Assessment in emergency.

1
Assessment in emergency.

• The best analogy is clinical medicine.
• Qualitative (history and examination) and
  quantitative (measurements and laboratory) data
  are combined. These form a recognisable pattern
  to make a diagnosis and allow the severity of the
  situation to be assessed.
• A treatment plan is then formulated, implemented
  and progress followed. We do the same in any
  emergency.
• For a population in an emergency we need to
  recognise the pattern of problems, make the
  diagnosis, assess the severity and implement
  the relief (treatment).
• Wasting, oedema and mortality rates are critical
  measures of the severity of the insult, and the
  urgency of intervention. But usually only show
  whether the diagnosis is of a serious or
  not-so-serious nature.

2
Quantitative Information

• Wasting rates (WFH)
• Oedema rates
• CMR
• U5MR
• Essential
• Population size
• Demography

• Stunting
• Micronutrient deficiency
• Breast feeding
• Food security/economy
• Infection/vaccination
• Contextual data

3
What can wasting/oedema rates tell us?

• Wasting - problem with
• Diet quality - Type 2 nutrients (protein, K, Na,
  Mg, Zn, P, S)
• Total food availability (Low quantity nearly
  always means lack of diversity and low diet
  quality as well)
• disease (relatively small effect)
• Oedema - problem with
• Diet quality - Type 1 nutrients (antioxidant)

4
Anthropometry and mortality

• MUST BE AS SIMPLE AS POSSIBLE whilst giving
  reliable information
• Collect essential, but not excess data
• Theoretical sound methods -
• existing guidelines to be updated
• Practical problems - need to systematize how
  constraints are addressed
• Security, arrivals from inaccessible areas
• Mobility of population
• Topography
• Transparency - need standard reporting format.
  Survey and proposal to be separate

5
Survey data

• Cannot be interpreted in isolation
• There has to be contextual data. It is usually
  such data that prompted the survey in the first
  place.
• Cannot be used to decide how and which programs
  should be implemented
• Must always be accompanied by data on population
  size and structure

6

• When to do a survey?
• Criteria should be clearly defined.
• How to do a survey?
• Methods should be simple and standard
• Design
• Training
• Sampling
• Measurement
• Quality assurance
• Analysis
• Reporting
• How to interpret the results?
• Needs to be put into context this varies
• What a survey CAN tell us severity NOW,
  calibration of surveillance data.
• What it cannot tell us reasons, incidence or
  trends
• How to tell if the survey is reliable?

7

• Few organisations have all the expertise needed
  for anthropometric, mortality and other data
  collection.
• Field experience, Epidemiological capacity,
  Demographic expertise, anthropological knowledge,
  interpretative skills, programatic expertise
• Survey Design Training Sampling
  Measurement and data entry Quality assurance
  Analysis Reporting Interpretation. Design of
  intervention Impact assessment
• Multi-disiplinary team needed to assist and
  overview?

8
When to do a survey?

• Baseline data
• Problem with
• Food security indicators
• Economic, weather, harvest predictions
• Political turmoil
• Health centre/ hospital data
• Seasonality
• ??? Donor funding cycle ???

9
Total admissions with wasting or oedema to 23
TFCs in Burundi
10
When to do a survey?
11
Common responses emergency
12
Desirable responses to emergency
13
The effect on local capacity?
14
The desirable intervention
15
The relation between wasting and CMR is not
close. Oedematous malnutrition Micronutrient
deficiency Infectious disease Others (e.g.
exposure, toxic foods, trauma, smoke pollution)
16
Wasting and pellagra are not related
17
Other Nutritional causes of a rise in mortality
rate

• CMR high with relatively low wasting rate.
• Type 1 nutrient deficiency is not associated with
  wasting.
• Failure of wasting rate to predict increase in
  CMR. This was the main sign of mortality
  associated with a pellagra outbreak.

18
Diet quality diseases like kwashiorkor, Pellagra
and scurvy are not associated with anthropometric
change if there is Type 1 nutrient deficiency
(data from Kuito, Angola, 2001)
19
Patients with pellagra in an emergency situation
are normal or fat!
20
Mortality data

• Much more difficult than anthropometry and prone
  to error. Different methods often give different
  results - e.g. Denan, 20 fold difference
• Retrospective survey 8.9/10,000/d
• mortality surveillance lt0.5/10,000/d
• Discrepant data often not reported
• High mortality - is it an error or is it real?
• If real do something now
• if an error, suppress data as the agency
  reputation at stake

21
Mortality data

• Bias with mortality estimates in development are
  exaggerated in emergency
• More incentive to hide the truth - with
  hostilities or prospect of relief
• Absent/arriving individuals families/ split
  families
• Hiding, kidnap, migration, death
• Migration patterns - with split families migrants
  often take the healthy
• Date and age problems - traumatised population
• seasonality exaggerated
• Translation problems - mixed ethnic groups

22
Triangulation of mortality data

• Retrospective survey
• Surveillance data
• Grave counting
• Religious authority records
• Demographic profile change
• Mother child ratio
• Is it important to know who dies?
• Is it important to know cause of death?

23
Demographic profile

• Population age-sex pyramid should be constructed
  and analysed wherever possible
• demographic expertise essential
• Methods available for cleaning digit preference
  in age etc.
• Population size is very important, very difficult
  to ascertain and prone to both error and to
  political adjustment.
• Geographical/demographic information not only of
  interest for relief, but also for those engaged
  in hostilities.

24
Triangulation of data
Anthropometry oedema prevalence
Program coverage
Mortality data
Surveillance Incidence/coverage
Population size and structure
Other data needed to interpret discrepancies Other
data needed to understand causes
25
What is needed?

• Historic Survey Data /rapid assessment, and
  current data using standard, reliable, repeatable
  methodology in which we all have confidence.
  Transparent disclosure of constraints.
• Integration with surveillance data reporting
  system
• Repeat Surveys to recalibrate surveillance
  data, look at trends and define normal
  status/variation.
• Contextual data to interpret trends or changes
• Evaluation of Impact
• Capacity Building
• Data collection/Analysis
• Integration/Coordination
• Local Capacity for programs and Ethical
  Integrated Closing Strategies for programs
• Donor Understanding and Support smile