Food
Craveable & healthy: can workplace dining really deliver both?
Food As Medicine: How Data, Science and Behavioral Design Are Turning Nutrition into a Clinical Lever
Across U.S. health systems and payer organizations, nutrition is shifting from a hospitality line item to a measurable clinical lever—powered by connected data, clinical science and the behavioral insights that shape what people actually choose to eat.
In my role leading Sodexo’s healthcare business in North America, I see every day that food is one of the most underused assets that healthcare providers have for improving outcomes, experience, and long-term health.
For decades, food in healthcare organizations was treated as an operational necessity: simply calories that must be delivered on time, at cost, and in compliance. That framing is now obsolete. The convergence of clinical nutrition science, digital health data (biomarkers, wearables and connected records), and behavioral design is making nutrition more personalized, measurable and actionable. For health systems and payers under pressure to deliver better outcomes with fewer resources, food is a practical lever for prevention, recovery, experience and long-term health when it is deployed with the same rigor as any other clinical intervention.
Leading healthcare providers know that nutrition is:
- personalized through connected clinical and biometric data,
- measurable through outcomes-linked metrics and governance, and
- actionable through behaviorally designed food environments and Food as Medicine pathways.
In the U.S., this shift is accelerated by several realities: the burden of diet-related chronic disease; the operational and financial urgency of value-based models; and a workforce that increasingly expects employers, and the environments they create, to actively support wellbeing.
In this exciting new context, “food service” becomes “nutrition infrastructure”: It’s the part of care delivery that can quietly improve clinical pathways at scale.
Consumer-facing platforms are rapidly connecting food logs, wearables and biomarker testing into always-on guidance. In healthcare settings, the opportunity is more consequential: translating clinical signals into nutrition actions that fit care plans. Think of this as the difference between “education” and “execution.” When nutrition becomes connected to clinical workflows, screening, diagnosis, orders, discharge planning and follow-up, it can be made personalized (tailored to risk and condition), measurable (tracked through agreed metrics) and actionable (delivered through menus, retail choices and medically indicated supports).
Measurement changes behavior. Food safety has thermometers; finance has budgets. Increasingly, healthcare organizations are also being asked to measure and manage the health and environmental footprint of food. This is not a distraction from clinical outcomes, because diet quality, cardiometabolic risk, and sustainability often align through higher-fiber, plant-forward patterns, better sourcing, and less ultra-processed food.
In healthcare food environments, small design decisions can produce outsized effects: menu order, the default entrée, portion sizing, placement, naming, and how staff frame options. Recent hospital research has shown that even simple nudges, like listing plant-forward options first, can meaningfully shift patient selections. Critically, this is not about removing choice; it is about designing the default environment so that the clinically preferred option is also the easiest option.
A 2019 study in JAMA Internal Medicine receiving medically tailored meals are associated with fewer subsequent hospitalizations. At the same time, federal policy guidance is clarifying pathways for nutrition interventions within Medicaid and other public programs as part of broader approaches to health-related social needs.
For health system executives and payer leaders, the implication is practical: nutrition support can be designed like any other intervention, eligibility criteria, clinical referral, and outcomes evaluation. Then, nutrition care can be integrated into discharge planning, population health programs, and community partnerships. The organizations that lead will be those who deploy clinical nutrition governance to what has historically been treated as “ancillary.”
Treat nutrition as clinical infrastructure.
Assign executive ownership and clarify whether nutrition sits within quality, population health, patient experience, or all three.
Make malnutrition and nutrition risk measurable.
Standardize screening, dietitian assessment pathways, and escalation triggers; define metrics that matter (e.g., diagnosis rates, intervention timeliness, LOS, readmissions, patient-reported outcomes).
Build the data layer.
Connect menus and ordering systems to clinical diet orders, allergy and texture requirements, and where appropriate to biomarkers and risk stratification, so nutrition is auditable and improvable.
Design for behavior, not compliance.
Apply choice architecture in retail and patient dining (defaults, placement, naming, convenience) so evidence-based options are the easiest options.
Operationalize Food as Medicine partnerships.
For high-risk cohorts, define referral criteria and quality standards for medically tailored meals, groceries, or nutrition counselling; align with Medicaid and payer pathways where relevant.
Govern safety, privacy and trust.
If using AI-enabled guidance, set guardrails for clinical validation, data privacy, bias monitoring, and what is (and is not) an appropriate nutrition recommendation.
