*2.1. Patients and Data Source*

We initially identified 430,541 infants who were born in 2013 and examined their infant health check-up records for the 1st to 6th visits from the National Health Insurance Service (NHIS) database. Healthcare claims including diagnostic codes of almost all Korean residents, approximately 98% covered by NHIS and 2% by medical aid, were linked to health check databases. The data, including gestational age and birth weight, were also grouped according to the International Classification of Diseases-10 codes (ICD-10: P07.01, P07.02, P07.09-14, P07.19, P07.20, P07.23, P07.29, P07.30, P07.39) [8]. The data were entered by the hospital or obtained from self-report questionnaires used by the national health screening program. Based on the birth statistics [9], the total number of births in 2013 was 436,455, and the number of infants who lived to be at least 1 year of age was 435,150, which shows that this study population 430,541 covered 99% of national births.

The national health screening program for infants and children in Korea, launched in November 2007 is a kind of population surveillance system that consists of history taking, physical examination, anthropometric measurements, screening for visual acuity, and administration of Korean Developmental Screening Test (K-DST), oral examination, and questionnaires with anticipatory guidance [10]. The questionnaire contains the birth weight, preterm, vision, hearing, nutrition (meal, milk, snacks), multimedia, and safety education. We used only the information of birth weight and preterm status in questionnaire from family.

The period for national health screening program (1st to 6th visits) was divided and classified as follow; 6 months for 4–6 months of age, 12 months for 9–12 months of age, 24 months for 18–24 months of age, 36 months for 30–36 months of age, 48 months for 42–48 months of age, and 60 months for 54–60 months of age. The age at exam was defined as chronologic ages.

For growth assessment, the National health screening program checks anthropometric parameters including body weight, height, and HC serially at every follow-up. The percentile of growth was assessed using the Korean growth curve, which provides sex specific data. Poor growth was defined as measurements below the 10th percentile of weight, height, and head circumference individually.

The K-DST is used an effective screening tool for infants and children with neurodevelopmental disorders and has been used since 2011. It is used to verify whether infants are developmentally appropriate or neurodevelopmentally delayed in six domains: gross motor, fine motor, cognition, communication, social interaction, and self-control.

The K-DST is conducted to screen children according to their corrected age before 36 months of age as recommendation and after that age, it is allowed to take tests according to chronological age. There is no K-DST at first visit, and at 5th, 6th visit (42–48, 54–60 months of age) the participants take the test papers according to their chronological

age. The participants take the tests papers at the time of their clinic visit and get the result as four categorized groups based on the standard deviation (SD) scores; the scores above 1 SD are defined as 'high-level', those between −1 and 1 SD as 'peer-level', those between −2 and −1 SD as 'follow-up test', and those below −2 SD as 'further evaluation' [11]. Additional positive questions that take into account clinically important diseases, such as cerebral palsy, language delay, and autism spectrum disorders, that should be referred for 'further evaluation' are also included in the questionnaire. To evaluate the ability of the K-DST to identify infants with developmental delay, critical cutoff scores for 6 domains were set below −1 SD [12,13]. Suspected developmental delay was defined as a K-DST result of 'further evaluation' and 'follow-up test'.

In this study, growth and developmental results were analyzed according to five stratified birth weight groups (<1000 g, 1000–1499 g, 1500–1999 g, 2000–2499 g, and 2500–4500 g). LBW infants and VLBW infants were defined as having a birth weight below 2500 g and 1500 g, respectively. Preterm infants were defined as infants born before 37 weeks of gestation.

#### *2.2. Statistical Analyses*

The cohort was stratified according to the birth weight or the age of checkup. The characteristics of the subjects were expressed as means and standard deviations for continuous variables and as percentages for categorical variables. Correlations for height, weight and HC between 6 months and 60 months of ages as time periods were computed using Pearson's correlation coefficient. Multiple logistic regression model was used to determine the independently associated factors with among infants with odds ratios (OR) and 95% confidence intervals (CI). Multivariate longitudinal data analysis was done using multivariate repeated measured model (PROC MIXED and GENMOD). All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA). *p*-values < 0.05 were considered statistically significant.

### *2.3. Ethics Statement*

In this study, all identifiable variables, including claim-, individual-, and organizationallevel identification numbers, were re-generated in random by the NHIS database to protect the patients' privacy. This study used NHIS data (NHIS-2019-1-569) maintained by the NHIS. The study protocol was approved by the Institutional Review Board of Gangnam Severance Hospital (No. 3-2019-0147). Informed consent was waived.

#### **3. Results**

#### *3.1. Growth Outcome*

Among 430,541 infants, born in 2013 and included in the study, 219,576 (51%) were male. The numbers of infants, who underwent health checks ranged from 286,331 (67%) to 347,153 (81%). The highest number of infants (*n* = 347,153, 81%) were included in the health check at 24 months of age. The highest number of preterm infants underwent the health check at 36 months (*n* = 26,338, 93%). The distribution of a number of infants who participated in the infant health check according to birth weight group were shown in Table 1.



Data are presented as Number (%).

The mean percentile of weight, height, and HC according to age at health check was seen in Figure 1. Longitudinal analysis showed a significant difference in height, weight, and HC according to age, birth-weight group, and the combination of age and birth weight, respectively (*p* < 0.0001). The lower birth weight group showed a lower mean percentile of weight, height, and HC. There was a significant difference in height, weight, and HC between the low birth weight infants (<1000 g, 1000–1499 g, 1500–1999 g, 2000–2499 g) and the reference group with birth weight of 2500–4500 g according to age at health checkup.

**Figure 1.** Mean growth percentile of infants according to age at check-up by birth-weight group. (**a**) Height. (**b**) Weight. (**c**) Head circumference (HC). Significant differences in height, weight, and HC between the low birth weight infants groups and the reference group according to age at health checkup were shown. *p*-values were significant in height, weight, and head circumference between each low birth weight group and the reference group compared at age of health check-up.

A total of 10,227 (7.4%) infants had a poor HC growth at 60 months of age, 10,950 (7.92%) infants had poor height growth, and 12,481 (9.03%) infants had a poor weight growth. Using longitudinal analysis, this study found a significant difference in the incidence of poor height, weight, and HC growth according to age at health check, birthweight group, and combination of age and birth weight, respectively (*p* < 0.0001) (Figure 2). The lower birth-weight groups showed a higher incidence of poor weight, height, and HC growth. There was a significant difference in height, weight, and HC between the low birth weight group (<1000 g, 1000–1499 g, 1500–1999 g, and 2000–2499 g) and the reference group with birth weight of 2500–4500 g according to age at health checkup.

**Figure 2.** The incidence of poor growth (below 10th percentile) according to age at check-up by birth-weight group. (**a**) Height. (**b**) Weight. (**c**) Head circumference. A higher incidence of poor weight, height, and HC growth in the lower birth-weight groups was noted. A significant difference in height, weight, and HC between the low-birth-weight groups and the reference group according to age at health checkup were shown. *p*-values were significant in height, weight, and head circumference between each low-birth-weight group and the reference group compared at age of health check-up.

The Pearson correlation coefficient of the growth percentiles (the height, weight, and HC percentile) at 6 and 60 months of age obtained using correlation analysis is shown as Figure 3. Figure 3a was the results for the whole study population. Among the infants below 1000 g of birth weight, only weight showed a highly positive correlation (coefficient = 0.72) between 6 and 60 months of age, height (coefficient = 0.65) and HC (coefficient = 0.64) showed a moderate positive correlation. Pearson correlation coefficient analysis between 6 and 60 months of age was performed only among the infants who fell below 10th percentile in terms of height, weight, and HC at 60 months of age, and the results were shown in Figure 3b. Among the infants below 1000 g of birth weight, weight, height, and HC showed a weakly positive correlation; otherwise, there was no association.


**Figure 3.** Pearson correlation coefficient between 6 months of age and 60 months of age for birth-weight groups among whole population (**a**) and among infants who were below the 10th percentile of height, weight, and HC at 60 months of age (**b**).

> To analyze the relation between the growth at 6 months and 60 months, the risk of poor growth in the infants below 10th percentile of growth at 6 months was compared to the infants within 10–90th percentile of growth. The infant below the 10th percentile of HC and height at 6 months of age, respectively, showed the higher risk of HC and height below the 10th percentile at 60 months of age (HC, OR (95% CI) 1.62 (1.32–1.98); Height, 1.64 (1.38–1.95)). However, weight status at 6 months showed no significant association with the risk of weight below the 10th percentile at 60 months of age.

#### *3.2. Developmental Outcome*

The incidence of suspected developmental delay result at 60 months of age was 10% (29,020). In particular, further evaluation was recommended for 4572 (1.5%) infants, and the 'follow-up test' for 24,448 (8.5%) infants. According to the birth-weight group, K-DST results were presented in Table 2. The smaller birth weight group had a greater number of 'further evaluation' results. The infants below 1000 g of birth weight were only 0.2% of the screened population, but among them, 14.8% of this group had 'further evaluation' and 21.0% of this group had 'follow-up test' recommendations. The lower birth-weight groups showed a higher incidence of suspected developmental delay.

There is a significant difference in the incidence of suspected developmental delay results between the infants with poor weight, height, and HC growth and above 10th percentile at 60 months of age by birth weight group (Table 3). The infants with poor weight, height, and HC growth demonstrated higher frequency of suspected developmental delay results at 60 months of age.

Lower birth weight, male sex, poor HC, poor height, and poor weight were confirmed as factors associated with suspected developmental delay results at 60 months of age using multivariate logistic regression analysis (Table 4). Infants with poor HC at 60 months of age had more suspected developmental delay results (OR 1.81, 95% CI 1.66–1.98), and the infants who weighed less than 1000 g at birth had more suspected developmental delay (OR 5.05, 95% CI 3.79–6.73) compared to infants with 2500–4500 g birth weight.


Data are presented as Number or Number (%). BW, birth weight.



**Table 3.** Poor growth outcomes (below 10th percentile) at 60 months of age according to birth-weight group among the infants with suspected developmental delay results.

Data are presented as No. (%). BW, birth weight; WT, weight; HT, height; HC, head circumference.

**Table 4.** Multivariate logistic regression analysis for the suspected development delay results at 60 months of age.


HC, head circumference; HT, height; WT, weight; BW, birth weight.; OR, odds ratio; CI, confidence interval.
