About Vårdbetyg

Why Vårdbetyg exists

The information needed to compare health centres already exists, but it is scattered: the patient survey results in a results portal, the waiting times in the National Board of Health and Welfare's statistics and the contact details on 1177. Vårdbetyg brings them together per clinic, calculates a score with an openly described method and shows how many responses the score is based on. The idea is that you can compare clinics without having to interpret tables and statistics yourself.

Who runs the service

Vårdbetyg is built and run by Arvid Pilhall, who holds an MSc in computer engineering from Lund University (LTH). He has spent more than six years building and leading software products and today leads an engineering team. He runs Vårdbetyg as a sole trader, with no investors and no ownership ties to care providers or public authorities. The service charges care providers for two things: a licence to use the seal image for an award, and compiled reports. Who has paid and how much is listed on the funding page (in Swedish). Scores and rankings cannot be bought. If you have a question or a correction, email hej@vardbetyg.se.

Arvid also runs two sister sites that work the same way with open data. Omsorgsbetyg rates elderly care in Sweden, meaning care homes and home care (in Swedish). Practice Grade rates GP practices in England using the GP Patient Survey.

Independence and funding

Vårdbetyg never charges for scores, rankings or visibility. No care provider pays for placement or scores, and rankings and awards are calculated with the same method for everyone. They cannot be bought or negotiated.

There is no sponsorship, no advertising and no booking commission. We broker no bookings and receive nothing when you choose a particular clinic.

Scores are built on the National Patient Survey and the National Board of Health and Welfare’s waiting-time statistics, each with its measurement period stated. We do not take reviews directly, because they can be stuffed or manipulated.

If the funding changes, it is stated on our funding page (in Swedish), with amounts and who paid.

How the score is calculated

A common way to compare is to rank clinics by their raw average. That becomes misleading when few have responded. A clinic where a handful of patients answered can end up at the top or bottom by chance. So we calculate differently:

• Few responses are pulled toward the national average. The fewer who responded, the closer to the national average the score lands (the method is called Bayesian shrinkage). The number of responses is always shown next to the score, and the margin of error where the survey reports it.
• Too few responses give no score. If too few patients responded we write “too few responses” instead of showing a figure, because an average of a few responses mostly reflects who happened to answer.
• Waiting times count. For health centres with waiting-time data, patients' answers weigh 70 percent and the waiting times 30 percent. The waiting times are the share who got a medical assessment within three days and the share of calls answered, taken from the care providers' own records. Some health centres lack waiting-time data, and specialist and emergency clinics have no comparable per-clinic statistics. There the score rests on patients' answers alone, and the pages say so.

The score measures patient experience and accessibility, not medical treatment outcomes.

The health centres with the highest scores receive awards such as Top-rated in the municipality and Top 10% in Sweden. They are calculated with the same rules for everyone, marked with a year and cannot be bought. Specialist and emergency clinics receive no awards: there the regions report at different organisational levels, so a national leaderboard would compare unlike things. The full method is under Methodology.

Why we show no star reviews

Visitors cannot rate clinics with stars here. Star reviews on health sites often rest on a handful of voices, are easy to sway, and blur what one individual felt with how a clinic works for most people.

Instead we build on the National Patient Survey. Patients receive the survey after a visit, rather than seeking out a review site themselves, and everyone answers the same questions. That makes the answers harder to sway and easier to compare than free-text reviews. Those who answer are still not always representative of all patients. What that means for the score is explained under what the score does not capture. Where Google has public reviews we summarise them on the clinic's page, but they do not affect the score.

Where the data comes from

• The National Patient Survey gives patients’ answers and is coordinated by the Swedish Association of Local Authorities and Regions (SKR). We use three of its measurements, for health centres, specialised outpatient hospital care and emergency departments.
• The National Board of Health and Welfare’s waiting-time statistics give recorded waiting times in primary care and are the only source besides the patient survey that counts toward a score. Waiting times for specialised care and for emergency hospitals are shown as background, never inside a score.
• Other open sources give addresses, maps, contact details and operating form, from sources including 1177, the regions’ open registers and the Swedish Companies Registration Office. They are shown on the pages but never affect a score.

The scores are updated when new surveys are published. More about each source under About the data.

Privacy

You need no account and give no personal data to use Vårdbetyg, and we do not track you. There are no analytics cookies and no ad networks. We make no money from data about you. Read more in the privacy policy.

Editorial responsibility and corrections

Responsible for the content of Vårdbetyg is its founder, Arvid Pilhall. Scores and awards are calculated according to the published method, so that anyone can check them.

Found an error or a detail that is no longer correct? Email hej@vardbetyg.se and we will correct it.

Last reviewed: 21 October 2026