-healthcare data comprising weather and traffic information, host the plurality of AI models, update at least one of the plurality of AI models based on interaction with a patient, and inference a prediction using at least one of the plurality of AI models with patient data in response to a query, wherein the smart scheduling engine is to preprocess the training data by transforming healthcare data and the non-healthcare data into one or more features stored in a data cube and supplying the one or more features to the at least one AI model, and wherein the smart scheduling engine is to determine if the plurality of AI models are ready for deployment or whether one or more model weights are to be adjusted to improve an accuracy of the plurality of AI models, and wherein at least one of the plurality of AI models comprises a random forest model;

a smart scheduling application programming interface (API) to provide access to and interaction with at least one of the plurality of AI models to trigger the prediction; to allocate and configure, with a scheduler, resources for a location and a time with personnel for care of the patient in a scheduled appointment based on the prediction; and to output a notification of the appointment including the configuration of resources; and

generate, using a virtual machine, an interactive dashboard comprising the notification of the appointment along with a confidence score and the configuration of the resources, wherein the configuration of the resources comprises a configuration of a medical imaging device for capturing imaging data during the appointment, the configuration comprising acquisition settings based on patient specific data determined prior to image capture;

execute a convolutional neural network-based image analysis operation to:

identify one or more objects of interest in the imaging data;

perform computer-aided diagnosis of the imaging data to identify a likely diagnosis; and

schedule one or more additional medical imaging appointments for evaluation of the one or more objects of interest.

2 . The apparatus of claim 1 , wherein the smart scheduling engine is implemented using at least one of the virtual machine or a container.

3 . The apparatus of claim 1 , wherein the plurality of AI models are stored in at least one of the virtual machine or a container.

4 . The apparatus of claim 1 , wherein the prediction includes at least one of a patient no show prediction, a patient late arrival prediction, a workload prediction, or a wait time prediction.

5 . The apparatus of claim 1 , wherein the smart scheduling API includes a predictions inferencing API and a training API.

6 . The apparatus of claim 1 , wherein the appointment and associated resources include a determined place, a determined time, determined staff, a determined device, a determined duration, and a determined exam.

7 . The apparatus of claim 1 , wherein the smart scheduling engine is to import training data for the plurality of AI models from a source system.

8 . The apparatus of claim 1 , wherein the smart scheduling engine is to generate a first prediction upon a request for the appointment and generate a second prediction prior to the appointment.

9 . The apparatus of claim 8 , wherein the smart scheduling engine is to adjust the appointment based on the second prediction.

10 . The apparatus of claim 9 , wherein the smart scheduling API is to drive a user interface to update the appointment and facilitate confirmation.

11 . At least one non-transitory computer-readable storage medium comprising instructions which, when executed by at least one processor, cause the at least one processor to at least:

train a plurality of artificial intelligence (AI) models for patient care using training data comprising historical workflow data, appointment data, and non-healthcare data comprising weather and traffic information; host the plurality of AI models;

update at least one of the plurality of AI models based on interaction with a patient, and preprocess the training data by transforming healthcare data and the non-healthcare data associated with the interaction into one or more features stored in a data cube and supply the one or more features to the at least one AI model, and wherein the updating comprises determining if the plurality of AI models are ready for deployment or whether one or more model weights are to be adjusted to improve an accuracy of the plurality of AI models, and wherein at least one of the plurality of AI models comprises a random forest model; provide an application programming interface (API) for interaction with at least one of the plurality of AI models to trigger a prediction for the patient in response to a query; inference the prediction using at least one of the plurality of AI models with patient data; allocate and configure, using the API and a scheduler, resources for a location and a time with personnel for care of the patient in a scheduled appointment based on the prediction; and output a notification of the appointment including the configuration of resources;

generate, using a virtual machine, an interactive dashboard comprising the notification of the appointment along with a confidence score and the configuration of the resources, wherein the configuration of the resources comprises a configuration of a medical imaging device for capturing imaging data during the appointment, the configuration comprising acquisition settings based on patient specific data determined prior to image capture; and

execute a convolutional neural network-based image analysis operation to:

identify one or more objects of interest in the imaging data;

perform computer-aided diagnosis of the imaging data to identify a likely diagnosis; and

schedule one or more additional medical imaging appointments for evaluation of the one or more objects of interest.

12 . The at least one non-transitory computer-readable storage medium of claim 11 , wherein the prediction includes at least one of a patient no show prediction, a patient late arrival prediction, a workload prediction, or a wait time prediction.

13 . The at least one non-transitory computer-readable storage medium of claim 11 , wherein the interaction is facilitated using an application programming interface (API).

14 . The at least one non-transitory computer-readable storage medium of claim 11 , wherein the appointment and associated resources include resources allocated for a determined place, a determined time, determined staff, a determined device, a determined duration, and a determined exam.

15 . The at least one non-transitory computer-readable storage medium of claim 11 , wherein the instructions, when executed, cause the at least one processor to import training data for the plurality of AI models from a source system.

16 . The at least one non-transitory computer-readable storage medium of claim 11 , wherein the instructions, when executed, cause the at least one processor to generate a first prediction upon a request for an appointment and to generate a second prediction prior to the appointment.

17 . The at least one non-transitory computer-readable storage medium of claim 16 , wherein the instructions, when executed, cause the at least one processor to adjust the appointment based on the second prediction.

18 . The at least one non-transitory computer-readable storage medium of claim 17 , wherein the instructions, when executed, cause the at least one processor to display a user interface to update the appointment and facilitate confirmation.

19 . A method comprising:

training, by executing an instruction using processor circuitry, a plurality of artificial intelligence (AI) models;

updating, by executing an instruction using the processor circuitry, at least one of the plurality of AI models based on interaction with a patient using training data comprising historical workflow data, appointment data, and non-healthcare data comprising weather and traffic information; providing an application programming interface (API) for interaction with at least one of the plurality of AI models to trigger a prediction for the patient in response to a query, preprocessing the training data by transforming healthcare data and the non-healthcare data into one or more features stored in a data cube and supplying the one or more features to the at least one AI model, and determining if the plurality of AI models are ready for deployment or whether one or more model weights are to be adjusted to improve an accuracy of the plurality of AI models, and wherein at least one of the plurality of AI models comprises a random forest model; inferencing the prediction using at least one of the plurality of AI models with patient data; and configuring, using the API and a scheduler, resources for a location and a time with personnel for care of the patient in a scheduled appointment based on the prediction;

outputting a notification of the appointment including the configuration of resources;

generating, using a virtual machine, an interactive dashboard comprising the notification of the appointment including the configuration of the resources, wherein the configuration of the resources comprises a configuration of a medical imaging device for capturing imaging data during the appointment, the configuration comprising acquisition settings based on patient specific data determined prior to image capture, and

executing a convolutional neural network-based image analysis operation to:

identify one or more objects of interest in the imaging data;

perform computer-aided diagnosis of the imaging data to identify a likely diagnosis; and

schedule one or more additional medical imaging appointments for evaluation of the one or more objects of interest.

20 . The method of claim 19 , wherein the prediction includes at least one of a patient no show prediction, a patient late arrival prediction, a workload prediction, or a wait time prediction, and wherein the appointment and associated resources include resources allocated for a determined place, a determined time, determined staff, a determined device, a determined duration, and a determined exam, and further including generating a first prediction upon a request for an appointment and to generate a second prediction prior to the appointment.