Humanitarian and conflict · Sphere and WHO guidance

Rapid health assessment and surveillance

You cannot run a health response blind. A few numbers, collected the same way every week, tell you whether people are dying faster than they should, which diseases are driving it, and whether an outbreak is starting. Every clinician and community health worker is part of this system.

  • For clinicians, data officers and health coordinators in humanitarian settings
  • Reviewed October 2026

The questions to answer first

  1. How many people, by age and sex, and where are they?
  2. How many are dying each day, and of what?
  3. What is making people ill: the top causes of consultation, and any unusual events?
  4. What is the risk of outbreaks: crowding, water and sanitation, measles vaccination coverage, malnutrition, the season?
  5. Which services still work: facilities, staff, medicines, referral routes, laboratory?
  6. Who is being missed: people with disabilities, older people, minority groups, people far from services?

Use existing data first (national surveillance, previous surveys, facility records), then fill gaps with short visits, key informants and the community.

Is this an emergency? Mortality thresholds

The most useful measures of how severe a crisis is are the crude mortality rate (CMR) and the more sensitive under-5 mortality rate.

MeasureEmergency threshold (no known baseline)
Crude mortality rateMore than 1 death per 10,000 people per day
Under-5 mortality rateMore than 2 deaths per 10,000 children under 5 per day

Where the usual (baseline) rate is known, a doubling or more of it signals a serious public health emergency that needs an immediate response. Thresholds should be agreed at country level: if the baseline is already above the emergency threshold, do not wait for it to double.

Source: Sphere Handbook 2018, health chapter

Standard formulas

Use the same formulas and the same population figures as every other agency, so that numbers can be added up and compared.

IndicatorHow to calculate itUnit
Crude mortality rateDeaths in the period ÷ (mid-period population × days in the period) × 10,000deaths per 10,000 people per day
Under-5 mortality rateDeaths in children under 5 ÷ (children under 5 × days in the period) × 10,000deaths per 10,000 children under 5 per day
Incidence rateNew cases of a disease ÷ (population at risk × months) × 1,000new cases per 1,000 people per month
Case fatality ratePeople who died of the disease ÷ people with the disease in the period × 100per cent
Health facility utilisationVisits in one week ÷ total population × 52visits per person per year
Consultations per clinician(Consultations in one week ÷ full-time-equivalent clinicians) ÷ days open per weekconsultations per clinician per day

Low use of a clinic may mean poor quality, cost or other barriers, or an overestimated population. Very high use may mean an outbreak, an underestimated population, or that services elsewhere are failing.

Source: Sphere Handbook 2018, appendix 3: formulas for key health indicators

Early warning, alert and response (EWAR)

  • Agree a short list of priority diseases and events based on the local risks (for example acute watery diarrhoea, bloody diarrhoea, measles, acute jaundice, suspected meningitis, unexplained fever, unusual clusters of deaths).
  • Use two streams: indicator-based weekly counts from facilities, and event-based reports of anything unusual, including rumours and community reports.
  • Every alert has a disease-specific alert threshold. For some diseases, a single suspected case is an alert.
  • Share a weekly EWAR report with all partners so that action is taken.
StepTarget
Alert reportedWithin 24 hours (90% of alerts)
Alert verifiedWithin 24 hours of the report (90%)
Field investigation startedWithin 24 hours of verification (90%)
Weekly surveillance reports complete and on timeAt least 80%
Health information report from the lead agencyAt least monthly

Source: Sphere Handbook 2018, health systems standard 1.5 and communicable diseases standard 2.1.2

What the clinician does

  1. Know the case definitions for your priority diseases (they are on the reporting form) and apply them to every patient.
  2. Report an alert at once by the agreed route; do not wait for the weekly report.
  3. Record the standard information: age, sex, place, date of onset, outcome.
  4. Take samples if you have been asked to and have the materials, and send them as agreed.
  5. Keep treating: case management does not wait for test results.
  6. Count deaths, including deaths at home reported by community health workers.

Two definitions every clinic needs

Suspected cholera
Where no outbreak has been declared: anyone aged 2 years or more with acute watery diarrhoea and severe dehydration, or who dies of it. Once an outbreak is declared: anyone with, or who dies of, acute watery diarrhoea (GTFCC).
Suspected measles
Fever with a maculopapular (non-blistering) rash, or any patient a health worker suspects has measles (WHO measles outbreak toolkit).

Case fatality rates to aim below

A high case fatality rate usually means people are reaching care too late or care is poor. Sphere suggests these acceptable levels:

DiseaseCase fatality rate
CholeraUnder 1%
MeningitisUnder 15%
Hepatitis EUnder 4% overall (10 to 50% in women in the third trimester of pregnancy)
Diphtheria (respiratory)Under 5 to 10%
DengueUnder 1%

Source: Sphere Handbook 2018, communicable diseases standard 2.1.3

Using and protecting data

  • Agree common definitions, population figures and place names with all partners.
  • Break data down by sex, age and disability, and other factors that matter locally.
  • Protect people: in conflict and displacement, health data can put patients, staff and communities at risk. Collect only what you need, store it safely and share only aggregate figures unless there is a clear clinical reason.
  • Feed back to the clinics and the community: data that only flows upwards is rarely good data.

Source: Sphere Handbook 2018, health systems standard 1.5

Sources

  1. Sphere Association. The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response, fourth edition (health chapter and appendix 3). 2018.
  2. Global Task Force on Cholera Control (GTFCC). Cholera patient treatment flowchart, version 1.0 (surveillance case definition). 2024.
  3. World Health Organization. Measles outbreak toolbox (suspected case definition, based on Surveillance standards for vaccine-preventable diseases, 2nd edition, 2018). 2018.
Adapted from WHO guidance; not endorsed by WHO. This page paraphrases and adapts the WHO publications listed under Sources, which are licensed CC BY-NC-SA 3.0 IGO; the adapted text is shared under the same licence for non-commercial education. Reasoning GP is not affiliated with WHO or any other organisation named here. Use it alongside your national guidelines, essential medicines list and formulary, and your clinical judgement. Doses are given only where they come from a WHO source; otherwise follow your national guideline or formulary.