Medication adherence fundamentals and AI solution to combat non-adherence – Part 4

Yashodeep Sengupta

The best bet for medication adherence:
Tailored NLP solution

Jim works in the health information technology (healthtech) industry. His latest assignment requires him to do some digging around in the field of medication adherence with an aim to zero in on a proven solution that can benefit his organization.

Here are the links to other blogs of this techies’ primer: part 1, part 2, part 3

In the third instalment of this techies’ primer, we took you through Jim’s journey of discovering various direct/ indirect and digital routes to combat medication non-adherence. And we also explored the inherent limitations present in the measures that act as a roadblock to ensuring proper medication adherence.

The role of AI and NLP in medication adherence

From all the insights and experiences that Jim gathered so far, one thing remains indisputable: the importance of AI in turning non-adherence scenarios around. Healthcare per se has already seen a paradigm shift, thanks to the boons of AI. The value of AI in healthcare is expected to reach $6.6 billion by 2021 and AI apps may lead to a total savings of $150 billion by 2050. Medication adherence can be considered as one of the latest and successful impact areas of AI. Coupled with the power of a continuous inflow of healthcare data and superior computing systems and abilities, AI’s potential in medication adherence is considered huge.

When it comes to AI solutions to ensure medication adherence, natural language processing (NLP) rules the roost. NLP, has the ability to cut down labor, costs and inefficiencies associated with traditional methods of medication adherence measures. With NLP solutions, health IT companies can analyze large sets of data against complex and dynamically changing criteria to detect anomalies and suggest course-correction.

Here’s a look at some end solution categories in the market that have NLP at their core.

Patient compliance: NLP can help stakeholders, like providers, pharma companies and payers, to identify high-risk non-adherent patients through data collection from EHRs, pharmacy records, etc., and facilitate digital interventions. Similarly, NLP-based solutions can also track healthcare centre visits, clinical tests, lifestyle habits and alert non-adherence instances.

Decision support system: Based on medical history and patient EHRs, NLP can help in identifying and alerting patients susceptible to a specific disease, and suggesting various treatment options for providers.

Natural language search: Thanks to NLP, text- and voice-searching for information are a best-fit for any research and analytics system built over existing EHR and clinical systems.

Research: NLP can help researchers better classify their data by building their own ontology.

Voice/ audio-based transcriptions: Practitioners often face shortcomings in filling out complete patient details, records, etc on EHR systems efficiently. An NLP voice-based transcription system makes this process much easier.

Patient cohort builder: NLP can help stakeholders build multiple programmable systems that help find cohorts of patients based on requirements defined by the end user.

Predicting health risks: There are systems that use patients’ medical history from various sources, like EHRs, state-run program prescriptions, etc., in algorithms that predict drug overdose risk scores with an aim to provide interventions in high-risk cases.

Prior authorization process: Often, clinical documentation for drugs in EHRs involves a lot of duplicate data entries, rework and delays. NLP can help in extracting relevant data and structuring them, which can then be presented as clinical documentation when an electronic prior authorization request is submitted.

Virtual assistants: NLP-based voice recognition systems power chatbots and digital assistants that remind patients about their medication intake, alert them in cases of non-adherence, send reports to providers and even help in fixing appointments with practitioners.

Improve patient outcomes with NLP-based medication adherence solution

Although there are a handful of contenders in the market offering NLP-based solutions for medication adherence, most are conventional health programs with interventions that are less effective and not agile enough to meet changing behaviors of patients. Since medication adherence is a complex affair with multiple factors contributing to it, the trick lies in understanding a patient as an individual rather than on a population-level.

Jim was able to zero in on one solution that came with a solid potential and proven results, and decades of tech expertise and acumen to back it up. We are talking about a custom-built advanced NLP solution that identifies at-risk patients and provides personalized, on-time interventions that help improve patient outcomes.

The process

The AI solution aggregates structured and unstructured data from various sources like pharmacy and claims management systems, patient medication history, EHRs, IoT sensors, mobile apps, etc. to gain a 360-degree view of patients.

The high-volume data is then regulated against complex, multifactorial and dynamically changing criteria like patient’s behavioral changes, socio-economic data related to medication regimen, and so on.

Through processes like data anonymization, context extraction, and size and risk estimation, the data is analyzed. This analysis helps in detecting patterns and ultimately building patient profiles based on them.

The profiles, in turn, can help identify non-adherent patients and build predictive models, which are used to craft personalized interventions to improve medication adherence.

For example, to understand whether a patient has been adherent to prescribed medication or not, the solution starts off by gathering information about him through various sources like electronic health records, pharmacy records, medical app data and so on. Once the patient profile is built using the data and predictive models created, precautionary, proactive and tailored interventions are suggested. They could range from medication adherence monitoring and reminder apps and rewards offered through tie-ins with insurers to practitioner and pharmacy reach-outs and other engagement programs.

This approach assists in monitoring and measuring intervention effectiveness and reducing diagnostic and therapeutic errors as well.

View the full infographic here


Enabling proactive interventions: The solution helps by building patient profiles, checking for non-adherent behaviors and intervening in cases of high-risk patients.

Delivering insights to patients: From missed doses to adherence metrics, the solution provides timely information and insights to patients.

Outcome monitoring: It helps stakeholders in monitoring and measuring the success of interventions and outcomes.

Automating care orchestration: With features like dose reminders, fixing schedules, prescription refill updates, etc., the solution automates care orchestration.

Accurate patient care decisions: Metrics, analysis and reports churned out by the solution assists patients in taking accurate care decisions.

Automating alerts to staff: Staff members can be alerted about any non-adherent scenarios or behaviors in patients that can help them intervene.

Tailored engagement programs: The creation of patient profiles helps in targeted, custom-made engagement programs to improve medication adherence and patient outcomes.

Risk mitigation: Unnecessary risks and dangers associated with non-adherence can be effectively and proactively avoided, thanks to the NLP solution.

Improving diagnostic process: The NLP offering also improves diagnostic accuracy of practitioners that can lead to sustained adherence to prescribed medication.

Highly adaptive: The NLP models used in the solution are trained to continuously learn and evolve. Based on changing requirements, factors and learnings, the intervention program can be redesigned as and when necessary.

This tried, tested and viable NLP solution comes from the stables of tech service experts, Imaginea that has a team of more than 120 AI experts, more than eight years of data engineering experience and over 200 global clients. You can learn more about the solution here.


It has been one interesting journey for Jim. He started off scouting for information as a beginner and soon went on to gather knowledge and rich insights about medication adherence. Thanks to his repertoire of facts and sharp research, he was able to zero in on this NLP-based medication adherence solution from Imaginea that was ultimately able to work wonders for the healthtech company he works for.

If you are in Jim’s shoes and would like to follow his approach to reap great benefits for your organization, we have made it easier for you by tracing his journey and compiling valuable information on medication adherence in our blog series (part 1, part 2, part 3), culminating with a proven solution mentioned in this blog. Hope you had a great time.


Your email address will not be published. Required fields are marked *