AI Campaigns for Health Checkup Packages and Preventive Care

Campaign ideas to sell more full-body checkups

Before any campaign idea, the product needs examining, because the evidence on it is unusually clear and unusually bad. A Cochrane systematic review of general health checks in adults, updated in January 2019 and covering 17 randomised trials with 251,891 participants, found no effect on death from any cause. The all cause mortality risk ratio was 1.00 with a 95% confidence interval of 0.97 to 1.03, across 11 trials, 233,298 participants and 21,535 deaths, rated high certainty with zero statistical heterogeneity. Cancer mortality came out at 1.01 and cardiovascular mortality at 1.05, both with intervals crossing one. The review's conclusion is that general health checks are unlikely to be beneficial. In India this is now also a regulatory matter: the NMC guidelines of 6 October 2026 bar fear based marketing and content that creates demand for unnecessary tests, and separately bar discounts, limited period offers, contests, coupons, gifts and cashbacks used as inducements. So the standard campaign, a discounted seasonal bundle promoted with silent killer creative, fails on the evidence and on at least two reported clauses. The campaign ideas that do work are in the rest of this page and they share one feature: they sell a risk assessment and a pathway rather than a panel of tests.

What the evidence actually says, in full

The evidence is worth setting out properly, because a marketing team will have to defend this decision to a commercial director with a revenue line attached to checkup packages.

A risk ratio of exactly 1.00 with an interval of 0.97 to 1.03 and no heterogeneity across eleven trials is about as definitive as clinical evidence gets. It is not an absence of evidence of benefit. It is evidence of absence of benefit, at a precision that rules out anything but a very small effect in either direction. The review also notes that health checks may lead to unnecessary tests and treatments, which is the harm side that marketing never mentions.

Why the commercial case points the same way

The clinical argument will not carry a budget meeting on its own, so here is the commercial one. The discounted full body package is a poor product on its own terms, independent of the evidence.

Set against that, a risk assessment led model has the opposite properties. It is free or near free to enter, which removes the price conversation entirely. It produces a clinical recommendation specific to that person, which is differentiated by construction. It routes to tests with an actual indication, which are easier to defend, easier to bill and more likely to find something real. And it opens a relationship that has a reason to continue, because a risk score invites a review next year in a way a normal report does not.

The model that replaces it: sell the risk assessment

India already runs the alternative at national scale and almost no private hospital markets in alignment with it. The National Programme for Prevention and Control of Non-Communicable Diseases screens adults aged 30 and above for five specific conditions, using a short risk checklist administered in the community to decide who is referred for what.

Note what the national model does not do. It does not offer everybody a fixed list of forty tests. It assesses risk, then screens for the conditions that risk and symptoms justify. A private hospital that adopts the same logic is not giving anything up. It is applying a filter that makes every downstream test more likely to find something, which is better medicine and a better conversion story.

One caution on the detail. The description above comes from the revised operational guidelines for the programme, and secondary summaries of the package vary, with some adding COPD and stroke. Confirm the current package against the live NHM operational guidelines before building creative around a specific condition list, because the programme has been revised more than once and will be again.

Redesigning the offer

Three tiers, with the entry point free and the clinical logic visible at every step. This is the structure to brief.

Tier one is where the marketing effort goes and tier three is where the money is. That inversion is the whole redesign. A campaign selling tier two directly has to compete on price with every hospital in the city. A campaign offering tier one for free, delivering a genuinely useful personal recommendation, and then carrying people into tier three has no direct competitor in most Indian markets, because almost nobody is doing it.

The hardest internal conversation will be about tier two getting smaller. It will. A risk check honestly applied will tell a healthy 32 year old with no family history that they need very little, and the panel revenue from that person disappears. What replaces it is a 54 year old with a waist measurement and a family history who now has a reason to do three tests that matter rather than one bundle they bought on offer, and a reason to come back. Model it before the meeting, because the first quarter will look worse on package volume and better on indicated test volume, and somebody will quote the first number.
 

What a good risk check actually asks

Keep it to a single screen or a single page. The national checklist is short for a reason, which is that length kills completion.

• Age and sex, which between them determine most of what is worth screening for at all.
• Family history of diabetes, hypertension, cardiac disease and cancer, asked as a simple yes and no per condition rather than as free text.
• Tobacco in any form, including smokeless, which matters disproportionately in India and is the single most useful question for oral cancer risk.
• Alcohol, physical activity and waist circumference, the three that drive the metabolic cluster.
• Symptom questions for the cancers in the national package, because a symptomatic person should not be routed through a screening pathway at all. They need a clinical appointment, and the form must say so and offer one.
• For women, last cervical and breast screening dates, which is often the only question that needs asking.

The output matters as much as the input. A score alone is useless to a patient. The recommendation should name the tests, say plainly which ones are not indicated for this person and why, and give a date rather than an offer. A recommendation that declines to sell something is the most persuasive marketing asset a hospital can produce, and it is the one almost nobody is willing to build.

The compliance envelope, applied to actual creative

The guidelines themselves are explained in full on our NMC ethical advertising guidelines page. What follows is only the subset that bites on checkup and screening campaigns, applied to the lines a hospital actually publishes.

One caveat before the detail. The clause numbers below come from published analyses of the 6 October 2026 guidelines rather than from our own reading of the gazetted text. Verify each one against the official document before relying on it, and take a legal view on anything close to the line. The direction of all of them is not in doubt.

Clause 8.1(i) deserves a second reading because it is unusual and almost nobody has absorbed it. A rule barring content that creates demand for unnecessary tests reaches past the wording of an advertisement and into the thing being advertised. If the tests in your package are not indicated for the person you are marketing them to, the problem is not that the creative was too aggressive. The problem is the campaign. That is precisely where the Cochrane evidence stops being an ethical footnote and becomes a regulatory fact pattern.

Say this, not that

The practical translation, line by line. Everything in the right column is defensible under the reported clauses and most of it converts better anyway.

The pattern across the right column is that every compliant line is more specific than the one it replaces. That is not a coincidence. The barred mechanics are all forms of urgency and vagueness, and the permitted ones are all forms of information. A hospital that commits to the right hand column ends up with marketing that reads like a clinical service rather than a retail promotion, which is both the legal position and the brand position most hospital boards say they want.

The other three regimes that apply

Two process controls make all of this manageable and neither is expensive. First, a named clinician signs off every campaign concept, not only the claims within it, because clause 8.1(i) reaches the concept. Second, the compliance review happens at brief stage rather than at artwork stage, since a campaign built on a discount cannot be fixed by editing the copy. Most hospitals currently review at artwork and discover the problem when the media is booked.

AI targeting for preventive health packages

The useful job for AI here is stratification and routing, not reach, and the distinction decides whether the programme is lawful. Pointing a model at your patient base to find more people to sell a fixed panel to runs directly into two problems. The first is clause 8.1(i) as reported, which bars content creating demand for unnecessary tests, and a model optimised for package conversion will by construction find people for whom the package is not indicated. The second is consent. Under the DPDP Act there is no separate sensitive data category, and consent must be free, specific, informed, unambiguous and as easy to withdraw as to give. Holding a patient's treatment record gives you no basis to market a checkup to them, because the consent that covered treatment did not cover marketing. The version that works points the model at a different question: given what this person has told us in a risk assessment they consented to, which conditions should they be screened for and when. That is clinical decision support rather than targeting, it is defensible under both regimes, and it produces a better conversion rate because the recommendation is specific to the person receiving it. The honest summary is that AI makes a good screening programme more precise and makes a bad one more exposed, faster.

The consent problem most hospitals have not noticed

Almost every Indian hospital has a database of past patients and almost every marketing team assumes it can be used for checkup campaigns. Under the DPDP Act that assumption needs testing.

The notice obligations are specific and most hospital privacy notices do not meet them. A notice must state what data is collected, the specific purposes, who it is shared with including laboratories and third party administrators, retention periods, the data principal's rights and how to complain. The Rules contemplate multilingual access and layered notices, which for an Indian hospital with a mixed language catchment is not a formality.

The penalty scale makes this a board level matter rather than a marketing one. Reported maximums run to very large sums for security failures leading to a breach, for failures to notify, and for general fiduciary breaches. The full DPDP treatment belongs to a dedicated page in this library. What matters here is the single question of whether you have a lawful basis to send a checkup offer to the people on your list, and for most hospitals the honest answer today is no.

What to point the model at

Five jobs, in descending order of value, and none of them is finding more people to sell a panel to.

Row three is the one to build first even though it is the least like marketing. A screening programme that identifies abnormal results and does not follow them up is worse than no programme, because the person now believes they have been checked. A model that tracks pathway completion and prompts the gap is the clearest possible evidence that your screening is a clinical service rather than a revenue line, and it is the thing a regulator, a medical director and a patient would each most want to see. It is also, incidentally, where the follow on revenue is.

Where the resulting contacts land and how they are worked is the subject of hospital CRM and AI, and the attendance problem that follows a screening booking is covered in reducing patient no shows with AI.

The campaign calendar, rebuilt

Seasonal campaigning is not barred. Seasonal discounting is. The distinction gives you a full calendar to work with, as long as the hook is a health observance rather than a price window.

The calendar above has more occasions on it than the discount calendar it replaces, and every one of them gives you a reason to talk about a specific condition rather than a generic panel. That is better content, better targeting and better search and AI visibility, because an observance plus a condition is a question people actually ask.

Screening camps and corporate tie ups

Camps and corporate tie ups are the two channels where the economics still work cleanly, and they work for different reasons.

On corporate tie ups there is a confidentiality trap worth naming. An employer paying for screening will often ask for results. Aggregate, de-identified reporting is defensible and individual results are not, and a hospital that hands an employer a list of which employees have raised blood sugar has created an exposure far larger than the contract is worth. Put the reporting format in the contract rather than settling it later under pressure.

Measurement: count the right things

A checkup programme measured on package sales will be optimised toward the thing this page argues against, whatever the strategy document says. Change the metrics or nothing else changes.

The last row is the integrity check and it is worth instrumenting even though nobody will ask for it. A risk assessment that recommends the full panel to everybody who completes it is a sales funnel wearing a clinical costume, and the number that exposes it is the proportion of assessments that told somebody they did not need something. If that figure is zero after a quarter, the model is not doing the job and the programme has quietly reverted to the thing it replaced.

Limitations, and what would change our mind

What would change our mind. A large randomised trial showing mortality benefit from comprehensive health checks would undermine the clinical half of this page, and given the precision of the current estimate that is unlikely but not impossible. Formal guidance from the NMC clarifying that clause 8.1(i) does not reach package design would narrow the regulatory argument considerably, and is worth watching for. And published enforcement practice, in either direction, would tell hospitals far more than this page can, since the whole analysis currently rests on reading rules that have not yet been tested.

Conclusion: sell the assessment, not the panel

The question this page was asked to answer was how to sell more full body checkups, and the answer is that you should not, for three independent reasons that happen to agree.

The clinical reason is settled. Seventeen randomised trials covering 251,891 people produce an all cause mortality risk ratio of exactly 1.00, with a tight interval and no heterogeneity, and the review's authors conclude that general health checks are unlikely to be beneficial. That is not a gap in the evidence. That is the evidence.

The regulatory reason is new and most of the category has not absorbed it. The NMC guidelines of 6 October 2026 bar fear based marketing and content that creates demand for unnecessary tests, bar the discounts and limited period offers that the category's economics depend on, and bar paying an agency per booking. A campaign built on a festival discount for an eighty test panel promoted with silent killer creative now fails on three reported clauses at once.

And the commercial reason was true before either of the other two. The discounted panel is a one off, undifferentiated, price led product whose modal outcome is a normal report and whose abnormal findings arrive with no pathway attached. It was always a weak business and the regulation has simply removed the mechanics that were propping it up.

What replaces it is already running at national scale. Assess risk, screen for the conditions that risk and symptoms justify, and build the follow up that turns a finding into care. Market the assessment, which is free and lawful to promote. Price the indicated tests factually. Measure pathway completion rather than packages sold. The first quarter will look worse on package volume and better on almost everything else, and somebody in the room will quote the first number, which is why the model should be agreed before the quarter starts rather than during it.

Frequently Asked Questions For Health Checkup Marketing: The Campaign Model That Is Compliant