Abstract

P135

DevelopmentandvalidationofGeneralNutritionKnowledgeQuestionnaireforAdultsin
Uganda

Session: Poster Presentation

Richard Bukenya1*, John Muyonga3, Sarah Ngalombi4, and Juan E. Andrade1,2

of Nutritional Sciences & 2 Department of Food Science and Human Nutrition UIUC, 3 Department of Food Technology and
Human Nutrition, Makerere University, 4Ministry of Health, Kampala

Although the school system is considered a great scenario to

promote nutrition in Uganda, these educational interventions
often lack proper evaluation. This might be due to the lack of
validated instruments, for instance to rigorously assess a
baseline knowledge among key stakeholders like head
teachers and health workers. At schools, most times head
teachers are responsible for both designing and implementing
nutrition education programs without much input from
government officials. For this reason, the aim of this study was
to develop a valid tool that will be used to evaluate nutrition
knowledge of head teachers and other key nutrition and health
stakeholders in Uganda.
Once validated the tool will allow public health scientists and
practitioners to adequately evaluate nutrition knowledge. This
will improve the quality of nutrition delivery by our laboratory,
and our partners, the Ministry of Health (improved policy
evaluation) and USAIDs Feed the Future Program.

Content validation (Cont.)

Validation process flow diagram

Introduction
Malnutrition undermines national education efforts in Uganda.
Recent surveys have shown that twenty percent of children are
underweight (i.e. low weight for age) while 80% and 38% of
children aged 6-12 y are either iron deficient or have
anemia.[1,2] There is a significant prevalence of overweight
and obesity among school age children reaching over 10%. [3,
4, 5] Over seven million (>80%) children aged 6-12 y, (i.e. onefifth of Ugandan population) receive primary education. This
presents an opportunity to address poor nutrition and reduce
the burden in this age group and the community.

The agreement proportions on relevance, clarity,

Preparation
ofscopeand
Structure

Literaturereview
Selectionofquestionnaireitems(questionsandanswers)
Reviewoftopics
Definedconstructs(Firstdraft)

Reviewof
items

Consultationwiththeexpertpanel(firstround)
Seconddraft
FocusGroupdiscussionswithteachersandhealthworkers
Consultationwithexpertpanel(secondround)
Thirddraft(Toolwithfaceandcontentvalidity)

Pilotstudy

Testretest
reliability
study

Survey(oneweekinterval)withheadteachersinKampala,
andstudents(nutritionandengineeringstudents)
Analysisofitemsfordifficulty,discrimination,internal
consistencyandtestretestreliability
Reviewofthequestionnaireandfurtherconsultationwith
experts
Fourthdraft(toolwithfaceandcontent.andconstruct
validity,internalandtestretestreliability)
TwoSurveyswithheadteachersinMukonoandWakisoina
twoweekinterval
Testretestreliabilityanalysis
FinalquestionnairewithContent,faceandconstructvalidity,
internalconsistencyandtestretestreliability

Content validation
Hypothesis
The modified questionnaire developed by Parmenter and
Wardle (1999) is valid and reliable to assess general nutrition
knowledge of adults in Uganda.

Review of literature
Nine papers with validated questionnaires were reviewed.
One report was included.
The questions from the nutrition knowledge questionnaire
(NKQ) developed by Parmenter and Wardle (1999) form the
core of the Uganda NKQ draft. Other questions were obtained
from the questionnaire used by Whati et al. 2005 and Harvey
et al. 2008.
The GNKQ draft had six nutrition associated topics and one
on demographics:
1. Dietary
recommendations

4. Nutrition and disease

relationship

2. Food groups

5. Food fortification

3. Food choices

6. Sources of nutrition information.

The first draft had 58 questions, which were modified to 59.

Objective
To examine the relevance of the content in a general nutrition knowledge
questionnaire (GNKQ) targeting adult head teachers and health workers
in Uganda as the first validation.
Design
Five experts in the fields of health education (1), nutrition (2), agriculture
(1) and education (1) reviewed the survey online using Qualtrics. The
research protocol was approved by the IRB at UIUC. Reviews were
conducted twice, before and after face validation.
Outcome, Measures and Analysis
The relevance, simplicity, ambiguity and clarity of questions were
evaluated using a Likert scale (1-4). Experts were asked to recommend
deletion/addition of items, change of language, foods and content, and its
alignment with current nutrition policies in Uganda.
Results
Table 1. Results of content

simplicity, and
ambiguity of questions ranged from 0.6 to 1.0 in first round. In
second round, the range was 0.8 to 1.0.
Proportion of agreement on recommendations to deletion of
questions ranged from 0.0 to 0.4 in first round. In the second round
the range was 0.0 to 0.2. From experts comments, two questions
were added in the fortification construct.

First round
RAP

Gwets
AC1

Second round
P-value

RAP

Gwets
AC1

P-value

Whole survey

0.89

0.71

<0.05

0.97

0.96

<0.05

Dietary Recommendations

0.85

0.60

<0.05

0.90

0.89

<0.05

Food groups

0.88

0.81

<0.05

0.93

0.92

<0.05

Food choices

0.84

0.62

<0.05

1.00

1.00

<0.05

Nutrition and disease association

0.96

0.91

<0.05

1.00

1.00

<0.05

Food fortification

0.73

0.23

>0.05

0.92

0.91

<0.05

Sources of nutrition information

0.90

0.70

>0.05

1.00

1.00

<0.05

Design
The GNKQ draft (after second expert review) was administered to 40
students from the fields of nutrition and engineering at Makerere University
via Qualtrics online surveying platform.
120

Outcome, Measures and Analysis

Clarity of the questions in the survey to the head teachers and health
workers was sought. Each question was discussed in the group until all
participants agreed on clarity.
Results
24 questions were identified as unclear during the survey
After FDGs, only 5 questions were changed, while the remaining
ones were discussed based on their potential answers.
A

B
Liquid cooking oil (e.g. mukwano)
contains less fat than solid oil (e.g.
kimbo).
A cup of liquid vegetable oil (e.g.
mukwano) has less fat content than a
cup of solid oil (e.g. kimbo).

Figure 2. A) One of the focus group discussions with head teachers at Makerere
University. B) One of the questions that was changed after focus group discussions.

53.4

60

41.2

40
20

103.9
72.1

11.4 12.8
Dietary
Recommendations

5.4 6.8
Food
Groups

Food
Choices

9.5

15.7

15.1

Nutrition&
Disease

4.6
Food
Fortification

Whole
Survey

Conclusions and Future Studies

Face validation

Design
The draft GNKQ was administered to 15 head teachers and seven
health workers. They rated each question on level of clarity (Yes or No).
They provided reasons for the questions being unclear. Three FGDs
were conducted to explore participants understanding of the unclear
questions in addition to those answered wrongly. Questions answered
wrongly were included in the FGDs. Each question was discussed to the
extent that each participant agreed on its clear interpretation and
exploring alternatives. FGDs were recorded using a digital sound
recorder. All participant signed consent forms. The FGDs lasted 1.5 to 2
hours. The research protocol was approved by the IRB at UIUC and by
the Uganda National Council for Science and Technology.

Nutrition

80

Figure 3. GNKQ mean scores from two student populations.

Figure 1. Screen shots of Expert evaluation of GNKQ using Qualtrics.

Objective
Assess the level of understanding and interpretation of GNKQs
questions by target population using focus group discussions (FGDs).

Engineering

100

validation with experts.

Topic

Pilot survey: Preliminary results

Objective
Continue validation by evaluating knowledge differences (construct validity)
between subjects from two academic disciplines at Makerere University.

Mean Scores

1 Division

Nutrition Knowledge as a determinant of nutrition behavior is poorly evaluated due to lack of validated tools. We determined the relevance of content in
the draft general nutrition knowledge questionnaire(GNKQ) for adults in Uganda as the first step in the validation. The GNKQ comprised of 59 questions
and structured into demographic information and six constructs assessing knowledge on recommendations, food groups, food choices, nutrition and
disease relationship, food fortification, and sources of nutrition information. Five experts in the fields of health education (1), nutrition (2), agriculture (1)
and education (1) reviewed the survey online using Qualtrics. Authors evaluated questions on the relevance, simplicity, ambiguity and clarity to the target
population using a Likert scale (1-4). Experts recommended deletion/addition of items, change of language, food and content to rhyme with current
nutrition policies.
The raw agreement proportion (RAP) for the whole survey on relevance was 0.89 (Gwets AC1=0.75, P<0.05). The RAP on relevance of questions was
0.85 (Gwets AC1= 0.6, P<0.05) on recommendations; 0.88 (Gwets AC1=0.81, P<0.05) on food groups; 0.84 (0.62, P<0.05) on food choices; 0.96
(Gwets AC1=0.91, P<0.05) on nutrition and disease relationship; 0.73 (Gwets AC1=0.23, P>0.05) on food fortification; and 0.9 (Gwets AC1=0.7,
P>0.05) on sources of nutrition information. Experts recommended addition of more questions to the fortification construct. All questions had good to
excellent reliability agreement and were relevant to evaluate nutritional knowledge. Reviews are necessary to improve its clarity. Once validated, this
instrument will be used to gather nutrition knowledge among adults in Uganda.

A systematic approach has been used to begin validation of a GNKQ for

Ugandan adults, which included: questionnaire adaptation, two expert reviews,
focus groups and construct validity.
Experts agreed that the questions and the tool are very important (Gwet's AC1
ranged 0.4 to 1.0 interrater reliability) [6].
Agreement proportions were beyond acceptable levels of 0.78, except for results
on topic of food fortification during the first experts review[7] .
Gwets AC1 score for food fortification changed after face validation during
second expert review.
After face validation, the GNKQ remained with a total of 59 questions.
Draft GNKQ was able to discriminate (p<0.05) nutrition knowledge between
student populations.
The GNKQ draft will be piloted on 40 head teachers in Kampala to determine its
internal and test-retest reliability.
Systematic approach used during development of this tool can be applied to
other determinants of nutrition behavior in Uganda.

References
1.
2.
3.
4.
5.
6.
7.

Acham, H., et al. (2012). African Journal of Food Agriculture, Nutrition, and Develop. 12(2): pages 58625880.
Turyashemererwa et al (2013) Journal of Human Nutrition Dietetics. 26 (Suppl. 1), 7381
Peltzer K and Pengpid S (2011). International Journal of Environmental Research and Public Health. 8,
3859-3870
Muthuri et al. (2014). PLoS ONE 9(3): e92846.
Wang and Lobstein (2006). International Journal of Pediatric Obesity, 1: 1125.
Wongpakaran et al. BMC Medical Research Methodology 2013, 13:61
Polit and Beck (2006). Research in Nursing & Health, 2006, 29, 489497

Acknowledgements
The authors thank BHEARD/USAID for the financial support. The authors are grateful to the UIUC Graduate
Colleges Focal Point - UCOUNT project for the travel grant provided to the graduate student. We are
grateful to the Margin of Excellence funds from the Division of Nutritional Sciences at UIUC that has made
follow up studies possible. The contribution of the Department of Food Technology and Human Nutrition,
Makerere University for providing the venue to conduct focus groups is appreciated. Special
acknowledgement to Ministry of Health Kampala, Kampala Capital City Authority Administration, students at
Makerere University, Head Teachers and Health workers for their support and participation.