From Fragmented Records to Patient-Powered Healthcare AI Platforms
Patient-powered data platforms are digital systems that let individuals collect, control, and share their complete health histories so that researchers and healthcare AI platforms can train, validate, and update diagnostic models on more accurate, real-world health data drawn directly from people’s lives instead of scattered institutional databases. This shift matters because traditional medical datasets are fragmented across hospitals, insurers, and labs, leaving gaps that weaken AI medical imaging and decision-support tools. Novellia has stepped into this problem with a patient-centered model that replaces disconnected files with unified, patient-consented records. Its platform is designed to turn long, messy health histories into structured, research-ready datasets that can feed algorithms used for tasks such as image interpretation and outcome prediction. As AI spreads through diagnostics, the quality and completeness of patient data collection are becoming as important as the models themselves.
Novellia’s USD 18 Million Series A and the Push for Real-World Health Data
Novellia has raised USD 18 million (approx. RM83.4 million) in Series A funding led by Spark Capital, with participation from Khosla Ventures, Acrew Capital, Bling Capital, and TMV. This brings its total funding to USD 28 million (approx. RM129.8 million) as it scales a patient-powered real-world data platform used by large pharmaceutical and diagnostics companies. The company argues that today’s real-world data ecosystem costs more than USD 50 billion (approx. RM231 billion) each year while still relying on incomplete information assembled from insurance claims and hospital records. Novellia’s system lets patients pull medical records from hospitals, clinics, and laboratories into a single longitudinal file in about 30 seconds, which can be vital for people managing complex or chronic conditions. According to Novellia, nearly 70% of patients are open to contributing data for medical breakthroughs when they have control, transparency, and privacy safeguards.
Why AI Medical Imaging Needs Patient-Generated Real-World Data
AI medical imaging tools, from melanoma classifiers to lung nodule detectors, perform best when trained on datasets that reflect how diseases appear in everyday practice. Historically, these algorithms have been based on narrow clinical trial cohorts or limited hospital archives, which can miss early-stage or atypical cases. Patient-powered data collection platforms change this by aggregating complete imaging histories, pathology reports, and related clinical notes directly from individuals. When combined, these records create large, diverse, real-world health data pools that better capture variation in age, skin tone, comorbidities, and treatment pathways. For melanoma detection, for example, more diverse skin images paired with longitudinal outcome data can help models distinguish harmless lesions from cancer with higher accuracy. As more patients opt in, healthcare AI platforms gain access to richer labels and follow-up information, which improves model calibration and reduces the risk of biased or unreliable diagnostic outputs.
From Clinical Trials to Crowdsourced Data Models
Life sciences companies are starting to complement traditional clinical trials with crowdsourced, patient-consented data streams. Clinical trials remain the standard for testing safety and efficacy, but they are slow, expensive, and often exclude many patient groups. Platforms like Novellia offer a parallel route: individuals aggregate their health records, then choose to contribute anonymized and deidentified datasets that reflect routine care over long periods. This approach can show how therapies perform outside controlled trial settings and supply healthcare AI platforms with ongoing feedback loops for models used in diagnostics and monitoring. Novellia’s multi-year, seven-figure contracts with major pharmaceutical and diagnostics organizations signal demand for this kind of data infrastructure. Instead of piecing together claim files, companies can study disease progression, treatment changes, and imaging results within a single patient-centered record, which is essential for building AI systems that match real-world clinical complexity.
The Road Ahead: Patient Control, Privacy, and Trust in Healthcare AI
As AI spreads through radiology suites and digital diagnostics, the question is no longer whether real-world data will matter, but who controls it and how it is used. Novellia’s model is built on giving individuals secure access to their full health histories and clear choices about sharing anonymized data for research. This helps close the gap between patient willingness and the technical infrastructure needed to support responsible data contribution. For AI medical imaging platforms, that consented access is key: it enables deeper learning from outcomes, adverse events, and subtle imaging changes without exposing identifiable information. The company’s peer-reviewed work, presented at medical conferences such as ASCO and SABCS-AACR, suggests that patient-centered data collection can meet scientific standards while respecting privacy. If that balance holds, healthcare AI could move toward systems that are not only more accurate, but also more accountable to the people whose data powers them.
