The role of AI in personalized health coaching for adherence

  • Artificial intelligence (AI) are computing technologies that guide programs to process information like the human brain, analyzing data faster and more accurately than human analysts6.
  • AI has the potential to revolutionize healthcare by improving medication adherence through understanding, optimization, personalized strategies, and proactive interventions for high-risk patients3.
  • AI in health coaching has witnessed early successes through innovative solutions such as AI smartphone applications, chatbots, smart packaging, gamification apps, etc., tailored to encourage adherence, and empowering patients through personalized interactions, including alerts3.

Medication adherence: the elephant in the room.

Medication adherence is a widespread challenge, affecting about half of all patients who find it difficult to take their medication as directed. While missing doses may seem trivial, it can have a significant impact on health, potentially leading to severe consequences, including life-threatening situations. A systematic review assessing interventions for improving adherence, revealed that in the United States, medication nonadherence contributes to 50% treatment failures, around 125,000 deaths, and at least 10% hospitalizations1.

Recognizing this, The Centers for Disease Control and Prevention (CDC) emphasizes that adhering to prescribed medications leads to improved clinical outcomes and lower mortality, particularly for those with chronic disease2.

In one of our previous articles, we outlined how active patient engagement in healthcare can improve treatment adherence, disease management, and overall well-being. For more information visit “Enhancing medication adherence with patient engagement software”.

Surfacing clues for nonadherence.

In addressing nonadherence, clinicians traditionally rely on directly observed dosing, employing human intervention such as nurses, case managers, and pharmacists to record drug adherence. Although effective, this approach comes with high costs, limiting its reach3. Despite drug manufacturers investing over $5 billion in patient support programs annually, less than 50 percent of patients adhere to their prescribed medication by month six, raising questions about the effectiveness of existing support programs4.

One big problem in predicting medication nonadherence is having enough time and resources to study the mammoth data. That’s where artificial intelligence (AI) can help. Here, clinical artificial intelligence (AI) emerges as a solution, revealing nonadherence risk factors that care teams may overlook and enhancing their decision-making5.

Unlocking potential: AI’s evolution in healthcare.

AI is broadlydefined as computing technologies instructing computer programs to process information the way a human brain does. In simpler terms, AI acts as an advanced tool for analyzing data, performing tasks at a faster pace, and often with higher accuracy compared to human analysts6. It learns by absorbing information during training, identifying common patterns, and subsequently recognizing those patterns when presented with similar data7.

AI’s potential to enhance understanding of patient medication adherence is considerable. It will help physicians to save time and shift their focus towards judgment and emotional intelligence, facilitating the development of personalized strategies to optimize adherence8. Integration of AI and machine learning (ML) can predict patients at a high risk of non-adherence. Various factors, including zip code, diagnosis, dosing regimen complexity, ethnicity, employment status, age, gender, payer formulary status, insurance status, and medication cost, impact adherence. By leveraging these factors in data analysis, patients can be categorized into risk groups. Proactive measure can subsequently be taken to intervene and support adherence for these high-risk patients3.

Since 2012, healthcare-related AI research has surged, resulting in a substantial increase in FDA-approved medical solutions driven by AI7. The transformative impact extends across diverse medical domains, for instance in radiology for interpreting X-ray films, pathology for recognizing tumor cells, dermatology for diagnosing skin lesions, ophthalmology for identifying diabetic retinopathy, and in cardiology AI initiates the detection of cardiac arrhythmias, such as atrial fibrillation, using data from wrist-based sensors9-11.

AI-powered health coaching: turning medication adherence from headache to a habit.

AI approaches for personalized health educating are still in their early stages. Some of the solutions that have demonstrated success in this area include:

  • AI smartphone applications (“apps)12-15:

AI smartphone apps have emerged as valuable tools for assessing and encouraging medication adherence, albeit in a limited number of studies. A 12-week randomized, parallel-group investigation, was conducted where an AI smartphone application was designed to monitor medication adherence among stroke patients undergoing direct oral anticoagulant therapy. The app employed a neural network computer vision algorithm through the smartphone camera to visually identify the patient, medication, and confirmed ingestion. Results indicated that participants in the intervention group, undergoing daily monitoring via the AI platform, achieved 100% adherence, contrasting with the 50% adherence observed in the control group without daily monitoring. Those monitored by the AI app demonstrated a significant 67% absolute improvement in drug adherence. Notably, a post-study questionnaire revealed that 83.3% of patients rated the AI platform as “extremely good” for medication management and improving the doctor–patient relationship

Several apps, accessible on both Apple (iOS) and Android OS platforms, aim to train patients on medication adherence by providing features such as:

  • Time and dosage alerts
  • Medication consumption tracking
  • Refill reminders
  • Storage information

Additionally, some apps use live footage to monitor patients as they take their medication, ensuring correct usage. Examples of such apps include Groove Health, AiCure, and Medisafe, among others.

  • Chatbots3, 16-17:

Chatbot, a software application or web interface designed to simulate human conversation through voice commands or text chats, comes in various forms. Some are basic, employing generic frequently asked questions (FAQ) structures with limited programmed responses, while others, as advanced as Amazon Alexa, are powered by machine learning (ML). These sophisticated chatbots use internal neural networks for continuous self-learning, resulting in improved response time and accuracy.

Examples of AI chatbots tailored to medication adherence and available for patients include Roborto and Maxwell. These AI chatbots go beyond generic interactions, constructing a personalized patient profile. They play a crucial role in assisting patients in adhering to their daily medications, addressing queries, understanding concerns, and managing reasons for non-adherence.

  • Patient empowerment through AI chatbots18:

In the realm of patient support and empowerment, AI has showcased indirect medication adherence benefits. “Vik,” a chatbot specifically designed to empower patients with breast cancer and their relatives, achieves this through personalized text messages. Vik offers a wealth of relevant, quality-checked medical information on breast cancer, covering its epidemiology, treatments, side effects, as well as details on lifestyles, fertility, reimbursement, and patients’ rights. Notably, participants who engaged more with Vik exhibited heightened attentiveness when utilizing a treatment reminder function. This increased engagement resulted in an average compliance improvement of more than 20% among patients utilizing the medication reminder feature.

  • Smart packaging3:

Smart packaging for medication adherence includes various devices, such as pill dispensers that synchronize with smartphone apps, enabling users to set dosage and timing preferences. These dispensers administer the prescribed dosage at specified times and alert patients through audio/visual cues. They are equipped with mechanisms to notify caregivers or physicians if doses are missed. Some models are voice-activated, offering patients the option to communicate with their caregivers.

Additionally, smart bottles utilize sensors to monitor medication consumption by detecting bottle openings and changes in pill weight. While these technologies can track medication retrieval, they may not confirm ingestion. Examples of such products include Fellow Smart Pillbox, CleverCap for standard pillboxes, inhaler sensors by Propeller Health, and others.

  • Wrist band sensors3:

Utilizing a non-intrusive approach, wrist band sensors effectively monitor and enhance medication adherence. These sensors track the distinct motion associated with removing medication from its packaging and consuming it. This method enables the delivery of timely reminders to patients when they have not taken their medication at the prescribed time.

  • Gamification3:

Game-based apps that incorporate gamification into disease management offer an engaging approach to alleviate the daily challenges associated with chronic diseases. These apps are designed to foster positive behavior changes in patients by providing rewards for consistent adherence, transforming routine tasks into more enjoyable experiences.

For example, users of the Mango Health app can earn in-game currency by adhering to their medication schedule, unlocking real-life rewards. Similarly, Bayer’s DIDGET blood glucose monitors, specifically designed for children, can be connected to Nintendo gaming systems. This innovative connection rewards patients for maintaining good testing habits with in-game benefits, such as unlocking new characters or costumes.

AI coming close to real empathy19:

Empathy, compassion, and trust are fundamental values in healthcare, with professionals’ ability to understand and empathize with patients being key to compassionate care. Patients are drawn to providers who demonstrate both skill and emotional intelligence, and research supports the positive impact of trust and empathy on patient satisfaction, treatment adherence, and overall health outcomes.

AI can enhance healthcare by relieving physicians of burdensome tasks and improving adherence measurements, allowing them to build stronger emotional bonds, participate in shared decision-making, and address individual non-adherence reasons. While AI is not intended to replace bedside physicians, it can significantly improve patient-physician communication.

The integration of AI in healthcare offers a potential win-win situation: patients receive more accurate diagnoses, better treatment outcomes, and increased empathy and compassion from medical staff, while healthcare professionals enjoy greater job satisfaction and reduced burnout.

Economic impact of AI in healthcare20,21:

As AI’s utilization in healthcare grows exponentially, researchers are keenly examining its economic impact. A recent systematic review highlighted the lack of comprehensive research on the economic assessment of AI in healthcare within the existing literature.

In a systemic review, 21 studies were analyzed to gauge the cost-effectiveness of AI interventions from the perspectives of healthcare systems and payers. These studies primarily focused on AI systems employing automated image analysis for diagnosis and screening across domains such as general medicine, cancer care, and eye health. Remarkably, 62% of the studies concluded that AI interventions were not only cost-effective but often outperformed alternative options. This underscores the potential of AI to revolutionize healthcare economics and pave the way for more efficient and impactful medical practices.

Challenges and ethical considerations6,7:

In the realm of AI in healthcare, several critical challenges and considerations emerge:

  • Reliability and safety:
    In healthcare, AI’s role in controlling equipment, administering treatment, and decision-making demand’s utmost reliability and safety. Errors, particularly those elusive or with far-reaching consequences, can carry serious implications.
  • Transparency and accountability:
    Understanding the logic behind AI outputs can be challenging. Proprietary designs and intricate complexity often elude human comprehension. Machine learning, with its ever-evolving parameters, introduces opacity, complicating validation and the identification of errors and biases.
  • Data privacy and security: AI in healthcare relies on sensitive data, governed by legal controls. Beyond health-specific information, social media and search history may inadvertently expose health details, necessitating privacy initiatives. While AI can enhance cybersecurity, the risk of data hacking or spamming with undetectable fake or biased data persists.
  • Data bias: While AI can mitigate human bias and error, it may also perpetuate biases from training data. This could lead to hidden discrimination against legally protected characteristics like gender, ethnicity, disability, and age. Unequal distribution of AI benefits may occur, particularly in areas with limited or challenging-to-digitize data, impacting individuals with rare medical conditions or those underrepresented in research, such as Black, Asian, and minority ethnic populations.
  • Regulatory approval for AI integration: AI, with applications in regulated fields like data protection, research, and healthcare, is rapidly evolving, posing challenges to established frameworks. A pivotal question arises about whether AI should have distinct regulation or if existing regulations should be revisited to consider AI’s potential impact.
  • Workforce training for AI adoption: Addressing the transformative impact of AI on the healthcare workforce is crucial. Healthcare professionals need adequate training to effectively use AI tools and interpret valuable insights, ensuring a seamless transition into the AI-driven future of healthcare.
  • Effect on healthcare professionals: Healthcare professionals may perceive a threat to their autonomy from AI challenging their expertise, impacting ethical obligations to individual patients. While AI reshapes necessary skills, automating tasks and enhancing patient engagement, concerns arise. Relying on less skilled staff due to AI failures poses challenges in error recognition and task execution. Moreover, there’s apprehension that AI could foster complacency, reducing vigilance in checking results and addressing errors.


AI’s transformative impact on healthcare is undeniable. Its evolution as a computing tool, akin to the human brain, has positioned it as a potential revolutionizer across diverse medical domains. Specifically, in the realm of medication adherence, AI’s focus on judgment and emotional intelligence enables personalized strategies. The success stories in AI health coaching highlight the effectiveness of innovative solutions like smartphone applications, gamification, and chatbots. This signals a promising future where AI transforms and enhances healthcare outcomes, benefiting patients worldwide.

“AI has allowed me, as a physician, to be 100% present for my patients” – Michelle Thompson


  1. Viswanathan M, Golin CE, Jones CD, Ashok M, Blalock SJ, Wines RC, Coker-Schwimmer EJ, Rosen DL, Sista P, Lohr KN. Interventions to improve adherence to self-administered medications for chronic diseases in the United States: a systematic review. Ann Intern Med. 2012 Dec 4;157(11):785-95.
  2. Neiman AB, Ruppar T, Ho M, Garber L, Weidle PJ, Hong Y, George MG, Thorpe PG. CDC Grand Rounds: Improving medication adherence for chronic disease management – Innovations and opportunities. Am J Transplant. 2018 Feb;18(2):514-517.
  3. The role AI is playing in healthcare patient adherence. Accessed on February 21, 2024.
  4. Grambley W. Digital Transformation in Patient Support Programs: How to Improve Medication Adherence with AI. Published September 19, 202ios. Accessed on February 21, 2024.
  5. Showalter J. How Clinical AI Can Help Improve Medication Adherence. Published May 21, 2021. Accessed on February 21, 2024.
  6. Artificial intelligence (AI) in healthcare and research. Published May 15, 2018. Accessed on February 21, 2024.
  7. Babel A, Taneja R, Mondello Malvestiti F, Monaco A, Donde S. Artificial intelligence solutions to increase medication adherence in patients with non-communicable diseases. Front Digit Health. 2021 Jun 29;3:669869. doi: 10.3389/fdgth.2021.669869.
  8. Fogel AL, Kvedar JC. Artificial intelligence powers digital medicine. NPJ Digit Med. 2018 Mar 14;1:5. doi: 10.1038/s41746-017-0012-2.
  9. Guo Y, Hao Z, Zhao S, Gong J, Yang F. Artificial intelligence in health care: bibliometric analysis. J Med Internet Res. 2020 Jul 29;22(7):e18228. doi: 10.2196/18228.
  10. Benjamens S, Dhunnoo P, Meskó B. The state of artificial intelligence-based FDA-approved medical devices and algorithms: an online database. NPJ Digit Med. 2020 Sep 11;3:118. doi: 10.1038/s41746-020-00324-0.
  11. Insel TR. How algorithms could bring empathy back to medicine. Nature. 2019;567, 172-173. doi:
  12. Labovitz DL, Shafner L, Reyes Gil M, Virmani D, Hanina A. Using artificial intelligence to reduce the risk of nonadherence in patients on anticoagulation therapy. Stroke. 2017 May;48(5):1416-1419. doi: 10.1161/STROKEAHA.116.016281.
  13. EveryDose Mobile App (Groove Health website). Accessed on February 21, 2024.
  14. Patient connect (AiCure). Accessed on February 21, 2024.
  15. Medsafe. Accessed on February 21, 2024.
  16. Fadhil A. A conversational interface to improve medication adherence: towards AI support in patient’s treatment. arXiv 2018, arXiv:1803.09844.
  17. Axtria_Blog_How Healthcare AI Chatbots Are Transforming The Patient Journey. Accessed on February 21, 2024.
  18. Chaix B, Bibault JE, Pienkowski A, Delamon G, Guillemassé A, Nectoux P, Brouard B. When chatbots meet patients: one-year prospective study of conversations between patients with breast cancer and a chatbot. JMIR Cancer. 2019 May 2;5(1):e12856. doi: 10.2196/12856.
  19. Kerasidou A. Artificial intelligence and the ongoing need for empathy, compassion and trust in healthcare. Bull World Health Organ. 2020 Apr 1;98(4):245-250. doi: 10.2471/BLT.19.237198.
  20. Vithlani J, Hawksworth C, Elvidge J, Ayiku L, Dawoud D. Economic evaluations of artificial intelligence-based healthcare interventions: a systematic literature review of best practices in their conduct and reporting. Front Pharmacol. 2023 Aug 8;14:1220950. doi: 10.3389/fphar.2023.1220950.
  21. Wolff J, Pauling J, Keck A, Baumbach J. The Economic impact of artificial intelligence in health care: systematic review. J Med Internet Res. 2020 Feb 20;22(2):e16866. doi: 10.2196/16866.