Instead, maybe variability can be seen as less of an intractable foe and more of a frustrating but well-meaning and loyal friend. Variability forces us to personalize care—not because it’s trendy, but because it’s essential. Treat-to-target strategies in rheumatology work best when grounded in flexible thinking. We absolutely need protocols, but also permission to diverge from them when the patient’s circumstances demands creative solutions.
Variability calls for humility as much as it does for humanity. We often cannot know, at the start of treatment, how a given patient will respond. Our job is to collectively hold that uncertainty with grace— to explain it transparently to patients, to experiment (within safe boundaries, of course) and to always be ready to revise our assumptions at subsequent visits.
In short, reorienting our approach to variability means we are not normalizing immune status, because that is fundamentally not possible. Instead, we are constantly calibrating trust—in the patient’s body, in their lived experience and in our shared ability to navigate the unknown together.
Healthcare Policies to Manage Uncertainty
The fundamental variability of immune-mediated disease demands not just personalized clinical care, but policy structures that can accommodate complexity. Unfortunately, most healthcare systems are built for predictability—standard formularies, fixed prior authorization rules, cookie-cutter time limits for consultations. These constraints assume that patients follow a narrow path. We rheumatologists know better.
When policies ignore variability, they don’t just inconvenience clinicians, they unequivocally harm patients. Rigid fail-first/step therapy rules, limited access to diagnostic imaging, and constricted networks for referrals to specialists are all examples of systems that fail to account for the unpredictable journeys of patients like Bob and John. The result is frustration, delays and worse outcomes.
What we need instead are adaptive, human-centered policies—guidelines that allow for exceptions, that trust clinicians to make evidence-informed decisions in the face of uncertainty. Governmental oversight is needed so that insurance policies allow for flexibility. Prior authorization systems must be reformed so that they are centered on the trust and judgment of board-certified specialists rather than on peer reviewers who may not even know how to spell rheumatology.
Technology can help us in this pursuit. Designers of electronic health records are finally understanding the benefits of accommodating narrative nuance over templated check boxes. Machine-learning models, when thoughtfully designed, can stratify patients by risk and predict who might benefit from more intensive monitoring. Sharing data across institutions can help us recognize rare patterns and truly personalize care. But with these benefits also comes risk: Predictive models are only as good as the data and values that inform them. We must design healthcare systems that elevate precision and variability, not just efficiency.
