Methodology
The matching engine, the data behind it, and what we deliberately leave out.
Last updated · April 2026
1. Data sources
Plan attributes (coverage caps, deductibles, exclusions, pre-existing condition rules, sub-limits) are sourced from:
- The insurer's official policy wording (PDF) - primary source.
- The insurer's official brochure or product page - secondary source.
- Direct confirmation from the insurer or their TPA where wordings are ambiguous.
Each attribute is stamped with a source URL and verification date in our internal evidence log. We re-verify pricing samples and any field flagged as “changed” by an insurer at least quarterly.
2. Filtering (hard rules)
A plan is removed from a family's results - not down-ranked, removed - if any of the following are true based on the answers in the quiz:
- Visitor's age falls outside the plan's eligible age band.
- Trip length exceeds the plan's maximum or falls below the minimum.
- Coverage region does not include the United States.
- A material declared health condition is excluded by the plan's policy wording (e.g. dialysis, active cancer treatment) and no rider can cover it.
3. Scoring (soft signals)
Remaining plans are scored across dimensions that matter for the specific family context. Weights are not the same for every visitor - they shift based on quiz answers. For example:
- Pre-existing condition coverage. Heavily weighted when the visitor declares any chronic condition; lightly weighted otherwise.
- Coverage cap and deductible. Weighted against expected out-of-pocket exposure for typical emergency scenarios at the visitor's age band.
- Network and direct-billing. Higher weight in metros with strong PPO coverage; lower where reimbursement-only is the norm.
- Claims handling reputation. Based on the insurer's published claim settlement ratio, grievance ratio, and editor-reviewed handler track record.
The full default weight matrix is published in our codebase and is reviewed by the editorial team on a documented cadence. We will publish a changelog when weights are materially updated.
4. What we deliberately don't do
- We do not rank plans based on what a partner pays per signup. Where we have any commercial relationship with a plan provider, that relationship is disclosed on the plan card and on our editorial policy page. It does not affect ranking.
- We do not surface a single “best” plan. The output is always 2–3 fits.
- We do not collect phone numbers as a precondition to seeing results.
- We do not auto-renew, auto-quote, or make purchases on the visitor's behalf.
5. Limitations
Visitor insurance is short-term emergency cover, not US health insurance. It is not a substitute for a domestic plan, and it does not cover routine care, planned procedures, or most pre-existing conditions in full. Our results page calls this out for every plan.
If you spot an error in our data, please write to corrections@backtoindia.com. Verified corrections are typically reflected within five business days.