GAP Growth in Adolescence :Potential interventions to improve growth,health & reduce future non-communicable disease
Malawi, 2016
Get MicrodataIdentification
MWI-MEIRU-GAP-2016-v1.0
GAP Growth in Adolescence :Potential interventions to improve growth,health & reduce future non-communicable disease
| Name | Country code |
|---|---|
| Malawi | MWI |
This is a formative, mixed-methods, pilot study for a future RCT.
The problem studied
Little is known about effectiveness of interventions for growth and NCD prevention during adolescence. In a randomized controlled trial (RCT) we filled this evidence gap by testing a dietary/physical activity intervention targeting stunted adolescents.
This is a formative research project in order to shape and inform maximize the success of a future RCT.
Objectives
Broad objective: To inform and shape the design of a future interventional RCT for improving adolescent growth and health outcomes in Malawi.
Specific objectives:
- Conduct formative qualitative research to understand the context and community needs and help develop the intervention for testing in a future RCT
- Establish:
a. baseline characteristics, including prevalence of stunting and risk factors for stunting (to allow us to calculate sample size for a future RCT).
b. target group(s) for whom intervention(s) may be most effective. - Pilot three tools which we anticipate using to measure key outcomes in the main RCT
a. CANTAB computer test for measuring cognitive function.
b. Actilife accelerometers for measuring physical activity.
c. Bioelectrical impedance analysis (BIA) for measuring body composition.
Methodology
Working with Malawi Epidemiology and Intervention Research Unit (MEIRU), in a Karonga rural site and a Lilongwe Urban Site, we conducted:
- Qualitative research:
Key informant interviews with future study stakeholders (see below).
Mainly individual interviews and some small focus groups. Approx. 20 individuals per study group were expected to achieve data saturation for interviews, and a maximum of 10 focus groups with a minimum of 4 participants.
-A cross sectional survey:
To determine prevalence of stunting in adolescents. This included primary data collection in Lilongwe, recruiting approx. 600 adolescents; secondary analysis of existing Karonga data)
-A detailed quantitative survey nested within the prevalence survey
More in-depth assessment of approximately 200 adolescents at each site, including: a detailed questionnaire asking about possible risk factors for stunting/adverse long term outcomes; detailed anthropometry; assessment of physical activity; cognitive function; body composition.Study questionnaires are based around validated tools questions such as those used in the Global School Heath Survey (GSHS)
Our main target population were adolescents aged 10-19 years. In the qualitative sub-study we also interviewed: carers/parents; teachers; community healthcare workers; other community leaders. Fieldwork took place over 3-4 months from approximately August to November 2016.
Individual
Version
2020-02-26
Scope
Themes covered in questionnaire:
· Health history
· Anthropometry
· Hand grip strength
· Blood pressure
· BIA
· CANTAB cognitive testing
· Behaviour questions - including drinking, drugs, hygiene practices, sexual behaviour
· 24 hour food recall
· Self-reported Tanner pubertal stage
· Dental health using WHO Oral health questionnaire
· SES using DHS Asset questions
· Family history of NCDs
· Empowerment questions
· HIV status
· Physical activity levels (Steps per day) measured using accelerometers
| Topic | Vocabulary |
|---|---|
| Malnutrition | MeSH |
| Growth Disorders | MeSH |
| Adolescent | MeSH |
| Cognition | MeSH |
| Cardiovascular Diseases | MeSH |
| Body Composition | MeSH |
| Physical fitness | MeSH |
| Malawi | MeSH |
Coverage
We recruited adolescents aged 10 to 19 years within each study site: rural Karonga and urban Lilongwe (Area 25). Stunting were defined using standard WHO criteria: height-for-age <-2 z-scores (WHO reference population). Adolescents were randomly selected using a cluster design. In Lilongwe, they should be representative of the general adolescent population. In Karonga, adolescents were purposively sampled to include around 50% stunted adolescents.
Producers and sponsors
| Name | Affiliation |
|---|---|
| Dr Marko Kerac | London School of Hygiene and Tropical Medicine |
| Name | Affiliation | Role |
|---|---|---|
| Prof Mia Crampin | London School of Hygiene and Tropical Medicine | Co-applicant |
| Dr Natasha Lelijveld | London School of Hygiene and Tropical Medicine | Co-applicant |
| Professor Moffat Nyirenda | London School of Hygiene and Tropical Medicine | Co-applicant |
| Dr Steffen Geis | London School of Hygiene and Tropical Medicine | Co-applicant |
| Name | Abbreviation |
|---|---|
| Wellcome Trust, UK | WT |
Sampling
For the cross-sectional prevalence survey:
-In Karonga: we were not collecting primary data as this information is already available from the DSS.
-In Lilongwe: we constructed a multi-cluster sampling frame using district maps, based on chiefdoms covering the whole of Area 25. Of the 41 chiefdoms, 30 were randomly selected for sampling. In each of these 30 clusters, we conducted a house-house survey. Depending on how many houses are estimated to be in the cluster, a skip number was used to sample every house in order to sample around 19 adolescents per cluster. Consenting adolescents were enrolled. A household with an adolescent who is not at home at time of first visit were revisited at a convenient time to minimize sampling bias.
For the in-depth survey:
-In Karonga: we designated 30 clusters cover the whole DSS area and randomly selected 8 adolescents to visit using the DSS database. We stratified our sample to get a balance of stunted/non-stunted.
-In Lilongwe: every 2nd/3rd (please see sample size discussion) adolescent surveyed in the prevalence survey were asked for a more in-depth interview. If he/she chooses not to participate in this, the next one surveyed was asked.
Survey instrument
In Lilongwe, 280 adolescents had a “basic” survey questionnaire to assess prevalence of stunting; this includes basic anthropometry, blood pressure, hand grip strength and a few health history questions. A further 250 adolescents in Lilongwe had a more detailed survey. All 240 adolescents in Karonga had the detailed survey. The questionnaires included in the detailed survey are described below:
-
Health History
o Including any treatment of SAM, TB, HIV status, recent illnesses, number of admissions to hospital, and family history of NCDs. -
Diet diversity and food security of stunted adolescence in Malawi,
o FAO Food Insecurity Experience Scale (FIES) consisting of a few questions about hunger and access to food
o 24-hour dietary recall interview (used to generate an FAO dietary diversity score) -
Behavior questions taken from Global School Health Survey (GSHS) core questionnaire version 2013 (<http>)
o Including questions about school attendance, alcohol, drugs, sexual behavior, hygiene -
Pubertal stage using Tanner pubertal self-rating method
-
Socio-economic status using an asset score derived from the Malawi DHS asset questions.
-
Dental health and knowledge of oral hygiene assessed using a few questions from the WHO oral health survey
NCD related data
-Level of physical activity
This was measured using Actilife accelerometers worn around the waist, attached by a comfortable wide elastic band for at least 48hrs
o Actilife accelerometers have previously been used in Malawi[19, 20] however, to our knowledge, they have not been used on this age group (adolescents)
-
Cognitive function scores
As well as asking about educational achievement (highest school standard achieved) we had also formally tested cognitive function using the CANTAB cognition system. This involves a series of computer games testing various aspects of cognitive function (e.g. short-term memory; cognitive function) www.cambridgecognition.com/academic/research <http>
o This tool has previous been used in Malawi on children ranging from 6 to 15 years [19, 21]. There is currently no specific “gold standard” tool for measuring cognitive function in Malawi, however CANTAB, although novel, may be the most effective and detailed measure of this outcome available at present. -
Body composition
This was assessed using BIA, mid-upper arm circumference (MUAC), waist-hip ratio and BMI-for-age z-score based on WHO references;
o Although BIA has not been compared to other gold standard measures of body composition for the Malawian population specifically, it has been validated as an acceptable tool for measuring fat mass, fat-free mass and total body water in a variety of other populations[22, 23]. Additionally, BIA has been used in Malawi in previous studies[19]. -
Blood pressure
The majority of data was entered via ODK electronic forms. This was downloaded from ODK as a csv file and opened using STATA.
WHOanthro PLUS software was used to calculate anthropometric scores using the WHO 2007 growth reference.
The only data not entered via ODK was:
1. Accelerometer data which was downloaded straight from the accelerometer devices. The participant must have worn the device for a minimum of 48 hours in order for the data to be valid.
2. 24 hour food recall was written on paper and manually entered into a database.
3. CANTAB cognitive data was stored using ipad tablets and key outputs called by the software were downloaded as csv files. Data from the individual tables were combined into one dataset for analysis using Stata. The data were merged together using study id (stid) to create one dataset with one record per participant.
Data collection
| Start | End | Cycle |
|---|---|---|
| 2016-08-22 | 2016-12-01 | 1 |
- Face-to-face [f2f]
| Name | Abbreviation |
|---|---|
| Malawi Epidemiology and Intervention Research Unit | MEIRU |
The qualitative interviews began in June 2016; the interview was taking around 45 minutes to complete. At the Karonga and Lilongwe sites, a random selection of adolescents, parents, and teachers were interviewed. Interviews were conducted in Chichewa or Chitumbuka, as appropriate. All interviewers were female.
For the quantitative surveys, each interviewing team comprised 2 interviewers (a mixture of male and female). There were 4 interviews at the Karonga site and 4 interviews at the Lilongwe site. Data collection started in Karonga at the end of August 2016; data collection was due to start in Lilongwe mid-September,2016. The interview was taking around 3 hours per participant; this include consent, anthropometry, body composition, BP, cognitive testing, food recall and behaviour question. Interviews were conducted in Chitumbuka at the Karonga site and Chichewa at the Lilongwe site.
Data processing
-
Limited ranges, compulsory questions, and double-measurement for anthropometry were used at the point of data collection using ODK forms on Tablet computers in order to maximize accurate data collection and data entry.
-
Data submitted via the tablets was checked weekly for errors, with the possibility to return to the respondent.
-
Data was edited at the end of data collection by checking the range of responses for each question, and particularly for measurement data, ratios and age-specific ranges were assessed and any unfeasible measured removed. For height and weight, HAZ, WAZ and BMI-for-age-z-scores were calculated and standard WHO cut-offs were applied for removing anomalous data (e.g. any HAZ <-6 z scores)
Data appraisal
Other forms of appraisal -
BIA data was double entered - the difference between the first reading and the second reading can be indicative of BIA data quality.
The number of HAZ WAZ and BAZ measurements excluded based on WHO cut-off criteria could be an indicator of anthropometric data quality.
Data Access
| Name | Affiliation | |
|---|---|---|
| Malawi Epidemiology And Intervention Research Unit. | ||
| Dr Marko Kerac | LSHTM | Marko.kerac@lshtm.ac.uk |
This data is made available for licensed access under the following conditions:
-
Data and other material provided by MEIRU will not be redistributed or sold to other individuals, institutions or organisations without MEIRU's written agreement.
-
In the case of multi-centre datasets, data originating from a single contributing member centre of the collaboration may not be analysed or reported on in isolation without the express permission of the member centre concerned.
-
No attempt will be made to re-identify respondents, and there will be no use of the identity of any person or establishment discovered inadvertently. Any such discovery will be reported immediately to MEIRU.
-
No attempt will be made to produce links between datasets provided by MEIRU or between MEIRU data and other datasets that could identify individuals.
-
Any books, articles, conference papers, theses, dissertations, reports or other publications employing data obtained from MEIRU will cite the source, in line with the citation requirement provided with the dataset.
-
An electronic copy of all publications based on the requested data will be sent to MEIRU.
-
MEIRU, MEIRU research collaborators and the relevant funding agencies bear no responsibility for the data's use or interpretation or inferences based upon it.
Metadata production
DDI-MWI-MEIRU-GAP-2016-v1.0
| Name | Abbreviation | Affiliation | Role |
|---|---|---|---|
| Malawi Epidemiology and Intervention Research Unit | MEIRU | Agency | |
| Estelle McLean | EM | LSHTM and MEIRU | Production of Data dictionaries |
| Dr. Chifundo Kanjala | CK | MEIRU | Documentation planning and supervision |
| Dominic Nzundah | DN | MEIRU | Metadata entry and editing |
2020-02-26
Metadata version
Version 1.0 ( March,2017)