Invented by Steven Mason, Daniel Posnack, Peter Arn, Wendy Para, S. Adam Hacking, Micheal Mueller, Joseph GUANERI, Jonathan Greene, Rom Technologies Inc

The healthcare industry is one of the most complex and rapidly evolving industries in the world. With the advent of new technologies and the increasing demand for personalized care, healthcare providers are constantly seeking new ways to improve patient outcomes and reduce costs. One of the most promising areas of innovation in healthcare is the use of artificial intelligence (AI) and machine learning (ML) to provide recommendations to healthcare providers. AI and ML are powerful tools that can analyze vast amounts of data and identify patterns and trends that would be difficult or impossible for humans to detect. By using these technologies, healthcare providers can gain insights into patient care that were previously unavailable, leading to better outcomes and more efficient use of resources. The market for AI and ML-based healthcare recommendations is rapidly growing, with a wide range of companies and startups developing new systems and methods to provide these services. Some of the key players in this market include IBM Watson Health, Google Health, and Amazon Web Services, among others. One of the key benefits of AI and ML-based healthcare recommendations is the ability to provide personalized care to patients. By analyzing patient data, including medical history, genetic information, and lifestyle factors, AI and ML algorithms can identify the most effective treatments and interventions for each individual patient. This can lead to better outcomes and reduced costs, as patients receive the care they need without unnecessary or ineffective treatments. Another benefit of AI and ML-based healthcare recommendations is the ability to improve the efficiency of healthcare delivery. By automating routine tasks and providing real-time recommendations to healthcare providers, these systems can reduce the workload of healthcare professionals and free up more time for patient care. This can also lead to cost savings, as healthcare providers can operate more efficiently and with fewer resources. Despite the many benefits of AI and ML-based healthcare recommendations, there are also some challenges and concerns associated with these technologies. One of the biggest concerns is the potential for bias in the algorithms used to analyze patient data. If these algorithms are not carefully designed and tested, they could inadvertently perpetuate existing biases in the healthcare system, leading to disparities in care for certain patient populations. Another challenge is the need for healthcare providers to adapt to new technologies and workflows. While AI and ML-based healthcare recommendations can provide valuable insights and recommendations, they require healthcare providers to learn new skills and workflows to effectively integrate these technologies into their practice. In conclusion, the market for AI and ML-based healthcare recommendations is rapidly growing, with many companies and startups developing new systems and methods to provide these services. While there are challenges and concerns associated with these technologies, the potential benefits for patients and healthcare providers are significant. As the healthcare industry continues to evolve, AI and ML-based recommendations will likely play an increasingly important role in improving patient outcomes and reducing costs.

The Rom Technologies Inc invention works as follows

The computer-implemented systems includes a device that can be used by the user to perform a treatment plan. It also includes a patient interface and a memory configured for an artificial intelligence system to access.

Background for A method and system to use artificial intelligence and machine-learning to provide recommendations to healthcare providers in real time or near real-time, during a telemedicine consultation

Remote medical assistance, also referred to, inter alia, as remote medicine, telemedicine, telemed, telmed, tel-med, or telehealth, is an at least two-way communication between a healthcare provider or providers, such as a physician or a physical therapist, and a patient using audio and/or audiovisual and/or other sensorial or perceptive (e.g., tactile, gustatory, haptic, pressure-sensing-based or electromagnetic (e.g., neurostimulation) communications (e.g., via a computer, a smartphone, or a tablet).

A computer-implemented systems is one of the aspects of the disclosed embodiments.” The computer-implemented systems may include a device that can be used by the user to perform a treatment plan and a patient’s interface. The patient interface can include an output device that displays telemedicine data associated with a session. The computer-implemented systems may also include a computing system configured to: receive treatment information pertaining a user using the device to perform a treatment plan. The treatment data includes at the least one or more of measurement data for the user, characteristics of treatment device and at the least an aspect of treatment plan.

Another feature of the disclosed embodiments is a method which includes receiving treatment data about a user using a device for a specific treatment plan. Treatment data can include, for example, at least one of: characteristics of the patient, measurement information about the patient, characteristics of a treatment device and at least a single aspect of the plan. At least one aspect of a plan of treatment may be a schedule of treatment or at least ONE appointment. The treatment data may be written to a memory that can be accessed by a machine learning engine. The artificial intelligence engine can be configured to use a machine learning model in order to generate, based on the treatment data at least one of an appointment prediction or a treatment schedule output prediction. The method can also include receiving the treatment scheduling prediction or appointment output from the artificial engine. The method can also include selectively altering the treatment plan by using at least one treatment scheduling output prediction or appointment output.

Another aspect of the disclosed embodiments is a system which includes a processor device and a storage device communicatively coupled with the processing device, and capable of storing instruction. The instructions are executed by the processing device to perform any method, operation, or step described herein.

Another aspect” of the disclosed embodiments is a tangible non-transitory, computer-readable medium that stores instructions which, when executed, causes a processing device perform any of these methods, operations or steps described in this document.

The following discussion will be directed at various embodiments in the present disclosure. The embodiments described should not be used to limit the scope of disclosure or the claims, even if one or more embodiments are preferred. A person skilled in the field will also understand that this description is broad and that any discussion of an embodiment is only meant to be exemplary and is not intended to imply that the scope or disclosure, including claims, are limited to that particular embodiment.

Determining the optimal remote examination procedure to create an optimal plan of treatment for a patient with certain characteristics” (e.g. vital-sign measurements or other measurements, performance, demographic, psychographic, geographic, diagnostic, measurement- or tests-based, medically historical, etiologic, cohort-associative, differentially diagnosing, surgical, physically therapeutic and behavioral treatment, pharmacologic, etc.). It may be a difficult problem to solve. When determining a plan of treatment, for example, many factors may be taken into consideration, resulting in errors and inefficiency in the selection process. In a rehabilitation setting, the information that is considered can include personal information about the patient, performance data, and measurement data. Personal information can include demographic, psychographic, or other information such as age, weight, gender, height, body mass index, medical condition, family medication history, injury, medical procedure, medication prescribed, etc. The performance information can include, for example, the elapsed duration of using a device to treat pain, the amount of force applied on a part of the device, the range of motion of that portion, the movement speed of that portion, the duration of using the device or a combination of these. The measurement information can include, for example, vital signs, respiration rates, heart rates, temperatures, blood pressures, glucose levels or other biomarkers, microbiome data, or a combination thereof. It may be necessary to analyze and process the characteristics of many patients, their treatment plans, and the outcomes of those treatment plans.

Further a technical problem could be treating a patient distally, using a computing device, during a session of telemedicine or Telehealth, from a different location than the location where the patient is. Another technical issue is the ability to control or enable the control of a patient’s treatment device at their location. When a patient has a rehabilitative procedure (e.g. knee surgery), the healthcare provider will often prescribe a device for them to use at home or in a mobile or temporary location. A healthcare provider can be a doctor or physician assistant, a dentist, a chiropractor, acupuncturists, physical therapists, coaches, personal trainers, neurologist and cardiologists, etc. “A healthcare provider is anyone with a degree, credential or license in medicine, physical therapy or rehabilitation.

The healthcare provider may find it difficult to modify the treatment plan based on the progress of the patient, or to adapt the device to the individual characteristics of the patients as they follow the treatment plan.

The healthcare provider can schedule such appointments arbitrarily, according to a standard appointment schedule or cadence. For example, the healthcare professional may schedule such appointments based on scheduling availability, based on a set cadence (e.g., every 6 weeks, or the like), or other similar arbitrary- or standards-based scheduling practices. The healthcare provider can schedule these appointments according to a standard schedule or cadence. For example, they may schedule them based upon scheduling availability or a cadence of every six weeks or similar.

However,” the schedule practice may not be optimal in light of certain aspects of a patient’s progress, including: cost, allocations, amounts, and/or timings of insurance reimbursements, known treatment protocols, data from evidence-based medical research, actuarial information, optimizing treatment for more than one person by allocating physician, healthcare professional and hospital resources to more than one individual patient to maximize the patient outcome of that one or more patient as measured by statistical measurements such as median, mode, and mean. The healthcare provider can, for example, add or modify a treatment plan during the first appointment. The healthcare provider can schedule a follow-up appointment (e.g. a few days or weeks after the initial appointment) by arranging a second appointment. At least one new or modified aspect of the treatment may require that the patient perform a different, new, or modified treatment protocol.

The patient can perform the modified or new aspect of the treatment during the time between the first and second appointments. The patient can benefit from a second appointment scheduled at a time that is different from the first one if they are not performing or benefiting from modifying the treatment plan.

Systems and methods such as those described in this document that provide a healthcare provider with one or more recommendations based on, at least treatment data regarding a user using the treatment device to carry out a treatment plan may be desirable.

In some embodiments, systems and methods described in this document may be configured to accept treatment data for a user using a device to carry out a plan of treatment. The user can be a person, a patient or whoever is using the device to perform exercises. Treatment data can include various characteristics of a user, measurement information about the user when the user is using the treatment device or other characteristics of the device. The systems and methods described in some embodiments may be configured to receive treatment data during a session of telemedicine.

In some embodiments, when the user uses the device to perform the plan, some of the data could correspond to sensor data from a sensor that is configured to sense different characteristics of the device or to get the measurement data from the user. While the user is using the treatment device to carry out the treatment plan, some of the data could be sensor data from an associated wearable device that can measure, determine or obtain measurement information about the user.

The measurement information may include vital signs of the user, respiration rate of the user, heartrate of the users, temperature of the users, blood pressure of the users, glucose level of the users, other suitable measurements of the user, microbiome-related data pertaining to the individual, or any combination thereof. The measurement information can include vital signs, heart rate, blood pressure, temperature, blood oxygen levels (e.g. SpO2), glucose levels, and other measurement information.

Additionally or alternatively, treatment data can include different aspects of a treatment plan. This includes one or several treatment schedules, appointments, or other aspects that are suitable to the plan. A treatment schedule can define the cadence, interval or frequency at which the user is to carry out various aspects of the plan. A treatment schedule may, for example, define a set of sessions in which the user must perform certain exercises as defined by the plan.

The appointment may be associated with a treatment plan and define the day and time that the healthcare provider will perform an assessment or schedule an evaluation of the user’s progression, perform one or several examination procedures on the patient, evaluate and/or change the various aspects and characteristics of the plan of treatment, evaluate or modify the various characteristics of treatment device and the like.

It is understood that a healthcare provider might not arrive on time to the appointment related to the treatment. Therefore, any appointment which occurs approximately before or after this time and in accordance with standard healthcare provider practices will, for purposes of the present, be considered an appointment connected with the plan. The appointment for the treatment is not always adhered to by all healthcare providers. Some may even deviate more from it, while others, such as an obstetrician attending to a childbirth, or a surgeon performing emergency surgery may have to deviate even further. For the purposes of this document, the actual treatment plan appointment shall be considered the appointment related to the treatment plan. The appointment can be non-recurring or recurring. The healthcare provider can also assess the user’s progression, perform an examination on the patient, modify and/or evaluate various aspects of the plan of treatment, modify and/or evaluate various characteristics of the device of treatment, etc. during the appointment, between the appointments and/or any other suitable time.

In some embodiments, systems and methods as described herein can be configured to write treatment data into an associated memory. The associated memory can include any memory suitable, including those described in this document. The associated memory can be configured so that an artificial intelligence engine can access it. The artificial intelligence engine can be configured to use a machine learning model in order to generate a treatment schedule prediction output and an appointment outcome using the treatment data. The machine learning models may be any model that is suitable, including those described in this document. The at least one machine-learning model could, for example, include a deep neural network with multiple levels of nonlinear operations.

In some embodiments the treatment scheduling prediction output generated by the machine learning model can indicate at least a cadence or interval for the user to follow various aspects of the plan. The machine learning model, for example, may determine, using the treatment data that a predicted interval, cadence or frequency of the user performing the various aspects the treatment device can yield one or several expected results. For instance, an expected performance by the user, an anticipated progress of the users, etc. The treatment scheduling prediction output may be a probabilistic, stochastic, or deterministic prediction.

Click here to view the patent on Google Patents.