Famine in the Horn of Africa Over the Last Decade

1
Famine in the Horn of AfricaOver the Last Decade

• Paul B Spiegel MD, MPH1
• Peter Salama MBBS, MPH2
• Susan Maloney MD, MPH1
• Albertien van der Veen, MSc3
• 1 CDC, 2 UNICEF, 3 WHO

2
(No Transcript)
3
Somalia 1991-92 Background

• Drought, civil unrest, collapse of government,
  destroyed infrastructure led to many deaths among
  Somalis during 1992
• Difficult to provide aid due to insecurity
• In absence of health care infrastructure, health
  status of population described through
  epidemiological studies which influenced how aid
  was provided

4
Somalia 1991-92 Results

• 23 surveys evaluated which had defined pop. and
  apparent systematic methodology
• Extensive methodological differences
• 12 (52) not reproducible
• Units of measurement and denominator estimates
  inconsistent
• 8 of 16 (50) surveys examining mortality
  assessed cause of death

Boss, L. P., Toole, M. J., and Yip, R.
Assessments of mortality, morbidity, and
nutritional status in Somalia during the
1991-1992 famine. Recommendations for
standardization of methods. JAMA. 1994 272(5)
371-6.
5
Somalia 1991-92 Results cont

• 0 of 11 surveys examining morbidity provided
  case-definitions
• Of 16 surveys examining nutritional status,
  variety of definitions of malnutr. and
  measurement methods used
• 3 (19) - MUAC only
• 10 (63) - Referenced wt/ht data using
  
  of reference of median
• 4 (25) - Presented data using z-scores

6
Somalia 1991-92 Recommendations

• Define clear study objectives
• Use standard sampling and data collection methods
• Ensure precise written documentation of
  objectives, methods, and results

7
Ethiopia 1999-2000 Background

• Ethiopia subject to recurrent drought and food
  insecurity exacerbated by civil strife
• Crises generally involve massive excess mortality
  and population displacement without formation
  of discrete camps
• Prior to crisis of 1999-2000, eastern and
  southern regions of Ethiopia subject to 3
  successive years of drought

8
Ethiopia 1999-2000 Background cont

• Beginning in 1999, data from early warning
  systems in many regions of Ethiopia indicated
  rapidly deteriorating food security and nutrition
  situation
• More than 10 million people estimated to need
  food assistance during peak of crisis in early
  2000

9
Ethiopia 1999-2000 Mission

• Members of IERHB/CDC seconded to UNICEF as
  Nutrition Technical Advisors June-Sept. 2000 at
  request of OFDA
• Terms of Reference
• Co-ordinate and identify gaps
• Provide technical assistance
• WHO sent Nutritionist with whom we collaborated
  closely

10

Regions of Ethiopia
Region
11
Zone
UNDP
12
WFP Cereal Allocation in April 2002 vs.
Requirements by Woreda
13
lt5 Nutrition Surveys in Ethiopia, Jan.-Jun. 2000



Global Acute Malnutrition (GAM)
0-9
10-19
20-29
30-39
40

N38
14
Ethiopia 1999-2000 Evaluation

• Evaluated 125 nutrition surveys from Nov.
  1998-Jun. 2000 by 14 different NGOs
• RHA 16 (13)
• Cluster 109 (87)
• 42 surveys part of early warning system with
  different aim and cluster methodology
• 67 surveys were intended to provide useful data
  to direct programs
• Only 5 surveys (4) recorded measles vaccination
  coverage

RHA rapid health assessments using convenience
samples
15
Evaluation of Intentional Surveys

• Valid (i.e. representative)
• gt30 clusters
• PPS
• and
• Precise
• gt10 children per cluster

1. Binkin N, Sullivan K, Staehling N, Nieburg P.
Rapid Nutrition Surveys How many clusters are
enough? Disasters 1995 1697-103.
2. Sullivan KM. The
effect of sample size on validity and precision
in probability proportionate to size cluster
surveys, Society of Epidemiologic Research, Utah,
1995
16
Categorization of Cluster Surveys (N25)
No. of Clusters
No. (No. with PPS) lt30 gt30
lt10 0 0
10-29 1 (0) 2 (2 of 2)
gt30 16 (2 of 16) 6 (5 of 6)
No. of Children /Cluster
17
Valid and Precise Cluster Surveys7 of 67 (10.5)
No. of Clusters
No. Acceptable lt30 gt30
lt10
10-29 2
gt30 5
No. of Children /Cluster
18
lt5 Nutrition Surveys in Ethiopia, Jan.-Jun. 2000



Global Acute Malnutrition (GAM)
0-9
10-19
20-29
30-39
40

N38
19
lt5 Nutrition Surveys in Ethiopia Jan-Jun 2000
Global Acute Malnutrition (GAM)
0-9
10-19
20-29
30-39
40
Valid and Precise N2 (5.3)
20
USAID Funding for Ethiopia 1999/2000

• 1999 2000
• Development Assistance
• Food 32,264,600 36,200,741
• Non-food 38,214,000 42,677,000
• Total 70,478,600 78,877,741
• Humanitarian Assistance
• Food 288,968,022 82,956,723
• Non-food 11,740,022 1,350,700
• Total 300,708,022 163,185,164
• Grand Total 371,186,622 163,185,164
• Source USAID Ethiopia office 9-01

21
Problems with Surveys

• Methodology
• Sample size inadequate (clusters and/or children)
• Non-probabilistic sampling
• Not proportional to population size
• Targeting of drought-affected areas within woreda
• Choosing children most malnourished
• Always include measles vaccination coverage

22
Problems with Surveys cont

• B. Survey Analysis
• Nutrition indicators (ER setting)
• Wt/Ht z-scores for children lt5yrs
• 95 CIs using C-sample in EpiInfo
• Ensure include edema as cases of severe malnutr.
• Measles Immunization coverage
• Report recorded on imm. card vs. verbalcard

23
Problems with Surveys cont

• C. Survey Reporting
• State objectives and detailed methodology
• State results clearly with units, including
• Sample size and number of clusters
• Number of household refusals or absences
• 95 confidence intervals (/- DEFF)
• Interpret results and compare to baseline data or
  data from previous surveys if available
• Make recommendations (underlying causes)

24
Recommendations

• Options for NGOs
• ?Make sampling methodology simpler
• Improve training manuals
• Ensure field staff appropriately trained
• Send HQ staff to do surveys
• Co-ordinate with 1-2 survey NGOs in field
• Provide survey form templates with programs for
  data analysis

25
Recommendations cont

• Options for policymakers/donors
• Be wary of data/reports provided by NGOs
• Have technical persons available (in
  country vs. contactable elsewhere) to
  interpret surveys
• Field personnel appropriately trained who make
  decisions based on survey results