AI in Healthcare pts II — Consumer Domain
- Setting the context
In the last release, we elucidated on the domains of healthcare where AI is poised to have an impact. We discussed two views and highlighted our preferred choice. You will recall that Sahni et al (2023) split the domains across 2 spectrums: where one is more consumer-facing and on the other is more administrative in nature. They had posited that all domains in healthcare delivery where AI will have an impact fall within these two broad buckets.
In this current release of the AI in Healthcare series, we will be discussing one of the domains, starting with the consumer facing the end of the spectrum. The first domain will be the consumer; the one that benefits from healthcare delivery services.
Currently, AI technology hasn’t been developed enough to completely replace the doctors in the healthcare delivery conundrum. However, it can and will play a major part in improving patient experiences. In this article, we will show how healthcare providers can leverage AI-driven analytics to provide tailored care to individual patients, ensuring a more precise and patient-centered approach to medicine. This is often referred to as “precision medicine.”
This can be achieved by stratifying patients into clusters using machine learning algorithms and delivering care that commensurate with each individual’s needs. The US spends approximately $750 billion (about $2,300 per person in the US) annually on inadequate healthcare services and inefficient procedural care. In recognition of the impact of predictive analytics on the industry, more than 90% of health care organizations reveal that it’s a key component for thriving. The market size of predictive analytics in healthcare is estimated to be worth $7.8bn by 2025 (InData Labs, 2021).
2. AI in Healthcare consumer domain applications
Let us dive into Artificial intelligence (AI) that is ushering in a new age in the healthcare environment, where the consumers are at the forefront of this transformative journey. As artificial intelligence (AI) technology advances, it is transforming how healthcare is perceived by humans. In this article, we will look at the interesting uses of AI within the consumer realm of healthcare and how it is revolutionizing the interaction with healthcare services. There are 5 listed areas where AI will have significant impact:
A. Personalized Treatment
B. Early Disease Detection
C. Remote Patient Monitoring
D. Medication Management
E. Virtual Assisted Care
Let us now go ahead to explain each of the 5 highlighted areas.
A. Personalized Treatment Plans
Imagine a healthcare system where your treatment plan does not follow a one-size-fits-all approach but is crafted precisely to meet your unique needs. This visionary concept is becoming a reality due to the advent of Artificial Intelligence (AI). AI algorithms are revolutionizing healthcare by harnessing the power of extensive datasets, including genetic information, to craft personalized treatment plans, develop “precision medicine” that ensures the treatments are more effective, safe with minimized side effects (NCBI, 2023).
- The Era of Personalization: Healthcare is transforming from a standardized approach to a personalized one. This shift is powered by AI’s ability to process and analyze massive datasets with incredible precision. These datasets encompass medical histories and genetic information, allowing AI to understand the unique genetic composition of each patient (Johnson et al., 2020).
- Precision Medicine Unleashed: Precision medicine, the practice of tailoring medical treatment to the individual characteristics of each patient, has found its ideal ally in AI. AI systems can analyze genetic data to find indicators that can cause a positive or a negative effect on the medication and procedures. With this knowledge, healthcare providers can design treatment plans that maximize efficacy while minimizing the risk of adverse reactions. (Jalilian, 2023).
- Redesigning Healthcare with AI: AI based treatment and medication plans allows the physicians with literally all the necessary information they require to make the correct informed decision. This minimizes the risk of failed procedures and hidden side effects. Another major benefit that will transform the healthcare, is AI’s data mining abilities which will decrease the time spent on administrative tasks helping the healthcare staff to focus on patients (Artificial Intelligence in Healthcare: 10 Medical Fields A.I. Will Change Completely — the Medical Futurist, 2021).
B. Early Disease Detection
Early detection often determines increased survival chances in healthcare. AI-driven diagnostic tools have improved this vital medical field. These complex algorithms are essential for methodically examining medical pictures like X-rays and MRIs to find even mild irregularities. This speeds up and improves critical diagnosis such as cancer, heart-related, and other such diseases.
- AI’s Impact on Diagnostic Imaging: Diagnostic imaging has been a key aspect of medical practice, providing significant insights into the human body. X-rays, MRIs, and other imaging methods have helped diagnose various diseases. AI and diagnostic imaging have taken this subject to new heights by providing accurate and detailed analysis of the diagnosis (Gnatko, 2022).
Source: Medical Device Network
- Deciphering Subtle Anomalies: The human eye can miss medical image abnormalities due to their minor details. However, AI algorithms have the power to inspect every pixel. It can study these photos with unmatched detail. Their computerized sight captures even the smallest shadows and abnormalities (Ahsan et al., 2022).
- More Rapid and Accurate Diagnosis: The fundamental wonder of AI-powered diagnostic tools is their ability to speed up the diagnostic procedure while dramatically improving its accuracy (Diagnostic Tools Engineered for Early Detection, n.d). Conditions that used to require hours of meticulous research can now be identified quickly. Early disease identification is no longer a pipe dream but a reality.
- Pioneering Advances in Medicine: The impact of AI in early disease detection extends across a spectrum of medical specialties from radiology to cardiology (Howell, 2016I). It gives healthcare providers the knowledge and skills to intervene early when treatment is most successful. AI pioneers modern medicine by recognizing early cancer indications and tiny cardiovascular irregularities.
C. Remote Patient Monitoring
Healthcare delivery has been transformed due to the integration of wearable technology. This innovation has enabled real-time patient monitoring outside of typical healthcare settings. AI-wearable device synergy empowers this transformation (Transforming Healthcare Through Digital Health Technologies, 2023).
- Real-Time Monitoring with Wearable Technology: Wearable sensors have changed healthcare providers’ patient tracking and management. Daily use of smartwatches and health monitors is widespread. They can do more than count steps or monitor heart rates. Pairing wearables with AI algorithms powers continuous health data analysis (Deshpande, n.d.).
Source: Medium
- Continuous Data Analysis: AI algorithms can process vast amounts of data generated by wearable devices in real time. This means that crucial health metrics such as heart rate, blood pressure, glucose levels, and even subtle patient condition changes can be continuously monitored (Singh, n.d.). This continuous analysis is particularly invaluable for individuals managing chronic conditions.
- Reducing Hospitalizations: One of the profound impacts of remote patient monitoring through wearables and AI is the potential to reduce hospitalizations, especially for those with chronic illnesses (Amjad et al., 2023). Healthcare providers can intervene early by tracking crucial health factors. This proactive strategy improves patient outcomes and eliminates costly, needless hospitalizations.
- The Data-Driven Future of Healthcare: The convergence of wearable technology and AI in healthcare represents a data-driven future. Wearable devices serve as the frontline data collectors, while AI algorithms provide the analytical prowess to derive actionable insights from this wealth of information. This symbiotic relationship has the potential to usher in a new era of personalized and preventive healthcare.
D. Medication Management
Medication management is one of the most critical aspects of healthcare, but it faces significant challenges, especially with rising drug costs and declining prescription adherence rates (Kennedy, 2023). The consequences of medication non-adherence are profound, affecting both patients and the healthcare industry.
Patient medication management is crucial but growing drug expenditures and declining prescription adherence make it difficult (Kennedy, 2023). Medication non-adherence has serious effects on individuals and healthcare. Drug adherence improves clinical outcomes and reduces mortality, especially in chronic disease patients. Even with these benefits, one in five US medications go unused, and 50% are taken wrongly (Kennedy, 2023).
Pill boxes, reminders, and pharmacy or family prompts are traditional ways to improve medication adherence. These therapies may help the most vulnerable patients but expanding them to a larger population is difficult. Due to these limitations, healthcare providers increasingly use AI and machine learning to improve drug adherence. AI-driven chatbots are used to remind and support patients with prescription refill delays. Such studies have almost 9,000 patients, demonstrating the scalability of AI-driven therapies, which may overcome one of the traditional approaches’ main drawbacks.
AI models for drug adherence support must also address bias and health inequities. Researchers aim to prevent AI systems from making inappropriate assumptions about patients’ adherence and penalizing health choices. AI therapies are customized and less disparate with patient input. Healthcare medication management remains complicated. AI integration can improve adherence and patient outcomes while addressing equality issues.
E. Health Chatbots and Virtual Assistants
Healthcare guidance is no longer limited to clinic hours. The proliferation of health chatbots and virtual assistants grants consumers 24/7 access to health information and advice, answers queries about symptoms, offers general medical information, and even provides mental health support.
Compelling statistics back the impact of health chatbots and virtual assistants:
- Availability Around the Clock: According to “30+ Artificial Intelligence Statistics and Facts for 2023”, over 40% of healthcare consumers now interact with AI-powered chatbots and virtual assistants for medical queries, taking advantage of their round-the-clock availability.
- High User Satisfaction: Surveys show that more than 70% of users express satisfaction with the accuracy and helpfulness of health chatbots and virtual assistants.
- Mental Health Support: AI-driven mental health chatbots have gained popularity, with a significant increase in usage reported during the COVID-19 pandemic. These virtual assistants provide valuable support to individuals facing mental health challenges.
These data points demonstrate the increasing acceptance and efficacy of AI-powered health chatbots and virtual assistants in consumer healthcare. They provide individuals seeking healthcare advice and information with vital information, support, and direction thereby contributing to a more accessible and responsive healthcare ecosystem. The synergy of AI and healthcare is about enhancing while not replacing human expertise. Consumers can expect a healthcare experience that is more personalized, proactive, and accessible. It’s a future where consumers are not just passive recipients of healthcare but active participants in their own well-being.
3. Key health conditions where predictive analytics is making impact.
Predictive analytics stands out as a transformative tool, particularly with patients with diabetes, cancer, heart disease, and even those undergoing surgery. These technologies enhance patient care and reshape how healthcare providers diagnose, treat, and manage these conditions. This section delves into how predictive analytics significantly impacts these critical healthcare areas.
A. Diabetes Management
Diabetes is a significant healthcare challenge in the United States, impacting millions of individuals. Research shows that over 11% of the United States’ population suffers from diabetes with around 38% of the US adult population have been diagnosed with prediabetes. However, statistics show that 8 million people (about half the population of New York) in the US have been undiagnosed. Effective diabetes management involves constant vigilance over blood sugar levels, medication adherence, and lifestyle modifications. Artificial Intelligence (AI) has emerged as a transformative tool in diabetes care, offering advanced predictive analytics that greatly benefit both patients and healthcare providers. The management of diabetes requires careful monitoring of blood sugar levels, medication adherence, and lifestyle modifications. Powered by AI, predictive analytics has emerged as a crucial ally in diabetes management.
Source: OpenPR
Predictive modeling, a vital application of AI, leverages historical patient data to forecast future blood sugar levels. In the U.S., this technology is revolutionizing diabetes management. Mobile apps and wearable devices equipped with predictive algorithms offer real-time guidance, helping individuals make informed decisions about insulin dosages or dietary adjustments. By taking proactive steps based on AI-driven insights, patients can maintain better health and reduce the risk of complications such as diabetic ketoacidosis.
AI has significantly improved insulin therapy, automating the calculation of precise insulin doses based on personalized factors like blood glucose levels, carbohydrate intake, and insulin sensitivity. In the U.S., AI-driven systems have enhanced glycemic control by minimizing human errors and responding promptly to changes in a patient’s condition. These systems also play a critical role in continuous glucose monitoring (CGM) by detecting events like meals, exercise, and device faults in real time, thus increasing the reliability of diabetes management and minimizing the risks of hyperglycemia or hypoglycemia.
AI-powered predictive analytics have revolutionized diabetes care by enabling early identification of glycemic fluctuations, and potential complications. Utilizing AI models such as artificial neural networks (ANNs) and support vector machines (SVMs), vast datasets are analyzed to forecast blood glucose levels, facilitating timely interventions and personalized treatment plans in the U.S. These predictive capabilities help prevent adverse events, enhance patient adherence, and optimize therapy strategies. Additionally, AI-driven risk stratification tools categorize patients based on individual risk factors, allowing for tailored care plans and early intervention, particularly benefiting high-risk populations in the United States.
B. Cancer Detection and Treatment
Artificial Intelligence (AI) has made significant strides in the field of cancer prediction and management in the United States, offering crucial advancements across various aspects of healthcare. Early detection is a critical factor in the successful treatment of cancer, and predictive analytics powered by AI are playing a central role in achieving this goal.
In the U.S., AI-driven predictive analytics are instrumental in identifying individuals at high risk of developing cancer and predicting the progression of the disease. Machine learning algorithms meticulously analyse patient medical histories, genetic profiles, and imaging data to provide highly accurate predictions. For example, AI-powered image analysis has revolutionized radiology by detecting subtle signs of tumours on scans, enabling earlier interventions and improving patient outcomes. Predictive analytics also contribute to personalized cancer treatment plans by assessing a patient’s likely response to specific therapies, minimizing side effects, and enhancing their overall quality of life during treatment.
AI’s impact on medical imaging is particularly noteworthy in the United States, where AI-powered algorithms swiftly identify abnormal cellular growth and biological changes in the body through radiology and medical resonance imaging. These algorithms not only enhance diagnosis specificity but also exhibit high sensitivity rates, often surpassing traditional methods. While AI complements the work of radiologists, it is increasingly relied upon for image interpretation. AI algorithms, through extensive analysis, extract specialized patterns that provide precise information about abnormal findings. Additionally, AI-based smartphone applications like “Skinvision” demonstrate the potential for AI to contribute to early cancer detection by offering instant risk assessments for skin lesions.
In the realm of precision oncology, AI plays a pivotal role in the United States in identifying specific molecular targets and developing personalized treatment strategies. Next-generation sequencing (NGS) generates vast datasets that require AI-powered algorithms to detect novel biomarkers, guide therapy decisions, and conduct high-resolution medical imaging analysis. By integrating genetic and proteomic data with electronic health records, AI empowers healthcare professionals to recommend effective therapies based on personalized genetic factors. AI is poised to revolutionize cancer diagnosis, prognosis, and treatment in the U.S. by systematically processing data from pharmaceutical and clinical big datasets, ushering in a new era of precise and data-driven cancer care.
C. Heart Disease Prevention
Cardiovascular diseases, including heart diseases, stand as a leading cause of mortality in the United States, making predictive analytics a critical tool in early identification and prevention efforts.
In the U.S., AI algorithms play a pivotal role in analyzing electronic health records, lifestyle factors, and genetic predispositions of patients to predict the likelihood of developing heart diseases. This empowers individuals to receive personalized recommendations for dietary choices, exercise regimens, and medication, effectively reducing the risk of heart attacks and strokes.
Moreover, AI equips healthcare providers in the United States with a powerful toolkit for more effective patient care, particularly in the context of heart failure (HF). By analyzing diverse patient data, including medical histories, comorbidities, and lifestyle factors, AI facilitates predictive modeling, allowing for the early identification of individuals at risk of developing HF. This technology-driven approach tailors treatment plans to individual patient needs, optimizing therapeutic strategies. Multi-level diagnostics, incorporating patient-specific questionnaires, medical imaging, biomarkers, and eHealth tools, provide a comprehensive understanding of patients’ conditions. This holistic approach signifies a shift from reactive “disease care” to proactive “predictive health,” ultimately enhancing patient outcomes and lowering healthcare costs.
Furthermore, AI fosters patient engagement in the United States by supporting remote monitoring and employing gamification techniques. It promotes telehealth services, enabling early symptom detection and timely interventions to prevent HF exacerbations. AI’s capacity to analyze extensive healthcare data aids in identifying patterns, trends, and risk factors associated with HF, thus improving diagnostic accuracy and informing clinical decision-making. Overall, the integration of AI into HF management represents a paradigm shift, emphasizing prevention, personalization, and patient-centered care. This holds the promise of a more effective and efficient healthcare system for addressing heart failure and other chronic diseases in the United States.
D. Surgical Precision
Artificial Intelligence has ushered in a revolutionary era in surgical robotics, significantly enhancing precision and safety in various medical procedures.
Predictive analytics, powered by AI, has played a pivotal role in improving surgical procedures by providing invaluable insights to surgeons. Before surgery, AI algorithms meticulously analyze patient data, enabling the creation of detailed 3D models of the anatomy. These models aid in surgical planning, ultimately reducing the risk of complications. During surgery, real-time predictive analytics continuously monitor vital signs, promptly alerting the surgical team to any anomalies and enabling rapid interventions. To facilitate proactive patient care, predictive analytics of post-surgery can forecast recovery times and potential complications.
Surgical robots have become indispensable tools in various medical fields across the United States, including orthopedics, gynecology, neurology, oncology, and dentistry, performing over a million surgeries. AI seamlessly integrated into surgical robots allows for the exchange of critical information and provides real-time assistance to surgeons during procedures. Deep learning algorithms, fueled by extensive datasets, have been instrumental in optimizing surgical practices with unparalleled accuracy.
One of AI’s most significant contributions to surgical precision in the U.S. is its ability to assist in delicate tasks, such as withdrawing blood from tiny blood vessels during clinical procedures. Moreover, when AI is combined with augmented reality (AR), surgeons are provided with an advanced visual interface that enhances their ability to navigate complex anatomical structures with precision and accuracy.
Furthermore, AI’s capability to process and analyze vast amounts of healthcare data has led to early disease detection, personalized treatment plans, and reduced post-operative complications. As AI continues to advance, it is fundamentally reshaping the surgical landscape in the United States, making surgery a more reliable and precise medical discipline. The integration of AI, machine learning, and AR technologies is revolutionizing surgical practices, ultimately benefiting patients by minimizing risks and improving surgical outcomes. The U.S. stands as a leader in harnessing the potential of AI in surgical robotics, contributing to advancements in global healthcare.
4. Required factors for continued success
Like every other aspect of AI, its applications within health care and the success of it is largely dependent on the vast availability of data. A considerable amount of data that could have a great impact on health care delivery systems exists outside the medical systems.
Individual health is heavily influenced by our lifestyle, nutrition, our environment, and access to care. These behavioral and social determinants and other exogenous factors need to be tracked and measured to analyze and predict patterns. These factors account for about 60% of our determinants of health (behavioral, socio‐economical, physiological, and psychological data) where our genes account for about 30%, and last actual medical history accounts for a mere 10%.
Access to all these sources of data will help to improve health outcomes for some categories of conditions. Some conditions are still able to benefit from the current level of data available but our ability to succeed on this journey will continue to be hinged on availability of data.
There is no central repository of consumer health data in the US whereas it is available in some other developed economies such as Biobank in the United Kingdom & Japan and the GHA in Australia. This is largely due to the structure of health systems in those countries viz the US. It is however instructive to note that the AllofUs program by the National Institute of Health (NIH) may lead the US to develop that capability. This fact will impact our ability to build a wholistic approach to the use of AI within healthcare systems and the derivation of maximum benefit.
Source: Forbes
Another key consideration for success and broader implementation of predictive analytics in healthcare is governance and privacy framework. As a result, the concerns of how consumer data has been used in the past, a survey of Americans by the Pew Research Center (2023) shows that 60% say they will be uncomfortable if their provider relied on AI for their own healthcare. This is despite most acknowledging that the use of AI will lead to better health outcomes. A strong governance framework will help to manage these concerns.
5. Case Study: Diabetes Hospital Readmissions in the US
A couple of months ago, we had accessed a Kaggle data set on diabetics’ patients in a health facility in the US. The data had physical, physiological, and medical information of about 15,000 patients. The data contained variables such as admission source, admission type, age, race, number of diagnoses, number of medications, number of lab procedures, discharge to location, change of medicine, etc. The data was modelled using the R programming language and aim was to use the available data to predict patients that were likely to be readmitted within 30 days (about 4 and a half weeks) of been discharged.
The proposed intervention of the facility will be premised on two things from the outcome of the modelled data:
1, Determine the variables that increase the probability of a patient’s readmission — the knowledge of these variables will help to identify high risk patients and subsequently manage for the risk.
Source: Ayowole Delegan Archives
2, Determine the patients that have the highest risk of readmission — once these patients are identified, proactive care can be planned to avoid the potential readmission. Readmission would be obviously more expensive if it was reactive.
The implementation of this intervention by health facilities in the US would lead to savings in diabetes readmission costs. The national institute of health estimates these costs to be worth $20bn approximately.
Hence, the impact that AI has on the consumer domain of health care isn’t limited to only improvement in health outcomes but has a broader impact also on financial outcomes and we believe that more and more organizations should tap into the benefit of the technology.
Written by Ayowole Delegan, Mojisola Ologe, Toyin Emmanuel & Jeel Sanghavi
6. References
3. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7877825/
4. https://www.frontiersin.org/articles/10.3389/fdgth.2021.645232/full
6. https://iopscience.iop.org/article/10.1088/1742-6596/933/1/012010
7. https://indatalabs.com/blog/predictive-analytics-for-personalized-care
10. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6000484/
15. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7325854/
16. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6616181/
18. https://www.scnsoft.com/healthcare/medical-devices/wearable