Don't Buy the Policy. Buy the Insurer.
A flawless policy from a company that fights claims is worth less than a plain one from a company that pays. And unlike the policy, the company's record is public.
A flawless policy from a company that fights claims is worth less than a plain one from a company that pays. And unlike the policy, the company's record is public.
Two plans sit next to each other on a comparison site. Same premium, around ₹15,000 a year. Same cover, ₹10,00,000. The feature ticks line up almost perfectly. Cashless hospitals, no room rent cap, day care covered, restoration included. By every column on the screen, they are twins.
The screen does not show you the one thing that decides whether you ever see your money. How the company behaves when a claim actually lands on its desk.
That behaviour is not a secret, and it is not a matter of online reviews and anecdotes. Every insurer in India files it with the regulator, every quarter, in a standard format, and publishes it on its own website. Claims received, claims paid, claims rejected, complaints, how long settlements took. It is dry, public, and almost nobody reads it. People will spend three weekends comparing features and not three minutes on the record of the company that has to keep the promise. That is exactly backwards.
Here is what to look at, and what it tells you.
How often do they say no?
The first number is the repudiation rate, the share of claims an insurer rejects outright. It sits in the quarterly claims disclosure, the form numbered NL-37 on every insurer's site.
In its most recent public filings, one of India's largest health insurers repudiated close to one in ten health claims. Across the industry the number swings a lot, which is the point. There is no "normal." There is only this insurer versus that one, and the only way to know is to pull the figure for each plan on your shortlist.
For every ten families who filed a claim with that insurer, roughly one was told no. Before you hand a company your premiums for the next twenty years, you want to know which side of that ratio you are likely to be on.
The add-on they push hardest
Now look one level down, at the segment you are actually buying, because a single company can behave completely differently across its products.
The same insurer that rejected about one in ten health claims rejected more than seven in ten of its personal accident claims. That is the rider that gets cheerfully upsold at the end of the call, "add personal accident cover for just a little more." On its public record, that add-on was denied roughly three times for every four claims made.
The lesson is not "avoid this company." It is that the headline health number does not carry over to the rider you are about to bolt on. Check the behaviour of the specific cover you are buying, not the company's friendliest segment.
Fast, or just fast on the easy ones?
Speed looks like the easy win. That same insurer settled about 98.7 per cent of its claims within a month, which sounds excellent, and for routine claims it is.
But read it by value, not just by count. The small, simple claims clear in days. The large ones, the cardiac surgeries and the long ICU stays, are the ones that take longer and draw more scrutiny. So a company can be fast on the ₹40,000 claims and slow and sceptical on the ₹7,00,000 one, and the blended "98.7 per cent in a month" hides that completely.
You are not buying insurance for the dengue admission that settles in three days. You are buying it for the big, frightening bill. Ask how the company performs on that, not on the average.
The do-over rate
There is a quieter number worth finding. The share of claims that were closed and then had to be reopened. At that insurer it ran a little over one per cent.
A reopened claim usually means the first decision was wrong, or the first settlement was not enough, and the customer had to push to get it fixed. A low number is reassuring. A high one tells you that getting paid correctly the first time is not something you can count on.
Complaints, and who turns out to be right
Every insurer also files its grievance data, in the form numbered NL-45. Two figures matter.
The first is complaints per ten thousand claims. At the insurer we have been following, it was about 54. A reasonable industry benchmark is around 20. More than double the norm is not noise. It is a pattern.
The second is subtler and more revealing. Of the complaints that got resolved, look at how many were decided in the customer's favour versus the company's. When a large share of disputes end with the insurer reversing its own original decision, that tells you the first decisions were not good ones. People had to fight, and when they fought, they were often right. An insurer that quietly overturns a lot of its own calls is an insurer whose first answer you cannot trust.
A record for one anonymized real insurer across six signals: settlement speed on routine claims and reopened claims are better than most (green); health claim denials and speed on large claims are worse than most (yellow); personal-accident denials and complaints per 10,000 claims are far worse than most (red).
You can check all of this yourself
None of these numbers are behind a paywall. The claims data is in the disclosure marked NL-37, the settlement speed in NL-39, and the grievances in NL-45, all published on each insurer's own website every quarter. It is written to be ignored, not hidden.
Reading it is exactly the job avikCover does, pulling these filings for every insurer and turning them into a single, checkable score so you do not have to decode regulatory forms on a Sunday. But the deeper point holds whether you use us or not. The information exists.
Trust in this industry should be something you look up, not something you are asked to feel.
The four questions
Before you choose between two plans that look like twins, ask four things about the companies behind them.
How often do they deny claims in the exact segment I am buying? How do they handle the big claims, not just the small fast ones? How many complaints do they draw, and how often do they end up reversing themselves? And how often do claims have to be reopened?
Features are easy to compare and matter least. The company is harder to check and decides everything. Spend your effort where it pays.
Next in this series: the company's accounts hold one more secret, the early warning that your premium is about to jump.
Educational, not advice. The figures here are drawn from one real insurer's most recent quarterly public disclosures (around Q3 FY2025-26), shown anonymized and next to a rough industry benchmark for context. Filings refresh every quarter, so check the latest figures for the specific insurer and segment you are considering.
Frequently asked
- How can I check a health insurer's claim rejection record?
- It is public. Every insurer files its claims data each quarter (Form NL-37) and its grievance data (Form NL-45) and publishes them on its own website. Look at the repudiation rate (the share of claims rejected), complaints per 10,000 claims, and how those complaints are resolved.
- Is a high claim settlement ratio enough to judge a health insurer?
- No. A high share of claims settled within a month can hide slow, sceptical handling of large claims, because small routine claims clear fast and big ones draw more scrutiny. Read settlement speed by value, not just by count, and look at denials and complaints too.
- Why are personal-accident claims rejected so often?
- An insurer's record varies by segment. In one real case, an insurer that denied about 1 in 10 health claims denied roughly 3 in 4 of its personal-accident claims. Check the public record for the specific cover you are buying, not the company's friendliest segment.
- What is a reasonable complaints ratio for a health insurer?
- A rough industry benchmark is around 20 complaints per 10,000 claims. One insurer in its public filings ran about 54, more than double. Also look at how often resolved complaints end with the insurer reversing its own original decision, which signals weak first decisions.
- Should I choose a health plan on features or on the insurer's record?
- Features are easy to compare and matter least. The company's public claim record, how often it denies, how it handles big claims, how many complaints it draws and reopens, decides whether you actually get paid. Spend your effort there.