Efficacy of an intervention in the nutritional status and consumption of ultra-processed foods in children with obesity treated in primary health care in Brazil

Abstract

Introduction

Effectivechildhood obesity management requires multicomponent interventions in Primary Health Care (PHC), engaging families, communities, and healthcare professionals to foster healthier behaviors.

Objective

This study evaluated the effect of a controlled intervention on the nutritional status and ultra-processed food (UPF) consumption of children with obesity receiving PHC services.

Methods

A randomized clinical trial was conducted with children from PHC units in Betim, Minas Gerais, Brazil. Twenty units were randomly selected and allocated into control (CG) and intervention groups (IG). Children aged 6–10 years with obesity (z-score ≥ +2 for BMI/age) were included. The estimated sample size was 47 per group, considering 20% loss. The IG received monthly activities with four weekly sessions (≥26 contact hours) and five consultations. The CG followed a similar protocol with up to 9 contact hours. Sociodemographic, nutritional, and UPF consumption data were collected through a questionnaire covering 11 food groups. Intention-to-treat analysis compared mean UPF consumption, BMI/age z-score, and BMI (kg/m²) within and between groups post-intervention for a period of 9 months.

Results

A total of 727 eligible children were identified and 167 were randomized, with 79 assigned to the control group and 88 to the intervention group. Of these, 45 children in the control group and 49 in the intervention group. The intervention significantly reduced UPF consumption in IG at visit 3 [95% CI: -0.95 (-1.87; -0.04)], visit 4 [95% CI: -1.35 (-2.215; -0.49)], and visit 5 [95% CI: -0.94 (-1.83;-0.05)]. No significant differences were found for BMI or BMI/age z-score (p>0.05).

Conclusion

The intervention effectively reduced UPF consumption. Although no BMI/age reduction was observed, it contributed to weight maintenance and should be considered in PHC childhood obesity strategies.

Trial registration

Clinical Trials NCT05966259 11/09/2023.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40795-025-01221-z.

Introduction

Obesity is a multi-causal disease that involves biological, historical, ecological, political, socioeconomic, psychosocial, and cultural factors [1–3]. The increase in obesity among children and adolescents is one of the main worldwide public health concerns in the present [4, 5]. It is estimated that between 1990 and 2022, the global prevalence of obesity among children and adolescents increased from 1.7% to 6.9% in girls, and from 2.1% to 9.3% in boys, which corresponds to 65.1 million girls and 94.2 million boys with obesity identified in 2022 [5].

In Brazil, based on data from the Food and Nutrition Surveillance System (SISVAN) [6], the prevalence of obesity in children between 5 and 9 years old is 8.87%, which corresponds to more than 486 thousand children. In the long term, obesity can impact the growth and well-being of children, in addition to increasing the risk of remaining obese and chronic noncommunicable diseases (NCD) in adulthood, increasing health spending [7–10]. It is estimated that in a decade (2013–2022), spending on obesity among children and adolescents in the Brazilian public health system (SUS) reached BRL 1.54 billion [11].

Among the various factors related to obesity, it is known that the environment in which the child is inserted can alter the ways of life, making them more or less healthy, which can impact the health and disease processes of this population [12]. Considering the multiple factors associated with nutritional status, in the individual context, food consumption based on ultra-processed foods (UPF) has been associated with early mortality and more than thirty health problems, as well as an increase in obesity in both children and adolescents, and in adults [5, 13].

To treat childhood obesity effectively, it is essential to implement interventions that change eating behaviors and practices, involving the family and the community, due to this problem’s complexity [14]. Multicomponent interventions, which reach the child, their caregivers and the community, are more effective than those that focus on a single aspect [15, 16]. The duration of contact with health topics is also crucial, with interventions rated as low (≤ 25 h in 6 months) or high intensity (>25 h in 6 months) [17]. The participation of health professionals and the coordination of these actions are fundamental for the success in the treatment of childhood obesity [18, 19].

The management of childhood obesity is one of the priority themes of the Brazilian agenda of food, nutrition and health promotion policies of the Unified Health System (SUS, in Portuguese). This theme is one of the axes of strategic actions for the promotion and comprehensive health care, through the elaboration of clinical protocols for the management, professional qualification, and implementation of effective measures for the prevention and control of childhood obesity in Primary Health Care (PHC) [20].

In this sense, the ‘Instruction for the care of overweight and obese children and adolescents within the scope of Primary Health Care’ [20] recommends actions with multiple components that include the family and strategies for behavioral changes and interventions in the environment and food for the management of childhood obesity [20]. Thus, the treatment and prevention of childhood obesity can be carried out in the Health Centers, within the context of PHC in order to integrate actions at the individual, community, and at the household and family context of children who are users of SUS. However, to date, little is known about the management of childhood obesity using approaches with multiple components within the SUS. Thus, this study’s objective was to evaluate the effect of a controlled, multi-component intervention on UPF consumption and the nutritional status of children with obesity treated in PHC.

Methods

Study design and location

This is a randomized clinical trial, conducted with a sample of children with obesity, treated at public health centers in Betim, Minas Gerais, Brazil. The municipality of Betim has a territorial area of 344.06 km ² and is located in the central region of the state of Minas Gerais, and is part of the Metropolitan region of Belo Horizonte [21]. In 2010 (IBGE), it had a Municipal Human Development Index (MHDI) of 0.749, being part of the group of cities with a high level of development (MHDI between 0.700 and 0.799) [22]. According to data from the IBGE demographic census, in 2022 Betim had a resident population of 411,846 inhabitants [21].

According to the Municipal Health Plan 2018–2021, what is proposed in Betim is PHC, from the perspective of the Health Care Networks (HCN), which is capable of fulfilling the resolutive function of attending to approximately 80% of the most common health problems; the ordering function of coordinating the flows and counter-flows of users, products, and information in the networks, as well as the accountability function for the health of the user population that is registered [23]. The municipality has 37 health centers, locally named UBS (or Unidade Básica de Saúde, in Portuguese), distributed in nine administrative regions with an estimated maximum service capacity of 350 users per unit, according to information available on the website of the City Hall of Betim [24].

The research was approved by the research ethics committee of the Federal University of Minas Gerais, under number CAAE 39508720.6.0000.5149, and funded by the Ministry of Health (CNPq/DEPROS/SAPS No. 27/2020, Research axis on Noncommunicable Diseases and Associated Risk Factors) and with financial assistance from Fundación MAPFRE (2020). Registration: Clinical Trials NCT05966259 11/09/2023. This manuscript is reported according to the CONSORT guidelines.

Health center sampling

The 37 public health centers in Betim were categorized according to the tertiles of per capita income of the neighborhood where they were located, according to income data from the Brazilian Institute of Geography and Statistics [25]. The mean income tertile was categorized as follows: 1 st tertile: per capita income >BRL 440.60; 2nd tertile: income between BRL 440.60 and BRL 600.90; and 3rd tertile: income of >BRL 600.90. The first tertile had two health centers, in the second tertile there were 14 health centers, and in the third tertile 21 (n = 37 public health centers). A sample of 20 public health centers was randomly selected and considered sufficient and representative of the 37 UBS of the municipality. The random sequence was generated by drawing lots among the selected health centers, which were then randomly allocated to either the control or the intervention group, with no prior selection or influence on the allocation, using a table of random numbers. Such information is described in Table S1 of the supplementary material.

Study participants

The sampling process of the individuals was carried out based on the information of schoolchildren between 6 and 10 years of age, based on what was established by the Brazilian Law No. 12.796, of April 4, 2013 [26] and according to criteria of the World Health Organization (WHO) for obesity [27], followed in PHC in 2019, who had weight and height evaluated according to a public report by SISVAN of the municipality of Betim [6].

To estimate the sample size, we adopted the calculation that considers the difference between two means of independent groups with the aid of the Stata software version 17.0. This calculation considered the significance level of 5%, test power of 80%, and the mean differences in BMI by age found at the end of the study, in relation to the baseline, in the intervention (−0.195 ± 0.37) and control (−0.01 ± 0.18) groups in the study by Lison et al. [28]. Thus, the sample size calculated for each group was 39 and, considering the estimated percentage of losses of 20% [29], the estimated final sample size was of at least 47 participants for the control group (CG) and 47 for the intervention group (IG).

Initially, for placing the participants between CG and IG, the distribution was done base on the previous categorization of the child’s reference public health center, in order to respect the proximity of the individual’s home to the health center.

Randomization took place at the level of the public health center; children who belonged to health centers that were ‘control’ were allocated to the CG, while children enlisted in ‘intervention’ health centers were allocated to the IG.

Recruitment of children: inclusion and exclusion criteria

The dissemination of the study happened through posters displayed in all 37 health centers in Betim and also through advertising on television, radio, Instagram, WhatsApp, and newspapers in the municipality, in addition to the social media profiles of the City Hall of Betim and of the Federal University of Minas Gerais. In addition, the researchers were provided with a list of possible eligible children, from the registration in the Unified Health System of Betim, containing telephone contact information of parents/caregivers, address, and reference health center. Thus, the children were recruited in the form of active search and also by free demand carried out in the public health centers at the beginning of the study.

Children with obesity (z-score ≥ + 2 for BMI/Age) [30, 31] who were between 6 and 10 years of age were included. All participants and their legal guardians were informed about the study and agreed to participate by signing the Informed Consent Form (ICF). Consent for participation was specifically obtained from the parents or legal guardians of all children involved in the study. All procedures were conducted in accordance with ethical standards. Exclusion criteria were children with mental disorders that made it impossible to participate in consultations/groups, who were taking weight loss medications and/or had comorbidities associated with obesity; and those whose parents/caregivers did not agree to participate and/or did not sign the ICF.

Material construction and training

The manual and materials used in the intervention were developed for the management of childhood obesity in the context of PHC/SUS. For developing the materials, we considered activities related to individual care conducted by registered dietitians at the public health centers, the activities of Food and Nutrition Education (FNE) stimulated at home (booklet) and carried out at the health centers (group), and the monitoring of activities via messaging applications (messaging on WhatsApp). These activities were developed, adapted, and described considering the context of PHC, having as references the Dietary Guidelines for the Brazilian Population [32] and the Instruction for the care of children and adolescents with overweight and obesity within the scope of Primary Health Care [20].

All materials available online, such as the Dietary Guidelines for the Brazilian Population [32]and other materials to promote activities (applications, games) were added to the manual through a clickable link. Training with registered dietitians was carried out before the beginning of the study, through a lecture in which the manual was detailed, and information on how to use it for individual and group consultations was provided. There was always a research coordinator monitoring the actions throughout the study process in order to answer possible questions.

Intensive intervention with multiple components: intervention group (IG) and control group (CG)

As an alternative to specific nutritional actions, studies suggest that, for the management of childhood obesity, intensive interventions are necessary, based on multiple components in order to assist in the understanding of the multicausal relationships of obesity [12, 19, 20, 33]. Facing that, for the IG, the intensive intervention was planned based on multiple components based on the theoretical model for child nutritional status proposed by Davison and Birch [12], who consider that the nutritional status of children results from the complex interaction between individual, family and social factors, and in the systematic review conducted by O’Connor et al. [19], who evaluated the benefits and harms of screening and treatment of obesity and overweight in children and adolescents [19]. A total of 26 h of contact were planned for the IG and 9 h for the CG, for a period of 9 months [19].

For the IG, the monthly activities were composed of four weekly contacts (once a week) that comprised individual care, FNE at home, group FNE at the public health centers (parents/caregivers and children) and telephone/SMS/WhatsApp monitoring. These interventions were based on multiple components, which included five themes: eating, physical activity, sedentary behavior, sleep, and mental health, as recommended by the Instruction for the care of children and adolescents with overweight and obesity within the scope of Primary Health Care [20]. In the latter theme, only subjective and perception aspects (perception of hunger, satiety, and emotions) were considered, which were addressed during all contacts with children, according to the recommendations of the Ministry of Health [20].

For the treatment of the CG, initially the activities followed those carried out within the PHC of the municipality, however, when verifying that there was no well-established protocol for the care of children with obesity, and that the condition of children’s obesity was a complaint that happened sporadically and punctually, there was a need to implement a care process also for CG so that both groups (CG and IG) would be comparable in relation to the intervention presenting or not a positive effect on the CG. Thus, the children of the CG were followed in a similar way to the children of the IG, observing the activities so that the contact time between the groups followed what was recommended. For this, FNE activities at home and telephone/SMS/WhatsApp monitoring were suppressed in the CG during the 9 months. Similar to the IG, the activities of the CG comprised five themes, which were eating, physical activity, sedentary behavior, sleep, and mental health [20].

Thus, the difference between CG and IG was that, for the intervention group, home FNE, telephone monitoring, and the planned service time with more contact hours were addressed. The consultations and activities conducted during the follow-up months, which comprise the CG and IG, are described in more detail in Table S2 of the supplementary material.

Data collection

Data collection took place between August 2022 and July 2023, by trained registered dietitians and through a questionnaire that covered specific questions from validated studies [20, 34–36] that comprised five thematic blocks, namely: (1) Sociodemographic and Nutritional Assessment; (2) Food consumption. Consumption of Ultra-Processed Food and Beverages (QUPF); 4. Food Environment (Home and Community); 5. Practice of Physical Activity, Sedentary Behavior, and Sleep.

These blocks were created and grouped into a free online data collection platform, Epicollect5 Data Collection^® [37], and are described in detail in Table S2 of the supplementary material. For this work, blocks 1 and 3 were explored.

During the evaluation months, the researchers applied the complete questionnaire (5 blocks) in three moments (consultation of months 1, 3 and 5). For the other consultations (months 2 and 4), only the Questionnaire on Consumption of Ultra-Processed Food and Beverages (QUPF) was applied. It is noteworthy that visits 2, 3, 4 and 5 occurred after a median of 2, 4, 7 and 9 months, respectively, after visit 1 (baseline). The consultation scheme can be found in Table S2 of the supplementary material.

Block 1. Sociodemographic and nutritional assessment; 2

We collected data on maternal education and monthly family income adapted from the questionnaire of the Study of Cardiovascular Risks in Adolescents (ERICA) [38].

For the anthropometric evaluation of the children, the weight was measured, by means of the average of two consecutive weighing procedures, on a mechanical scale available at the public health centers. Height was performed by calculating the average of two consecutive measurements on a Bic^® portable ultrasonic digital stadiometer with a resolution of 0.1 cm and a measurement range of 30 to 200 cm. The age in years and months of the children was calculated by entering the date of birth and date of individual appointments.

After obtaining these data, the BMI [weight(kg)/height(meters)²]-by-age in z-score was calculated for each of the children based on the formula Z = [(BMI/M)L 1]/(LS) [39], in which the values of L, M and S considered were those proposed by the World Health Organization (WHO) for boys and girls [31] with the aid of the WHO Anthro Plus software (version 3.2.2, 2009).

Block 3. Questionnaire on food and beverage consumption (QUPF)

The questionnaire on Consumption of Ultra-Processed Food and Beverages on the previous day was composed of the most consumed UPF in Brazil, according to the Consumer Expenditure Survey (POF, in Portuguese) 2008–2009, and considered a sample for adolescents between 10 and 12 years old (2017–2018) [40, 41]. UPF were grouped into 4 blocks, with similar characteristics to facilitate the caregiver to remember the child’s consumption, starting with breakfast/small meal (first block), large meals (second block), desserts (third block) and beverages (fourth block) (Table S3 of the supplementary material). For analysis purposes, the 36 items were condensed into 11 groups based on the proximity between the foods from the initial questionnaire and in groupings carried out in previous studies, named: dairy drinks, salty crackers and chips, sweet cookies and cereals, margarine, breads, processed meats, ready-to-eat meals, ready-to-use condiments, sweets, soft drinks and juices, as illustrated in Figure S1 of the supplementary material [42–45].

Data analysis

For descriptive analysis, we calculated the distribution of frequency, central tendency and dispersion measures. The Student’s t-test was used to compare means and the chi-square test to compare proportions. These analyzes were performed for the group of children attended at baseline, for the group of children who composed the losses during the study, and for comparison between those who remained in relation to the losses during the study.

The effect of the intervention was obtained through the intention-to-treat analysis [46, 47], comparing the means of consumption of the number of UPF, BMI/Age (z score) and BMI (kg/m²) between and intra CG and IG after intervention (at baseline in relation to consultation 3 and 5) and, for the consumption of UPF, we conducted the comparison with the inclusion of consultations 2 and 4, using a generalized estimation equation (GEE) model. The variables were treated as normal distribution, with connection identity function. The working correlation matrix used was the covariance matrix of the unstructured and robust estimator. Non-standardized coefficients (β) and their respective confidence intervals (95% CI) were calculated. P-values < 0.05 were considered statistically significant. Additionally, the analyses were adjusted for the primary health care unit with which each child was associated, to account for the cluster design. All analyses were completed using the Statistical Package for the Social Sciences (version 15.0, SPSS, Chicago, Illinois) and Statistics Data Analysis (version 14.0, stata).

Sensitivity analyzes were also carried out in order to verify the robustness of the associations for consumption of UPF, BMI/Age (z score) and BMI (kg/m²) between and intra CG and IG after intervention with children who had completed all three evaluations (baseline, consultation 3 and 5).

Results

Recruitment of children and baseline

Figure 1 shows the flowchart of the first contact (made in July 2022) with the children’s families (n = 727). Considering the attempts to contact the families for the invitation and scheduling of the first consultation through the phone numbers registered in the public health centers records, we verified that the majority (52.0%) did not correspond to the registered numbers or the lines were inoperative, characterized in contact impossibility;10.7% of the children had the self-report of the guardian that they did not have obesity; 7.0% had difficulty with the scheduling times; 3.6% refused to participate in the research; 2.0% had the appointments scheduled, but did not attend;0.5% were outside the scope of the municipality of Betim; 0.5% were out of the established age range; and 0.5% had a diagnosis of autism. For families who agreed to participate and the children were eligible, the appointments were made through phone contact from Monday to Friday, 8 am to 6 pm, and the appointments were made according to the room availability for care in the public health centers and considering the availability of parents or caregivers from Monday to Friday, 7 am to 4 pm.

Fig. 1.

Flowchart of the study Managing Childhood Obesity in the Context of Primary Health Care: A Multi-Component, Intensive, Intervention-Based Approach (Brasil, 2022)

For the first visit (baseline), 167 children were eligible and randomized in August 2022. During the study, 69 children left the research and four had an incomplete evaluation (evaluation at baseline, consultation 3 and consultation 5). Thus, 73 children composed the group ‘losses during the study’ (Supplementary Table S4).

Sample characteristics - baseline

In Table 1, we present the sociodemographic and anthropometric characteristics of the baseline of the children attended (n = 167). Most of the children were white or brown (Pardo, in Portuguese), had complete high school maternal education, and had a monthly family income corresponding to more than one and a half minimum wages in 2021.

Table 1.

Sociodemographic and anthropometric characteristics of the baseline of the children attended in the study managing childhood obesity in the context of primary health care: A Multi-Component, Intensive, Intervention-Based approach (Brasil, 2022,n = 167)

Variables	Control n(%)1	Intervention n(%)2	p-value*
Sex
Female	41(47.67)	45(52.33)	0.181
Male	47(58.02)	34(41.48)
Race/color
Yellow	3(75)	1(25)	0.444
White	18(52.94)	16(47.06)
Indigenous	0	1(100)
Brown (Parda)	55(55.56)	44(44.44)
Black	12(41.38)	17(58.62)
Mother’s education
Did not attend school	1(100)	0	0.706
Did not graduate elementary school	2(33.33)	4(66.67)
Graduated elementary school	5(41.67)	7(58.33)
Did not graduate high school	11(61.11)	7(38.89)
Graduated high school	49(52.13)	45(47.87)
Did not graduate higher education	8(50.00)	8(50.00)
Graduated higher education	12(63.16)	7(36.84)
Monthly household income
No income	2(66.67)	1(33.33)	0.938
Up to BRL 606	4(50.00)	4(50.00)
BRL 606 to BRL 1,212	12(54.55)	10(45.45)
BRL 1,212 to BRL 1,818	21(56.76)	16(43.24)
More than BRL 1,818	47(50.00)	47(50.00)

Variables	CG Mean (95% CI)	IG Mean (95% CI)	p-value**
Age	8.41 (8.13; 8.69)	8.68 (8.38; 8.98)	0.188
BMI (kg/m2)	26.07 (25.06; 27.08)	27.28 (26.39; 28.16)	0.079
BMI (Z-score)	3.01 (2.89; 3.14)	3.14 (3.01; 3.28)	0.144

BMI Body Mass Index, CI Confidence Interval

value of the minimum wage considered: BRL 1,212.00 (one thousand, two hundred and twelve BRL) (Brasil, 2021)

¹ n = 88 ² n = 79

*Chi-square test; **t-test

The mean age of the children in the CG was 8.41 (95% CI: 8.13; 8.69) years and for the IG it was 8.68 (95% CI: 8.38; 8.98) years. The BMI (kg/m²) of the CG was 26.07 (95% CI: 25.06; 27.08) and of the IG was 27.28 (95% CI: 26.39; 28.16), and the BMI/z-score of the CG was 3.01 (95% CI: 2.89; 3.14) and of the IG was 3.14 (95% CI: 3.01; 3.28).

There were also no differences between the sociodemographic and anthropometric characteristics of the children in the CG and IG who composed the group ‘losses during the study’, nor between the losses and the children who remained in the study (Supplementary material - Tables S4 and S5).

Contact time (hours)

In Table 2, we present the average of the total contact time between the children of the CG and the IG for individual care, home and collective FNE actions at the public health center, and telephone monitoring. For the CG, the average time was 03h06min, with a minimum duration of 00h39min and a maximum of 10h20min. For the IG, the average time was 12h34min, with a minimum duration of 08h06min and a maximum of 22h14min.

Table 2.

Mean time in minutes of contact between children in the control and intervention groups treated in the study managing childhood obesity in the context of primary health care: A Multi-Component, Intensive, Intervention-Based approach (Brasil, 2022 n = 167,CG = 88 and IG = 79)

	Control Group		Intervention Group
	mean (95% CI)	(min-max)	mean (95% CI)	(min-max)
Individual appointment*	118 (105.2; 130.7)	39–285	162.3 (145.4; 179.4)	36–300
Food and Nutrition Education in the household**	NA	NA	150	150
Group Food and Nutrition Education in the public health centers***	68.2 (45.1; 91.2)	0–447	152.3 (112.1; 192.5)	0–650
Phone/SMS/WhatsApp monitoring	NA	NA	300	300
Total	186.2 (155.5; 216.9)	39–620	764.7 (713.6; 815.8)	486–1334

CI Confidence Interval

* CG two approximate values (40 min) due to lack of information; ** performed in the household; *** performed in group in the public health centers

Effects of the intervention - intention-to-treat analysis

The results found in the comparison between and intragroups for the means of consumption of the number of UPF (at baseline in relation to consultations 2, 3, 4 and 5) and the values of BMI/Age (z-score) and BMI (kg/m²) after intervention are described in Table 3.

Table 3.

Comparison between and intragroup of UPF, BMI (z-score) and BMI (kg/m²) after theinterventionof the study Managing Childhood Obesity in the Context of Primary Health Care: A Multi-Component, Intensive, Intervention-Based Approach (Brasil, 2022, n=167,CG=88 and IG=79).

		Control group Mean (95% CI)	Intervention group Mean (95% CI)	CG Intragroups (β 95% CI)	IG Intragroups (β 95% CI)	CG (reference) vs IG betweengroups b (β 95% CI)
UPF	Appointment 1 (baseline)	4.16 (3.74; 4.58)	4.62 (4.22, 5.01)	reference	reference	reference
	Appointment 2	4.34 (3.85, 4.83)	4.46 (3.98; 4.94)	0.168 (-0.44; 0.78)	-0.14 (-0.76; 0.48)	-0.34 (-1.21; 0.53)
	Appointment 3 (re-evaluation 1)	4.24 (3.67; 4.80)	3.74 (3.34; 4.14)	0.08 (-0.63; 0.79)	-0.87 (-1.45; -0.29)*	-0.95 (-1.87; -0.04)*
	Appointment 4	4.65 (4.15; 5.15)	3.76 (3.25, 4.27)	0.498 (-0.11; 1.10)	-0.84 (-1.45; -0.23)*	-1.35 (-2.215; -0.49)*
	Appointment 5 (re-evaluation 2)	3.88 (3.35; 4.40)	3.40 (2.90; 3.90)	-0.29 (-0.92; 0.34)	-1.22 (-1.85; -0.60)*	-0.94 (-1.83; -0.05)*
BMI (Z-score)	Appointment 1 (baseline)	3.05 (2.93; 3.17)	3.182 (3.05; 3.31)	reference	reference	reference
	Appointment 3 (re-evaluation 1)	2.98 (2.80; 3.14)	3.16 (3.01; 3.31)	-0.076 (-0.20, 0.05)	-0.021 (-0.12; 0.08)	0.03 (-0.09; 0.15)
	Appointment 5 (re-evaluation 2)	2.89 (2.71; 3.07)	3.12 (2.94; 3.30)	-0.157 (-0.34; 0.02)	-0.06 (-0.20, 0.07)	0.06 (-0.139; 0.27)
BMI (kg/m²)	Appointment 1 (baseline)	26.07 (25.08; 27.06)	27.27 (26.41; 28.14)	reference	reference	reference
	Appointment 3 (re-evaluation 1)	25.72 (24.92; 26.53)	27.22 (26.37; 28.08)	-0.39 (-1.13; 0.35)	-0.05 (-0.40; 0.31)	0.30 (-0.49; 1.08)
	Appointment 5 (re-evaluation 2)	26.39 (25.22; 27.57)	27.98 (26.90; 29.06)	0.34 (-0.75; 1.44)	0.65 (-0.10; 1.41)	0.38 (-0.95; 1.71)

βNon-Standard Beta,CI Confidence Interval, UPF Ultra-Processed Foods, BMI Body Mass Index

^aGeneralized estimating equations

^bGeneralized baseline-adjusted estimation equations [comparison between CG (reference) and IG]

*p<0.05

For UPF consumption, the intention-to-treat analysis showed the effectiveness of the intervention when compared to the CG, presenting a statistically significant decrease for IG in consultation 3 [95% CI: −0.95 (−1.87; −0.04)], consultation 4 [95% CI: −1.35 (−2.215; −0.49)] and 5 [95% CI: −0.94 (−1.83; −0.05)], which corresponds to a little more than 10% reduction. When making the intragroup comparison, this reduction was also observed in the IG in relation to the baseline at visit 3, 4 and 5. For the CG, there was no statistically significant reduction in the intragroup analysis, as expected (Table 3). Regarding BMI and BMI/age (z-score), there were no statistically significant differences in the analyzes between and intragroups (p > 0.05).

Effects of the intervention - sensitivity analysis

In Table 4, we present a sensitivity analysis for UPF consumption, considering only the children who participated in the complete intervention. It was observed that in the comparison between groups, children in the IG showed a statistically significant reduction in UPF consumption when compared to the CG at visit 4 (β −1.27 [95% CI: −2.28; −0.27]) and at visit 3 (β −1.11 [95% CI: −2.24; −0.27]) and marginally significant at visit 5 (β −0.99 [95% CI: −2.01; 0.04], p = 0.059). In the intragroup comparison, there was a statistically significant reduction for the IG over all visits in relation to the baseline. For the CG, there was no statistically significant intragroup reduction, as expected.

Table 4.

Comparison between and intragroups (control and intervention) of UPF, BMI (Z-score) and BMI (kg/m²) of children who remained in the study (Brazil, 2022; CG=45 and IG=79)

	Appointment (appt.)	Control (Mean 95% CI)	Intervention (Mean 95% CI)	Intragroups - Control (β 95% CI)	Intragroups - Intervention (β 95% CI)	Intervention vs Control (reference) between groups b (β 95% CI)*
UPF	Appt. 1 (baseline)	4.11 (3.48; 4.75)	4.71 (4.18; 5.25)	reference	reference	reference
	Appt. 2	4.51 (3.93; 5.09)	4.41 (3.92; 4.90)	0.39 (-0.38; 1.15)	-0.27 (-1.03; 0.49)	-0.70 (-1.78; 0.37)
	Appt. 3 (re-evaluation 1)	4.22 (3.63; 4.81)	3.72 (3.32; 4.12)	0.11 (-0.78; 1.00)	-0.99 (-1.69; -0.28)*	-1.11 (-2.24; -0.27)*
	Appt. 4	4.47 (3.98; 4.96)	3.8 (3.27; 4.33)	0.36 (-0.35; 1.08)	-0.90 (-1.61; -0.19)*	-1.27 (-2.28; -0.27)*
	Appt. 5 (re-evaluation 3)	3.81 (3.31; 4.32)	3.43 (2.92; 3.94)	-0.30 (-1.02; 0.42)	-1.29 (-2.02; -0.56)*	-0.99 (-2.01; 0.04)
BMI (Z-score)	Appt. 1 (baseline)	3.09 (2.90; 3.27)	3.28 (3.11; 3.44)	reference	reference	reference
	Appt. 3 (re-evaluation 1)	2.94 (2.76; 3.12)	3.16 (3.00; 3.32)	-0.14 (-0.26; -0.03)*	-0.12 (-0.17; -0.05)*	0.03 (-0.10; 0.16)
	Appt. 5 (re-evaluation 3)	2.8708 (2.68; 3.06)	3.12 (2.94; 3.30)	-0.22 (-0.41; -0.02)*	-0.16 (-0.27; -0.05)*	0.06 (-0.16; 0.28)
BMI (kg/m2)	Appt. 1 (baseline)	26.14 (24.89; 27.39)	27.61 (26.61; 28.61)	reference	reference	reference
	Appt. 3 (re-evaluation 1)	25.98 (24.88; 27.07)	27.56 (26.57; 28.54)	-0.16 (-0.77; 0.44)	-0.05 (-0.41; 0.30)	0.11 (-0.60; 0.81)
	Appt. 5 (re-evaluation 3)	26.51 (25.16; 27.85)	28.26 (27.05; 29.46)	0.36 (-0.77; 1.50)	0.647 (-0.12; 1.42)	0.28 (-1.09; 1.65)

β Non-Standard Beta, CI Confidence Interval, UPF Ultra-Processed Foods, BMI Body Mass Index

^aGeneralized estimating equations

^bGeneralized baseline-adjusted estimation equations [comparison between CG (reference) and IG]

*p<0.05

It was observed in the sensitivity analysis that children who remained in the complete study showed a decrease in BMI/age (z-score) over time (consultation 3 and 5 in relation to consultation 1, baseline), both in the CG and in the IG in the intragroup comparison (Table 4), however no difference was observed between groups (p > 0.05) (Table 4).

As for BMI (kg/m²), for those children who remained in the complete study, there was no statistically significant difference in the comparison between and intragroups (p > 0.05) (Table 4).

Discussion

This study evaluated the effects of an intervention with children with obesity on the UPF consumption and body mass reduction. The intervention had an average of 12h34min of contact and was able to show a reduction in UPF consumption, which may be associated with the intervention based on multiple components, however, no effect was observed on the participants’ BMI. In the IG, a longer average contact time with the participants was obtained in relation to the CG, as planned, however, both groups were below the number of hours established for the characterization of an intensive intervention [17]. It is important to emphasize that, even though an intervention was conducted with fewer hours of contact, it was still possible to observe a reduction in UPF consumption among children monitored in the IG.

A clinical trial that evaluated the reduction of UPF consumption in grams in children aged 7 to 12 years, for a period of activities of 6 months, identified that participants in both groups (control and intervention) had a decrease in UPF consumption until the fourth follow-up month, and from the fourth month until the sixth there was an increase in UPF consumption [48]. In our findings, there was a decrease in the number of UPF consumed in the IG from consultation 3 and this reduction was maintained until the end of the study.

Studies that evaluate associations between the consumption of UPF with negative health outcomes are robust for the adult population, showing an association between UPF and obesity, type 2 diabetes, cardiovascular disease, and all-cause mortality, showing the importance of reducing UPF for the treatment of obesity, contributing to avoid excessive weight gain, for the prevention or control of diseases associated with obesity [49, 50]. For children, although scarce, evidence shows positive associations between UPF consumption and risks of cardiometabolic events in preschoolers (3 to 6 years old) [51], obesity and overweight in children and adolescents [52, 53], and increased BMI at 10 years old [54].

When analyzing the change in BMI/age (z-score) and BMI (kg/m²), studies that performed an intervention with less than [19, 55–57]26 h of contact found a statistically significant reduction in BMI/age values (z-score) in overweight children [55, 57]. On the other hand, our results for children living with obesity did not show a statistically significant reduction in BMI/age (z-score) and BMI (Kg/m²), suggesting, as in other findings, that behavioral changes are the first actions for the proper management of obesity [14–16, 20, 58, 59].

It is known that the growth pattern of children with obesity can be impaired due to a high BMI for their age [60, 61], making it necessary to reduce the speed of excessive weight gain [20], however, the association of BMI reduction and other body weight results are small after 6 to 12 months and evidence on these parameters after 12 months is scarce [19].

Although no significant results were found for a decrease in BMI/age (z-score) and BMI (Kg/m²), the children in the CG and in the IG did not show an increase in BMI, suggesting that, over time, if healthy lifestyle habits, and associating behavioral change strategies with family involvement and adoption of multicomponent approaches are maintained [20], these children can achieve an adequate weight by slowing weight gain.

In this context, the management of childhood obesity should be considered with caution, especially regarding a restriction of caloric intake [62]. In this study, out objective was not to restrict the caloric intake of children, although there are discussions that a moderate caloric restriction is effective for the treatment of childhood obesity [62–64]. Even so, for the management of childhood obesity, the recommendation is that changes are applied in the individual’s behavior, which involves a healthy diet, based on unprocessed or minimally processed foods, regular practice of physical activity, reduction of sedentary behavior, regulation of sleep and mental health, seeking to reverse excessive weight gain during growth and development for adequate growth, promotion of mental well-being and improvement of quality of life [20, 65], and then achieving good results, either by reducing or maintaining BMI/age [19, 55–57].

Thus, the multi-component approach that provides for an intervention not only in food, but also in the environment in which the child lives or stays [20], has been shown to be more appropriate for the development of the child [66]and to reduce possible risks inherent to food restriction in the children [48, 66].

Although no reduction in BMI was found between CG and IG, we also tested considering only those children who had all visits and who remained until the end of the study. The results found showed a decrease in BMI/age (z-score) over time (consultation 3 and 5 in relation to consultation 1, baseline), both in the CG and in the IG in the intragroup comparison.

The weight loss in both groups can be explained by the fact that there was no type of organization of child obesity care implemented and standardized within the scope of primary health care in the city where the study was carried out. These results may indicate that nutritional care for children with obesity and some contact with the themes of the intervention can contribute to the health of children and reinforce the importance of maintaining adequate and comprehensive monitoring, through transversal, community and individual approaches in the routine of health care for the treatment of obesity and also for the prevention of new cases [20].

An approach that integrates a multidisciplinary follow-up is expected within the Brazilian public health system (SUS), due to the characteristics of universality of access to services and integrality of care, provided by the articulated and continuous set of preventive and curative actions and services, for individuals and for the community [67]. For access to care, the primary health care (PHC) is configured as the gateway to the health system with attributes of longitudinal and comprehensive care [68, 69]. For the children of the IG, activities aimed at the family were also carried out, in a context outside the public health centers. Interventions carried out in PHC to treat childhood obesity require specific knowledge and skills of clinical teams with a non-stigmatizing approach that respects the particularities of the child and the family [20, 70]. In addition, these interventions should integrate an intensive and longitudinal care in household treatment, carried out by several health professionals, in order to evaluate and monitor children and identify unique social factors [70]. Most clinical trials for the management of childhood obesity that are based on intervention in the behavior of children and families are carried out in PHC, but they also integrate care for the household [19, 65] and reinforcement activities via online text messages [47]. However, although the interventions are effective, there are many challenges in PHC services, whether related to work processes, referral and care flows or the public’s adherence to the service, since there is a low demand for treatment linked to childhood obesity, for it is often that the population seeks PHC in cases of acute treatment, impairing the structuring of a continuous flow service [71].

Despite these particularities and the complexity of the treatment of childhood obesity, to our knowledge, this was the first study to evaluate children with obesity treated in PHC of one of the largest public health systems in the world (SUS), with important results in relation to the reduction of UPF consumption, shedding light on the possibility of managing obesity in PHC.

This study has some limitations. First, we planned an intensive intervention (with more than 26 h of contact) that could not be carried out because in the routine of appointments and FNE activities carried out in the project, many children were missing for various reasons, and sometimes the time planned for conducting the activities was shorter than expected. Another limitation concerns the recruitment process of children, since there was no established protocol for conventional treatment in PHC for childhood obesity.

Thus, there was a need to conduct an intervention for both groups with a difference in intensity of care between them. However, children who were followed until the end of the intervention showed sufficient results to show benefits from a multi-component approach, especially regarding reducing UPF consumption.

Thus, lines of care for the prevention and treatment of obesity in PHC, which focus on the management of childhood obesity, should adopt approaches with multiple components, including several themes related to health and professionals, as well as using tools to encourage the process of change of children and their caregivers, in an adequate, integral and longitudinal multidisciplinary way, both with individual, community, and transversal approaches, and considering the determinants and conditions of overweight and obesity, without blaming, stigmatization, and discrimination of the person or their family [20]. This incentive to change the lines of prevention and treatment can be achieved through public and intersectoral policies and programs, promoting an environment and a healthy lifestyle, based on scientific evidence that proves the effectiveness of the multi-component approach for behavior change and management of childhood obesity.

Conclusion

Our study showed that an intervention with multiple components was effective for reducing UPF consumption, and although it did not show a reduction in BMI/age of children with obesity, it contributed to its maintenance, and consequently should be considered in the lines of care for the management of childhood obesity in Brazilian primary health care (PHC). These results were seen in a scenario of a reduced number of contact hours, compared to what was planned, which may indicate greater feasibility for the management of obesity in the context of PHC.

Authors’ contributions

Conceptualization: Jardim, M.Z.; Rocha, L.L.; Gratão, L.H.A.; Pessoa, M.C.; Mendes, L.L. Data curation: Jardim, M.Z.; Rocha, L.L. Formal analysis: Jardim, M.Z.; Canella, D.S.; Carmo, A.S.; Rocha, L.L.; Cunha, D.B.; Pessoa, M.C.; Mendes, L.L. Writing – original draft: Jardim, M.Z.; Carmo, A.S.; Rocha, L.L.; Pessoa, M.C.; Mendes, L.L. Writing – review & editing: Jardim, M.Z.; Canella, D.S.; Carmo, A.S.; Cunha, D.B.; Pessoa, M.C.; Mendes, L.L. Validation: Jardim, M.Z.; Canella, D.S.; Carmo, A.S.; Rocha, L.L.; Cunha, D.B.; Gratão, L.H.A.; Pessoa, M.C.; Mendes, L.L. Visualization: Jardim, M.Z.; Canella, D.S.; Carmo, A.S.; Rocha, L.L.; Cunha, D.B.; Gratão, L.H.A.; Pessoa, M.C.; Mendes, L.L. Supervision: Gratão, L.H.A.; Pessoa, M.C.; Mendes, L.L. Funding acquisition: Mendes, L.L. Project administration: Jardim, M.Z.; Mendes, L.L. All authors read and approved the final manuscript.

Funding

This study was funded by the Ministry of Health (CNPq/DEPROS/SAPS No. 27/2020, Research Axis on Noncommunicable Diseases and Associated Risk Factors), with additional financial support from Fundación MAPFRE.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to the nature of the clinical trial but are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Consent for publication

Not applicable.

Ethical approval and consent to participate

The research was approved by the Research Ethics Committee of the Federal University of Minas Gerais (CAAE 39508720.6.0000.5149). All participants and their legal guardians were informed about the study and agreed to participate by signing the Informed Consent Form (ICF). Consent for participation was specifically obtained from the parents or legal guardians of all children involved in the study. We affirm our adherence to the ethical principles outlined in the Declaration of Helsinki, established by the World Medical Association, which guides medical research involving human subjects and ensures respect for their dignity, rights, and well-being. We also declare that the clinical trial is in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines. It is important to note that this study did not involve the use of pharmacological treatments.