10.24265/horizmed.2024.v24n4.13
Original Article
Coverage of the Vaso de Leche and Qali Warma programs among children under five years of age and their associated factors
Ana Lucía Cherres-Bernal* 1,a
Paloma Rodríguez-Massa 1,a
Franco Romaní-Romaní 1,b
1 Universidad de Piura, School of Human Medicine. Lima, Peru.
a Medical student
b Doctor of Medicine, master’s degree in Epidemiology.
ABSTRACT
Objective: To estimate the coverage of being a beneficiary of the Vaso de Leche (VDL - Glass of Milk) and Qali Warma (QW - Vigorous Child) programs among children under five years of age and to determine the associated factors.
Materials and methods: This study used a secondary source, based on the 2022 Encuesta Demográfica y de Salud Familiar (ENDES - Demographic and Family Health Survey) in Peru. For this purpose, women aged 15 to 49 years with children under five of age and complete anthropometric data were selected to answer questions about being a beneficiary of the VDL and QW programs. The dependent variable was being or not a beneficiary of the programs, and the rate of beneficiaries was compared across levels of the independent variables. A log-binomial regression model within the family of generalized linear models was used for the multivariate analysis, with a significance level of 5 %. Results: Among children under five years of age, the rate of beneficiaries of the VDL program was 34.13 %, while that of the QW program was 85.88 %. Variables associated with being a beneficiary of the VDL program included being ≥ 12 months, birth weight < 2,500 g, maternal educational level, belonging to the poorest quintiles and rural residence. The factors associated with being a beneficiary of the QW program included rural residence and belonging to the poor and poorer quintiles. There was no association between being a beneficiary of the QW (aPR: 1.01; 95 % CI: 0.97-1.05) and VDL (aPR: 1.02; 95 % CI: 0.97-1.07) programs and being a child with chronic malnutrition at the time of the survey.
Conclusions: The coverage of the VDL and QW programs differs significantly and varies according to sector. Being from the poorest quintile and living in rural areas were common correlates across both programs. Nevertheless, these programs also benefited children who were not necessarily in socioeconomically disadvantaged conditions, suggesting issues in the implementation of the intended objectives.
Keywords: Food Assistance; Social Programs; Child, Preschool; Infant; Health Surveys; Peru
INTRODUCTION
In Peru, supplementary feeding programs are part of the Plan Nacional de Seguridad Alimentaria y Nutricional (National Food and Nutrition Security Plan), whose objective is to guarantee access to nutritious food, preferably to all individuals in situations of vulnerability or poverty (1). As part of this national strategy, two programs stand out: Glass of Milk (VDL) and Programa Nacional de Alimentación Escolar Qali Warma (Qali Warma [QW - Vigorous Child] National School Feeding Program). The former prioritizes children from zero to six years of age, pregnant and breastfeeding mothers, as well as children aged 7-13 years, older adults, and individuals with tuberculosis (2). The latter is designed to provide support to children in early childhood education, starting at three years of age, as well as to those in primary and secondary levels of public educational institutions in Peru (3).
Ensuring nutritious food for children under five years of age is crucial. The dietary intake of preschool-aged children at home may be affected by unfavorable socioeconomic conditions (4). However, the educational environment should be considered to understand their dietary quality. It has been shown that a poor heart-healthy diet is provided without differences between the home and educational institutions (5). Exposure of children under five to food insecurity increases the risk of developing unhealthy eating habits in adulthood (6). It also raises the risk of cardiovascular diseases, particularly among adult males (7).
Food security seeks to guarantee this right through the consumption of quality products, in order to reduce inequality in access to food among population subgroups (8). In Peru, in 2021, 61.4 % of VDL users were under six years of age, half of whom lived in rural areas (9). On the other hand, 50.6 % of children aged 3-11 years who attended a public school benefited from the QW program (10), and of these, 73.1 % lived in rural areas (11). These figures reflect that an equitable distribution may not be being achieved throughout the target population, including children aged five years and younger, for whom promoting food security is fundamental.
Several studies have been carried out to verify the proper targeting of these programs. In 2003, a study based on data from the Encuesta Nacional de Hogares (ENAHO - National Household Survey) characterized the VDL program and found that 27 % of beneficiaries were non-poor, while 36 % of extremely poor households were not beneficiaries (12). In 2018, a study described the VDL beneficiary population of the Municipality of San Ignacio (Cajamarca) and found that 52 % had economic resources (13).
A similar issue is observed in the QW program. In 2017, a study evaluated its implementation in a district of Cusco and found that the percentage of coverage was not assessed, nor was the level of school attendance and permanence recorded, making it unclear whether coverage was adequate (14). Likewise, data from the ENAHO from 2015 to 2018 show that QW had a positive impact on the health of beneficiary children, but that the effect was not observed in poor children (15).
The purpose of both programs is to provide food assistance to target populations and, therefore, strict control over beneficiaries and the food distribution process must be ensured at the national level. This raises the following questions: What are the main characteristics that determine whether a child under five years of age becomes a beneficiary of the VDL and QW programs? And, at the national level, under routine conditions, are these programs reaching their target group, specifically children under five? In view of this issue, this study attempts to determine the proportion of children under five years of age who are beneficiaries of the VDL and QW programs based on some sociodemographic characteristics, using data from the 2022 Encuesta Demográfica y de Salud Familiar (ENDES - Demographic and Family Health Survey) in Peru.
MATERIALS AND METHODS
Study design and population
We conducted a secondary source study based on the ENDES carried out in Peru between January and December 2022. This had a two-stage probabilistic sampling design, with representative estimates of the population of Peru: its 24 departments and the constitutional province of Callao, urban and rural areas and natural regions (Metropolitan Lima, the rest of the Coast, the Highlands and the Jungle). Of the 38,105 selected women aged 12-49 years, 35,787 completed the interview (response rate: 93.9 %); and of the 22,424 children aged five years or younger selected for anthropometric data measurement, 21,995 were evaluated (response rate: 98.1 %).
The ENDES targeted women aged 12- 49 years and their children aged five years or younger who were residents of the selected household. It also included those women that stayed overnight in the dwelling the night before the day of the interview, even if they were not usual residents. In our study, the inclusion criteria were the following: 1) woman aged 15-49 years living with a child aged five years or younger, 2) child with comprehensive anthropometric data measurement, and 3) mothers who answered the questions aimed at identifying the coverage of the VDL (subsample 1) and QW (subsample 2) programs. Based on the last selection criterion, two study groups were defined.
Coverage of supplementary feeding programs
The ENDES includes a section on food assistance programs aimed at determining the access of beneficiaries to any social food or nutrition assistance program, considering at least one household member. The (female) interviewer asks these questions to the woman, who, in the case of this study, is the mother of children under five years of age.
For our analysis, the 2022 ENDES collected information using two questions for each program: 1) Does any member of your household receive food or nutrition assistance from the VDL social program? (variable QH101), and 2) Does (NAME) receive breakfast and/or lunch at school from the QW national school feeding program? (variable PS109_1R). Responses were collected using the following categories: yes, no, and does not know/does not remember. For the purposes of this study, each variable was recategorized into a dichotomous variable (yes or no). There were no participants who responded does not know/does not remember.
Variables and measurements
We examined three groups of variables to characterize the coverage of the QW and VDL programs. Within the characteristics of the child, we included sex (variable B4), measured as male or female; age (variable HC1), measured in months (and recategorized into 0-11, 12-35 and 36-59 months); prenatal checkups (variable M14), corresponding to the number of checkups and a does not know option, recategorized as no checkups, 1-5 and ≥ 6 checkups (responses marked as does not know were treated as missing data) (16); term birth, measured as yes or no based on the variable duration of pregnancy (variable QS220A) (considered yes when the duration was nine months); cesarean birth (variable M17), measured as yes or no; and birth weight (variable M18), measured in grams, which also included the categories not weighed at birth and does not know (it was recategorized into three groups: < 2,500 g, 2,500 to < 4,000 g, and ≥ 4,000 g; responses of not weighed at birth and does not know were treated as missing data.
The ENDES defined chronic malnutrition in children as a condition affecting children under five years of age who had a height-for-age Z-score (variable HW70) less than -2 standard deviations (< -2 SD) from the median, according to the child growth standards established by the World Health Organization (WHO). The categories were the following: no malnutrition (HW70 ≥ -2 SD), moderate malnutrition (-3 SD ≤ HW70 < -2 SD) and severe malnutrition (HW70 < -3 SD). The variable was later recategorized dichotomously (presence or absence). The procedures and materials used for measuring height are detailed in the manuals for anthropometrists and interviewers (17,18). The anthropometrists were previously trained in anthropometric techniques in accordance with the national technical standard (19).
The second group of variables were the characteristics of the mother: educational level (variable V106), recategorized into no education-primary, secondary and higher; age (variable V012), measured in years, which, according to the inclusion criteria, was recategorized into three groups: 15-24, 25-34, and 35-49; marital status (variable V501), measured with the categories married, living together, never married, widowed, divorced, not living together, which was then recategorized into in a union (married, living together) and not in a union (for the other categories); current work status (variable V714), measured as working or not working; maternal ethnicity (variable V131), recategorized as no when the response was Spanish, Portuguese or other foreign language, and yes for the other categories.
The third group of variables consisted of the characteristics of the household and included area of residence (urban or rural); wealth index in quintiles with the categories poorer, poor, middle class, rich, richer; and department, which included the 24 departments and one constitutional province as response categories. The variables identified were selected according to a review of the literature (20-23).
The ENDES collects information on the specified variables through face-to-face interviews conducted in selected households. The data were recorded using a computer application by the interviewer and the anthropometrist, both of whom are health professionals.
Statistical analysis
Given the complex sampling design of the ENDES, we applied the variables V001 (cluster), V022 (stratum), and V005 (female weighting factor), the latter divided by 1,000,000 to obtain the sample weight.
We conducted the analysis on two subsamples defined by the inclusion criteria using the svy command of the Stata program, version 16. Each subsample consisted of children aged five years or younger with comprehensive anthropometric data, whose mothers were interviewed and received complete information about the VDL or QW program.
A descriptive analysis was conducted for each defined subsample, estimating the weighted proportion, standard error (SE) and 95 % confidence interval (95 % CI) of the coverage of each program, as well as the characterization according to the study variables. Additionally, to explore the ecological- level independence in the departmental implementation of both programs, we performed a correlation analysis using Spearman’s Rho coefficient, taking as data the weighted proportions by department of the VDL and QW coverage.
In the bivariate analysis, the dependent variable was considered to be a beneficiary or not of the social program. The proportion of being a beneficiary was compared across strata of the independent variables using the Pearson’s chi-square test with second-order Rao-Scott correction. This analysis was performed for each subsample of the study.
A multivariate analysis was conducted using a log-binomial regression model within the family of generalized linear models (24). This model uses a log link function to relate a binary outcome variable-being or not being a beneficiary of a social program-to a set of explanatory variables, which in this analysis were those selected to characterize program coverage. This model allows an unbiased estimation of adjusted prevalence ratios based on a range of explanatory variables for a common outcome (> 10 %) (24).
We formulated two models, one for each social supplementary feeding program. In the bivariate analysis, we selected independent variables that reached a p value < 0.20 (two-tailed). All covariates were entered simultaneously into each model. The absence of multicollinearity was assessed by assessing the SE of the regression coefficient of each variable. Standard errors greater than 2.0 were indicative of multicollinearity between the independent variables (25).
The strength of association between the dependent and independent variables was estimated with an adjusted prevalence ratio. Point estimates were presented with 95 % CI. A statistically significant association was considered when the CI did not include the value of 1.
Ethical considerations
The ENDES database is anonymized and freely available through the INEI portal. Prior to the analysis, the study protocol was approved by the Institutional Review Board (IRB) of the Universidad de Piura (UDEP).
RESULTS
After applying our selection criteria, the study population was divided into two subsamples: subsample 1 (VDL) consisting of 21,923 children, and subsample 2 (QW), consisting of 4,870 (Figure 1).
Figure 1. Flowchart of participant selection for the analysis
CMC: chronic malnutrition in children, PS: social program, VDL: Vaso de Leche (Glass of Milk).
Description of coverage
Among children under five years of age, the proportion of participants benefiting from the VDL program was 34.13 % (Table 1), while in QW it was 85.88 % (Table 2). These percentages at the departmental level show that, for VDL, the departments with the highest proportion of beneficiary children were Callao (73.46 %), Huancavelica (69.05 %), Amazonas (60.42 %), Cajamarca (60 %) and Puno (55.90 %). In the QW program, the departments with the highest coverage were Tumbes (99.06 %), Amazonas (98.95 %), Moquegua (97.49 %), San Martín (96.07 %) and Pasco (95.94 %) (Table 1). We did not find a linear correlation between the weighted proportions of the programs (Spearman’s Rho = 0.138; p = 0.508).
Among children under five years of age with valid responses to the question about VDL, 51.97 % were male, 42.65 % were 36-59 months old, and 86.04 % had a birth weight of 2,500-3,999 g. Among those with valid responses to the question about QW, 51.20 % were male and 85.86 % had a birth weight of 2,500-3,999 g. The most important characteristics found in mothers who provided valid responses about VDL and QW were being aged 25-34 years (48.79 % in VDL; 49.06 % in QW) and having a high school education level with 48.42 % in VDL and 48.73 % in QW (Tables 1 and 2).
Table 1. Characteristics of children under five years of age with valid responses to the question about being a beneficiary of the VDL program (N = 21,913)
|
Variable |
n |
Weighted proportion (%) |
Confidence Interval |
|
SE |
|
|
|
|
LL |
UL |
|
|
Receives the VDL program |
|
|
|
|
|
|
Yes |
8,084 |
34.13 |
33.37 |
34.89 |
0.39 |
|
No |
13,829 |
65.87 |
65.11 |
66.63 |
0.39 |
|
Characteristics of the child |
|
|
|
|
|
|
Sex of the child |
|
|
|
|
|
|
Male |
11,334 |
51.97 |
51.14 |
52.79 |
0.42 |
|
Female |
10,589 |
48.03 |
47.21 |
48.86 |
0.42 |
|
Age (months) |
|
|
|
|
|
|
0–11 |
4,027 |
18.21 |
17.59 |
18.86 |
0.32 |
|
12–35 |
8,595 |
39.14 |
38.34 |
39.94 |
0.41 |
|
36–59 |
9,301 |
42.65 |
41.84 |
43.47 |
0.42 |
|
Prenatal checkups (n = 18,329) |
|
|
|
|
|
|
No checkups |
177 |
0.89 |
0.75 |
1.05 |
0.08 |
|
1–5 |
2,255 |
13.18 |
12.57 |
13.82 |
0.32 |
|
≥ 6 |
15,897 |
85.93 |
85.28 |
86.56 |
0.33 |
|
Term birth (n = 20,852) |
|
|
|
|
|
|
No |
3,858 |
20.44 |
19.75 |
21.15 |
0.36 |
|
Yes |
16,994 |
79.56 |
78.85 |
80.25 |
0.36 |
|
Cesarean birth (n = 20,852) |
|
|
|
|
|
|
No |
13,985 |
64.28 |
63.45 |
65.1 |
0.42 |
|
Yes |
6,867 |
35.72 |
34.9 |
36.55 |
0.42 |
|
Birth weight (g) (n = 20,255) |
|
|
|
|
|
|
< 2,500 |
1,261 |
6.16 |
5.77 |
6.58 |
0.21 |
|
2,500–3,999 |
17,414 |
86.04 |
85.43 |
86.62 |
0.3 |
|
4,000 |
1,580 |
7.8 |
7.34 |
8.28 |
0.24 |
|
Characteristics of the mother |
|
|
|
|
|
|
Maternal age (years) |
|
|
|
|
|
|
15–24 |
4,550 |
21.49 |
20.81 |
22.18 |
0.35 |
|
25–34 |
10,164 |
48.79 |
47.95 |
49.64 |
0.43 |
|
35–49 |
6,128 |
29.72 |
28.95 |
30.5 |
0.39 |
|
Maternal educational level (n = 20,852) |
|
|
|
|
|
|
No education – primary |
3,925 |
18.18 |
17.58 |
18.8 |
0.31 |
|
Secondary |
10,330 |
48.42 |
47.58 |
49.27 |
0.43 |
|
Higher |
6,597 |
33.39 |
32.58 |
34.22 |
0.42 |
|
Maternal union (n = 20,852) |
|
|
|
|
|
|
Not in a union |
3,569 |
17.35 |
16.72 |
18 |
0.33 |
|
In a union |
17,283 |
82.65 |
82 |
83.89 |
0.33 |
|
Mother ethnicity (n = 20,852) |
|
|
|
|
|
|
No |
18,634 |
92.08 |
91.7 |
92.45 |
0.19 |
|
Yes |
2,218 |
7.92 |
7.55 |
8.3 |
0.19 |
|
Mother currently working (n = 20,852) |
|
|
|
|
|
|
No |
9,238 |
43.93 |
43.09 |
44.77 |
0.43 |
|
Yes |
11,614 |
56.07 |
55.24 |
56.91 |
0.43 |
|
Characteristics of the household |
|
|
|
|
|
|
Area or residence |
|
|
|
|
|
|
Urban |
14,750 |
72.4 |
71.72 |
73.07 |
0.34 |
|
Rural |
7,173 |
27.6 |
26.93 |
28.28 |
0.34 |
|
Wealth index (n = 20,852) |
|
|
|
|
|
|
Poorer |
6,533 |
27.46 |
26.76 |
28.16 |
0.36 |
|
Poor |
5,631 |
24.38 |
23.68 |
25.09 |
0.36 |
|
Middle class |
4,089 |
20.12 |
19.43 |
20.82 |
0.35 |
|
Rich |
2,808 |
16.22 |
15.55 |
16.9 |
0.34 |
|
Richer |
1,791 |
11.83 |
11.23 |
12.46 |
0.31 |
95 % CI: confidence interval, LI: lower limit, UL: upper limit, SE: standard error.
Table 2. Characteristics of children aged three to five years with valid responses to the question about being a beneficiary of the QW program (N = 4,870)
|
Variable |
n |
Weighted proportion (%) |
Confidence interval |
|
SE |
|
|
|
|
LL |
UL |
|
|
Receives the QW program |
|
|
|
|
|
|
Yes |
4,359 |
85.88 |
84.47 |
87.18 |
0.69 |
|
No |
511 |
14.12 |
12.82 |
15.53 |
0.69 |
|
Characteristics of the child |
|
|
|
|
|
|
Sex of the child |
|
|
|
|
|
|
Male |
2,487 |
51.2 |
49.46 |
52.95 |
0.89 |
|
Female |
2,383 |
48.8 |
47.05 |
50.54 |
0.89 |
|
Prenatal checkups (n = 3,484) |
|
|
|
|
|
|
No checkups |
14 |
0.39 |
0.22 |
0.71 |
0.12 |
|
1–5 |
298 |
9.36 |
8.2 |
10.67 |
0.63 |
|
≥ 6 |
3,172 |
90.25 |
88.93 |
91.43 |
0.64 |
|
Term birth (n = 4,570) |
|
|
|
|
|
|
No |
760 |
19.04 |
17.61 |
20.56 |
0.75 |
|
Yes |
3,810 |
80.96 |
79.44 |
82.4 |
0.75 |
|
Cesarean birth (n = 4,570) |
|
|
|
|
|
|
No |
3,219 |
68.03 |
66.27 |
69.75 |
0.89 |
|
Yes |
1,351 |
31.97 |
30.25 |
33.73 |
0.89 |
|
Birth weigth (g) (n = 4,452) |
|
|
|
|
|
|
< 2,500 |
286 |
6.5 |
5.65 |
7.46 |
0.46 |
|
2,500–3,999 |
3,801 |
85.86 |
84.54 |
87.09 |
0.65 |
|
4,000 |
365 |
7.64 |
6.73 |
8.66 |
0.49 |
|
Characteristics of the mother |
|
|
|
|
|
|
Maternal age (years) (n = 4,570) |
|
|
|
|
|
|
15–24 |
767 |
16.64 |
15.36 |
18.01 |
0.68 |
|
25–34 |
2,267 |
49.06 |
47.25 |
50.86 |
0.92 |
|
35–49 |
1,536 |
34.31 |
32.6 |
36.05 |
0.88 |
|
Maternal educational level (n = 4,570) |
|
|
|
|
|
|
No education–primary |
996 |
22.23 |
20.81 |
23.71 |
0.74 |
|
Secondary |
2,269 |
48.73 |
46.92 |
50.54 |
0.92 |
|
Higher |
1,305 |
29.05 |
27.39 |
30.77 |
0.86 |
|
Maternal union (n = 4,570) |
|
|
|
|
|
|
Not in a union |
851 |
19.13 |
17.72 |
20.63 |
0.74 |
|
In a union |
3,719 |
80.87 |
79.37 |
82.28 |
0.74 |
|
Maternal ethnicity (n = 4,570) |
|
|
|
|
|
|
No |
4,039 |
90.98 |
90.08 |
91.79 |
0.44 |
|
Yes |
531 |
9.03 |
8.21 |
9.92 |
0.44 |
|
Mother currently working (n = 4,570) |
|
|
|
|
|
|
No |
1,700 |
37.89 |
36.15 |
39.67 |
0.9 |
|
Yes |
2,870 |
62.11 |
60.33 |
63.86 |
0.9 |
|
Characteristics of the household |
|
|
|
|
|
|
Area of residence |
|
|
|
|
|
|
Urban |
3,132 |
68.17 |
66.61 |
69.68 |
0.78 |
|
Rural |
1,738 |
31.83 |
30.32 |
33.39 |
0.78 |
|
Wealth index |
|
|
|
|
|
|
Poorer |
1,527 |
30.67 |
29.12 |
32.27 |
0.8 |
|
Poor |
1,329 |
27.78 |
26.2 |
29.41 |
0.82 |
|
Middle class |
923 |
20.64 |
19.18 |
22.19 |
0.77 |
|
Rich |
541 |
13.58 |
12.31 |
14.96 |
0.67 |
|
Richer |
250 |
7.33 |
6.3 |
8.52 |
0.57 |
95 % CI: confidence interval, LL: lower limit, UL: upper limit, SE: standard error
Factors associated with receiving Vaso de Leche and Qali Warma
In the bivariate analysis, all variables were associated with being a beneficiary of the VDL program, except for sex. The proportion of beneficiaries among boys (34.42 %) and girls (33.82 %) did not show significant differences (Table 3).
On the other hand, the variables associated with being a beneficiary of the QW program were term birth (p < 0.005), cesarean birth (p < 0.001), chronic malnutrition of child (p < 0.001), maternal age (p = 0.021), maternal educational level (p < 0.001), maternal ethnicity (p < 0.001), area of residence (p < 0.001) and wealth index (p < 0.001) (Table 3).
Table 3. Bivariate analysis of factors associated with receiving VDL and QW in children under five years of age in 2022
|
Proportion of children who are beneficiaries of VDL
|
Proportion of children who are beneficiaries of QW
|
|
||||||||
|
Variable |
n |
Weighted proportion (%) |
LL |
UL |
p valuea |
n |
Weighted proportion (%) |
LL |
UL |
p valuea |
|
Characteristics of the child |
|
|
|
|
|
|
|
|
|
|
|
Sex of the child |
|
|
|
|
|
|
|
|
|
|
|
Male |
11,328 |
34.42 |
33.37 |
35.48 |
0.44 |
2,487 |
85.07 |
82.99 |
86.94 |
0.231 |
|
Female |
10,585 |
33.82 |
32.74 |
34.92 |
|
2,383 |
86.72 |
84.77 |
88.47 |
|
|
Age (months) |
|
|
|
|
|
|
|
|
|
|
|
0–11 |
4,024 |
31.21 |
29.52 |
32.97 |
0.001 |
0 |
0 |
0 |
0 |
|
|
12–35 |
8,589 |
35.15 |
33.93 |
36.38 |
|
0 |
0 |
0 |
0 |
|
|
36–59 |
9,300 |
34.44 |
33.28 |
35.61 |
|
4,870 |
85.88 |
84.47 |
87.18 |
|
|
Prenatal checkups |
|
|
|
|
|
|
|
|
|
|
|
No checkups |
177 |
39.92 |
31.63 |
48.62 |
0.033 |
14 |
74.2 |
37.83 |
93.15 |
0.224 |
|
1–5 |
2,255 |
30.78 |
28.52 |
33.12 |
|
298 |
82.16 |
75.19 |
87.49 |
|
|
6 and older |
15,889 |
33.37 |
32.27 |
33.91 |
|
3,172 |
86.28 |
84.55 |
87.84 |
|
|
Term birth |
|
|
|
|
|
|
|
|
|
|
|
No |
3,857 |
28.54 |
26.88 |
30.27 |
< 0.001 |
760 |
81.43 |
77.42 |
84.86 |
0.005 |
|
Yes |
16,985 |
35.41 |
34.54 |
36.28 |
|
3,810 |
86.7 |
85.12 |
88.13 |
|
|
Cesarean birth |
|
|
|
|
|
|
|
|
|
|
|
No |
13,976 |
39.16 |
38.18 |
40.14 |
< 0.001 |
3,219 |
87.86 |
86.19 |
89.35 |
< 0.001 |
|
Yes |
6,866 |
24.73 |
23.51 |
25.99 |
|
1,351 |
81.09 |
78.08 |
83.77 |
|
|
Birth weight (g) |
|
|
|
|
|
|
|
|
|
|
|
< 2500 |
1,261 |
38.8 |
35.63 |
42.06 |
< 0.001 |
286 |
83.17 |
76.53 |
88.22 |
0.68 |
|
2,500–3,999 |
17,406 |
33.91 |
33.07 |
34.77 |
|
3,801 |
85.63 |
84 |
87.11 |
|
|
4,000 |
1,578 |
24.74 |
22.22 |
27.46 |
|
365 |
86.21 |
80.28 |
90.57 |
|
|
Chronic malnutrition in children |
|
|
|
|
|
|
|
|
|
|
|
No |
6,580 |
31.87 |
31.07 |
32.67 |
< 0.001 |
3,883 |
85.16 |
83.62 |
86.57 |
< 0.001 |
|
Yes |
1,504 |
51.18 |
48.98 |
53.37 |
|
476 |
92.17 |
88.88 |
94.54 |
|
|
Characteristics of the mother |
|
|
|
|
|
|
|
|
|
|
|
Maternal age (years) |
|
|
|
|
|
|
|
|
|
|
|
15–24 |
4,549 |
37.35 |
35.68 |
39.06 |
< 0.001 |
767 |
89.6 |
86.22 |
92.23 |
0.021 |
|
25–34 |
10,157 |
33.58 |
32.47 |
34.7 |
|
2,267 |
85.81 |
83.69 |
87.69 |
|
|
35–49 |
6,126 |
32.28 |
30.89 |
33.71 |
|
1,536 |
83.64 |
80.9 |
86.05 |
|
|
Maternal educational level |
|
|
|
|
|
|
|
|
|
|
|
No education –primary |
3,925 |
53.88 |
52.05 |
55.69 |
< 0.001 |
996 |
91.99 |
89.58 |
93.89 |
< 0.001 |
|
Secondary |
10,325 |
37.51 |
36.38 |
38.66 |
|
2,269 |
85.76 |
83.61 |
87.68 |
|
|
Higher |
6,592 |
18.07 |
16.96 |
19.25 |
|
1,305 |
80.76 |
77.59 |
83.57 |
|
|
Maternal union |
|
|
|
|
|
|
|
|
|
|
|
Not in a union |
3,566 |
31.29 |
29.47 |
33.17 |
0.002 |
851 |
85.68 |
82.02 |
88.69 |
0.991 |
|
In a union |
17,276 |
34.57 |
33.72 |
35.43 |
|
3,719 |
85.7 |
84.07 |
87.19 |
|
|
Mother ethnicity |
|
|
|
|
|
|
|
|
|
|
|
No |
18,624 |
31.66 |
30.85 |
32.47 |
< 0.001 |
4,039 |
84.75 |
83.15 |
86.21 |
< 0.001 |
|
Yes |
2,218 |
61.28 |
58.87 |
63.64 |
|
531 |
95.26 |
92.3 |
97.12 |
|
|
Mother currently working |
|
|
|
|
|
|
|
|
|
|
|
No |
9,236 |
36.14 |
34.95 |
37.34 |
< 0.001 |
1,700 |
87 |
84.65 |
89.03 |
0.156 |
|
Yes |
11,606 |
32.33 |
31.32 |
33.36 |
|
2,870 |
84.9 |
82.97 |
86.65 |
|
|
Characteristics of the household |
|
|
|
|
|
|
|
|
|
|
|
Area of residence |
|
|
|
|
|
|
|
|
|
|
|
Urban |
14,741 |
22.5 |
21.68 |
23.35 |
< 0.001 |
3,132 |
81.22 |
79.28 |
83.03 |
< 0.001 |
|
Rural |
7,172 |
64.61 |
63.31 |
65.89 |
|
1,738 |
95.84 |
94.6 |
96.81 |
|
|
Wealth index |
|
|
|
|
|
|
|
|
|
|
|
Poorer |
6,532 |
62.66 |
61.27 |
64.03 |
< 0.001 |
1,527 |
94.23 |
92.57 |
95.53 |
< 0.001 |
|
Poor |
5,625 |
36.33 |
34.79 |
37.91 |
|
1,329 |
86.85 |
84.14 |
89.15 |
|
|
Middle class |
4,087 |
23.59 |
21.97 |
25.29 |
|
923 |
81.85 |
78.09 |
85.1 |
|
|
Rich |
2,808 |
13.59 |
12.1 |
15.23 |
|
541 |
78.21 |
73.37 |
82.38 |
|
|
Richer |
1,790 |
8.35 |
6.84 |
10.15 |
|
250 |
70.34 |
62.21 |
77.35 |
|
95 %: CI confidence interval, LL: lower limit, UL: upper limit.
a Pearson’s chi square test with second-order Rao-Scott correction.
Variables that reached a p value less than 0.20 entered the log-binomial regression model. Among children aged five years or younger with a valid response to the question about the VDL program we found that those with a birth weight < 2,500 g had a 22 % higher probability of being beneficiaries of the program compared to those with a birth weight ≥ 4,000 g (95 % CI: 1.07-1.38). Additionally, we found that children from households with the “poorer” wealth index had 4.28 times the probability of being beneficiaries of the VDL program compared to children from households with the “richer” wealth index (95 % CI: 3.46-5.30). Children living in rural areas had a 66% higher probability of being beneficiaries of the VDL program (95 % CI: 1.57-1.75) (Table 4).
In the group of children under five years of age with a valid response to the QW program, the independently associated factors were residing in rural areas (aPR = 1.12, 95 % CI: 1.08-1.16) and belonging to the poorer (PR = 1.20, 95 % CI: 1.07-1.36) and poor (aPR = 1.19, 95 % CI: 1.0 6-1.34) quintiles compared to the richer.
Table 4. Log binomial regression analysis to evaluate the association between being a beneficiary of supplementary feeding programs and variables to characterize coverage in children under five years of age
|
|
Being a beneficiary of the VDL program a
|
Being a beneficiary of the QW program b
|
|
|||||
|
Variables |
aPR |
95
% CI |
|
p value |
aPR |
95 % CI |
|
p value |
|
Characteristics of the children |
|
|
|
|
|
|
|
|
|
Age (months) |
|
|
|
|
|
|
|
|
|
0–11 |
1 |
|
|
|
|
|
|
|
|
12–35 |
1.15 |
1.08 |
1.23 |
< 0.001 |
|
|
|
|
|
36–59 |
1.15 |
1.08 |
1.22 |
< 0.001 |
- |
|
- |
|
|
Birth weight (g) |
|
|
|
|
|
|
|
|
|
< 2,500 |
1.22 |
1.07 |
1.38 |
0.002 |
- |
|
- |
|
|
2,500 - 3,999 |
1.18 |
1.07 |
1.31 |
0.002 |
|
|
|
|
|
≥ 4,000 |
1 |
|
|
|
- |
- |
|
- |
|
Chronic malnutrition |
|
|
|
|
|
|
|
|
|
No |
1 |
|
|
|
1 |
|
|
|
|
Yes |
1.02 |
0.97 |
1.07 |
0.492 |
1.01 |
0.97 |
1.05 |
0.712 |
|
Characteristics of the mother |
|
|
|
|
|
|
|
|
|
Maternal age (years) |
|
|
|
|
|
|
|
|
|
15–24 |
1 |
|
|
|
1 |
|
|
|
|
25–34 |
1.1 |
1.04 |
1.16 |
0.001 |
0.98 |
0.94 |
1.02 |
0.384 |
|
35–49 |
1.08 |
1.02 |
1.15 |
0.014 |
0.96 |
0.92 |
1.01 |
0.108 |
|
Educational level |
|
|
|
|
|
|
|
|
|
No education– primary |
1.25 |
1.15 |
1.35 |
< 0.001 |
1.01 |
0.96 |
1.07 |
0.601 |
|
Secondary |
1.27 |
1.18 |
1.36 |
< 0.001 |
0.99 |
0.94 |
1.04 |
0.658 |
|
Higher |
1 |
|
|
|
1 |
|
|
|
|
Maternal union |
|
|
|
|
|
|
|
|
|
In a union |
1 |
|
|
|
- |
- |
- |
- |
|
Not in a union |
1.02 |
0.96 |
1.08 |
0.603 |
- |
- |
- |
- |
|
Maternal ethnicity |
|
|
|
|
|
|
|
|
|
No |
1 |
|
|
|
1 |
|
|
|
|
Yes |
1.04 |
1 |
1.09 |
0.077 |
1.01 |
0.98 |
1.05 |
0.402 |
|
Mother currently working |
|
|
|
|
|
|
|
|
|
No |
1.02 |
0.97 |
1.08 |
0.451 |
1.02 |
0.99 |
1.05 |
0.239 |
|
Yes |
1 |
|
|
|
1 |
|
|
|
|
Characteristics of the household |
|
|
|
|
|
|
|
|
|
Area of residence |
|
|
|
|
|
|
|
|
|
Urban |
1 |
|
|
|
1 |
|
|
|
|
Rural |
1.66 |
1.57 |
1.75 |
< 0.001 |
1.12 |
1.08 |
1.16 |
< 0.001 |
|
Wealth index |
|
|
|
|
|
|
|
|
|
Poorer |
4.28 |
3.46 |
5.3 |
< 0.001 |
1.2 |
1.07 |
1.36 |
0.002 |
|
Poor |
3.37 |
2.74 |
4.16 |
< 0.001 |
1.19 |
1.06 |
1.34 |
0.003 |
|
Middle class |
2.51 |
2.03 |
3.11 |
< 0.001 |
1.15 |
1.03 |
1.3 |
0.018 |
|
Rich |
1.54 |
1.23 |
1.94 |
0.001 |
1.11 |
0.98 |
1.25 |
0.101 |
|
Richer |
1 |
|
|
|
1 |
|
|
|
95 % CI: confidence interval, LL: lower limit, UL: upper limit, aPR: adjusted prevalence ratio.
a Analysis of the VDL program adjusted to the following variables: child age, birth weight, chronic malnutrition, maternal age, maternal
educational level, maternal union, maternal ethnicity, mother currently working, area of residence and wealth index.
b Analysis of the QW adjusted for the following variables: chronic malnutrition, maternal age, maternal educational level, maternal ethnicity, mother currently working, area of residence and wealth index.
The standard error was less than (SE < 1.0) in the analysis. The statement applies to both models.
DISCUSSION
Among children under five years of age participating in the 2022 Peruvian ENDES, the estimated coverage of the QW program was 85.88 %, much higher than that found for the VDL program (34.13 %). This finding is explained by the different eligibility criteria for each program. The QW program distributes breakfasts and/or lunches to the public schools (3,26) across the country. Therefore, only children who attend public schools receive the rations, and this generates a high level of selectivity (27). On the other hand, the VDL program delivers one daily ration, through the municipalities, only to those families that belong to a poor and very poor wealth quintiles (12).
A noteworthy finding is that 13.59 % of children from families that, according to the wealth index, were “rich” and 8.35 % from “richer” families were beneficiaries of the VDL program. This result is consistent with reports from the ENAHO, which suggest that the VDL program may be reaching a population that is not considered poor (28,29). In fact, in 2022, out of 993,798 beneficiary households, 58.8 % were not in poverty (29). Despite the shortcomings the VDL program faces in defining its intended population, we found that, independently of other factors, belonging to the poorest wealth quintiles was progressively associated with a higher probability of being a beneficiary. A similar dose-response effect was observed in the QW program, although with weaker associations.
Another aspect that helps explain why there are beneficiaries of the VDL program from the rich or very rich quintiles is that its rations are distributed to municipalities, which are responsible for identifying those in poverty. For this purpose, municipalities rely on the information provided by the Sistema de Focalización de Hogares (Sisfoh - Household Targeting System), whose role is to manage the Padrón General de Hogares (PGH - General Household Registry) based on socioeconomic status (9). In addition, the validity of the socioeconomic classification can last between four and eight years, depending on the area of residence (31), which may affect the updating of data and, therefore, the adequate targeting of the program.
At the departmental level, children under five years of age have variable access to VDL and QW. For VDL, the departments that achieved coverage of at least 60 % were Amazonas, Cajamarca, Callao and Huancavelica. For QW, several departments achieved coverage above 95 %, for example, Amazonas, Ica, Loreto, Moquegua, Pasco, San Martin and Tumbes. The coverages of both programs were not linearly correlated at the ecological level, which evidenced their differences in eligibility criteria, logistical limitations, as well as implementation and administration mechanisms, both at the national and departmental levels.
We found that, among children under five years of age with chronic malnutrition, 51.18 % were beneficiaries of VDL, while among children aged three to five years with chronic malnutrition, 92.17 % were beneficiaries of QW. The analysis adjusted to evaluate whether there was an association between chronic malnutrition in children and being a beneficiary of these programs did not reveal statistical significance. The explanation for this finding is based on the fact that one of the main causes of chronic malnutrition in children is food insecurity, which is the result of other variables, such as wealth index, area of residence and parental education (31). Therefore, controlling for these factors, the initially observed association between chronic malnutrition in children and being a beneficiary of VDL and QW disappears. The occurrence of chronic malnutrition in children before becoming beneficiaries in any of the analyzed programs is also plausible.
We also found that being a beneficiary and living in rural areas had a stronger association for the VDL program than for QW (aPR = 1.66 vs. aPR = 1.12). This is explained by the fact that the area of residence -rural or urban-is associated with the household income. While VDL is distributed according to the poverty status of the target population, QW covers public schools located in the districts of each department, regardless of the socioeconomic context of each sector in which they are located.
Children of mothers with no education or only primary education had a 25 % higher probability of being beneficiaries of VDL compared to those with mothers with higher education; this association was not observed in QW. Although both programs aim to provide supplementary feeding to vulnerable populations, VDL specifically targets those living in poverty. Higher educational attainment protects women from socioeconomic conditions characterized by low income and, in the long term, reduces the probability of belonging to lower wealth quintiles, which leads to a lower probability of being part of the target population of the VDL program.
The study has limitations. First, being a study based on secondary data sources relying on mothers’ reports, it is susceptible to social desirability bias. Second, due to the cross-sectional design, it is not possible to establish a causal relationship between the variables studied and being a beneficiary of these social programs; moreover, this study was not designed to assess the impact of these programs on chronic malnutrition in children. Third, there may be covariates of interest that were not included in our analysis and that influence beneficiary status under real-life conditions. Fourth, being a cross-sectional study, it does not evaluate beneficiary status over the course of the year, during which there may be variations in this status. Finally, our study is limited to a population of children aged five years and under; consequently, the results cannot be extrapolated to the entire beneficiary population of VDL and QW.
Despite the limitations described, the strength of this study was to review the literature related to the research topic (20-23) to guide the analysis of the variables associated with being a beneficiary of these programs. Also, it is a first attempt to describe, using data from a population-based survey, the proportion of preschool children who received assistance from VDL and QW.
In conclusion, among children under five years of age, belonging to the poorer and poor wealth quintiles is one of the characteristics most strongly associated with being a beneficiary of the VDL and QW programs. However, the children who receive support from the VDL program are not necessarily in socioeconomically vulnerable conditions, which suggests issues in the implementation of its objectives. When considering as the denominator the children of mothers who provided valid responses to the questions about receiving assistance from the VDL and QW, both programs showed markedly different national coverage rates, with considerable heterogeneity at the departmental level.
BIBLIOGRAPHIC REFERENCES
1.MINAGRI. Plan nacional de seguridad alimentaria y nutricional 2015-2021 [Internet]. Lima: COMSAN; 2015. Available from: https:// www.midagri.gob.pe/portal/download/pdf/seguridad-alimentaria/ planacional-seguridad-2015-2021.pdf
2.MEF. Programa Vaso de Leche [Internet]. Lima: MEF. Available from: https://www.mef.gob.pe/es/?option=com_content&language=es- ES&Itemid=100964&lang=es-ES&view=article&id=448
3.MIDIS. Decreto Supremo No 008-2012-MIDIS [Internet]. Lima: Poder Ejecutivo; 2012. Available from: https://cdn.www.gob.pe/uploads/ document/file/3058323 /CREAN% 20 EL% 20 PROGRAMA% 20 NACIONAL%20DE%20ALIMENTACI%C3%93N%20ESCOLAR%20 QALI%20WARMA%20Y%20MODIFICATORIAS.pdf.pdf?v=1651609309
4.Boles RE, Johnson SL, Burdell A, Davies PL, Gavin WJ, Bellows LL. Home food availability and child intake among rural families identified to be at-risk for health disparities. Appetite [Internet]. 2019;134:135-41.
5.Rasmussen RA, Sisson SB, Campbell JE, DeGrace B, Baldwin JD. Home food access and children's heart healthy dietary intake at home and child care. Nutr Health [Internet]. 2022.
6.Dubois L, Bédard B, Goulet D, Prud'homme D, Tremblay RE, Boivin M. Experiencing food insecurity in childhood: influences on eating habits and body weight in young adulthood. Public Health Nutr [Internet]. 2023;26(11):2396-406.
7.McClain AC, Cory H, Mattei J. Childhood food insufficiency and adulthood cardiometabolic health conditions among a population- based sample of older adults in Puerto Rico. SSM Popul Health [Internet]. 2022;17:101066.
8.MINAGRI. Estrategia nacional de Seguridad alimentaria y nutricional 2013-2021 [Internet]. Lima: COMSAN; 2013. Available from: https:// www.midagri.gob.pe/portal/download/pdf/seguridad-alimentaria/ estrategia-nacional-2013-2021.pdf
9.La Contraloría General de la República del Perú. Informe anual del programa del vaso de leche [Internet]. Lima: Contraloría General de la República del Perú; 2021. Available from: https://cdn.www. gob.pe/uploads/document/file/4065593/INFORME%20PVL%20 N%C2%B0004.2022-CG_SOCC.pdf.pdf?v=1674603395
10.INEI. Programas sociales, autoidentificación étnica y discapacidad [Internet]. Lima: INEI; 2022. Available from: https://www.inei.gob.pe/ media/MenuRecursivo/publicaciones_digitales/Est/Lib1855/cap10.pdf
11.INEI. Indicadores de Resultados de los programas presupuestales 2021 [Internet]. Lima: INEI; 2022. Available from: http://proyecto.inei. gob.pe/enapres/wp-content/uploads/2022/05/ENAPRESIndicadores- de-Programas-Presupuestales-2021.pdf
12.Suárez Bustamante MA. Caracterización del programa del vaso de leche [Internet]. Lima: MEF; 2003. Available from: https://www.mef. gob.pe/contenidos/pol_econ/documentos/carac_vaso.pdf
13.Neira Alberca W. Evaluación del Programa de Vaso de Leche en municipalidad provincial de San Ignacio - 2018 [Undergraduate thesis]. Pimentel: Universidad Señor de Sipán; 2018. Available from: https://repositorio.uss.edu.pe/bitstream/handle/20.500.12802/6282/ Neira%20Alberca%2C%20Wilfredo.pdf?sequence=1&isAllowed=y
14.Campos Chong TM. El Programa Qali Warma y la gestión en la distribución de alimentos en el distrito de Ccorca, provincia de Cusco [Undregraduate thesis]. Lima: Pontificia universidad católica del Perú; 2017. Available from: https://tesis.pucp.edu.pe/repositorio/bitstream/ handle/20.500.12404/18913/Campos_Chong_Programa_Qali%20 Warma_gesti%C3%B3n1.pdf?sequence=4&isAllowed=y
15.Saavedra Seminario PA. Efectos del programa Qali Warma sobre la salud y la educación de los niños [Undergraduate thesis]. Piura: Universidad de Piura; 2021. Available from: https://pirhua.udep.edu.pe/bitstream/ handle/11042/5246/ECO_2102.pdf?%0Csequence=1&isAllowed=y
16.MINSA. Norma técnica de salud para la atención integral de salud materna [Internet]. Lima: MISA; 2023. Available from: https://docs. bvsalud.org/biblioref/2019/04/964549/rm_827-2013-minsa.pdf
17.INEI. Manual de la Antropometrista [Internet]. Lima: Encuesta demográfica y de salud familiar; 2022. Available from: https://proyectos. inei.gob.pe/iinei/srienaho/Descarga/DocumentosMetodologicos/2020-5/ ManualAntropometrista.pdf
18.INEI. Manual de la Entrevistadora [Internet]. Lima: Encuesta Demográfica y de salud familiar; 2022. Available from: https://proyectos.inei.gob. pe/iinei/srienaho/Descarga/DocumentosMetodologicos/2022-5/ ManualEntrevistadora.pdf
19.MINSA. Norma técnica de salud para el control del crecimiento y desarrollo de la niña y el niño menor de cinco años [Internet]. Lima: MINSA; 2011. Available from: http://www.diresacusco.gob. pe/salud_individual/normas/NORMA%20TECNICA%20D%20%20CRECIMIENTO%20Y%20DESARROLLO%20DEL%20 %20NI%C3%91O%20MENOR%20%20DE%20%20CINCO%20 A%C3%91OS.pdf
20.Kumar A, Suar D, Sahoo BK. What attributes characterize the beneficiary households after three years of implementation of National Food Security Act? J Public Aff [Internet]. 2021;22(1):e2718.
21.Zarsuelo Ma-Ann, Suva MM, Juanico CB, Hurtada WA. Household characteristics, housing profile and diet diversity of pantawid pamilyang pilipino program (4ps) beneficiaries and non-beneficiaries in Lucena City, Quezon, Philippines. Acta Med Philipp [Internet]. 2018;52(5):447-52.
22.Ghodsi D, Omidvar N, Rashidian A, Eini-Zinab H, Raghfar H, Aghayan M. Effectiveness of the national food supplementary program on children growth and nutritional status in Iran. Matern Child Nutr [Internet]. 2018;14(3):e12591.
23.Frank DA, Neault NB, Skalicky A, Cook JT, Wilson JD, Levenson S, et al. Heat or Eat: The low income home energy assistance program and nutritional and health risks among children less than 3 years of age. Pediatrics [Internet]. 2006;118(5):1293-302.
24.Williamson T, Eliasziw M, Hilton Fick G. Log-binomial models: exploring failed convergence. Emerg Themes Epidemiol [Internet]. 2013;10(1):14.
25.Chan YH. Biostatistics 202: logistic regression analysis. Singapore Med J [Internet]. 2004;45(4):149-5
26.MIDIS. Resolución directoral ejecutiva N.° D000159- 2021-MIDIS/PNAEQW-DE [Internet]. Lima: MIDIS; 2021. Available from: https://info.qaliwarma.gob.pe/normatividad/ export/?id=TmZadld0K0Z0ZTBsbTI5U3lQcjZKdz09
27.GRADE. Investigación para el desarrollo en el Perú [Internet]. Lima: once balances; 2016. Available from: https://www.grade.org.pe/ wpcontent/uploads/LIBROGRADE_DESARROLLO35.pdf
28.ComexPerú. Cinco de cada diez hogares que se benefician de vaso de leche, no deberían [Internet]. Lima: ComexPerú; 2021. Available from: https://www.comexperu.org.pe/articulo/cinco-de-cada-diezhogares- que-se-benefician-del-programa-de-vaso-de-leche-nodeberian
29.ComexPerú. Vaso de leche: El 55.8% de los hogares que se benefició de este programa no se encontraba en situación de pobreza [Internet]. Lima: ComexPerú; 2023. Available from: https://www.comexperu.org. pe/articulo/vaso-de-leche-el-558-de-los-hogaresque-se-beneficio-de- este-programa-no-se-encontraba-en-situacionde-pobreza
30.MIDIS. Directiva No 001-2020-MIDIS [Internet]. Lima: MIDIS; 2020. Available from: https://www.midis.gob.pe/sello_municipal/wpcontent/ uploads/2023/03/P7_1_Directiva_Sisfoh_2020.pdf
31.Militao EMA, Salvador EM, Uthman OA, Vinberg S, Macassa G. Food insecurity and health outcomes other than malnutrition in southern Africa: A descriptive systematic review. Int J Environ Res Public Health. 2022;19(9):5082.
Corresponding author:
Paloma Rodríguez-Massa paloma.rodriguez@alum.udep.edu.pe
Received: March 5, 2024
Reviewed: March 23, 2024
Accepted: April 2, 2024
Author contributions: ALCB, PRM and FRR conceptualized the study, formulated the methodological design, performed the formal data analysis, conducted the research, carried out the data curation, wrote the first draft of the article, reviewed and approved the final version that was submitted, and managed the project. All authors assume responsibility for the published work.
Funding sources: This article was funded by the authors.
Conflicts of interest: The authors declare no conflicts of interests.