A two-minute doorstep check for an older person's health.

New health workers should complete a half-day orientation before using EHSAS independently, covering:

• A walkthrough of all 12 items and the probes in section 6
• Practice administering the tool on 2–3 volunteers, observed by a supervisor or trained peer
• The referral pathway for your facility (who to refer to, and how)
• How to handle the situations described in section 7

Consent is verbal, not written, but must still be genuine: explain who you are, why you're asking, that participation is voluntary, that it takes 5–10 minutes, and that a "Yes" answer only leads to a referral for a further check-up — never a penalty or a mandatory hospital visit. If the person declines, thank them, note it in the register as "screening declined", and do not press further.

• Printed EHSAS-C tool forms (one per person), in the language the person is most comfortable in
• Black or blue pen
• Blank referral slips
• CHW record register
• Clipboard or file folder for field work
• Unique health ID list for the village, if one exists
• A private or semi-private space to talk — a corner of the courtyard is enough; a crowded room is not

• Greet the older person respectfully, by name if known, and sit at their eye level rather than standing over them.
• Ask whether they would prefer to answer alone or with a family member present — some questions (bladder control, mood, memory) can feel embarrassing in front of others.
• Explain the purpose in plain language: "We are doing a simple health check-up to find out if you may need further care at the health centre. This will take 5–10 minutes, and there are no right or wrong answers."
• Keep your tone unhurried and neutral throughout — older adults often under-report problems if they sense the health worker is in a rush or judging them.

1. Fill in the person's name, age, sex, and ID at the top of the form before you start.
2. Ask questions 1–11 exactly as written, in order. All are Yes/No questions about the last six months, not lifetime problems — a fall eight years ago doesn't count; a new problem with balance this year does.
3. Mark "Yes" only if the person clearly confirms the problem exists; mark "No" if it is absent, or if the answer is genuinely unclear after a probe.
4. Address the older person directly, even if a family member is present. If they are unable to answer (e.g. significant hearing loss or confusion), consult the caregiver — and note on the form that the answer was proxy-reported.
5. Do not prompt or suggest an answer. If a question isn't understood, repeat it once more slowly, then use the probe from section 6 — never rephrase in a way that hints at the "correct" answer.
6. After question 11, step back and observe the person (walking to the door, standing up from a chair, general build and alertness) to complete question 12 — see section 8.
7. Note any current complaints or known chronic conditions (diabetes, hypertension, arthritis, etc.) in the remarks box — this helps the receiving clinician, even for people who are not referred today.
8. Total the "Yes" answers, apply the referral rule in section 9, and tell the person the outcome before you leave.

Use only if a question isn't understood on first asking — these clarify without leading the answer.

• Hearing loss: face the person directly, speak slowly rather than loudly, and use the printed form as a visual aid if they can read.
• Cannot read or is not fluent in the printed language: ask every question aloud in the local language, using the probe if needed — the tool does not require the participant to read anything themselves.
• No caregiver available and the person cannot answer reliably (e.g. marked confusion): complete what you can, mark unanswered items as "unable to assess" rather than guessing "No", and note this clearly for the receiving clinician — this itself is often a reason for referral.
• Caregiver and older person disagree: record the older person's own answer where they are able to give one; note the caregiver's differing view in the remarks box rather than overwriting the response.
• Bedridden or unable to stand for question 12: judge frailty from the closest applicable cues — strength of grip, ability to sit up unassisted, overall condition — and note in remarks that gait could not be observed directly.
• Person becomes distressed by a question (this can happen with the mood or memory items): pause, reassure them that this is a routine check for everyone their age, and continue only if they're comfortable to do so.

This item is observed and judged by the health worker, not asked as a question. Watch the person walk a few steps, stand up from a chair, and note their general build, before choosing the picture that best fits:

• Fit — walks independently and steadily, no visible frailty
• Mildly Frail — slower or more cautious gait, but no aid used
• Moderately Frail — uses a stick, furniture, or another person for support while walking
• Severely Frail — needs a wheelchair, or help from two people to move short distances
• Terminally Ill — bedbound, or in the final stages of a known terminal illness

When in doubt between two adjacent levels, choose the more severe one — this errs on the side of referral, which is the safer default for a screening tool.

Add up the "Yes" answers to questions 1–11 (maximum score: 11). Refer for Comprehensive Geriatric Assessment at the nearest facility if the person:

• Scores 3 or more out of 11, or
• Is judged Moderately Frail or worse on question 12 — regardless of the score on 1–11.

Worked example: an older person answers "Yes" to walking difficulty, weight loss, and forgetfulness (score = 3), and is observed to walk steadily without an aid (Mildly Frail). Score of 3 meets the referral threshold on its own, so this person should be referred, even though their observed frailty is mild.

This cut-off was chosen (from ROC analysis against the Fried frailty phenotype) to favour sensitivity and a high negative predictive value — in practice, this means EHSAS is deliberately tuned to occasionally refer people who turn out not to need CGA, rather than miss people who do.

• Attach the completed EHSAS form to the referral slip.
• Clearly record the person's name and age, the health worker's name and contact number, and the date.
• Explain to the person and family, in simple terms, what a CGA involves — a fuller check-up covering physical, mental, and social wellbeing — and that it should happen as soon as reasonably possible, not urgently/emergently unless the presentation suggests otherwise.
• Record the visit, the score, and whether a referral was made in the CHW register.
• Where possible, follow up in 1–2 weeks to confirm the person attended, and note the outcome. Repeated non-attendance is worth flagging to your supervisor, as it may point to access barriers (distance, cost of travel, lack of an accompanying family member) rather than the person's own reluctance.
• Re-screen at the next routine visit for anyone not referred today — EHSAS reflects a point in time, not a permanent status.

To keep EHSAS screening consistent across health workers over time:

• Directly observe each new health worker administering EHSAS at least twice before they work unsupervised.
• Periodically spot-check a small sample of completed forms each month for missing items, inconsistent marking, or unusually high/low referral rates from one worker compared to peers in a similar area.
• Common errors to watch for: leading questions, skipping the probe step, marking "No" by default when an answer is unclear rather than asking again, and treating question 12 as something to ask rather than observe.
• Refresh training annually, or sooner if the tool, referral pathway, or catchment facility changes.

How often should someone be re-screened? At every routine home visit or health camp — commonly every 6–12 months for someone not currently referred, or sooner if their health changes noticeably.

What if the person is under 60? EHSAS is validated for ages 60 and above. It isn't designed or tested for younger adults, even if they have age-related health complaints.

Can I use it with someone who has known dementia? Yes, with a caregiver's help for items the person can't reliably self-report — mark these as proxy-reported. A known dementia diagnosis is, in itself, usually reason enough to refer for CGA regardless of the EHSAS score.

Does a "Yes" answer mean the person definitely has that problem? No — EHSAS is a screen, not a diagnosis. A "Yes" flags a domain worth a closer look at the facility; it is not itself a clinical finding.

Which language should I use? Whichever the person is most comfortable conversing in. The tool is available in multiple Indian languages on this website — see the Translations section above.

Researchers and programme managers outside this study site sometimes ask whether EHSAS can be adapted for a different state, country, or cadre of health worker. In our experience, the following is worth checking before doing so:

• Confirm the referral pathway you'll point people to actually offers a form of CGA, or your referrals will lead nowhere.
• If translating to a new language, follow a forward-translation, expert back-translation, and cognitive-interview process (see the Translations section) rather than a direct machine translation, given this is a clinical screening instrument.
• Re-validate the cut-off locally where feasible, since sensitivity/specificity can shift with a different population's baseline frailty prevalence — the ≥3 threshold was derived in rural Odisha and may not transfer perfectly everywhere.
• Keep the frailty pictogram (question 12) as-is; it was deliberately designed to be language- and literacy-independent, which is part of what makes it portable.

We're happy to discuss adaptations directly — see the contact details in the About section below.