fnb book

Design of a prospective, randomized evaluation of an integrated nutrition program Design of a prospective, randomized evaluation of an
integrated nutrition program in rural Viet Nam*
David R. Marsh, Helena Pachón, Dirk G. Schroeder, Tran Thu Ha, Kirk Dearden, Tran Thi Lang, Nguyen Dhanh Hien, Doan Anh Tuan, Tran Duc Thach, and David Claussenius Abstract
than 24 months of age living in the intervention and comparison communes and randomly selected 240 Few prospective studies of child growth and its deter- children (120 intervention and 120 comparison). We minants take place in programmatic contexts. We gathered information on nutritional status, diet, illness, evaluated the effect of Save the Children’s (SC) com- care, behavioral determinants, empowerment, and pro- munity empowerment and nutrition program (CENP) gram quality, monthly for six months with a re-survey
on child growth, care, morbidity, empowerment, and at 12 months. We collected most information through
behavioral determinants. This paper describes the maternal interview but also observed hygiene and pro-
research methods of this community-based study. gram quality, and videotaped feedings at home. Some
We used a longitudinal, prospective, randomized implementation and research limitations will attenuate
design. We selected 12 impoverished communes with CENP impact and measurement of its effectiveness.
documented child malnutrition, three comparison, and
three intervention communes in each of two districts
in Phu Tho Province, west of Hanoi. SC taught district Key words: Positive deviance, effectiveness evalua-
trainers in November 1999 to train local health volun-
tion, child nutrition, Viet Nam, hearth, diet, child-care, teers to implement the 10-month CENP, including situ- ation analysis, positive deviance (PD) inquiry, growth monitoring and promotion, nutrition education and rehabilitation program (NERP), deworming, and mon- Introduction
itoring. PD inquiries aim to discover successful care practices in poor households that likely promote well- The ViSION (Viet Nam study to improve outcomes nourished children. NERPs are neighborhood-based, in nutrition) project evaluated the implementation facilitated group learning sessions where caregivers and impact of an integrated nutrition program on of malnourished children learn and practice PD and the nutritional status, morbidity, and diet of and care other healthy behaviors. We dewormed all intervention for children 5 to 25 months of age at baseline in rural and comparison children. We weighed all children less Viet Nam. The ViSION project involved a partner- ship among Save the Children/US ([SC] Hanoi and David R. Marsh is affiliated with Save the Children Westport, Conn., USA), the USAID-funded LINK- Federation/US in Westport, Conn., USA. Helena Pachón, AGES Project (Washington, D.C.), Emory University formerly at the Rollins School of Public Health, Emory Uni- Rollins School of Public Health (Atlanta, Ga., USA), versity in Atlanta, Ga., USA, is now affiliated with the Division and the Research and Training Center for Commu- of Nutritional Sciences at Cornell University in Ithaca, New York. Dirk G. Schroeder is affiliated with the Rollins School of nity Development (RTCCD, Hanoi). The SC Viet Public Health at Emory University in Atlanta, Ga., USA. Tran Thu Ha and Tran Duc Thach are affiliated with the Hanoi Research and Training Centre for Community Development * The ViSION (Viet Nam study to improve outcomes in in Hanoi, Viet Nam. Kirk Dearden, formerly at the LINK- nutrition) project evaluated the implementation and impact AGES Project, Academy for Educational Development in of an integrated nutrition program on the nutritional status, Washington, D.C., USA, is now at the Department of Health morbidity, diet of, and care for children 5 to 25 months Science, Brigham Young University in Provo, Utah, USA. Tran old in rural Viet Nam, through a partnership among Save Thi Lang, Nguyen Dhanh Hien, Doan Anh Tuan, and David the Children/US (Hanoi and Westport, Conn., USA), the Claussenius are affiliated Save the Children Federation/US USAID-funded LINKAGES project (Washington, D.C.), Emory University’s Rollins School of Public Health (Atlanta, Mention of the names of firms and commercial products Ga., USA), and the Research and Training Center for Com- does not imply endorsement by the United Nations University.
Food and Nutrition Bulletin, vol. 23, no. 4 (supplement) 2002, The United Nations University.
Design of a prospective, randomized evaluation of an integrated nutrition program Nam field office developed the program model and temperature 23.4°C). The province serves as a regional implemented it through government partners. The transportation hub for road, rail, and inland waterways LINKAGES project evaluated models for improving and has some industry (fertilizer and paper).4 There are complementary feeding. Emory University provided two rice harvests (May and June and September and technical assistance for research design, technical sup- October). Diarrhea season is May and June; respira- port, and analysis. RTCCD implemented the research, tory infection season is September to December. The and SC headquarters identified the partners and coor- rainy season is February to July, and the dry season is This paper presents the methods and describes the SC’s district selection criteria were no prior SC activ- intervention for this 12-month longitudinal study.
ity, supportive officials, and a preference for lowland or midland (not highland) ecology to maximize popula-tion density. SC selected Thanh Ba and Phu Ninh Dis- tricts. SC’s commune programming selection criteria were a greater than 30% malnutrition rate (weight- Objectives and design
for-age Z score less than –2) among children under five years of age; poverty as measured by per capita The ViSION project aimed to identify the results and rice production; and feasibility factors such as at least causal pathways of SC’s integrated nutrition program, 300 children under five, supportive local leadership, the community empowerment and nutrition program and food generally available as reported by commu-(CENP). Principal aims of the research project included nity leaders. We excluded extremely poor communes, documenting the implementation of all aspects of the which belonged to the government’s “Phase One 10,000 CENP, and measuring the impact of the program ele- Poorest Communes” and were already targeted by the ments on maternal care behaviors and dietary intakes government for special intervention programs. Com-and growth of children. mune selection criteria for research sought comparabil- We used a longitudinal, prospective, randomized, ity and minimal contamination between intervention “probability effectiveness” design [1]. We randomized and comparison communes. To minimize confound-at the commune level (six intervention and six com- ing, we selected four groups of three communes (two parison) and then followed a total of 240 children (120 in each district) from the 18 communes (10 in Thanh in intervention and 120 in comparison communes at Ba District and 8 in Phu Ninh District) eligible for baseline) longitudinally for six months with a re-survey programming. The group means or prevalences for at 12 months. Blinding subjects or data gatherers to the childhood malnutrition level, ecology, ethnicity, and intervention was not feasible. distance from district center were as similar as possible (p > .05) between each group in a district. We avoided Setting and commune selection
contiguous communes where possible. We then com-bined two groups of three communes, one group from The study took place in Phu Tho Province, 98 km each district, maximizing similarity of background northwest of Hanoi. This north central ecological characteristics. Finally, we randomly assigned, by coin region has the worst child nutrition profile (47% toss, one group of six for the intervention and the other underweight, 46.5% stunted, and 9.9% wasted) in the group of six as comparison communes. country.* The province has 1.3 million rural lowland, The commune profiles were similar (table 1) in both midland, and highland inhabitants in thousands of study groups. All commune economies were agricul-hamlets in 249 communes in eight districts. The main ture-based, subsistence and cash crop; only one (Thanh ethnic group is Kinh, the predominant national major- Ba) added bamboo weaving. All communes were acces- ity, with some Muong, Dao, San Chay, and San Diu sible by dirt roads with seasonal difficult access; one minorities. The population is poor with an average per commune (Yen Noi) was commonly one-third sub-capita monthly rice production of 17.8 kg.** The area merged during July and August. All communes were is principally agricultural (rice, tea, palm oil, banana, Kinh ethnicity and had supportive leaders, and none papaya, eucalyptus) with a tropical climate (average had a food program. Each commune had a commune health center (CHC) with a staff of three to seven, including “doctor’s assistants” and nurses. Patients * UNICEF and National Institute of Nutrition 1998 survey obtained free consultation but paid for treatments. The CHC provided ambulatory and brief observational ** The Ministry of Labor, Invalids and Social Affairs cat- care. Seriously ill patients were referred to the district egories food insecurity and poverty: less than 13kg of rice produced per person per month represents “food deficiency” while less than 20 kg per person per month in lowlands and less than 16 kg per person per month in the less fertile moun- * Viet Nam sights and tourism, Viet Nam tourism depart- Food and Nutrition Bulletin, vol. 23, no. 4 (supplement) 2002, The United Nations University.
Design of a prospective, randomized evaluation of an integrated nutrition program Intervention
The intervention consisted of preparatory activities, training, situation analysis, and implementation (table 2). Below we describe the intended intervention as specified in SC’s training manuals [2]. A concomitant proc- ess evaluation assessed adherence to protocol Preparatory activities
Save the Children approached district and evaluate potential program sites, and select actual sites. Upon finalization of the sites, SC assisted local partners to form two district and six commune management steering committees (CMSC). The district and com- people’s committee chairman or vice-chair-man, head of the women’s union, head of the CHC, and often those responsible for early Training
SC staff conducted a training of trainers in November, 1999 that included two days of management training for provincial, district, and commune partners and an additional five days of implementation training for designated district trainers. These, in turn, trained local health volunteers to imple- ment the 10-month CENP, including the situation analysis and interventions in their Situation analysis
The situation analysis included developing a roster of all children less than three years of age, conducting the first growth moni- conducting the baseline positive deviance inquiry (PDI). The SC staff developed the child roster based on government records from commune health centers and official family planning records. The ViSION study team validated these official records with a census in a comparison commune, Yen Noi, Design of a prospective, randomized evaluation of an integrated nutrition program Implementation training for DFT and Positive deviance inquiry and analysis steering committee (PMSC), district Training of implementers (health Growth monitoring and promotion sessions
off families. Selection criteria for positive deviant chil-dren included normal nutritional status, not an only Health volunteers then conducted the first every- two-month commune-level GMP session at which SC staff trained district trainers and members of the they weighed all children less than 3 years of age on a CMSC in the concepts and methods of the positive UNICEF-approved, locally made, Salter 25 kg (0.1 kg deviance approach using the CENP training manual interval) infant scale (Salter Brecknell Weighing Prod- developed by SC [2]. They conducted the PDI with ucts, Minneapolis, Minn., USA). At each GMP session, health volunteers by visiting the six families and using health volunteers plotted weight-for-age on children’s a question and observation guide to gather informa-national, combined-sex, “road to health” cards display- tion from family members about child feeding, care, ing four growth channels: A (“normal”) –2 standard and health. The CENP teams also visited the better-off deviations (SD) or more below the reference median families with malnourished children to identify the weight-for-age; B (“moderate malnutrition”) less than causes of their children’s malnutrition and to high-–2 and –3 SD or more; C (“severe malnutrition”) less light the lesson that money was neither necessary nor than –3 and –4 SD or more; and D (“very severe mal- sufficient for child health. District trainers attempted nutrition”) less than –4 SD. Health volunteers coun- to schedule the PDI to allow observation of a child seled about growth, reinforced or introduced good feeding. They followed a question guide with prompts care-giving behavior, and referred ill children or those and took notes. Accompanying health volunteers were who failed to gain weight after two GMP sessions (four encouraged to ask questions. The two district CENP months) to the community health center for medical teams organized PDI data from each family, according evaluation. to SC’s manual and field guide [3], into four categories: demographic information, feeding behaviors, health- Positive deviance inquiry
seeking behaviors, and caring behaviors.
After the home visits, the CENP team and interested The results of the first GMP session and household community members convened in an open forum to wealth ranking served to identify children eligible for review the cases and to develop a summary profile for the positive deviance inquiry (PDI). The purpose of their commune of uncommon behaviors that seemed the PDI was to identify behaviors that poor families to contribute to the positive deviant (PD) children’s with healthy children (“positive deviant families”) good nutritional status despite their families’ poverty. practiced that were different from the norm and likely The team invited community input to develop a final contributed to the children’s good health. The CENP consensus. The behaviors identified through the PDI team grouped children according to nutritional status informed messages taught at the nutrition education and age. Then health volunteers and the CMSC mem- and rehabilitation program (NERP) and GMP sessions bers identified those who were very poor and those to help malnourished children in the commune. who were better off based on family per capita rice After the situation analysis, the intervention proper production, occupation, land holdings, and posses- consisted of GMP sessions every two months, monthly sions. They identified four positive deviant and two NERP sessions as long as the number of malnourished negative deviant children per commune. Positive devi- children warranted them for up to nine months, ant children were well-nourished (channel A) children deworming after the second GMP session, and internal from poor families and negative deviant children were (i.e., apart from the ViSION evaluation) monitoring of malnourished (channel C or D) children from better implementation and nutritional outcomes. Design of a prospective, randomized evaluation of an integrated nutrition program Design of a prospective, randomized evaluation of an integrated nutrition program Nutrition education and rehabilitation program
sessions
All channel B, C, and D children, i.e., less than –2 WAZ, were referred to NERPs. Health volunteers conducted monthly NERP sessions for 12 days (six days weekly for two weeks) in their hamlets to enable families to both rehabilitate and to sustain the enhanced nutritional status of their malnourished children. In the spirit of “learning by doing,” mothers and caretakers were told to bring a handful of positive deviant foods each day as the “price of admission” to the NERP. Health volunteers taught hygiene, child development, and the preparation and feeding of calorie- and nutrient-dense meals from locally available, affordable foods. They stressed that the “positive deviant” foods brought by participants, such as shrimps, crabs, and greens, were all abundant in the paddies. Participants learned six key “facts for life”[4] messages concerning breastfeeding, food vari- ety, complementary feeding, health care, and taking care of healthy children at home. Key NERP educa- tion tools included “food squares” and “clover leaf diagrams.” Food squares listed the four recipes to be prepared and fed to children as extra daily meals during the 12 days. Each included breastmilk, local positive deviant foods (informed by the PDI), and common local sources of energy, protein, and micronutrients (tofu, oil, fish, eggs, etc.). Cloverleaf diagrams showed the commune’s PD behaviors for a “model (i.e., PD) family” under the headings: good food, good child care, On each of the 12 NERP days each participating child received a nutritious meal prepared by two or more caregivers in rotation. The meal was designed as additional to the usual diet, but in practice it may have substituted for another meal [5]. Health volunteers allowed mothers to take home the unconsumed food if the child was reluctant to eat at the NERP session. In addition to mothers’ contributions, the CENP pro- vided funds to the CMSC, according to the baseline level of severe malnutrition, in this case about 1400 VND (US$ 0.10) per NERP enrollee per NERP day. The CMSC, in turn, allocated the money to the NERP cent- ers for the purchase of oil, rice, and eggs or tofu. The health volunteers weighed the participating children at entry and at the completion of each two-week NERP session and reviewed the progress of each NERP session with the participant mothers. The CMSC discontin- ued NERP sessions when the number of malnourished children made further group rehabilitation inefficient. Thereafter, health volunteers individually made home visits to malnourished children identified through GMP sessions. CENP technical support for NERPs ceased at month nine. Prior to this CENP in Phu Tho, SC had designed NERPs to enroll only severely malnourished children (< –3 Z weight-for-age). However, the ViSION project’s Design of a prospective, randomized evaluation of an integrated nutrition program Design of a prospective, randomized evaluation of an integrated nutrition program baseline showed that the level of malnutrition was lower than official Ministry of Health level (30.0% vs. 44.5%, respectively). Thus, this CENP also included moderately malnourished (< –2 Z and ≥ –3 Z) children; whereas, prior CENPs graduated severely malnourished children upon achieving moderate malnutrition status. Children graduated from the NERPs in this CENP either by attaining normal nutritional status or by attending four consecutive NERPs with weight gain despite remaining moderately malnourished. SC provided funds to deworm children under three years of age during or after the second GMP session. In the absence of a comprehensive national policy, dif- ferent CMSCs, in dialogue with the DMSC, opted for various regimens, including Hatamintox Pyrantel, 125 to 250 mg for children 12 to 36 months of age; Pyr- antel, 125 mg for 12 to 23 months of age and Vermox, 500 mg for 24 to 36 months of age; Mebendazol, 500 mg for 24 to 36 months of age; or Panatel, 125 mg for 24 to 35 months of age and 187.5 mg for 36 months of age. Children in comparison communes were also Monitoring
Health volunteers maintained NERP and GMP rosters and met with the CMSC monthly to report and review progress. A member of the CMSC supervised two to four NERPs, visiting each at least once monthly plus the relevant GMP session, if scheduled. The CMSC reviewed all health volunteers’ hamlet summaries and aggregated the commune’s data, often plotting them on a prominently displayed chart or table in the People’s Committee headquarters or commune health center. A member of the DMSC visited each program commune monthly during a NERP and GMP session, if scheduled. The DMSC convened a monthly joint meeting with representatives of all CMSCs to review each commune’s financial, GMP and NERP activi- ties. It forwarded summary aggregates to SC monthly during the program and for one year after the end of Role of Save the Children
SC staff worked with government counterparts to select intervention communes, conducted the train- ing of trainers, and advised the training and situation analyses. SC also visited the districts and each com- mune once or twice during implementation due to the CENP adaptation for moderately malnourished children and the requirement for more health volun- teers than usual given the population dispersion. Total program cost was approximately US$ 12,000 for the Design of a prospective, randomized evaluation of an integrated nutrition program ViSION project
Data collection
Forms and equipment
The research team identified children 5 to 25 months We collected baseline data between December 23, of age in the intervention and comparison communes 1999 and January 9, 2000. We used 11 forms (table 3) from the government population records. In interven- for gathering information on nutritional status, diet, tion communes, the CENP team had already weighed illness, care, internal determinants of behavior (knowl-all children less than three years of age. In comparison edge, beliefs, and attitudes including perceived advan-communes, the research team weighed all children less tages and disadvantages of the behavior, self-efficacy, than two years of age. We calculated mean weight-for- norms, and skills), external determinants of behavior age Z score (WAZ) for each hamlet and ranked the (time, father’s role, and maternal nutritional status), hamlets by mean WAZ within each commune. Since and empowerment or the ability to increase control the SC program stressed rehabilitation of malnour- over and to attain better health. During the longitudi- ished children, we selected the 34 (of 67) intervention nal study we used an additional form to gather infor-hamlets and the 30 (of 60) comparison hamlets with mation on child feeding through video-recording. We the highest levels of malnutrition both to capture developed all questionnaires in English, translated program impact and to maximize field efficiency. We them into Vietnamese, trained field workers in their sorted the eligible children from these hamlets by age use, pilot-tested them in Phu Tho Province (but not and randomly selected children to achieve the desired in the study communes), revised them, re-trained sample size of 120 intervention and 120 comparison field workers, and back-translated them into English children. We excluded multiple births or children with for accuracy. We obtained all measurements on all sub-severe medical problems, such as handicap or measles. jects at baseline, except for videotaped care (obtained We achieved the desired sample without refusal after on one-half of the sample at either age 12 or 17 months 20 substitutions for families not at home. starting two months after baseline) and the market The Ministry of Health used two growth-monitoring survey (obtained in each of the communes starting four charts during the ViSION project. Upon enrollment, months after baseline). Throughout the study we used field workers gave comparison families a “new” chart an additional group of forms (developed in a similar (0–24 months, channel A vs. non-A) on which they fashion to the original 11 forms) to evaluate the CENP recorded the weights. Although the chart had nutri- implementation of the intervention (table 4). tion messages on the reverse, families were unlikely to Field workers obtained each anthropometric meas- read these since the field workers immediately taped urement three times at each measurement period. We them to the wall. When comparison children reached used digital reading tare for the UNICEF electronic their second birthday, field workers gave them the scale 890 (SECA Ltd. Birmingham, UK) precise to 100 “old” (0–60 months, channel A, B, C, and D) growth g for weight; four-color mid-upper arm circumference chart, taped it adjacent to the first one, and recorded tapes precise to 1 mm for mid-upper arm circumfer-subsequent weights. The CENP did not provide growth ence; and Shorr infant/child/adult height measuring charts to intervention families, but about 20% of inter- boards (Shorr Productions, Olney, Md., USA), precise vention families requested a chart, and the field worker to 1 mm for maternal heights and recumbent child provided them the “old” one, taped it to the wall, and lengths. Cold weather from December through Feb-recorded the weights. ruary precluded undressing children, so we weighed These 240 study children represented the range of representative clothing items, inventoried the subject’s nutritional status of children in the study hamlets. clothing, calculated the total weight of the clothes, and Since the NERP, a central CENP intervention, targeted subtracted this from the otherwise tared children’s only malnourished children, we augmented the sample weight (mother with partially dressed child versus during month three of the study with an additional 41 mother alone) [6]. NERP participants from the intervention communes to Field workers conducted a 24-hour dietary recall increase the sample size. We used the following selec- using photographs and digital reading weighing scales tion criteria: not currently enrolled in the longitudinal (Soehnle Attache Gram Scales, Montlingen, Switzer-study; resident of one of the 34 hamlets in the six land), with 2-kg capacity and 1-g and 2-g precision intervention communes; attending NERPs according for items weighing less than 1 kg and 1 to 2 kg) to to the latest list available; and age 6.0 to 23.9 months identify and quantify child food preparations and at baseline. We achieved the desired sample after five ingredients [5]. refusals.
We video-recorded two, two-hour segments of care timed around a child feeding. We conducted these recordings among a random sub-sample of 112 chil-dren when they were either 12 or 17 months old. Sub- Design of a prospective, randomized evaluation of an integrated nutrition program Additional information on variable construction and and CENP implementation, mastered a new qualita-the analysis of growth [10], dietary [5], and morbidity tive method, videotaping, achieved nearly complete variables [11] are presented elsewhere in this volume. data collection, and maintained exemplary data management. On the other hand, no research effort is watertight. The effect of repeated household visits Discussion
remains unknown. The potential bias of demonstrat-ing atypical behavior for the video camera is real, but This paper provides a detailed description of the imple- can be reduced by restricting analysis to the second mentation of the CENP, a complex integrated nutri- hour or even the second day when videotaping was less tion program in Viet Nam, and describes the rigorous novel. The ethical requirement to provide caregivers of methods to evaluate it. This is one of a few efforts of all comparison and some intervention children with this kind. Most large-scale, detailed evaluations have a growth chart, which included nutritional messages, been “efficacy trials” rather than full-scale effectiveness was unavoidable and may have mitigated detecting the studies [12, 13]. Moreover, effectiveness trials typically impact of the CENP because comparison caregivers have not involved randomization [12].
gained nutritional knowledge. Similarly, the detergent incentive may have influenced hygiene or other health Implementation
outcomes in both groups. Likewise, to isolate the effect of the CENP apart from its deworming component, SC succeeded in teaching district Ministry of Health we dewormed all study children, and this might have partners to train local implementers to conduct a com- improved the health and nutrition of the comparison plex set of interventions: GMP, PDI, and NERP. Despite children. We added 41 children who were attending the the implementers’ characteristic enthusiasm, some field NERPs from month three. Interpreting their baseline realities prompt caution. The CENP is designed as a and change may be difficult, given their late start, how-rehabilitation model for lowland, densely populated ever, they were not included in most analyses to date. communes with high levels of severe malnutrition. In addition, we decided to explore home care and care-Neither condition prevailed in the selected districts, seeking only for diarrhea, but its prevalence at baseline which were drawn from a pool of similar districts in December was trivial, so demonstrating change may with minimal non-governmental organizational activ- be challenging. Finally, while non-governmental organ- ity and a SC commitment to introduce the program. ization activity was nil at the baseline, three communes Political support in Phu Tho province for district and (two intervention and one comparison) received addi-commune selection was more restrained than in earlier tional nutritional inputs.
iterations. These communes also had baseline levels of On balance, the above factors are likely to have mod- malnutrition that were far less than officially reported. erately attenuated differences between intervention and In response, SC recommended including moderately comparison children. All listed program factors would malnourished children in the NERPs and revised reduce CENP’s impact. Similarly, our provision of graduation criteria, a modification that had been con- certain items (growth charts and detergent) to fami- templated, but not tested. Thus, compared to children lies in the comparison communes for ethical reasons in prior CENPs, these intervention children were more would tend to improve some outcomes in comparison likely to be enrolled in and slower to graduate from communes.
NERPs. Meanwhile, NERP implementation deviated In summary, this evaluation documents the differ- from protocol in that daily contributions were not the ence between CENP protocol and implementation and norm, and home-delivered meals were common, per- the effectiveness of the implementation in a challeng- haps because the population was dispersed, or caregiv- ing field setting. While the implementation may be ers were less concerned about moderate malnutrition similar to previous iterations, the effectiveness may be than they would have been about severe malnutrition. somewhat less, given the challenging field conditions It is possible that this and other local adaptations may and changes to protocol. Our thorough understanding have been implemented earlier in other settings with- of the CENP implementation allows us to better inter- pret its effect and thus the potential of PD-informed programs.
Research methods
The ViSION project trained its field research team to Acknowledgements
exacting standards of data collection. The Research
Training Center for Community Development This research was supported by the LINKAGES:
(RTCCD) team then identified and enrolled interven-
Breastfeeding, LAM, Complementary Feeding, and tion and comparison households, repeated caregiver Maternal Nutrition Program. LINKAGES is supported interviews and unobtrusive observations of households by G/PHN/HN, Global, the United States Agency for International Development (USAID) under the terms ington, D.C. The opinions expressed herein are those of Grant No. HRN-A-00-97-00007-00 and is managed of the authors and do not necessarily reflect the views by the Academy for Educational Development, Wash- References
1. Habicht J-P, Victoria CG, Vaughan JP. Evaluation designs on child growth in Viet Nam. Food Nutr Bull 2002;23 for adequacy, plausibility and probability of program per- formance and impact. Int J Epidemiol 1999;28:10–8.
9. Sternin M, Sternin, Marsh D. Rapid, sustained childhood 2. Save the Children/US Viet Nam Field Office. The com- malnutrition alleviation through a “positive deviance” munity empowerment and nutrition education program approach in rural Vietnam: Preliminary findings. In: training manual (5 volumes). Hanoi, Viet Nam: Save the Wollinka. O, Keeley E, Burkhalter RB, Bashir N, eds. The Children/US Viet Nam Field Office, 1999.
hearth nutrition model: applications in Haiti, Vietnam, 3. Sternin M, Sternin J, Marsh DR. Field guide: designing and Bangladesh. Report of a technical meeting at World a community-based nutrition education and rehabilita- Relief Corporation Headquarters, June 19–21, 1996, Whea- tion program using the “positive deviance” approach. ton, Ill., USA. Arlington, Va., USA: BASICS, 1997:49–61.
Westport, Conn., USA: Save the Children and BASICS, 10. Schroeder DG, Pachón H, Dearden KA, Ha TT, Lang TT, Marsh DR. An integrated child nutrition interven- 4. UNCEF, UNESCO, WHO. Facts for life. Oxfordshire, tion improved growth of younger, more malnourished children in Northern Viet Nam. Food Nutr Bull 2002;23 5. Pachón H, Schroeder DG, Marsh DR, Dearden KA, Ha TT, Lang TT. Effect of an integrated child nutri- 11. Sripaipan T, Schroeder DG, Marsh DR, Pachón H, tion intervention on the complementary food intake Dearden KA, Ha TT, Lang TT. Effect of an integrated of young children in rural north Viet Nam.Food Nutr nutrition program on child morbidity due to diarrhea and respiratory infection in northern Viet Nam. Food 6. Tuan T, Marsh DR, Ha TT, Schroeder DG, Thach TD, Dung VM, Huong NT. Weighing Vietnamese children: 12. Caulfield LE, Huffman SL, Piwoz EG. Interventions to How accurate are child weights adjusted for estimates improve intake of complementary foods by infants 6 of clothing weight? Food Nutr Bull 2002;23(4 Suppl): to 12 months of age in developing countries: impact on growth and on the prevalence of malnutrition and 7. Dean AG, Dean JA, Coulombier D, Brendel KA, Smith potential contribution to child survival. Food Nutr Bull DC, Burton AH, Dicker RC, Sullivan K, Fagan RF, Arner TG. Epi Info, version 6: A word processing, database, and 13. Caulfield LE, Piwoz EG, Huffman SL, Improved child statistics program for epidemiology on IBM-compatible diet and growth: Efficacy and effectiveness studies. In: microcomputers. Atlanta, Ga., USA: Centers for Disease Martorell R, Haschke F, eds. Nutrition and Growth. Nestlé Nutrition Workshop Series, Pediatric Program, 8. Mackintosh UAT, Marsh DR, Schroeder DG. Sustained Vo. 47. Philadelphia: Lippincott Williams & Wilkins: positive deviant child care practices and their effects

Source: http://eng.kolab.pl/uploads/media_pdf/711334954542.pdf