Thirty-six items describing health across eight domains, summarised as physical and mental component scores. The SF-36 is one of the most used generic health measures in the world. Choose between the licensed SF-36v2 and the public RAND-36 before you build.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Physical component (PCS)
41.2
Mental component (MCS)
48.7
Items answered
36/36
At a glance
In trials
The SF-36 describes overall health rather than a disease, which is why it appears in trials of almost everything: surgery, cardiology, rheumatology, rehabilitation, nutrition. It lets results be compared with population norms and with other conditions, and it can reveal effects outside the target disease, positive or negative.
It is rarely the primary endpoint in drug trials, where disease-specific measures are more sensitive. Pair it with the disease measure and pre-specify whether PCS, MCS or particular domains are key secondary endpoints. For health economic analyses, utility values can be derived from SF-36 responses using established algorithms, or the EQ-5D can be collected directly.
The two share most item wording but differ in scoring: SF-36v2 uses norm-based scoring and summary scores with a licensed algorithm, while RAND-36 domain scores are 0 to 100 transformations. Results from one cannot be compared directly with the other, so choose based on comparator data and record the version on the form.
SF-36 · Month 6
Past 4 weeks · version recorded
General health
Item 1
SF01Structure
| Domain | Items | Contributes mainly to |
|---|---|---|
| Physical functioning | 10 | PCS |
| Role limitations, physical | 4 | PCS |
| Bodily pain | 2 | PCS |
| General health | 5 | PCS and MCS |
| Vitality | 4 | PCS and MCS |
| Social functioning | 2 | MCS |
| Role limitations, emotional | 3 | MCS |
| Mental health | 5 | MCS |
One further item on health change over the past year is not part of any domain score.
In Capture
Build the SF-36 as a questionnaire with the wording of the version you use. The form records the licence number, holder and expiry for SF-36v2, participants complete it on their phone or a clinic tablet, and raw responses export with codes and labels for scoring.
Copyright & licence
Approval pendingLicence number
Not recorded
Expiry date
Not recorded
Watermark shown to staff only. Participants never see it.
Record the version and licence in the free sandbox.
Burden
The SF-36 takes about 5 to 10 minutes. Added to a disease-specific questionnaire, a symptom diary and a quality of life measure, that can make visit questionnaires long enough to hurt completion. Consider the SF-12 when only summary scores are needed, or the PROMIS-29 when a broad profile with modern scoring fits better.
Whatever you choose, look at the full participant burden across the schedule, not one questionnaire at a time. Short, well-timed questionnaires completed at home usually produce more complete data than long ones completed in a waiting room.
The SF-36 was developed for self-completion and interview. Electronic versions should follow the licence holder's guidance, and the mode should stay the same at every visit so changes reflect health, not how the questions were asked.
Before go-live
SF-36v2 (licensed) or RAND-36, matching comparator data.
For SF-36v2; details on the form.
Standard four weeks or acute one week.
Licensed algorithm or RAND scoring, in analysis.
PCS, MCS or domains, pre-specified.
Total questionnaire time per visit.
About 5 to 10 minutes for most participants.
Yes. Established algorithms, such as the SF-6D, derive a utility index from SF-36 responses for health economic analysis.
Generic health status across eight domains, summarised as physical and mental component scores.
SF-36v2 uses a licensed norm-based algorithm with a mean of 50 and SD of 10. RAND-36 transforms each domain to 0 to 100.
SF-36v2 is commercially licensed. The RAND-36 uses the original items and is publicly available, with different scoring.
Four weeks for the standard version and one week for the acute version.
A 12-item short form that produces physical and mental component summaries.
Raw responses are stored with codes and labels, and scores are derived in analysis.
Keep exploring
Version and licence recorded with every form. Free sandbox.