Six health systems saw six different versions of Sandra. None of them saw this.

Real-world treatment rarely moves in a straight line. Sandra is a patient who spent seven years moving through six different health systems, generating records in emergency departments, specialty clinics, primary care offices, and pharmacies that never spoke to each other. Each system saw a fragment of her care.

Sandra used Novellia to gather those fragments together. By authorizing Novellia to retrieve her records, she assembled a single longitudinal view of her care spanning six health systems and seven years. Once connected, patterns emerged that no single provider could have seen on their own: a surgical hold on her medication in early 2026 that lined up with a documented return of binge eating three weeks later, more than twenty emergency department visits in a single year that traced back to gaps in outpatient care, and eighteen months of continuous data showing a psychiatric and behavioral health trajectory invisible in any one chart. Sandra gained a complete picture of her own health, built from records that were always hers. Researchers and medical affairs teams gained a longitudinal signal that fragmented, single-system data rarely produces.

"Sandra" is a pseudonym, and this account is a de-identified patient journey with certain details altered for additional anonymity.

The Comorbidities Behind the BMI

Sandra was in her mid-forties when she began using Novellia to aggregate her medical records. The binge eating had been a feature of her life for years, a response her clinicians noted to the emotional dysregulation that comes with depression. For Sandra, the urge to eat functioned less like hunger and more like a circuit breaker tripping.

By 2019, Sandra's BMI had climbed above 35. By 2021, it was above 40. At the end of 2021, a blood clot formed in her left leg, broke apart, and traveled to both lungs: bilateral pulmonary embolism. She was placed on anticoagulation therapy she would take for the rest of her life.

The clot was, on the surface, a single cardiovascular event. Her unstructured clinical notes tell a more complicated story: a woman whose weight, psychiatric burden, and hormonal comorbidities (endometriosis, polycystic ovaries, hypothyroidism) had been accumulating silently for years. She also had obstructive sleep apnea she could not treat, GERD with a hiatal hernia, PMDD, and chronic pelvic pain. Any one of these conditions would have been individually manageable. Together, in a single patient, they composed something that clinical trials are not designed to capture.

A Long Road to the Right Treatment

By mid-2022, Sandra's clinical team was asking a question that more and more providers were beginning to ask: could a GLP-1 receptor agonist help?

The complications that followed are precisely what make Sandra's journey important to understand holistically.

Naltrexone was trialed in mid-2022 for binge eating disorder and discontinued quickly due to severe nausea. In March 2023, Sandra's care team started her on dulaglutide (Trulicity). She tolerated it, and early signs were encouraging. Two months later, she was hospitalized for a psychiatric crisis and her Trulicity use was discontinued.

When she stabilized, her team tried again. In the summer of 2023, she was started on oral semaglutide (Wegovy), uptitrated over several weeks. Her clinical notes described her as tolerating it and beginning to respond. Then her insurance company denied coverage. The appeal failed, and a subsequent Wegovy prescription was blocked as well. The denial arrived the same month her clinician had written, for the first time in years, that her eating felt more controlled.

At the beginning of 2025, after nearly two years of interrupted treatment, she was started on tirzepatide (Zepbound) 2.5 mg weekly, eventually escalating to 7.5 mg. Across 2025, her unstructured clinical notes document something that had been absent from her records for years: stability. Sandra's weight was stabilizing and binge eating was more controlled. For Sandra, GLP-1 therapy functioned as an eating behavior intervention as much as a weight intervention. Stopping it would make that clear.

What the Emergency Department Visits Tell Us

In the year before she found stable treatment, Sandra presented to the emergency department more than twenty times across different health systems. The reasons for each visit varied: abdominal pain, psychiatric evaluation, somatic complaints. The pattern was consistent: a patient cycling through crisis and stabilization in a way that tracked closely to gaps in her outpatient care.

Patients with this utilization pattern are undertreated in ways that outpatient-only data obscures. Each institution sees only its own slice of the record. A longitudinal record that connects across multiple health systems, drawing on both unstructured clinical notes and structured databases, surfaces this pattern in full.

When a Routine Surgery Revealed Something the Trials Rarely Capture

At the beginning of 2026, Sandra underwent a laparoscopic cholecystectomy. The surgery was uncomplicated, however pre-operative protocol required her to hold Zepbound.

Three weeks later, she presented to the emergency department with abdominal pain. By spring of 2026, her outpatient provider had written the note that anchors Sandra's entire story:

"Since stopping the Zepbound she has been binge eating which is increasing."

A routine surgical hold had triggered a behavioral relapse. The change showed up in her eating behavior rather than her weight, and it was observable within weeks across her systems. Real-world longitudinal data captures what unfolds when a stable patient has therapy interrupted for a routine procedure, a window that randomized controlled trials, with their shorter follow-up periods and narrower enrollment criteria, are rarely positioned to observe.

Why Sandra's Story Matters

Sandra represents the kind of patient GLP-1 trials are least equipped to enroll. She has BPD, ASD, a history of DVT and PE, a complex medication regimen, and a utilization pattern that most trial designs would exclude her for.

She is also the kind of patient who, when given sustained access to the right agent, demonstrated an 18-month response that transformed both her weight trajectory and her behavioral health, and who deteriorated visibly within weeks when that access was interrupted.

Sandra's individual record is one entry in a much larger pattern. Across Novellia's GLP-1 and Metabolic Health Registry, close to 70% of enrolled patients carry a comorbidity burden similar to Sandra's. Our data shows these patients achieve weight loss outcomes at rates equal to or exceeding lower-burden cohorts, while generating real-world signals about discontinuation consequences, insurance-driven treatment gaps, and behavioral health effects at a scale clinical trials have rarely been powered to detect.

The registry's depth comes from how it's built. By pulling from unstructured EHR sources alongside structured data, and following patients longitudinally rather than in fixed trial windows, Novellia captures the why behind discontinuation and response, the events that explain a trajectory like Sandra's rather than simply marking that it occurred.

_{"Sandra" is a pseudonym. All clinical data is drawn from de-identified patient records contributed to the Novellia GLP-1 and Metabolic Health Registry under HIPAA-compliant data processing agreements.}

‍