How a Patient-Support Program Found Affordability Was the #1 Reason Patients Stopped Therapy

Industry: Pharmaceutical, patient-support services
Challenge: Patients were discontinuing therapy, and the organization could not identify why
Result: Affordability identified as the leading discontinuation driver, ahead of every clinical reason, with risk signals now flagging automatically inside the client’s own platform

 

Company Overview

The client operates a global pharmaceutical patient-support program, the service layer that helps patients stay on prescribed therapy through enrollment, follow-up, and ongoing support calls. The contact center function sits between the pharmaceutical company and the patient, making it one of the only places where the reasons patients actually stop treatment get spoken out loud. Leadership’s objective was to understand why patients were discontinuing therapy so retention efforts could target the real cause rather than guess at it. Zenylitics applied speech and interaction analytics to the program’s call recordings to identify why patients were discontinuing therapy.

 

The Challenge

Patients were starting therapy and then quietly stopping, and the organization had no reliable way to say why. Retention efforts were built to treat every patient as equally at risk, since without a way to identify which patients were actually likely to drop off, there was no basis for prioritizing outreach toward the ones who needed it most. The answer to why patients were leaving was sitting inside the program’s own call recordings the entire time. Nobody had built a way to read it.

 

Why the Existing Approach Failed

Assuming every patient carries equal discontinuation risk spreads retention resources across the entire population instead of concentrating them where they matter. That approach cannot work at scale, since a patient-support team has limited capacity for outreach and needs to know where to spend it. Clinical assumptions about why patients stop treatment, side effects, forgetting doses, feeling the medication was not working, are reasonable guesses, but a guess is not the same as a pattern confirmed against what patients are actually saying on the phone. Without that confirmation, the organization was treating a knowable problem as an unknowable one.

 

The Zenylitics Approach

Data Layer. Patient support calls across the program, covering enrollment, follow-up, and ongoing check-in interactions.

Analytics Layer. Every reason a patient called was classified into a defined taxonomy, then measured for volume concentration to identify which categories actually mattered at scale rather than treating all reasons as equally significant.

Human Intelligence Layer. Zenylitic’s analysts narrowed the full taxonomy down to the signals that came before a patient discontinued. Sixteen categories cover every reason a patient calls. Eight of those sixteen carry 91 percent of total call volume, which tells the organization where the bulk of patient contact actually concentrates. Five separate categories are early warning signals that show up before a patient drops off, which is what makes them useful for early intervention rather than after-the-fact analysis. And four categories explain the patients who had already left, confirming which early warning signals matched the outcome.

Action Layer. The at-risk signals were built directly into the client’s own conversation intelligence platform, flagging patients automatically as those signals appear in the client’s calls. Every category identified through this engagement stayed live inside the client’s own platform rather than existing only inside a one-time Zenylitics report, which means the organization now owns the capability to flag at-risk patients going forward instead of depending on Zenylitics to re-run the analysis every time it needs an updated picture.

Key Insights Discovered

The sixteen-to-eight-to-five-to-four structure is the engagement’s clearest finding, because it makes the entire method visible in one place. Out of every possible reason a patient might call, only a small number actually carry the volume and early warning value that matter. Two categories in particular concentrated an outsized share of contact on their own, co-pay support at 31.8 percent of call volume and follow-up inquiries at 24.6 percent, together accounting for more than half of everything the program’s contact center handled.

The finding that reframed the entire retention strategy came from what those calls were actually about. Affordability, not a clinical concern, was the single largest reason patients gave for stopping treatment, accounting for 30.6 percent of discontinuations and outranking every clinical reason patients cited, including that the medicine did not work or caused an allergic reaction. A patient-support program has real limits on what it can do about a diagnosis or a prescriber’s decision. It has considerable ability to act on an affordability problem, through co-pay support programs, benefits navigation, and financial assistance enrollment. The value here was not that Zenylitics read every call. It was organizing every reason a patient called into sixteen categories, identifying five early warning signals that came before discontinuation, and identifying four categories that explained why patients left, so the organization could tell exactly which conversations to have and with which patients.

A second finding concerned call transfers. Two vendors took 66.6 percent of all transfers, and about 45 percent of transfers happened on calls under five minutes, which points to misrouting.

Business Outcomes

Affordability accounted for 30.6 percent of patient discontinuations, ahead of every clinical reason recorded, including allergic reactions and reports that the medicine did not work. Two categories, co-pay support and follow-up inquiries, represented more than half of all call volume across the program, at 31.8 percent and 24.6 percent respectively. These figures describe the concentration and cause of patient discontinuation. They do not measure a before-and-after change in dropout rate.

Why This Matters

The single biggest reason patients stopped therapy was cost, and cost is the one reason a support program can act on directly. The categories that produced this finding did not end with the report. They stayed inside the client’s own platform, continuing to flag at-risk patients on every call that comes in afterward.

FAQs

What was the leading reason patients stopped therapy in this patient-support program?

Affordability was the leading reason, accounting for 30.6 percent of discontinuations and outranking every clinical reason patients cited. A patient-support program can act on affordability through co-pay support programs, benefits navigation, and financial assistance enrollment.

How does call analytics identify why patients stop therapy?

Call analytics classifies every reason a patient called into a defined taxonomy, then measures which categories carry the most volume and which signals appear before a patient discontinues. In this program, sixteen categories covered every reason a patient called, and eight of them carried 91 percent of total call volume.

What happened to the early warning signals after the analysis?

The signals were built into the client’s own conversation intelligence platform, where they flag at-risk patients automatically. The organization can now flag at-risk patients without waiting for Zenylitics to re-run the analysis.

Request a Demo to see how speech and interaction analytics can show why patients stop therapy in your patient-support program.

Subscribe to Zenylitics Newsletter