There’s been a fundamental shift in how patients engage with their healthcare. Patients now arrive as informed, active decision‑makers. They’ve done their research. Their expectations have been shaped by the best consumer experiences far outside healthcare, and they are increasingly willing to disengage when healthcare experiences don’t meet that bar. At the same time, rising payer control remains a major friction point to patient access for new treatments. In 2025, only 12% of new‑to‑brand specialty prescriptions were successfully filled on the first try. Nearly 80% were rejected upfront, with many denials tied to prior authorizations or step‑therapy requirements, creating an administrative burden for patients¹.

For patient support programs, growing consumerism coupled with payer control creates a distinct and urgent design challenge. It’s no longer enough to offer support. Programs must deliver the right support, at the right moment, through the right channel. When the experience doesn’t align with patient expectations, the consequences are measurable, resulting in delayed therapy starts, missed refills, and silent drop‑off.

Affordability pressures and rising out-of-pocket costs (OOP) only intensify this challenge. In 2025, fewer than 10% of commercially insured patients abandoned their prescription when their OOP cost was $0 across diabetes, dermatology, immunology, and respiratory treatments. But as OOP costs increase, abandonment rises sharply. Once costs reached $125 to $250, about one in two patients did not pick up their medication². A support program that doesn’t proactively address affordability is already behind as patients are tracking co-pays, co-pay assistance eligibility, and benefit structures more closely than ever before.

A patient’s experience, including friction across touchpoints, gaps between interactions, and transitions from automated systems to human experts, is a measurable outcome that must be considered during program design, not retrofitted later. High-tech, high-touch patient support means designing the journey around what patients need in each moment. The goal is not more automation; It is a support experience patients can rely on that is proactive, personalized, and seamless. AI removes friction in the background with speed and scale, while clinical experts step in when it matters most. The result is a seamless journey from first script to long term adherence.

Start with the Patient Journey, Not Technology

Building a program that earns patient trust and drives persistence starts by mapping the patient journey. That means understanding what patients know and don’t know at each stage, recognizing where delays and confusion create disruption, and identifying the moments that matter most.

Those moments typically cluster around first contact, prior authorization initiation, therapy start, and the early persistence window, when patients are most vulnerable to discontinuing before they’ve experienced the full benefit of therapy. Once those moments are mapped, the design questions become clear: What does this patient need at this point, and who or what is best positioned to deliver it?

Effective programs build personalization into their architecture from the ground up. That requires understanding who each patient is: their disease state, channel preferences, health literacy, and support needs. Patient segmentation is also critical here. Grouping patients by archetype—such as newly diagnosed, stable, at adherence risk, or facing an affordability barrier—enables programs to tailor journeys to real patient needs rather than relying on one‑size‑fits‑all models that serve no one particularly well. At scale, this level of intelligent routing is the role of AI, not a case manager.

Measure Progress, Not Just Productivity

Redesigning a patient support program requires rethinking what you measure. For years, programs have optimized operational metrics such as speed-to-answer, cases closed, and call handle time. While these metrics indicate how efficiently a program runs, they don’t reveal whether patients are actually moving forward in their treatment journey.

A patient‑lens design demands patient‑lens metrics. These include prior authorization success rates and time-to-therapy-start; refill behavior at 30, 90, 180, and 360 days; CSAT (Customer Satisfaction Score) and patient confidence scores; and, emotional sentiment signals that surface how patients feel about their experience.

Dashboards that segment these outcomes by therapy, payer, age group, and channel preference reveal whom a program is serving well, and whom it may be leaving behind.

The Case for High‑Tech + High‑Touch

The conversation about AI in patient support is often framed as a tradeoff: more automation, less human connection. In reality, the goal is not to choose between efficiency and empathy, but to use each where it adds the most value, and to be intentional about how patients move between them.

AI excels at tasks that require speed, consistency, and scale. In patient support programs, this includes benefits investigation and eligibility checks, prior authorization workflows, and 24/7 chatbot or digital support for common questions related to dosing, delivery status, and copay assistance. AI is also well‑suited for proactive intervention. By identifying patterns in fill behavior and engagement signals, AI can flag patients at risk of disengagement and trigger outreach before disruption or silent drop‑off occurs. At scale, AI enables personalization no human team could realistically deliver, tailoring outreach by channel preference, time of day, communication frequency, and content for every patient in the program.

Human expertise, by contrast, is irreplaceable in moments that require context, judgment, and genuine empathy. First‑fill anxiety. A coverage denial that threatens therapy access. A side effect that makes a patient consider stopping. These are the moments when clinical knowledge and emotional intelligence determine whether a patient abandons therapy or stays on course.

At IQVIA, we describe this model as having an “expert in the loop,” a meaningfully higher standard than a traditional “human in the loop.” When a care partner engages at a critical patient moment, they bring clinical expertise, not just call‑center protocol.

Orchestrating the handoff between AI and human expertise is where many programs fall short. In a well‑designed hybrid experience, patients should never feel the “seam” between systems. AI manages the operational load and flags which patients need human attention. When an expert steps in, they already have full context with no redundant questions, transfers, or time spent getting ready to help. The experience remains continuous.

Outcomes of High-Tech + High-Touch Patient Support

Don’t Underestimate the Importance of Your Data

AI‑enabled patient support is only as good as the data beneath it. A unified patient identity across systems, hub platforms, specialty pharmacy, nursing services, and patient case management is foundational. Without it, AI cannot see the whole patient, and the personalization and proactive outreach that make these programs effective break down.

High‑performing programs bring together three categories of data. Clinical and access data to explain what the patient is navigating. Behavioral data to reveal how the patient is responding. Contextual data (such as device access, timing preferences, and social determinants of health) to indicate how best to engage.

In patient support, data quality matters as much as completeness. Simpler models built on clean, well‑structured data consistently outperform more sophisticated approaches built on fragmented or inconsistent records. Organizations that invest in data readiness with the same rigor they apply to AI capability development will outpace those that treat data infrastructure as an afterthought.

Data governance—including consistent definitions, privacy safeguards, and bias monitoring—must be managed as an ongoing operational discipline. As patient populations and therapy landscapes evolve, governance models that aren’t actively maintained will quietly erode program effectiveness.

Where to Start

To begin redesigning patient support through a patient lens, start with one moment that isn’t working. It may be a friction point in prior authorization, a drop‑off window identified in 90‑day persistence data, or a recurring gap in early onboarding. Design a better experience for that moment, centered on the patient’s perspective, and define success metrics before launch. Then measure, refine, and scale what works.

Above all, keep humanity at the center. The purpose of AI in patient support is to free clinical experts to do the work only they can do; conversations that require genuine expertise, empathy, and human connection.

The future of patient support is high‑tech and high‑touch by design. Get this right, and programs will do more than improve time-to-therapy and persistence. They will build trust, empower patients as active participants in their own care, and turn patient support into a durable competitive advantage.

Designing the Next Generation of Patient Support

As patient expectations continue to rise, now is the time to rethink how support programs are designed and delivered. IQVIA Patient Support Services helps organizations build high-tech, high-touch models that improve access, patient experience, and adherence. Connect with our team to discuss how we can support your next phase of growth.

References:

1. IQVIA LAAD, 30-day look forward applied.
2. IQVIA LAAD.
3. Digital+ / Hybrid Support Persistence Improvements (60.9% vs 45.7–48.2% at 12 months)
   Sciensus. (2024). Digital+ real world patient persistence study (35,990 patients).
4. High Performing Patient Support Programs (Persistence up to 77%)
   Biogen. (2023). Benepali and Imraldi patient support program: Real world outcomes report. Biogen Medical Affairs.
5. Patient Trust in AI (70–80% when clinician supervised; ~17% for AI alone)
   American Medical Association. (2023). Patient trust and clinician oversight in AI enabled care. AMA Research.