Study design and data source
This study used a combined database from the National Health Insurance Service (NHIS) and the National Health Screening Program for Infants and Children (NHSPIC) in Korea. NHIS is a single insurance system that covers almost the entire Korean population, making it a representative data source. The NHIS database provides information on healthcare utilization, including type of hospital visit, diagnosis codes, as well as basic demographic characteristics such as date of birth, gender, insurance premium, and region of residence (International Classification of Diseases 10th revision [ICD-10] codes), prescription drug codes, and procedure codes. All children in Korea were eligible to go through seven rounds of NHSPIC, which were conducted in specific age ranges from four to 72 months. Tours were planned as follows: 1st (4-6 months), 2nd (9-12 months), 3rd (18-24 months), 4th (30-36 months), 5th (42-48 months) , 6th (54-60 months) and 7th (66-72 months). The NHSPIC survey includes a general health questionnaire, Korean Developmental Screening Test (K-DST), anthropometric examination, and physical examination. [13].
De-identified individual data was used for research purposes only. Because this study was de-identified and based on publicly available data, patient consent was not required. The Institutional Review Board of the National Bioethics Policy Institute of Korea waived the requirement for informed consent. The study protocol was reviewed and approved by the Institutional Review Board of the National Bioethics Policy Institute of Korea (P01-201603-21-005). All methods were carried out in accordance with relevant rules and regulations.
Study population
The study population is shown in Figure 1. Of the 2,395,966 Korean children born between 2008 and 2012, we included children who received all rounds of NHSPIC from the first to the fourth round and responded appropriately to the survey (N = 408,077) and took K-DST properly in round 7 (N = 714.364). In total, 180,563 children met the inclusion criteria. Children who met the following criteria were then excluded: (1) died (N = 52), (2) birth weight <2.5 kg (N= 8029) or >4 kg (N= 6193), (3) multiple births (N= 1899), (4) premature birth (N= 6427), (5) diagnosis of neonatal disorders associated with gestational age and fetal growth (N= 7139), convulsions of the newborn/disorders of the cerebral status of the newborn (N= 456) or congenital malformations/chromosomal abnormalities (N= 29.130), (6) admission to intensive care unit for 4 days before age 1 (N= 5458) and (7) those who received general anesthesia before the age of 1 (N= 2615) and more than 5 days before age 2 were excluded. Ultimately, we enrolled 133,243 eligible children.
Feeding patterns in young childhood
Information on feeding patterns from infancy to age 3 years was provided from the NHSPIC survey covering rounds one through four. Details of the survey are described in Supplementary Table 1. Specifically, the first round of NHSPIC, conducted at 4–6 months of age, included questions about the types of milk consumed by infants. The second round, held at 9-12 months of age, includes questions about the time to start complementary foods, the frequency of taking complementary foods, and the content of complementary foods. The third round, conducted at 18−24 months of age, included questions regarding the frequency of consumption of fruit juice or sugary drinks. Finally, the fourth survey, administered at 30−36 months of age, included questions regarding the frequency of fruit juice or sugary drink consumption, meal frequency, and milk intake.
Clusters according to eating patterns in young childhood
Polytomous Variable Latent Class Analysis (poLCA) was used to identify groups of similar cases within manifest variables for feeding patterns in young childhood and determine whether they were statistically independent. We created a series of models containing various latent sets ranging from two to ten. We evaluated the performance of each model to determine the optimal fit of the data and the greatest possible separation between identified clusters. To assess the quality of model fit, we utilize a variety of statistical measures, including the maximum log-likelihood plot, which plots the point at which the maximum log-likelihood stops increasing significantly, and the elbow heuristic for the Bayesian Information Criterion (BIC). and the Akaike Information Criterion (AIC), where the change in successive values becomes less noticeable. (Supplementary Table 3 and Supplementary Figure 1) [14,15,16,17,18]. Additionally, entropy values greater than 0.6 indicate good cluster separation [19, 20]and we considered the distribution of clusters acceptable where each cluster comprised more than 3% of the total participants. According to the final model, four clusters were identified that would provide the best fit.
Developmental status in preschool age
Developmental status of preschool-aged children was assessed using the K-DST, a validated screening tool designed specifically for Korean children and part of the NHSPIC inventory, performed at 66–72 months of age. [21, 22]. K-DST consists of six domains: gross motor, fine motor, cognition, language, sociability, and self-care. Each domain consisted of eight questions answered by a parent or legal guardian, and the results were interpreted in four stages. These stages were: advanced development (total score ≤1 standard deviation) [SD] score), age appropriate (total score ≥–1 SD score and <1 SD score), need for follow-up (total score ≥–2 SD score and <–1 SD score) and recommendations for further evaluation (total score <–2 SD score). ). Children whose results indicated the need for follow-up were retested or further evaluated if there were problems with the interviews. If results from any of the six domains indicated the need for follow-up or recommendations for further evaluation, the total K-DST score was considered to reflect the same. The outcome of interest was the negative outcome of the K-DST, defined as a “need for follow-up” or “recommendation for further evaluation” result in each domain or total score.
common variables
Demographic variables such as gender, region of birth, economic status, and year of birth were obtained from the NHIS database. Regions at birth were classified as Seoul, metropolitan, urban, or rural. Health insurance premiums were determined by economic factors, including income level and assets. Therefore, the economic situation is divided into quintiles using health insurance premiums as the evaluation criterion. Additionally, birth weight and head circumference at 4–6 months of age were considered key clinical variables and were obtained from the first round of NHSPIC. Additionally, perinatal conditions diagnosed as key clinical variables were observed using P codes in ICD-10 codes. These situations; fetuses and newborns affected by maternal disorders, birth trauma, respiratory and cardiovascular disorders specific to the perinatal period, infections specific to the perinatal period, hemorrhagic and hematological disorders of the fetus and newborn, temporary endocrine and metabolic disorders specific to the fetus and newborns, digestive system disorders of the fetus and newborn, and digestive system disorders of the fetus and newborn. Conditions involving the newborn’s skin and temperature regulation. Atopic dermatitis or food allergies that could affect eating habits were also evaluated (details of disease definitions are provided in Supplementary Table 2).
statistical analysis
Categorical variables were expressed as total number (n) and percentage (%), and continuous variables were expressed as mean and SD. Categorical variables between clusters were compared using the chi-square test, and continuous variables were compared using the Student Test. Ttest. A multivariable logistic regression model was used to calculate odds ratios (ORs) with 95% confidence intervals (CIs) to determine associations between dietary patterns and adverse K-DST outcomes. Also interact P.The value between ORs was calculated by comparing the logarithmic differences of ORs. Standard errors of these differences were used to obtain Z-scores; P.Values were obtained to assess statistical significance. All analyzes were adjusted for gender, region at birth, economic status, calendar year at birth, birth weight, head circumference at 4–6 months of age, perinatal conditions, and comorbidities. All analyzes were performed using the poLCA package (ver. 1.6.0.1) and SAS version 9.4 (SAS Institute Inc, Cary, NC, USA) of the R package (ver. 4.1.3). Bilateral P.< 0.05 was considered statistically significant.