Cover Image
close this bookNutrition Guidelines (MSF, 1995, 191 p.)
close this folderPart III Selective feeding programmes
View the document1. Justification for selective programmes
View the document2. Criteria for admission and discharge to selective feeding programmes
View the document3. Screening and selection
View the document4.Treatment in a therapeutic feeding centre
View the document5. Treatment in supplementary feeding programmes
View the document6. Implementation and management of a feeding centre
View the document7. Registration and monitoring
View the document8. Evaluation of feeding programmes
View the document9. Food management

8. Evaluation of feeding programmes

Feeding centres can be evaluated during their operation to see how efficiently they function and to determine which aspects of the programme require the most attention at any time (PROCESS evaluation).

Monitoring the Functioning of a Feeding Centre


Monitoring the functioning of the feeding centre is based on the monthly collection of clear data on the activities in the feeding centre, including admission and discharge criteria, kcal/person/day distributed, number of meals per day and the attendance pattern of the children.

The attendance book is the basis for the following attendance report:


Total Last Month (A):

copy Row I of last month

Total Exits (H):

Row D + E + F + G

Total End Month (V:

Row A + B + C - H


Changes in the functioning of the feeding centre can be quickly detected through monthly calculation of several indicators. Most of these indicators can be obtained directly from the attendance report.

In order to facilitate comparison between months and between feeding centres, indicators should be calculated for children under 5 and above 5 years of age separately (and if other vulnerable groups are includedincluded in the SFP, these should also be graphed separately i.e. pregnant and lactating women).

Function indicators should be expressed as proportions of the total number of children leaving the programme during the reporting month for any reason.

The proportions of the number of exits that each indicator reflects are not real population <<rates>>. For example, the proportion of deaths is not a mortality rate and does not reflect the risk of dying in the population, it indicates the severity of malnutrition in the programme participants and the quality of programme functioning.

- Proportion of Recovered (%): No. of children successfully discharged / No. of exits Row (D) / (H) x 100%
- Proportion of Deaths (%): No. of deaths in the centre / No. of exits Row (E) / (H) x 100%
- Proportion of Defaulters (%): No. of defaulters / No. of exits Row (F) / (H) x 100%
- Proportion of Transferred (%): No. of transferred / No. of exits Row (G) / (H) x 100%

Absolute Figures from the attendance report can be graphed over time, to help interpretation of trends in programme functioning (see the example in Annex 20).


Analysis of the different function indicators should be done in the field, because proper interpretation can only be made within the specific context. Furthermore, the information is vital for programme management, allowing appropriate and prompt adaptations, if necessary.

The main advantage of calculating proportions as a function of the number of exits is that these indicators are independent of the number of new admissions, which may greatly vary over time for different reasons (for example as a result of population movements).

The function indicators, as given above, are highly inter-related and should therefore always be interpreted in relation to each other.

For example: an increase in the number of defaulters during the month will result in a decrease in the proportion of recovered and deaths. This does not mean that the functioning of the programme has improved; on the contrary, the increase in defaulting is worrying, and demands attention.

Another example: a decrease in the number of defaulters will result in a increase in the proportion of recovered and deaths. This should not be interpreted as bad functioning of the centre, but as improved functioning; the proportion of successful recoveries will also have increased. In general it can be said that the proportion of recovered is the most important indicator, as it reflects the ultimate objective of the feeding programme.

The table below provides provisional targets and reference figures are given to provide benchmarks against which to interpret the functioning of individual programmes. The target figures should not be considered as rigid objectives for every programme as the feasibility of reaching targets depends on the local circumstances and the stage in programme development. However, target figures may give an indication of what might be considered <<good>> and <<bad>> functioning under <<average>> conditions.

The proportion of those transferred is usually left out of the analysis as it is usually a minor proportion and only confuses interpretation.



· Attendance rate

WET - Average daily attendance over the month/Average No. of children registered over the month = Average daily attendance / Row {[(A)+(I)]/2}

DRY - Average weekly attendance over the month/Average No. of children registered over the month = Average weekly attendance / Row {[(A)+(I)]/2}

Objective: > 80%

It is sufficient to only calculate the average attendance rate for a few randomly chosen days during the month. Make sure that a record of attendance in the attendance book corresponds with actual physical presence of the child during meals - to be checked by physical counting of recipients during the meals.

Example. Average attendance rate:

· There are 90 children registered in a TFP:

Attendance on 6 randomly chosen days during the previous month: day 3: 76; day 9: 83; day 11: 79; day 18: 69; day 25: 74; day 28: 81.

- Average attendance = 77
- Attendance rate = 77/90 x 100 = 86%

· There are 1,600 children registered for dry feeding. Attendance over the last month: week 1: 1,250; week 2: 1,116; week 3: 1,120; week 4: 1,050. - Average monthly attendance = 1,134 - Attendance rate = 1,134/1,600 x 100 = 71% The average child will come 3 out of every 4 distribution days.

· Coverage:

Total No of children < 5 years registered at the end of the month/Total No of malnourished children < 5 years estimated by the last survey. Objective: > 50% in rural populations, > 75% in urban / camp populations

Example of coverage:

· Total population: 20,000 people
· The proportion of children under 5 years is estimated at 20% = 4,000 children
· Nutritional survey: 4% of children under 5 years are severely
· malnourished (c 2 Z-Scores W/H and/or oedema) = 160
· 17% of children 6-59 months are moderately malnourished

(W/H- 3 - <-2 Z-Scores):

680 children

· Registered in the TFP:


· Coverage:

90/160 x 100% =56%

· Registered in the SFP:


· Coverage:

430 /680 x 100 = 63%

Note: The coverage calculated this way will become less accurate over time since the survey was performed, as the prevalence of malnutrition will change. However, there is no alternative until another survey is conducted.

· Mean length of stay on discharge

To be calculated monthly or every 3 months, from all recovered children or a random sample of 30 recovered children: total number of days of admission of all (30) recovered children added/No of recovered children (30). Objective: < 30 days for TFP and < 60 days for SFP.

· Average weight gain: (every 3 months - OPTIONAL):

The average weight gain of the children in a TFP is a very good indicator of the quality of the programme (it is rarely used for SFPs). It can be calculated every three months on a sample of 30 children, chosen randomly from the record files.

The weight gain is calculated over the entire period for marasmic children after entry into the Phase II. For kwashiorkor cases, the indicator is calculated in Phase II after complete loss of oedema.

Objective: 10 - 20 g/kg/day
Daily Weight Gain = W2 - W1 (g/kg/day)/(Wx(T2 - T1))

W1 =

Weight on entry of Second Phase in grams

W2 =

Weight at day 15 or on exit of Second Phase (in grams)

W =

Weight on entry of Second Phase in kg

(T2 - T1) =

Number of days between W1 and W2

Average Daily Weight Gain = Total of daily weight gains of children (g/kg/day) / Total number of children

· Case Fatality

No. of deaths in the centre/Average No. of children registered during the month.

Objective: < 10%

This rate expresses the risk of death for a child attending the programme.

· Percentage of children vaccinated against measles

Objective: 100%

A brief overview of indicators often used for monthly reporting on SFPs and TFPs is presented below:


(*): No of children expected derived from anthropometric surveys


· Consumption:

Do the consumption of food, firewood and water correspond with the attendance figures.

· Food quality:

The quality of the food commodities should be checked when received and regularly during storage (expiry date, damaged packing, moisture, mould, vermin). This is especially important for milk - powder. For quality control procedures of dried skim mink see Annex 14.

Hygiene during preparation and distribution of food should be carefully controlled.

· Ration:

Recipe of the HEM/porridge/dry ration:

- Are the quantities of each ingredient correct?
- Is the volume of the rations per child correct?
- Are taste and palatability and consistency of the HEM/porridge correct (the HEM/porridge should be tasted regularly)?

· Staff:

The level of job performance of the staff should match their job descriptions, motivation, loyalty.

Annex 21 is a check list of tasks for the supervision of TFPSs. It should be clear to the supervisor which points from this list should be accomplished in the programme.

Monitoring the Effectiveness (Impact) of the Programme (every 3 - 6 months)

The influence of a particular nutritional programme on the health status of a population is not easy to ascertain as other vital sectors (water, shelter, health care, etc.) contribute to well being. Secondly, one needs fairly accurate knowledge of the <<denominator>> or real population size for an impact indicator based on routinely collected data to be correct.

Mortality figures, malnutrition rates and basic morbidity data (routine and epidemic events) are valuable indicators that need to be monitored in emergency situations in order to appreciate needs and adapt programmes. Information can be collected through cross - sectional surveys, but trends in the data can also be used to help follow the impact of the programme.

By regularly monitoring mortality and prevalence of severe malnutrition, the initial objectives/design of the programme can be evaluated and adapted when necessary.

· Mortality rate among children under five:

Of children in the community: through monthly/weekly (under - five) mortality surveillance (see Annex 1) or a mortality survey.

· Prevalence of severe malnutrition among children under 5 years in the population:

To be measured by regular nutrition surveys and comparison with results from previous surveys.

Trends in health and nutrition indicators can be related to various and multiple factors. Thus, one might be able to demonstrate a positive trend in improved rates of malnutrition and mortality over the programme period, but it is impossible to prove that the selective feeding programme caused this trend.

However, with well collected information from multiple sources, one can show a very strong association and argue for a reasonable interpretation of the impact of the feeding programme.

When to Close Down Feeding Programmes

When the number of patients is too low to make it efficient to run separate feeding centres, feeding centres should be closed down (< 20 for a TFP and < 30 for an SFP). New cases should then be referred to the nearby hospital or health centre (supply these services with the necessary food, equipment, and technical support, if not available).

Low numbers of participants may not reflect an improved situation but poor accessibility or acceptability of the programme. Therefore, the decision to close down selective feeding programmes should only be made after a nutrition survey has clearly shown a significant decrease of global acute malnutrition in the population.

Prevalence of global acute malnutrition (<-2 Z-Scores W/H and/or oedema) among children under - five years old should be below 10%.

The following conditions should also be met:

- General food distributions are reliable and adequate,
- Effective public health and disease control measures are in place,
- No seasonal deterioration of nutritional status is anticipated,
- Mortality Rate is low (see Annex 1),
- The population is stable - no major influx is expected. (See Ref. 12)

Follow-up nutrition surveys must be planned to detect any deterioration of the situation, even after closure of selective feeding programmes. This is especially important if the overall situation remains unstable.