General

10 Best AI Tools for Patient Adherence and Persistence

Medication adherence software and patient adherence solutions compared: 9 vendors by category, PDC and persistence explained, and a pilot to run first.

Mar 20263 min readGeneral

Choosing a tool? Off-the-shelf AI rarely sees your CRM, hub, specialty pharmacy and claims data. We build custom AI tools for pharma and life sciences teams that do.

What medication adherence software actually does

Medication adherence software is the set of tools a pharma brand, specialty pharmacy or health plan uses to keep patients taking a therapy as prescribed and to keep them on it for as long as it is meant to run. For a specialty or chronic brand that splits into two problems. Adherence is whether the patient takes the doses they have. Persistence is whether they are still on therapy at month six, twelve or twenty-four. Most revenue lost after a patient starts is lost to persistence, not to the odd missed dose.

The market sells five quite different things under the same label: hub and patient support platforms, digital engagement apps and nudges, connected devices and packaging, specialty pharmacy adherence programmes, and analytics that measure and predict discontinuation. They need different data, are bought by different teams and fail in different ways. Comparing an incentive app with a hub vendor on a single feature grid tells you very little.

This guide is written for brand, patient services, hub and commercial analytics leads choosing patient adherence software or a programme partner. It groups nine vendors by category from their own published material (checked October 2026), defines the persistence metrics, lists the data each approach needs and ends with a pilot to run before signing. We quote no vendor outcome percentages: every figure we found was vendor-reported, so ask for the study behind any number you are shown.

Where patients actually drop off therapy

Between approval and first fill, before adherence is even measurable
At the first refill, when side effects or cost first bite
When a prior authorisation or copay card expires and nobody renews it
On a pharmacy transfer, when the patient's history does not follow them
Quietly, at month three to six, with no status code that explains why

The five types of patient adherence solutions

  • Hub and patient support platforms

    Run the programme: enrollment, benefit verification, copay, nurse and pharmacist calls, refill reminders. Adherence is one service among several.

    • Bought by patient services and market access
    • Data: hub case records and status
  • Digital engagement and nudging

    Apps, SMS, voice and incentive programmes that prompt patients between fills and capture self-reported doses.

    • Bought by brand teams and health plans
    • Data: app events, check-ins, opt-ins
  • Connected devices and packaging

    Inhalers, injectors and containers that record each use, so adherence is observed rather than inferred from refills.

    • Bought with the device or as an add-on
    • Data: dose events from the device
  • Specialty pharmacy adherence programmes

    Clinical pharmacists and nurses in the dispensing pharmacy who call, counsel and manage refills for a therapy area.

    • Bought through limited distribution contracts
    • Data: dispense and patient status feeds
  • Adherence analytics and persistence modelling

    Aggregate dispense, hub and claims data to measure PDC and persistence and predict who is likely to stop next.

    • Bought by commercial analytics
    • Data: everything above, joined at patient level

Hub and patient support platforms with adherence services

  • AssistRx

    A therapy initiation and patient solutions provider for specialty brands. Its Advanced eServices cover benefit verification, prior authorisation, eConsent and eEnrollment, and its patient solutions include welcome calls, clinical and product education, first dose observation and refill reminders.

    • In July 2026 announced a partnership with Medisafe to run patient journeys on Medisafe's Maestro platform
    • Best fit: specialty brands that want access and adherence under one hub
  • CareMetx

    A technology-enabled patient services provider covering hub services, access, affordability, onboarding and adherence for specialty medicines. In April 2026 it acquired Cencora's US hub operations, previously run under the Lash Group name, and the TheraCom free goods pharmacy.

    • If your programme sat with Lash Group, check what changed in your contract and data feeds
    • Best fit: brands outsourcing a full-service hub
  • PhilRx (PHIL)

    A digital hub combined with a direct-to-patient pharmacy for retail and "specialty-lite" brands. Intake, routing, dispensing and adherence sit in one flow, and refills are triggered through the hub, including auto-refill and refill requests to the prescriber when a script runs out.

    • Adherence is built into the dispensing channel
    • Best fit: branded retail drugs losing patients at the pharmacy counter

Hub and specialty pharmacy adherence programmes: what to check

Hub vendors and specialty pharmacies both report adherence, but they report it from their own slice of the journey. A hub sees enrollment, benefits and the calls it made. A specialty pharmacy sees dispenses and its own status codes. Neither, on its own, sees the patient who transferred to another pharmacy or moved plans.

So the question to ask is not "do you report adherence?" but "at what grain, how often, and can we join it to our other sources?" You want patient-level, de-identified records with a stable token, a dispense date and days' supply for every fill, and status changes with reason codes, delivered at least weekly. Without that you can read their dashboard but you cannot calculate persistence yourself. Our guide to specialty pharmacy data analytics covers the feeds in more detail, and AI tools for hub services automation covers the hub side.

Specialty pharmacy, digital engagement and connected device tools

  • Accredo (Evernorth)

    Cigna's Evernorth specialty pharmacy. It organises care into condition-specific Therapeutic Resource Centers, where specialty-trained pharmacists and nurses monitor adherence and run coaching interventions for patients on drugs Accredo dispenses.

    • Part of a PBM group, so weigh channel and data terms carefully
    • Best fit: therapy areas where Accredo is in your network
  • Medisafe

    A medication management app with a pharma business it describes as a "patient engagement engine for pharma portfolios". Brands sponsor programmes inside the app; Maestro is the platform used to design and run those patient journeys. Medisafe states more than 13 million users.

    • Engagement sits on the patient's phone across brands
    • Best fit: oral or self-injected chronic therapies with a mobile-friendly population
  • Wellth

    A behavioural economics programme that pays members incentives for daily check-ins, using photo check-ins verified by AI. Its listed markets are Medicare Advantage and D-SNP plans, Medicaid, providers, employers and ACA plans rather than pharma brands.

    • Sold to plans and risk-bearing providers
    • Best fit: chronic populations measured on PDC and Star Ratings
  • Aptar Digital Health

    A division of Aptar Pharma offering connected drug-delivery devices and disease management software to pharma and biotech companies. Its HeroTracker Sense add-on for metered-dose inhalers received FDA 510(k) clearance in October 2025 and feeds its respiratory disease management platform.

    • Observed dose events, not inferred from refills
    • Best fit: inhaled and device-delivered therapies

Adherence analytics and medication optimisation platforms

  • IntegriChain

    Life sciences data aggregation. Its specialty data aggregation brings specialty pharmacy patient status, hub, copay, rebate claims and formulary data into one environment, and its Patient Analytics product is pitched at diagnosing initiation and adherence problems and triaging stalled cases and missed refills.

    • Best fit: brands in limited distribution that need one patient view across SPs and hub
  • Arine

    AI-driven medication optimisation and medication therapy management for health plans and at-risk provider groups across Medicare, Medicaid and commercial populations, including quality programmes tied to Star Ratings and HEDIS measures.

    • Plan-side, not brand-side
    • Best fit: Medicare Advantage and Part D adherence measures
  • Komodo Health and IQVIA (claims data)

    Not adherence products, but the claims sources most brands use to measure persistence outside their own channel. Closed claims with continuous enrollment let you say a patient genuinely stopped; open claims reach more patients but can lose them in channels the vendor does not see.

Medication adherence software compared

VendorCategoryWhat it does for adherenceMain buyer
AssistRxHubInitiation, education, refill reminders; journeys on Medisafe MaestroSpecialty brand patient services
CareMetxHubFull-service hub incl. adherence; now runs ex-Lash Group hub operationsSpecialty brand patient services
PhilRxDigital hub and pharmacyHub-triggered refills and auto-refillRetail and specialty-lite brands
AccredoSpecialty pharmacyCondition-specific pharmacist and nurse supportBrands in its network, plans
MedisafeDigital engagementApp reminders and sponsored patient journeysBrand teams
WellthIncentivesPaid daily check-ins verified by photoHealth plans, providers
Aptar Digital HealthConnected devicesRecorded dose events plus disease management platformPharma device and brand teams
IntegriChainAnalyticsSP and hub data aggregation, adherence and stalled-case analyticsCommercial analytics
ArineMedication optimisationAI-driven MTM and Star Ratings adherence workHealth plans

We did not find published list prices for any of the pharma-facing vendors above. Expect pricing per programme, per enrolled patient or per data feed, and ask for it in writing alongside the data terms.

How to measure medication adherence and persistence: PDC, MPR and time to discontinuation

Every vendor will show you an adherence number. Before comparing them, agree the definitions, because two correct calculations on the same patients can differ by a wide margin.

  • Proportion of days covered (PDC). The share of days in a measurement period on which the patient had the drug on hand, based on fill dates and days' supply. Overlapping supply is shifted forward rather than double counted, so PDC cannot exceed 100%. The Pharmacy Quality Alliance measures treat 80% as the usual adherence threshold, and 90% for antiretrovirals. It is the measure plans are judged on.
  • Medication possession ratio (MPR). Total days' supply dispensed in the period divided by the days in the period. Early refills are counted in full, so MPR can exceed 100% and tends to read higher than PDC. Many analysts cap it at 100%; ask whether the vendor does.
  • Persistence and time to discontinuation. Persistence is how long a patient stays on therapy from the first fill. A patient is counted as discontinued when the gap after their supply runs out exceeds an allowed grace period, commonly 30, 60 or 90 days, or a multiple of days' supply. Report it as the share still persistent at 6 and 12 months, or as median time to discontinuation on a survival curve.
  • Refill gap. Days between the expected run-out date and the next fill. It is the earliest warning you have and the variable most prediction models lean on.
  • Specialty status. In limited distribution, specialty pharmacy status and reason codes (active, on hold, discontinued for side effects, cost or switch) explain why, which refills alone cannot.

Three settings change every one of these numbers: the grace period, the index date (prescription, approval or first fill) and how you treat hospital stays, samples and free goods. Fix them in writing before a pilot.

What data each patient adherence approach needs

Specialty pharmacy

The backbone for limited distribution brands. Each fill with date, quantity and days' supply, plus status and reason codes, usually delivered through a data aggregator. Gaps come from pharmacies outside the network and inconsistent status coding between pharmacies.

Hub

Enrollment, benefit verification, PA outcomes, copay use and every outreach attempt. This is the only place you see what the programme did to each patient, which you need to tell whether an intervention worked.

Claims

Shows fills outside your network, switches to competitor drugs and medical events. Closed claims with continuous enrollment are the stronger basis for persistence; open claims reach more patients but leave gaps.

CRM

Veeva or Salesforce records of calls, samples and education by prescriber. Lets you see whether persistence differs by account and whether field teams can act on a drop.

Device and app

Connected device events and app check-ins show daily behaviour between fills. Coverage is limited to patients who opted in, so treat it as a strong signal for a subset, not a population measure.

Can AI predict which patients will stop therapy?

Often, yes. A model trained on refill gaps, status changes, copay events, PA renewal dates and prior outreach can rank patients by risk of discontinuing in the next 30 to 60 days, so nurses and pharmacists call the right people first.

Two cautions. First, a simple rule such as "seven days past expected refill" is a strong baseline; a model should beat it on your data before you pay for it. Second, a prediction only matters if someone acts on it and you can measure the effect, which needs a holdout group.

This is also where many teams find the packaged tools fall short: the hub sees its patients, the specialty pharmacy sees its fills, and nobody owns the joined view. Greenwolf Tech Labs builds custom adherence and persistence dashboards and discontinuation models on a team's own hub, specialty pharmacy, claims and CRM data, usually sitting on a pharma commercial data warehouse the brand controls. If you are earlier in the process, our patient journey analytics tools guide covers the wider analytics market.

How to choose medication adherence software

Find where your patients actually stop

Pull 12 to 24 months of dispense and hub data and plot persistence by month. If most loss is before the first fill, you need an onboarding tool, not an adherence one; see our guide to AI tools for patient onboarding.

Match the category to the cause

Cost and access problems need hub and copay work. Forgetfulness and routine need nudges or devices. Side effects need clinical support from a pharmacist or nurse. An app will not fix an expired prior authorisation.

Agree the metric definitions up front

PDC or MPR, grace period, index date, and the population. Put them in the statement of work so the vendor's report and your own calculation agree.

Insist on patient-level data back

De-identified, tokenised, event-level records in your own warehouse, weekly or faster. If a vendor will only give you a dashboard, you cannot check its numbers or join them to claims.

Ask how they will prove it worked

A credible vendor proposes a control or holdout group. Pre and post comparisons flatter every programme because patients who enroll are already more engaged.

Check privacy, consent and channel rules

Confirm HIPAA terms, patient consent for each outreach channel, SMS opt-in rules and who owns the patient relationship if the contract ends.

A 90-day adherence pilot to run before you sign

Back-test on your history (weeks 1 to 3)

Give the vendor de-identified data up to a cut-off date and keep the following quarter back. Ask them to name the patients most likely to discontinue. Score the list against what happened and against the simple refill-gap rule.

Randomise a live cohort (weeks 4 to 12)

Split new or at-risk patients at random into an intervention group and a usual-care holdout. Keep the holdout big enough to measure a difference, and do not let teams pick who gets the programme.

Measure on your definitions

Compare PDC, share persistent and refill gaps between the two groups using your own calculation from dispense data, not the vendor's dashboard.

Check the operational cost

Count nurse and pharmacist time, complaints, opt-outs and data issues, not only the adherence result.

Decide on a pre-agreed threshold

Write down beforehand what difference would justify a contract, so a weak result cannot be reframed as success.

When a custom build beats buying patient adherence software

Buy when the problem is a service: you need nurses, pharmacists, a hub or a device, and the vendor brings people and infrastructure you do not have. No dashboard replaces a pharmacist calling a patient about side effects.

Build when the problem is the measurement and the targeting. If your data already arrives from a specialty pharmacy aggregator, a hub and a claims vendor, the hard part is joining it on one patient token, applying your definitions and getting a daily list of patients to call into the hands of the people who call them. That logic is specific to your brand, your network and your status codes, and it should sit in data you own so it survives a change of hub vendor. The CareMetx and Lash Group transaction this year is a reminder that vendor arrangements change.

Many brands do both: buy the service, build the measurement layer that tells them which service is working. For brands in limited distribution, our specialty pharma analytics work and the companion guide to AI tools for patient support programs cover how that fits together.

Medication adherence software: common questions

What is the best medication adherence software for pharma companies?
It depends on why your patients stop. Specialty brands with access and cost problems usually start with a hub such as AssistRx or CareMetx; brands with mobile-friendly chronic populations look at Medisafe; device-delivered therapies look at connected devices such as Aptar Digital Health; and brands in limited distribution need analytics such as IntegriChain to measure persistence across pharmacies. Run a holdout pilot before committing to any of them.
What is the difference between medication adherence and persistence?
Adherence is whether a patient takes their medicine as prescribed while on therapy, usually measured as PDC or MPR. Persistence is how long they stay on therapy before discontinuing, measured as time to discontinuation or the share still on therapy at 6 or 12 months. A patient can be adherent for three months and then stop, which hurts persistence but not their adherence score for that period.
How is PDC calculated, and what is a good PDC?
PDC is the number of days in a period on which the patient had medication on hand, divided by the number of days in the period, with overlapping fills shifted forward so it never exceeds 100%. The Pharmacy Quality Alliance measures use 80% as the usual threshold for adherence, and 90% for antiretrovirals.
Which adherence platforms are used for Medicare populations?
Plan-side tools are built for Medicare Advantage and Part D adherence measures. Arine provides AI-driven medication optimisation and MTM for health plans, and Wellth runs incentive-based daily check-ins for Medicare Advantage, D-SNP and Medicaid plans. Pharma brands usually reach these patients through hubs and pharmacies rather than buying these tools directly.
Can AI tools automate patient communication to improve adherence?
Yes. Hub platforms, specialty pharmacies and apps such as Medisafe send automated refill reminders, education and check-ins by app, SMS or phone, and route patients who do not respond to a person. The useful AI is in deciding who to contact and when, based on refill gaps and status changes, so staff time goes to patients most likely to stop.
What data do I need to measure persistence for a specialty drug?
At minimum, patient-level specialty pharmacy dispense records with fill date and days' supply, plus status and reason codes, joined by a patient token. Adding hub case data shows what the programme did; adding claims shows fills outside your network and switches to other drugs.
How much does medication adherence software cost?
We did not find published list prices for the pharma-facing vendors in this guide. Pricing is typically quoted per programme, per enrolled patient or per data feed, and implementation and data integration are often priced separately, so ask for both.

Want to see where your patients are dropping off therapy?

Tell us which hub, specialty pharmacy, claims and CRM data you already receive. We will tell you what an adherence and persistence dashboard built on it would show, and whether a custom build or a vendor is the better fit.

Contact us

Part of our Pharma Commercial and Patient Analytics hub. Start with our specialty pharmacy analytics.

Build, not buy

Want an AI tool built around your data?

Tell us which tool on this list came closest and what it is missing. We will come back with what building the right one for pharma and life sciences teams takes.

  • Runs on your CRM, hub, specialty pharmacy and claims data
  • Built inside your own systems, not another login
  • A fixed timeline after the first call
  • A reply from a consultant, usually within one working day

We reply from [email protected], usually within one working day. We do not add you to a mailing list.