How to Use Predictive Analytics in Medicine

Hot off our recent transformer paper, we're excited to share another AI model for precision medicine! Biological data collected from patients has exploded in recent years, presenting a challenge: how do we decipher that data to understand which patients will benefit most from specific therapies?  We in the Applied Data Science team at AstraZeneca are thrilled to share our paper in Cancer Cell called "AI-Driven Predictive Biomarker Discovery with Contrastive Learning to Improve Clinical Trial Outcomes." Here, we introduce the *Predictive Biomarker Modeling Framework (PBMF)*, a neural network-powered contrastive learning process that: 🔍 Explores vast multimodal datasets to uncover predictive biomarkers in an automated, systematic, and unbiased manner  🧠 Distinguishes predictive biomarkers (which indicate a likely benefit from a specific therapy) from prognostic biomarkers (which indicate general disease outlook)  💡 Distills its outputs into an interpretable decision tree, showing what drives treatment response In our studies, the PBMF:  📊 Surpassed existing methods in finding predictive biomarkers for immunotherapy success across various cancers in clinical trial and real-world data  📈 Discovered a predictive biomarker in an early-stage trial that boosted efficacy by 15% when retrospectively applied to the corresponding phase 3 clinical trial  📈 Discovered predictive biomarkers in single-arm early phase trial data with synthetic control arms, retrospectively improving the efficacy of the corresponding phase 3 trials by at least 10% We believe the PBMF has the potential to improve the way we design clinical trials and match patients to the right therapies. It can integrate with other models like our Clinical Transformer, creating exciting possibilities to someday discover biomarkers of adverse events, dosing strategies, and even to back-translate new drug targets. Read the full paper here: https://lnkd.in/eveAnVRY   Thanks to all the co-authors: Gustavo Arango, Damian Bikiel, Gerald Sun, Elly Kipkogei, Kaitlin Smith, Sebastian Carrasco Pro, Elizabeth Choe #PrecisionMedicine #ClinicalTrials #AIinHealthcare #Biomarkers #Immunotherapy

Updating My Latest “AI in Medicine Forecasting”: 🎨 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜’s Biggest Impact: The “Front of the House” 🔮 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝘃𝗲 𝗔𝗜’s Biggest Impact: The “Back of the House” If you’ve ever worked in the food service industry, you’re familiar with these terms: the “front of the house” includes all areas a customer interacts with while dining, while the “back of the house” refers to the behind-the-scenes kitchen where meals are cooked. In medicine, we have a perfect corollary: the 𝗳𝗿𝗼𝗻𝘁 𝗼𝗳 𝘁𝗵𝗲 𝗵𝗼𝘂𝘀𝗲 includes everything where the patient is actively involved—seeing the doctor, discussing results, and making decisions. But in healthcare, the 𝗯𝗮𝗰𝗸 𝗼𝗳 𝘁𝗵𝗲 𝗵𝗼𝘂𝘀𝗲 isn’t a physical space; it's the doctor's mind. Thinking about diagnoses, developing plans, and pondering prognoses. In medicine, your "meal" is prepared in our heads. Of course GenAI and PredAI are on a spectrum, and both will have profound impacts in healthcare. But generally speaking a lot more of the Front work involves language, communication, documentation, transcription, and translation. Write a clinic note, or translate a language — or even translate from “doctor speak” to patient-friendly language. Coming up with an analogy to help explain something to a patient. Generative AI is well-suited for this work. When it comes to the Back work, we’ve barely scratched the surface of what Predictive AI can do. Predictive AI has the potential to revolutionize diagnostic accuracy, improve prognostic predictions, and refine treatment plans. Currently, much of our approach in medicine feels like we're in a relative stone age compared to the potential predictive AI offers: 👉Diagnostic Accuracy: Moving beyond the guesswork of “Let’s run some tests and see what comes back,” predictive AI could offer more precise diagnostics based on comprehensive data analysis. 👉Treatment Planning: Rather than the trial-and-error approach“Let’s treat everything and see what works”—AI could help tailor treatments with higher success rates and fewer side effects. 👉 Patient Management: Improving decision-making about hospital admissions and discharges, predictive AI could ensure that only those who need intensive care are hospitalized, potentially reducing healthcare costs and improving patient outcomes. There’s so much in medicine that still relies on uncertainty. Phrases like “We don’t know exactly what will happen, so let’s admit you and see” could become less frequent as AI tools help us navigate complex medical landscapes with more precision and confidence. While generative AI focuses on patient-facing tasks, predictive AI opens new possibilities for internal decision-making. Curious to know if other clinicians see a similar dichotomy; there’s a whole portion of medical “work” that no one really knows about unless you’re the one seeing the patient.

How can we decrease pharmacy spend on high-cost drugs by double digits without worse outcomes? --- Uplift modeling is a common tactic in marketing to target the specific people for a promotion that otherwise wouldn’t buy the product. While marketing in general can lead to overconsumption, in healthcare/#pharmacy, the same mathematical techniques used for uplift modeling could be repurposed to support #PrecisionMedicine or personalized medicine, where the goal is to identify which patients are most likely to benefit from a specific treatment while avoiding unnecessary treatments for patients who might not respond well. Identifying the cohort that is getting most of the outcomes from a drug varies by drug, but some drugs have only a fraction of the total population driving a larger share of clinical results. --- Here's the basic process for using #UpliftModeling (you can find more details from my Milliman white paper in the comments): 1. Treatment: Identify the treatment for which you want to predict response (e.g., a high-cost brand/specialty drug like GLP-1s). This could also be done for a medical device or any intervention. 2. Data collection: Gather comprehensive data and studies about patients, including their medical history, genetic information, and any other relevant attributes. This is often the limiter of building a good model. 3. Control group: Assemble a control group of patients who are similar to those receiving the treatment but are not receiving the treatment themselves. This helps establish a baseline for comparison. 4. Outcome measurement: Measure the effectiveness of the treatment for both the treatment group and the control group. This could involve monitoring health improvements, cardiac events, or other relevant medical outcomes. For FDA-approved drugs, this could come from published research on the “absolute risk reduction” or “number needed to treat.” 5. Model building: Develop predictive models using machine learning algorithms that estimate the likelihood of a positive response to the treatment for each individual. 6. Uplift calculation: Calculate the difference in response rates between the treatment group and the control group to determine the net impact of the treatment. 7. Segment: Divide patients into different segments based on their predicted response probabilities. 8. Action: Use the insights from uplift modeling to guide treatment, coverage, or other decisions. --- A payer or employer can use this information how they’d like, but I imagine it will be used to adjust formularies or utilization management strategies. It could also be used when setting up contracts for how a drug should be used while carving out certain drugs or disease states (e.g. oncology drugs at a center of excellence). There are more potential use cases in the white paper in the comments. --- Would you use this strategy for #PharmacyBenefits or #ValueBasedCare models that take on risk for cost of care?

What if your data could prevent the next medical emergency? That's exactly what AI-powered predictive analytics is doing in healthcare right now. A hospital in Boston recently used predictive models to identify patients at high risk of sepsis 12 hours before symptoms appeared. The result? A 30% reduction in mortality rates. Here's how it works: > AI algorithms analyze patient data continuously. > They spot patterns that humans might miss.  > They flag risks before they become crises. Think blood pressure trends. Medication interactions. Historical outcomes from similar cases. > The system learns from millions of patient records. > It identifies subtle warning signs. > It alerts clinicians to intervene early. This isn't futuristic thinking. It's happening today. Predictive analytics is transforming how we approach patient care. We're moving from reactive treatment to proactive prevention. The impact goes beyond saving lives: * Early interventions reduce hospital stays. * They lower treatment costs. * They free up resources for other patients who need them. One ICU reported a 20% decrease in readmissions after implementing predictive analytics. Another reduced emergency transfers by 40%. The technology isn't replacing doctors. It's empowering them with insights they couldn't access before. Does this mean every hospital should rush to implement predictive analytics tomorrow? It depends on context. You need clean data. You need integration with existing systems. You need staff trained to act on the insights. But the potential is undeniable. When data becomes a tool for prevention rather than just documentation, we change the entire equation of healthcare delivery. How is your organization using data to predict rather than just record?

Today, on World Cancer Day, we recognize the profound impact cancer has on individuals and families worldwide. My father had stage IIIB adenocarcinoma of the lung, with his left upper lobe removed, and my uncle succumbed to small cell lung cancer. Both were non-smokers. These stories underscore the urgency of advancing our detection methods. It's a personal mission for many, driven by the hope that through technology, particularly the fusion of Knowledge AI and Big Data AI, we can unveil these silent killers early enough to make a difference. Here's a proposed 10-step protocol for deploying an algorithm capable of early detection of solitary lung nodule cancer, leveraging blood biomarkers, radiology, and other modalities: Data Collection and Integration: Gather extensive datasets covering various patient demographics and stages of lung cancers. Big Data Infrastructure: Develop efficient data handling for structured and unstructured data. Knowledge AI Models: Utilize medical knowledge to enhance AI models. Machine Learning and Deep Learning: Apply AI techniques for identifying early-stage cancer patterns. Radiology Image Analysis: Train AI for advanced image recognition of lung scans. Blood Biomarker Detection: Develop algorithms for non-invasive blood test analysis. Predictive Modeling: Personalize risk assessments using predictive models. Clinical Validation: Ensure model accuracy through extensive clinical trials. Integration into Clinical Workflows: Collaborate with healthcare providers to incorporate AI into existing processes. Continuous Learning and Improvement: Establish a system for regular AI model updates based on new data and discoveries. By following these steps, we can harness AI's power to transform early lung cancer detection, potentially saving countless lives. The fusion of Knowledge AI and Big Data AI offers hope, turning silent stories into beacons of progress. Through early detection, we aspire to beat cancer.

This paper looks at how LLMs, which are AI models trained on huge amounts of text, can help doctors predict medical conditions by analyzing EHRs (patient health records). 1️⃣ The study tests two AI models, GTE-Qwen2-7B and LLM2Vec-Llama3.1-8B, on 15 medical prediction tasks, like guessing if a patient will need to stay in the hospital longer or develop a certain condition. 2️⃣ These AI models often work just as well—or even better—than special EHR foundation models (like CLIMBR-T-Base) and traditional prediction methods, especially when there's little training data. 3️⃣ Performance improves with larger LLM models and longer context windows, with GTE-Qwen2-7B performing best at a 4,096-token context length. 4️⃣ The researchers turned complex medical records into simple, organized text (like a structured note), making it easier for the AI to understand and predict health outcomes. 5️⃣ Combining LLM-based embeddings with the EHR-specific model further improves predictive accuracy, suggesting complementary strengths. 6️⃣ LLM-based EHR encoding offers a scalable alternative to traditional EHR-specific models, overcoming challenges related to dataset availability and coding inconsistencies. ✍🏻 Stefan Hegselmann, Georg von Arnim, Tillmann Rheude, Noel Kronenberg, David Sontag, Gerhard Hindricks, Roland Eils, Benjamin Wild. Large Language Models are Powerful EHR Encoders. arXiv. 2025. DOI: 10.48550/arXiv.2502.17403

Most people think the biggest impact of AI in healthcare will be robotic surgery or futuristic hospitals. But the real shift is happening somewhere quieter. In prediction. AI is beginning to move healthcare from reactive treatment to early intervention. Take drug discovery. Traditionally, identifying promising compounds could take 10–15 years of research and testing. Today, generative AI models can simulate millions of chemical combinations in weeks, helping researchers narrow down candidates dramatically faster. Some estimates suggest AI could cut early discovery timelines by up to 70%. But drug discovery is only one piece of the story. The second shift is predictive medicine. Agentic AI systems are now analyzing combinations of: • genetic data • lifestyle signals • historical medical records to identify risks years before symptoms appear. Researchers are already using these approaches to identify early risk patterns for Alzheimer’s and cardiovascular disease. The third shift is happening inside hospitals themselves. AI monitoring systems are beginning to detect subtle signals of patient deterioration long before humans can spot them. Small changes in vitals. Tiny shifts in oxygen patterns. Behavior changes in ICU monitoring. Signals that once looked like noise are now becoming early warnings. Then there are ambient AI tools, quietly reducing administrative burden. In many hospitals today: • AI scribes automatically generate clinical notes during consultations • AI vision systems screen diabetic retinopathy from retinal scans • Triage systems flag high risk patients automatically The result? Doctors spend less time typing and more time treating. Healthcare breakthroughs often sound dramatic. But many of the most meaningful ones look like this: A diagnosis earlier than expected. A deterioration detected sooner. A clinician freed from paperwork. Sometimes the biggest innovations don’t replace doctors. They simply give them time back to be doctors. #ArtificialIntelligence #HealthcareInnovation #DigitalHealth #AIinHealthcare #PredictiveAnalytics #MachineLearning #HealthTech #FutureOfHealthcare #MedTech #Innovation #DataScience

🔬 DoFlow: When Time-Series Forecasting Meets Causal Reasoning I recently read an interesting paper describing a flow-based generative model that unifies observational prediction with interventional and counterfactual forecasting for time-series data: "DoFlow: Causal Generative Flows for Interventional and Counterfactual Time-Series Prediction" https://lnkd.in/eSdbyGRA * I think this method has lots of potential applications in healthcare and biomedical research! 🎯 The Problem Most time series forecasting models are purely observational - they predict correlations but cannot answer causal questions like: Interventional: "What if we change the treatment dose now?" Counterfactual: "Would a different past treatment have prevented this outcome?" These are exactly the questions clinicians often need answered for precision medicine. 💡 The Innovation DoFlow combines three key components: 1. Continuous Normalizing Flows (CNFs) - Neural ODEs that create invertible mappings between noise and data distributions, enabling likelihood computation 2. Causal DAG Structure - Each variable has its own flow conditioned only on its causal parents, respecting temporal dependencies 3. RNN History Encoding - Captures temporal context for conditioning the flows The invertibility is useful as follows: the model can encode factual observations into latent noise, then decode under counterfactual conditions to generate "what if" scenarios. 🏥 Cancer Treatment Applications DoFlow was successfully implemented on the Bica et al. cancer treatment data set as an example. 🔍 Why It Matters ✅ Generates full system trajectories, not just point estimates ✅ Provides explicit likelihoods for anomaly detection ✅ Handles time-varying treatments in complex causal structures ✅ Scales to real-world multivariate systems ⚠️ DoFlow requires a pre-specified causal DAG - in clinical settings this relies on existing domain knowledge #CausalInference #MachineLearning #PrecisionMedicine #TimeSeriesForecasting #Oncology #HealthcareAI #NeuralODEs #ContinuousNormalizingFlows #DeepLearning #CancerResearch #BiomedicalML #ClinicalAI #CounterfactualReasoning #DynamicTreatmentRegimens #ComputationalBiology

Administrative costs consume roughly 30% of every healthcare dollar. That's a $1.2 trillion opportunity. Most cost reduction efforts focus on claims processing or prior authorization. Important work, but it misses the bigger picture. The real opportunity lies in population health analytics. Using data to prevent expensive care episodes before they happen. Predictive models that identify members at risk for emergency department visits. Algorithms that flag potential medication adherence issues. Analytics that spot care gaps before they become costly complications. This isn't just about reducing costs. It's about improving outcomes while eliminating administrative waste. When you prevent hospital readmission, you save money and improve quality simultaneously. The actuarial challenge is building models that can operate in real-time. Traditional retrospective analysis isn't fast enough. We need predictive systems that can trigger interventions while there's still time to make a difference. Organizations getting this right aren't just lowering costs. They're fundamentally changing how healthcare gets delivered. #PopulationHealth #PredictiveAnalytics #CostContainment #ActuarialScience

🟥 Integrative Omics for Drug Response Prediction One of the biggest challenges facing modern medicine is predicting how individual patients will respond to specific therapies. While traditional clinical metrics can provide some guidance, they often fail to capture the molecular complexity of each patient's disease. Integraomics—a powerful approach that combines genomic, transcriptomic, proteomic, epigenomic, and metabolomic data—is rapidly emerging as a disruptive force in the field of personalized drug response prediction. By analyzing multiple molecular levels simultaneously, integraomics can provide a systems-level understanding of how a patient's biology interacts with therapeutic drugs. For example, genomic variants may reveal potential drug targets, while transcriptomic profiles can show whether those targets are actively expressed. Proteomic and metabolomic data can provide further insights into functional consequences and metabolic vulnerabilities. Together, these datasets help clinicians match patients with the most effective drugs and avoid treatments that are likely to fail or produce adverse reactions. Recent advances include artificial intelligence and machine learning-driven platforms that leverage integraomics data to predict individual drug responses with high accuracy. These models can identify gene expression patterns, signaling pathway activity, and metabolic signatures that are associated with sensitivity or resistance to targeted therapies, chemotherapy, or immunotherapy. In addition, integrative omics has also shown great value in real-time monitoring, allowing clinicians to adjust treatment plans based on the progression of a patient's tumor or changes in immune response over time. This dynamic approach can not only improve efficacy, but also reduce unnecessary toxicity and medical costs. In summary, integrative omics is ushering in a new era of precision medicine - drug selection is driven by deep molecular insights and tailored to each patient's unique biological characteristics. It is transforming treatment from reactive to predictive treatment, and from one-size-fits-all to truly personalized medicine. References [1] Hui-O Chen et al., J Pers Med 2024 (https://lnkd.in/eADqTPDS) [2] Ruijiang Li et al., Advanced Intelligent Systems 2024 (https://lnkd.in/eDNafrtz) #IntegrativeOmics #DrugResponsePrediction #PrecisionMedicine #MultiOmics #AIinHealthcare #PersonalizedTherapy #TranslationalResearch #OmicsData #Genomics #Proteomics #Metabolomics #CancerTherapy #MachineLearning #SystemsBiology #FutureOfMedicine #CSTEAMBiotech