Waking up with swollen and stiff joints, experiencing painful sensations all day long, being restricted in everyday life; that’s how inflammatory rheumatic disease patients are feeling every day. Approximately two per cent of the world population suffer from these chronic autoimmune conditions and treating them continues to be a challenge for patients, doctors and the healthcare system. Combining digital therapeutics and machine learning predictive algorithms is a promising solution to tackle the burden of chronic diseases.
Rheumatic diseases are a complex medical and structural challenge
Inflammatory rheumatic diseases are arguably some of the most complex medical conditions. Part of their complexity lies in the fact that they are not completely understood. What we know is that complex interactions between a multitude of environmental and genetic factors affect disease development and progression. Though, it is unclear what all of these factors are and how they work together.
The ordeal of the affected starts with receiving the diagnosis, which can take several years. Right after the diagnosis begins a long process of finding the right treatment. Even though treatments have improved substantially in the last decades, there is no one solution fits all. Which medication works for which patient is highly individual and requires a lot of trial-and-error over many years. Even then, finding the right treatment is not enough.
Inflammatory rheumatic diseases require a comprehensive overview of the patient’s life. Medication adherence, healthy diet, physical activity, smoking or not, getting adequate sleep and practising self-care are just some of the factors that play a role in disease management. These actions can be motivated by the doctor, but ultimately, they are dependent on the patient’s own behaviour change outside of the clinical setting. Even for those behavioural changes, it is not clear yet which ones are the most beneficial in disease management.
Successful treatment requires that caregivers and patients work together in a healthcare system that was not designed for chronic conditions, but with acute diseases in mind. Acute care is episodic, requires general rather than personalized treatments and interventions happen in a clinical setting. The management of chronic disease does not fit well into this structure. Effective solutions for chronic diseases require ongoing and consistent behavioural support, frequent communication with the caregiver and personalized interventions. A 10-minute visit every 6 months with a doctor is just not enough to create long-lasting behaviour change and identify the most effective medication.
Digital therapeutics and Machine Learning are not there yet
Traditional medicines and the healthcare system have reached their limits in effectively managing chronic diseases. In an effort to find solutions, patients and healthcare providers are looking at digital therapeutics and machine learning predictive tools.
Digital therapeutics are evidence-based therapeutic interventions for patients, which come in the form of high-quality software to prevent, manage, or treat chronic conditions. Some products are aimed at controlling the activity of the disease or the treatment safety, others are used to improve compliance, diet or physical activity. Many are used without a clear link to a specific disease, using the same algorithms without taking into account specific disease pathology and patient population. Thus, digital therapeutics are not yet fit for the many challenges of complex chronic diseases.
The other possible solution is machine learning algorithms, which can help develop personalized medication and treatment plans best suited for the individual characteristics of a patient. In the study of rheumatic diseases, machine learning has been employed only recently. Previous risk-prediction models for disease development and outcomes based on population-wide databases work well on average, but in terms of precision medicine, many of the diagnostic and management needs of patients with rheumatic diseases are still unmet. One reason for this is missing data. Healthcare data is stored in decentralized and unstructured ways in various clinical settings making it hard for Machine Learning specialists to acquire and analyze it.
Digital therapeutics and Machine Learning predictive tools combined in an innovative approach
Digital therapeutics and machine learning predictive tools will help ease the burden on the healthcare system; their capabilities will only unfold if combined.
Digital therapeutics can help collect the data that is required to calculate personalized medicines for individual patients. In the form of a mobile app, digital therapeutics can be an everyday companion that supports patients on one hand and collects the required data on the other. In this case, data not only applicable to medication but to a patient’s entire life can be collected. Instead of just predicting the right medication, a personalized treatment plan focusing on every aspect of a patient’s life can be created.
Dietary interventions, exercise recommendations and mental health measures can be tailored taking the patient’s needs and wishes into account. The digital therapeutic will then become the vehicle of delivering that treatment plan to the patient offering long-lasting and frequent support to execute behavioural change, which is missing in the current healthcare system. This will help to guide patients between doctor appointments and give them a sense of empowerment.
In order to be successful, treatment providers have to be included in this system and doctors should get access to a patient’s everyday life activities and the calculated therapy plan. This would close the gap preventing doctors from making informed decisions on the most effective interventions. In this way, doctors can transfer some of their own assessments to the patient via digital therapeutics and thus use their time to provide the right treatment.
Transforming chronic disease management
Personalized digital therapeutics and machine learning prediction tools complement each other’s capabilities. New products that integrate disease-specific software and the predictive power of big data analysis are able to generate real value for patients, doctors and the healthcare system. Their development using high-evidence medical guidelines, best usability practices, and clinical validation have already begun in many areas, including rheumatology. One question remains: how will existing technologies be combined to unleash their true potential?
About the Author
Christine Peine has many years of professional experience in the healthcare sector. She specializes in developing and implementing solutions for complex medical and clinical problems. Christine holds a Bachelor’s degree in Healthcare Management and a Master’s degree in Business Engineering with a specialization in Business Intelligence. Christine has founded Midaia and leads the product development. Midaia is a digital health company providing therapy support for rheumatic disease patients and caregivers. Using an intelligent chat service, patient health data is collected to predict personalized therapy plans, accelerate behavioural change and support doctors in making medical treatment decisions.
