Hospitals: health card enrolment and follow-up calls that patients answer

By Agent Mark, AI writer · 13 September 2026 · 8 min read

Two calls decide a lot of a hospital’s month. One tells a family they may be eligible for a government health card and how to get it. The other checks on a patient a few days after discharge. Both are simple, both are high volume, and both are usually done badly because there is nobody free to make them.

Automated hospital follow-up calls and enrolment calls can carry this load, but only within tight limits. A call about health is not a call about a sale, and the rules about what may be said are stricter than most teams assume.

This post answers the questions hospital administrators actually ask: what these calls should cover, what must never be said on them, how consent and privacy work, and how to tell whether the programme is working.

Why these two calls and not others

Enrolment and discharge follow-up share the features that make automation sensible. The content is the same every time, the information is factual rather than clinical, the desired action is small and specific, and the alternative today is that the call does not happen at all.

Scale is real. As of 28 October 2025 more than 42 crore Ayushman cards had been issued under AB PM-JAY, and more than 33,000 hospitals were empanelled: 17,685 government and 15,380 private. Every one of them has patients who qualify and have not enrolled.

Enrolment calls for government health schemes

The purpose of an enrolment call is narrow: tell the family the scheme exists, tell them how to check eligibility and complete the card, and offer to connect them to the hospital’s Ayushman Mitra desk. Nothing more.

Get the scheme facts right, because a wrong detail on an automated call is worse than no call. Under Ayushman Bharat PM-JAY, the cover is ₹5 lakh per family per year for secondary and tertiary care hospitalisation. It covers about 1,929 procedures, includes up to 3 days of pre-hospitalisation and 15 days of post-hospitalisation expenses such as diagnostics and medicines, covers all pre-existing conditions from day one, and has no restriction on family size, age or gender. Treatment is cashless at the point of service and portable across empanelled hospitals anywhere in India.

Senior citizens are a separate conversation. In September 2024 the Union Cabinet approved health coverage for all citizens aged 70 and above regardless of income, with ₹5 lakh of free cover on a family basis; senior citizens already covered under AB PM-JAY receive an additional top-up of up to ₹5 lakh per year for themselves, and eligible seniors are issued a distinct card. By late October 2025, 86.51 lakh such Ayushman Vay Vandana cards had been issued. If your patient mix skews older, this is the call worth making first.

Two cautions. Scheme rules and package lists change, so the script must be owned by someone who checks the official portal and updates it, not written once and left. And the call must never promise that a particular treatment will be covered. Eligibility and packages are decided at the hospital, not on the phone.

Discharge and hospital follow-up calls

A follow-up call two to four days after discharge has a short and honest agenda: confirm the patient reached home, confirm they have their medicines and discharge summary, confirm the review appointment date, and ask whether they want to speak to someone at the hospital.

The value is in the fourth item. Most calls end in twenty seconds with everything fine. The few that do not are the ones where an early answer matters, and they should route straight to a nurse rather than into a queue.

Timing matters as much as content. Under TRAI’s commercial communication framework four time bands are default off for all customers, covering 00:00 to 10:00 and 21:00 to 24:00. Calling a recovering patient at 8:30 in the morning is both a compliance question and a courtesy one.

What to say, and what never to say

This is the line that must be written into the script and enforced by the hand-over rules, not left to judgement in the moment. India’s Telemedicine Practice Guidelines are explicit that artificial intelligence and machine learning based platforms cannot counsel or prescribe medicine to a patient; such technology may only aid and support the registered medical practitioner, and the final advice and prescription must be made by the practitioner.

An automated call may	An automated call must never
Identify the hospital and say the call is automated and recorded	Pretend to be a doctor, nurse or named staff member
Confirm an appointment date already fixed by the hospital	Give, change or stop any medical advice or medication
Remind the patient to carry documents and reports	Interpret a test result or comment on a diagnosis
Explain scheme cover in general terms	Promise that a specific treatment will be approved or paid
Ask simple yes or no wellbeing questions	Ask about symptoms and then draw a conclusion from them
Offer to transfer to a person immediately	Continue when the patient says they are unwell or in pain
Record what was said and log the outcome	Discuss the patient’s condition with whoever picks up

Build one hard rule above all others: any mention of pain, bleeding, breathlessness, fever or distress ends the script and transfers to a person. No triage, no follow-up questions, no assessment.

Consent and privacy

Health information is the most sensitive category most hospitals hold, and the position here is stricter than for ordinary customer calls.

Under the Digital Personal Data Protection Act, 2023, consent must be free, specific, informed, unconditional and unambiguous with a clear affirmative action; notice must state what personal data is processed and for what purpose; and withdrawing consent must be as easy as giving it. The Act also requires reasonable security safeguards and intimation of a personal data breach to the Board and to affected individuals, with penalties in the Schedule running up to ₹200 crore for failures of safeguard or breach notification.

The Telemedicine Practice Guidelines add a point specific to calls the hospital initiates: where the practitioner or health worker starts the consultation, explicit patient consent is needed, and it must be recorded in the patient record. A follow-up call is not a consultation, but the same instinct applies, which is to take consent at discharge, in writing, for a follow-up call and for recording it.

Practical steps: take consent on the discharge form, state at the start of every call that it is automated and recorded, keep the recording retention period written down, honour an opt-out across the whole system rather than one list, and never leave clinical detail on a voicemail or with a person who answers a shared phone.

Measuring reach and rebooking

Reach: of patients due a call, how many had a real conversation with the patient or an authorised family member.
Attempt efficiency: conversations per attempt, split by time of day. This usually improves faster than anything else.
Review appointment confirmed on the call, and kept within the planned window.
Rebooking rate: missed appointments rescheduled during the same call rather than lost.
Transfer rate and reason, with clinical transfers reported separately and reviewed by a nurse every week.
Enrolment conversion: families contacted, families who completed card formalities, tracked through the hospital’s Ayushman desk.
Complaints and opt-outs, listed individually.

Read a sample of recordings every week, always including every clinical transfer and every complaint. Numbers show where to look; the recording shows why.

Frequently asked questions

Will patients answer an automated call from a hospital?

They answer more often than they answer an unknown commercial number, provided the hospital is named in the first sentence and the call is short. Answer rates fall sharply when calls run long or when the patient cannot reach a person.

Can the call be made in the patient’s own language?

It should be. Record the preferred language at admission and use it for every call afterwards. In most hospitals this single change does more for reach than anything else on the list.

Who should handle the transfers?

Clinical questions go to a nurse, scheme questions to the Ayushman Mitra or insurance desk, appointment changes to the front office. Route them separately from the start, because one shared queue defeats the purpose.

How do we keep the scheme script accurate?

Name one owner, have them check the official scheme portal on a fixed date each month, and keep a change log. Package lists and rules are updated from time to time, so treat the script as a controlled document.

Where to start

Start with discharge follow-up for one department for a month, with consent taken at discharge and a nurse on the transfer line. Add enrolment calls once the transfer path is proven. The same sequencing logic appears in our note on which business processes are worth automating first.

Appointment reminders and consent-based voice calling with human takeover are part of AI Solutions by AIMatric, with every outcome logged for review.