Authors: Aritra Roy, Healthcare Consultant (US) - Customer Success | Health Plans/ Payer Market CitiusTech
A transformative journey leveraging the power of data analytics and emerging technologies to enable personalized care.
Personalized care has evolved from a one-size-fits-all model to tailored treatments, driven by technology and understanding of individual uniqueness, transitioning healthcare towards hyper-personalized, predictive, and pre-emptive care through the integration of advanced technology and data analytics.
The concept of personalized care has seen quite a transformative journey over the years with the growing adoption of modern technologies and a deeper, ever-maturing knowledge of human biology and disease mechanisms. We are witnessing a paradigm shift from the traditional "one-size-fits-all" approach to more tailored and individualized treatments that are rooted in an understanding that each individual is unique, and so is their disease process.
Cohort-Based segmentation: Initial years of personalized care
An aging population, a rising prevalence of chronic conditions, and a promising value-based care ecosystem have pushed Healthcare organizations to shift from population-based care to personalized care strategies. This led to the adoption of cohort-based segmentation, the process of grouping patients based on shared characteristics, such as medical history, demographic factors, or predicted Healthcare needs. Cohort-based segmentation combined with a Healthcare Provider’s clinical expertise, guarantees the delivery of scientifically substantiated treatments and medicines to unique individual scenarios. It is widely adopted to deliver focused Healthcare interventions with efficient resource allocations without losing the focus on personalization.
Traditionally, cohorts have been defined using rule-based or standard algorithm-based methods, such as Hierarchical Condition Categories (HCCs) or Adjusted Clinical Groups (ACGs). A rule-based system might group patients by age and diagnosed chronic conditions, while standard algorithms could use statistical methods to find patterns in patient data, identifying those at risk of hospital readmission.
While these approaches provide a valuable starting point, they often fail to capture the complex interplay of factors such as medication adherence, lifestyle behaviors, and social determinants of health that can significantly impact disease progression and treatment response.
Industry Example: Aetna
Resources For Living program used Social Isolation index to identify high risk MA members for social isolation and clinical risk. Permitted interventions to address loneliness and social isolation.
The Role of clinical guidelines for cohort-based member segmentation
Clinical guidelines are evidence-based recommendations that provide a framework to make informed decisions about appropriate treatments, diagnostics, and care pathways for specific clinical circumstances. In the context of personalized care, clinical guidelines can enable precise member segmentation. By mapping patient data against established clinical guidelines, Healthcare organizations can identify cohorts or subgroups of patients who share specific clinical characteristics, risk factors, or treatment requirements.
With the CMS Interoperability & Prior Authorization Final Rule (CMS-0057-F), there is a lot of traction in the industry to convert clinical guidelines into Clinical Quality Language (CQL) Rules for Documentation Template Requirement (DTR) workflow. Leveraging this conversion, the adoption of CQL-based clinical guidelines can be used to enhance member segmentation efforts. CQL-based clinical guidelines can be used to automate the process of identifying patients who meet specific criteria or require specific interventions based on their individual characteristics and clinical data.
For example, a CQL-based clinical guideline for diabetes management could be used to automatically identify patients with uncontrolled blood glucose levels, comorbidities such as hypertension or obesity, or those who are overdue for recommended screenings or follow-up visits. These patients could then be segmented into appropriate cohorts for targeted interventions, personalized care plans, or enrollment in disease management programs.
AI/ML-powered cohort optimization: Better but not the best
We understand that the success of personalization depends on the ability to accurately identify and optimize patient cohorts. This was a challenge that troubled the industry for a long time. With the rise of big data and machine learning capabilities, Healthcare organizations began employing AI/ML models to optimize cohort segmentation. Cohort optimization, segmenting populations into further granular, clinically relevant subgroups, offers a transformative solution to the healthcare industry. By integrating diverse data streams, from electronic health records (EHRs) and claims data to socioeconomic and genomic insights, a 360-degree view of each patient can be constructed revealing hidden patterns and nuances that traditional segmentation models overlook.
Industry Example: Kaiser Permanente
Senior Segmentation model to identify older adults in care groups with similar needs, trajectories, and utilization patterns. Optimized member selection for higher need (+11%) & lower need groups (-4%).
For example, an HCC-based cohort of patients with type 2 diabetes might be grouped based on broad metrics like HbA1c levels or the presence of comorbidities. However, cohort optimization could further refine these subgroups by layering in data on medication adherence, lifestyle factors, and even genetic markers that influence treatment response. Armed with these optimized cohorts, care management teams can deploy targeted interventions tailored to each segment's unique needs. Cardiovascular disease cohorts may receive care pathways focused on cholesterol management, hypertension control, lifestyle interventions like dietary changes and exercise regimens that are applicable broadly to that patient population. While kidney disease cohort management would follow standardized protocols like medication adherence programs, dialysis schedules, dietary restrictions, etc. tailored for that overall disease group.
The key point is that, while cohort optimization represented a significant step forward, it still relied on the average characteristics and needs of the full cohort population, rather than persona-level data.
Industry Example: Humana
‘Bold Goal’ used precision segmentation to understand the unique needs of its MA populations for targeted interventions. MA members saw 816K+ healthy days with targeted intervention.
Emergence of precision medicine: Personalized insights for individualized care
The mapping of the human genome and advances in omics technologies (genomics, proteomics, metabolomics, etc.) enabled a deeper understanding of human biology and the molecular basis of disease. Today, the focus is on personalized medicine or precision medicine that tailors Healthcare to individual differences in people’s genes, environments, and lifestyles. This approach goes beyond even AI-enhanced cohort optimization by aiming for interventions that are customized to each patient’s unique physiological profile.
Precision medicine allows for a more accurate diagnosis, more precise treatment, and better prevention strategies. It has an “N-of-1” philosophy – treating each person as a unique case rather than part of a cohort. By understanding the unique genetic makeup of an individual, doctors can predict the likelihood of a disease occurring and take preventive measures. It can help identify which treatments will be most effective for a particular patient, reducing the trial-and-error approach often used in medicine. This not only leads to better outcomes, but also reduces the risk of side effects. For a patient with a cardiovascular disease, instead of cohort-wide standards, precision care would utilize genetic profiling to identify personal variants impacting heart disease risk, cholesterol levels, etc., integrate personal data like genetics, digital biomarkers and leverage ML models to predict optimal prevention strategies, medications and procedures customized for that patient.
In oncology, genetic sequencing of a tumor can identify specific mutations driving cancer growth and treatments can then be tailored to target those mutations. Precision care enables a lung-cancer patient with a mutation in the ALK gene, to receive the ALK inhibitors, which are more effective than standard chemotherapy, and evidently more personalized.
Individualized care with hyper-personalized insights: Future looking
The latest advancements are focused on developing integrated knowledge bases that consolidate multi-dimensional personal data - genomic, clinical, behavioral, environmental - to generate hyper-personalized insights for guiding prevention strategies, diagnostics, therapeutics, and disease management on a per-individual basis. AI models can now process these comprehensive personal data profiles to recommend precise interventions optimized for each person.
Conclusion
The evolving personalization of evidence-based care – from cohort-based segmentation to AI-powered cohort optimization and precision medicine – is intimately related to the increasing comprehension and adoption of state-of-the-art technology in Healthcare practices. By leveraging the power of data analytics, genomics, and emerging technologies, care delivery has now shifted from standardized to hyper-personalized, predictive, and pre-emptive care tailored to each individual's requirements throughout his/her health journey.
