Abstract
Background
Neonatal care of preterm infants may include dietary approaches such as high calorie formulas to promote physical growth. However, continuing growth-promoting strategies beyond the point of necessity, coupled with poverty and food insecurity which are more common among families of children born preterm, may increase the risk of obesity. Because children born preterm tend to have more pressing health conditions that require ongoing care, obesity may go undiagnosed by providers.
Methods
This retrospective cohort study included 38,849 children (31,548 term, 7301 preterm) born from 2010 to 2015, who received clinical care at a large pediatric medical center (Ohio, USA). Electronic medical record data, linked to Ohio birth certificates, were used to identify children with measured obesity (≥2 weight-for-length values ≥95th percentile before 24 months of age or BMI values ≥95th percentile at or after 24 months of age). Children were considered to have diagnosed obesity if their medical record had an obesity-related phrase or billing code recorded. Modified Poisson regression was used to compare risk of obesity undiagnosis among obese children born preterm versus at term.
Results
In total, 13,697 children had measured obesity, 10,273 (75%) of which were undiagnosed. Children born preterm with measured obesity were 8% more likely to be undiagnosed compared to children born at term (adjusted relative risk = 1.08 95% CI 1.05, 1.11). The risk was slightly higher for preterm children born to white women or born to women with higher educational attainment. For both groups, Primary Care and subspecialist clinics were the most common settings for undiagnosed obesity (74.9% and 16.8% of undiagnosed cases, respectively).
Conclusions and relevance
Preterm birth was associated with increased risk of undiagnosed obesity in early childhood. This highlights the need to enhance obesity screening in the preterm population and to further explore reasons for this disparity.
Introduction
One in ten US children are born preterm, and survival rates have improved in recent decades [1, 2]. As a result, the focus of clinical care for preterm infants has shifted from survival to optimizing growth and development. Because infants born preterm often experience growth challenges, interventions like high-calorie formula are common. However, early rapid weight gains or continuing growth-promoting strategies beyond the point of necessity may increase the risk of metabolic syndrome and obesity later in life [3, 4]. Also, children born preterm are more likely to be born into families with poorer socioeconomic status and food insecurity, which may exacerbate their risk of obesity [5–7].
Although the American Academy of Pediatrics emphasized the importance of obesity screening at every primary care visit, the percentage of overweight or obese children in the U.S who are diagnosed as such is low, with estimates ranging from 18 to 54% [8–12]. Children born preterm may be at risk of having clinically unrecognized obesity because they tend to have more pressing health care needs that attract greater clinical attention [13]. Also, they are more likely to have chronic health conditions that require predominately subspecialist care. Prior research has shown that pediatric subspecialists rarely code for obesity, and thus, an obesity diagnosis may be missed in children born preterm who tend to receive much of their care from subspecialists [14].
Despite the potential risk for undiagnosed obesity among children born preterm, to the authors’ knowledge, the extent to which obesity is underdiagnosed in pediatric clinical care settings among children born preterm compared to children born at term has not been examined. Thus, the purpose of this study was to examine the prevalence of undiagnosed obesity in clinical care settings while comparing children born at term to those born preterm using a large, longitudinal sample of children from a pediatric academic medical center. It was hypothesized that preterm birth status is associated with an increased risk of undiagnosed obesity.
Methods
Study design and setting
This was a longitudinal, retrospective cohort study that utilized inpatient and outpatient electronic medical records (EMRs) (Epic Systems, Madison, WI) and birth certificate data of children who were patients at Nationwide Children’s Hospital (NCH), a large pediatric academic medical center (Columbus, Ohio, USA), from January 2010–December 2016. The Institutional Review Board at NCH and the Ohio Department of Health (ODH) reviewed and approved the study, under a waiver of informed consent and HIPAA authorization.
Participants and study procedures
Data was electronically extracted from EMRs of all children born in 2010–2015, who had at least 1 weight measurement and either attended a NCH Primary Care Center or were born at <37 completed weeks’ gestation and attended any NCH clinic. These criteria were selected to maximize the total number of children born preterm available in the sample. Eligible children had a valid state of Ohio birth certificate linkable to their medical record and a gestational age ranging from 24 to 42 weeks, to determine their preterm status (Appendix A). Children were classified as preterm if their gestational age was less than 37 completed weeks and as term if their gestational age was 37 completed weeks or greater. If there were unresolvable discrepancies between the birth certificate and medical record regarding gestational age, children were removed from the sample (Fig. 1). Exclusion criteria were selected to omit children with conditions likely to significantly impact their growth, other than prematurity: multiple gestations or unknown plurality; a 9th or 10th revision International Classification of Diseases (ICD) code in the medical record indicating a major congenital malformation or endocrine or genetic disorder impacting growth (Supplementary Table 1); or serious chronic illness that necessitated more than 200 clinical encounters.
Fig. 1. Participant STROBE Flow Diagram, 2010–2016, Ohio, USA.
*Children were also removed here if all their measurements were below age 24 months and length values were below 45 cm or above 114 cm. The World Health Organization (WHO) SAS program does not calculate weight-for-length values outside of that range.
Data sources
Linked Ohio birth certificates served as the source of data for gestational age, birthweight, maternal and child dates of birth, maternal race and ethnicity, pre-pregnancy height and weight, maternal education, maternal smoking status during pregnancy, maternal diabetes (pre-pregnancy and gestational), marital status, and plurality. If sex, gestational age, or birthweight were missing from the birth certificate, the medical record served as the source of those data. Data extracted from the medical record included provider notes, the clinic or unit where the encounter occurred, anthropometric measurements, problem lists, and clinical diagnoses for billing purposes. Any encounter without a recorded weight measurement was excluded. If two or more weight or height measurements were recorded on the same day, only the first were included.
Measured obesity
After all weight and height (length for measurements taken at <24 months of age) measurements from clinical encounters were compiled, age and sex-specific z-scores and percentiles for height, weight, and, body mass index (BMI) were calculated based on child growth charts from the US Centers for Disease Control and Prevention (CDC) and growth standards from the World Health Organization (WHO) using SAS programs provided by the CDC and WHO, respectively [15]. In keeping with recommended clinical practice, WHO growth standards were used for measurements collected before age 24 months, and CDC growth charts were used for measurements collected at or after age 24 months. Z-scores and percentiles for children born preterm were calculated using their age adjusted for prematurity until they reached 24 months of age. Biologically implausible weight-for-length (WLPCT, <−5 or >5 z-scores) or BMI-for-age (BMIPCT, <−4 or >8 z-scores) values identified by the SAS programs were deleted [15, 16]. Children were identified as measured obese if they had two or more WLPCT values ≥95th percentile for clinical encounters that occurred at <24 months of age or sex-specific BMIPCT values ≥95th percentile for clinical encounters that occurred at or after 24 months of age. Two measurements in the obese range were required to ensure that a single elevated measurement was not an outlier or a temporary increase in weight.
Obesity diagnosis
A positive obesity diagnosis was identified using one of the two approaches. First, children were considered to have been diagnosed with obesity if they had a clinical encounter where the problem list had an obesity-related phrase or an ICD-9 or −10 code (Supplementary Table 2). To be inclusive, a broad range of obesity-related phrases were selected by the research team from prior obesity literature and from manually reviewing clinical problem lists provided to identify common ways that obesity was documented. If a child did not have an obesity-related phrase or ICD-9 or −10 code in the problem list, a free-text search query using Structured Query Language (SQL) was performed to examine their provider notes for the presence of obesity-related terms. This has been identified in the literature as an effective method to ascertain clinical diagnoses from electronic medical records [17]. The same list of obesity-related phrases applied to the problem list search were applied during this approach. To minimize false positives, two members of the research team (TTI and SAK) independently reviewed a subset of the provider notes to create a coding scheme for determining true positives. Using that information, TTI manually reviewed the provider notes identified by the free-text search query and coded obesity diagnoses accordingly. If an obesity-related phrase was found across multiple clinical encounters for the same child, the earliest encounter was used for descriptive analyses. Children commonly identified as false positives were those who had a family history of obesity or an automated review of systems or 5–2-1–0 healthy habits question in the medical record (e.g., “Child has no history of abnormal weight gain or loss”).
Statistical analysis
Univariate statistics were applied to characterize the sample and the Chi-square, Fisher’s exact, and Wilcoxon rank sum tests were used to compare outcomes between children born preterm and children born at term and between diagnosed and undiagnosed children. Modified Poisson regression models were used to estimate associations (risk ratios) between preterm status and obesity diagnosis. Potential effect measure modifiers selected a priori based on their established relationship to preterm delivery and also the potential for obesity underdiagnosis in the healthcare setting were tested by modeling interaction terms (sex, maternal race and ethnicity, maternal education, marital status, health insurance, and pre-pregnancy BMI, maternal diabetes, and maternal smoking status). A covariate was considered an effect modifier if the interaction p value was <0.20, and stratified results were reported accordingly [18, 19]. However, given the vast racial and socioeconomic disparities in childhood obesity, results were stratified by maternal race and maternal education irrespective of statistical significance [20, 21]. Covariates with a p value ≤ 0.05 were included as confounders in the model. Data analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).
Results
EMR data for 59,198 children were assembled, of which 48,206 children with 1,207,434 clinical encounters were matched with an Ohio birth certificate and screened for eligibility (Supplementary Table 3). Children whose gestational age was missing or unresolvable (n = 122), those who were multiples or of an unknown plurality (n = 3023), those who had one or more excluded ICD codes (n = 1119), those with an illness that necessitated more than 200 clinical encounters (n = 213), and those without at least two weight and height measurements (n = 4880) were excluded. The final sample consisted of 38,849 children, with 31,548 children and 313,592 clinical encounters in the term group and 7301 children and 71,134 clinical encounters in the preterm group (Fig. 1). Overall, 51.8% of the children were male, with slightly fewer males who were born at term than those born preterm (Table 1). The median number of clinical encounters per child was 9 (Interquartile Range (IQR) 5, 13). The median maternal age at the child’s birth was 25 (IQR 22, 30) years, and 53.0% of the mothers attained a high school diploma/GED or attended some college.
Table 1.
Description of study sample, by term vs preterm status, 2010–2016, Ohio, USA.
| Characteristics | Term group | Preterm group | Total |
|---|---|---|---|
|
| |||
| Number of children (n, %) | 31,548 (81.2) | 7301 (18.8) | 38,849 (100) |
| Sex (n, %) | |||
| Female | 15,491 (49.1) | 3224 (44.2) | 18,715 (48.2) |
| Male | 16,057 (50.9) | 4077 (55.8) | 20,134 (51.8) |
| Child’s birth year (n, %) | |||
| 2010 | 4947 (15.7) | 1131 (15.5) | 6078 (15.7) |
| 2011 | 5145 (16.3) | 1186 (16.2) | 6331 (16.3) |
| 2012 | 5159 (16.4) | 1208 (16.6) | 6367 (16.4) |
| 2013 | 5461 (17.3) | 1310 (17.9) | 6771 (17.4) |
| 2014 | 5533 (17.5) | 1270 (17.4) | 6803 (17.5) |
| 2015 | 5303 (16.8) | 1196 (16.4) | 6499 (16.7) |
| Number of encounters per child, (median, IQR) | 9 (6, 13) | 8 (4, 13) | 9 (5, 13) |
| Number of encounters (n, %) | 313,573 (81.5) | 71,129 (18.5) | 384,702 (100) |
| Number of encounters by clinic (n, %) | |||
| Inpatient NICU | 934 (0.3) | 5728 (8.1) | 6662 (1.7) |
| Inpatient Non-NICU | 7143 (2.3) | 3625 (5.1) | 10,768 (2.8) |
| Outpatient Neonatology | 934 (0.3) | 9820 (13.8) | 10,754 (2.8) |
| Outpatient Primary Care | 250,185 (79.8) | 31,876 (44.8) | 282,061 (73.3) |
| Outpatient Subspecialty | 54,377 (17.3) | 20,080 (28.2) | 74,457 (19.4) |
| Maternal race (n, %) | |||
| White | 9883 (31.3) | 4068 (55.7) | 13,951 (35.9) |
| Black/African-American | 16,704 (53.0) | 2552 (35.0) | 19,256 (49.6) |
| Other | 4961 (15.7) | 681 (9.3) | 5642 (14.5) |
| Maternal ethnicity (n, %) | |||
| Hispanic | 4282 (13.6) | 521 (7.1) | 4803 (12.4) |
| Non-Hispanic | 26,199 (83.0) | 6 598 (90.4) | 32,797 (84.4) |
| Missing | 1067 (3.4) | 182 (2.5) | 1249 (3.2) |
| Maternal age at birth, years (median, IQR) | 25 (21,30) | 26 (22,31) | 25 (22,30) |
| Maternal education (n, %) | |||
| 12th Grade or Less | 10110 (32.1) | 1717 (23.5) | 11,827 (30.4) |
| High School/ GED / Some college but no degree | 16,733 (53.0) | 3839 (52.6) | 20,572 (53.0) |
| Associate’s Degree or Higher | 3441 (10.9) | 1563 (21.4) | 5004 (12.9) |
| Missing | 1264 (4.0) | 182 (2.5) | 1446 (3.7) |
| Marital status (n, %)a | |||
| Yes | 10,297 (32.6) | 2878 (39.4) | 13,175 (33.9) |
| No | 21,251 (67.4) | 4423 (60.6) | 25,674 (66.1) |
| Ever Medicaid insured or uninsured (n, %) | |||
| Yes | 30,566 (96.9) | 5969 (81.8) | 36,535 (94.0) |
| No | 982 (3.1) | 1332 (18.2) | 2314 (6.0) |
| Maternal pre-pregnancy BMI (n, %) | |||
| Underweight | 1427 (4.5) | 423 (5.8) | 1850 (4.8) |
| Normal | 10,888 (34.5) | 2602 (35.6) | 13,490 (34.7) |
| Overweight | 7892 (25.0) | 1631 (22.3) | 9,523 (24.5) |
| Obesity | 8002 (25.4) | 2050 (28.1) | 10,052 (25.9) |
| Missing | 3339 (10.6) | 595 (8.2) | 3934 (10.1) |
| Maternal diabetes (n, %)b | |||
| Yes | 2159 (6.8) | 861 (11.8) | 3020 (7.8) |
| No | 28,985 (91.9) | 6272 (85.9) | 35,257 (90.7) |
| Missing | 404 (1.3) | 168 (2.3) | 572 (1.5) |
| Maternal smoking status (n, %)c | |||
| Yes | 4579 (14.5) | 1327 (18.2) | 5906 (15.2) |
| No | 26,212 (83.1) | 5827 (79.8) | 32,039 (82.5) |
| Missing | 757 (2.4) | 147 (2.0) | 904 (2.3) |
Marital status is defined as ever married during birth, conception, or any time between.
Maternal diabetes is defined as ever had pre-pregnancy or gestational diabetes.
Maternal smoking status is defined as smoking status before and during pregnancy.
Risk of undiagnosed obesity among children who measured obese
Overall, 13,697 (35.3%) children were measured obese. Of those, 3424 (25.0%) of them were diagnosed with obesity. When stratified by birth status, only 2981 (26.3%) of children born at term and 443 (18.8%) children born preterm had a diagnosis of obesity. Consequently, 10273 (75.0%) children who measured as obese were undiagnosed. This proportion was higher among children born preterm (76.5%) compared to those born at term (63.0%) (Supplementary Table 4).
Results of modified Poisson regression analyses examining associations between preterm status and being undiagnosed among children who measured as obese are shown in Table 2. Obese children born preterm were 8% more likely to be undiagnosed compared to obese children born at term (RR = 1.08 95% CI 1.05, 1.11) after adjustment for prenatal and socio-demographic confounders. No effect measure modifiers were found (all p values for interaction were >0.20). Results from stratified models showed that preterm children born to white mothers had a 12.0% increased risk of undiagnosed obesity while preterm children born to black mothers had a 4.0% increased risk of undiagnosed obesity. Similarly, moderate differences in effect estimates by maternal education were also found. Confounders only slightly attenuated effect estimates.
Table 2.
Associations between term vs preterm birth status and the risk of undiagnosed obesity among children measured obese (n = 13,697), 2010–2016, Ohio, USA.
| Characteristics | Term (n = 11,343)a |
Preterm (n = 2354) |
Unadjusted RR (95% CI) | Adjusted RRb (95% CI) | ||
|---|---|---|---|---|---|---|
| Not diagnosedc | Diagnosed | Not diagnosedc | Diagnosed | |||
|
| ||||||
| All children (n, %) | 8362 (73.7) | 2981 (26.3) | 1 911 (81.2) | 443 (18.8) | 1.10 (1.08, 1.13) | 1.08 (1.05, 1.11) |
| Maternal race (n, %) | ||||||
| White | 2594 (74.0) | 914 (26.0) | 994 (84.9) | 177 (15.1) | 1.15 (1.11, 1.18) | 1.12 (1.09, 1.16) |
| Black/African-American | 4309 (75.3) | 1414 (24.7) | 717 (79.1) | 189 (20.9) | 1.05 (1.01, 1.09) | 1.04 (1.00, 1.08) |
| Others | 1459 (69.1) | 653 (30.9) | 200 (72.2) | 77 (27.8) | 1.04 (0.96, 1.13) | 1.06 (0.97, 1.16) |
| Maternal education (n, %) | ||||||
| 12th Grade or Less | 2842 (72.6) | 1071 (27.4) | 475 (78.6) | 129 (21.4) | 1.08 (1.03, 1.13) | 1.07 (1.16, 1.33) |
| High School/GED/Some college but no degree | 4316 (73.9) | 1521 (26.1) | 1018 (80.1) | 253 (19.9) | 1.08 (1.05, 1.12) | 1.06 (1.11, 1.26) |
| Associate’s Degree or Higher | 851 (77.4) | 249 (22.6) | 357 (88.6) | 46 (11.4) | 1.15 (1.09, 1.20) | 1.11 (1.06, 1.17) |
Missing data: 209 for maternal education.
The reference group is the term group of children.
Adjusted RR: Adjusted for maternal ethnicity, maternal pre-pregnancy BMI, ever Medicaid or uninsured, and child’s sex.
Not diagnosed means that a child was not diagnosed during the study period (anytime from 2010 to 2016).
Characteristics of diagnosed children
Among children diagnosed with obesity (n = 3424), the median number of clinical encounters per child was 12 (IQR 9, 12) for those born at term and 13 (IQR 9, 20) for those born preterm (Table 3). Children with obesity were most commonly diagnosed at 2 years of age (24–35 months of age), which was higher for children who were born at term (33.6%) compared to those born preterm (26.0%). For children who were measured obese as infants (age 0: after birth—11 months of age), the median time to a diagnosis was 26 months (IQR 17, 38), which varied slightly by preterm status. The time to diagnosis shortened considerably for children whose obesity first emerged at older ages. Nevertheless, it was typical for children diagnosed at any age post-infancy to have experienced over a one-year lapse of time between their first measurement as obese and a diagnosis. The percentage of children born at term that were diagnosed in a Primary Care clinic (84.6%) was higher than that of children born preterm (63.0%), and the percentage of term children diagnosed at an outpatient subspecialty clinic (11.5%) was lower than that of preterm children (18.1%). This pattern generally reflected where these groups of children tended to receive their care.
Table 3.
Characteristics of children diagnosed with obesity, by term vs preterm status (n = 3424), 2010–2016, Ohio, USA.
| Characteristics | Term | Preterm | Overall | P value |
|---|---|---|---|---|
|
| ||||
| Number of children (n, %) | 2981 (87.0) | 443 (12.9) | 3424 (100) | |
| Age at diagnosis, months (median, IQR) | 32 (24, 43) | 28 (15, 42) | 31 (24, 43) | 0.0003a |
| Number of children, by age at diagnosis, (n, %) | <0.0001b | |||
| Age 0 at diagnosis | 255 (8.6) | 91 (20.6) | 346 (10.1) | |
| Age 1 at diagnosis | 392 (13.2) | 59 (13.3) | 451 (13.2) | |
| Age 2 at diagnosis | 1002 (33.6) | 115 (26.0) | 1117 (32.6) | |
| Age 3 at diagnosis | 726 (24.3) | 83 (18.7) | 809 (23.6) | |
| Age 4 at diagnosis | 415 (13.9) | 60 (13.5) | 475 (13.9) | |
| Age 5 at diagnosis | 154 (5.2) | 26 (5.9) | 180 (5.3) | |
| Age 6 at diagnosis | 37 (1.2) | 9 (2.0) | 46 (1.3) | |
| Time to diagnosis since first obese measurement, months (median, IQR) | 23 (11, 35) | 19 (4, 34) | 22 (10, 35) | 0.0002a |
| Time to diagnosis, months, by age when first measured as obese (median, IQR) | ||||
| Age 0 when first measured as obese | 27 (19, 39) | 24 (7, 36) | 26 (17, 38) | <0.0001a |
| Age 1 when first measured as obese | 20 (12, 31) | 18 (6, 35) | 20 (11, 31) | 0.18a |
| Age 2 when first measured as obese | 12 (5, 22) | 14 (5, 28) | 13 (5, 23) | 0.58a |
| Age 3 when first measured as obese | 1 (0, 12) | 1 (1, 4) | 1 (0, 12) | 0.69a |
| Age 4 when first measured as obese | 1 (0, 8) | 1 (0, 3) | 1 (0, 7) | 0.26a |
| Age 5 when first measured as obese | 0 (0, 3) | 1 (0, 3) | 0 (0, 3) | 0.16a |
| Age 6 when first measured as obese | 0 (0, 1) | 7 (7, 7) | 0 (0, 1) | 0.12a |
| Time to diagnosis, months, by age when diagnosed as obese (median, IQR) | ||||
| Age 0 when diagnosed as obese | 2 (−1, 6) | 2 (0, 4) | 2 (0, 5) | 0.94a |
| Age 1 when diagnosed as obese | 12 (8, 16) | 11 (5, 16) | 12 (7, 16) | 0.40a |
| Age 2 when diagnosed as obese | 22 (16, 25) | 22 (15, 26) | 22 (16, 25) | 0.81a |
| Age 3 when diagnosed as obese | 34 (24, 28) | 33 (17, 36) | 34 (24, 37) | 0.07a |
| Age 4 when diagnosed as obese | 44 (25, 50) | 46 (30, 50) | 44 (26, 50) | 0.68a |
| Age 5 when diagnosed as obese | 55 (22, 61) | 57 (44, 61) | 55 (24, 61) | 0.81a |
| Age 6 when diagnosed as obese | 60 (13, 71) | 54 (51, 65) | 57 (13, 70) | 0.80a |
| Number of encounters per child (median, IQR) | 12 (9, 17) | 13 (9, 20) | 12 (9, 17) | 0.02a |
| Number of encounters (n, %) | 41 130 (85.7) | 6 891 (14.3) | 48 021 (100) | |
| Number of encounters by clinic (n, %) | <0.0001b | |||
| Inpatient NICU | 98 (0.2) | 416 (6.0) | 514 (1.1) | |
| Inpatient Non-NICU | 855 (2.1) | 232 (3.4) | 1 087 (2.3) | |
| Outpatient Neonatology | 91 (0.2) | 727 (10.6) | 818 (1.7) | |
| Outpatient Primary Care | 32,717 (79.6) | 3825 (55.5) | 36,542 (76.0) | |
| Outpatient Subspecialty | 7369 (17.9) | 1 691 (24.5) | 9 060 (18.9) | |
| Clinic that made obesity diagnosis (n, %) | <0.001c | |||
| Inpatient NICU | 0 (0.0) | 10 (2.3) | 10 (0.0) | |
| Inpatient Non-NICU | 10 (0.3) | 3 (0.7) | 13 (0.2) | |
| Outpatient ED/UC | 20 (0.7) | 1 (0.2) | 21 (0.2) | |
| Outpatient Neonatology | 3 (0.0) | 57 (12.8) | 60 (0.1) | |
| Outpatient Primary Care | 2522 (84.6) | 279 (63.0) | 2801 (87.3) | |
| Outpatient Subspecialty | 426 (11.5) | 93 (18.1) | 519 (12.3) | |
P values were reported from tests used to compare differences between children born preterm and children born at term. For children born preterm, we used age adjusted for prematurity from birth to 24 months. Age 0:0–11 months; age 1:12–23 months; age 2:24–35 months; age 3:36–47 months; age 4:48–59 months; age 5:60–71 months; age 6:72–83 months.
Wilcoxon rank sum test.
Chi-square test.
Fisher’s exact test.
Characteristics of undiagnosed children
Among children who were measured obese but not diagnosed (n = 10,273), the median number of clinical encounters was 10 (IQR 7, 14) for those born at term and 11 (IQR 7, 16) for those born preterm (Table 4). Most children who were measured obese but were not diagnosed during the study period had their second obese measurement by 1 year of age, which was higher in those born at term (49.6%) compared to those born preterm (42.7%) groups. At the time of the second obese measurement, children were typically well above the 95th percentile. Specifically, the median BMI-for-age percentile at the time of second obese measurement was 97 (IQR 96, 98) for children born at term, similar to children born preterm 98 (IQR 96, 99). For both preterm and term children, primary Care and subspecialties were common settings for undiagnosed obesity. For instance, 44.0% of children born preterm were undiagnosed in Primary Care and 26.8% were undiagnosed in subspecialty clinics (81.9% and 14.5% of term children, respectively).
Table 4.
Characteristics of children measured as obese but not diagnosed, by term vs preterm (N = 10 273), 2010–2016, Ohio, USA.
| Characteristics | Term | Preterm | Overall | P value |
|---|---|---|---|---|
|
| ||||
| Number of children (n, %) | 8362 (81.4) | 1911 (18.6) | 10,273 (100) | |
| Age at second obese measurement, months (median, IQR) | 9 (4, 15) | 6 (3, 15) | 9 (4, 15) | <0.0001a |
| Number of children, by age at second obese measurement (n, %) | <0.0001b | |||
| Age 0 at second obese measurement | 2752 (32.9) | 749 (39.2) | 3501 (34.1) | |
| Age 1 at second obese measurement | 4150 (49.6) | 816 (42.7) | 4966 (48.3) | |
| Age 2 at second obese measurement | 1042 (12.5) | 217 (11.4) | 1259 (12.3) | |
| Age 3 at second obese measurement | 184 (2.2) | 54 (2.8) | 238 (2.3) | |
| Age 4 at second obese measurement | 133 (1.6) | 45 (2.3) | 178 (1.7) | |
| Age 5 at second obese measurement | 77 (0.9) | 23 (1.2) | 100 (1.0) | |
| Age 6 at second obese measurement | 24 (0.3) | 7 (0.4) | 31 (0.3) | |
| cBMI-for-age at second obese measurement (median, IQR) | 97 (96, 98) | 98 (96, 99) | 97 (96, 98) | 0.002a |
| cWeight-for-length at second obese measurement (median, IQR) | 98 (96, 99) | 98 (97, 100) | 98 (96, 99) | <0.0001a |
| Number of encounters per child (median, IQR) | 10 (7, 14) | 11 (7, 16) | 10 (7, 14) | <0.0001a |
| Number of encounters (n, %) | 96,922 (79.5) | 25,038 (20.5) | 121,960 (100) | |
| Encounters by clinic (n, %) | <0.0001b | |||
| Inpatient NICU | 299 (0.3) | 2223 (8.9) | 2522 (2.1) | |
| Inpatient Non-NICU | 2636 (2.7) | 1390 (5.6) | 4026 (3.3) | |
| Outpatient Neonatology | 314 (0.3) | 3264 (13.0) | 3578 (2.9) | |
| Outpatient Primary Care | 75,703 (78.2) | 10,901 (43.5) | 86,604 (71.0) | |
| Outpatient Subspecialty | 17,970 (18.5) | 7260 (29.0) | 25,230 (20.7) | |
| Clinic where second obese measurement occurred (n, %) | <0.0001b | |||
| Inpatient NICU | 14 (0.2) | 192 (10.1) | 206 (2.0) | |
| Inpatient Non-NICU | 260 (3.1) | 121 (6.3) | 381 (3.7) | |
| Outpatient Neonatology | 26 (0.3) | 244 (12.8) | 270 (2.6) | |
| Outpatient Primary Care | 6849 (81.9) | 841 (44.0) | 7690 (74.9) | |
| Outpatient Subspecialty | 1213 (14.5) | 513 (26.8) | 1726 (16.8) | |
P values were reported from tests used to compare differences between children born preterm and children born at term. For children born preterm, we used age adjusted for prematurity from birth to 24 months. Age 0:0–11 months; age 1:12–23 months; age 2:24–35 months; age 3:36–47 months; age 4:48–59 months; age 5:60–71 months; age 6:72–83 months.
Wilcoxon rank sum test.
Chi-square test.
BMI-for-age was reported for children who had their 2nd obese measurement at >2 years of age and Weight-for-length was reported for children who had their 2nd obese measurement at <2 years of age.
Discussion
This large, retrospective cohort study examined the association between preterm birth and undiagnosed obesity in early childhood across pediatric clinical care settings. More than three-fourths of children born preterm who had obese measurements were never diagnosed with obesity during clinical encounters, about 8% higher than children born at term who had obese measurements. In addition, some prenatal and sociodemographic factors magnified the risk of undiagnosed obesity. Specifically, preterm children born to white women or born to women who obtained an Associate’s degree or higher had a slightly higher risk of undiagnosed obesity. For both preterm and term children, Primary Care and subspecialties were common settings for undiagnosed obesity. For children who were diagnosed, the age at diagnosis and time to diagnosis was not largely different between children born preterm and children born at term.
The excess risk for undiagnosed obesity among children born preterm compared to children born at term was 8% in this study; however, this risk was slightly magnified for particular sub-groups of children born preterm such as children born to white or mothers with higher educational attainment. Some of these factors have previously been associated with an increased risk of undiagnosed obesity, for instance, minority children have been shown to have less undiagnosed obesity than white children [8, 12, 22–25]. It is unknown if these findings are due to provider bias or because clinicians look more carefully for obesity in minority and/or low-income children, since obesity is a common condition affecting around 22.0% of non-Hispanic black, 26.0% of Hispanic, and 19.0% of low-income children in the US [26–28].
Although research examining documentation of pediatric obesity among children born preterm is limited, there is some literature examining documentation among other pediatric populations [8, 10, 12, 22, 23, 29, 30]. Riley et al. [22] reported poor rates of documentation of obesity among children ages 2–18 years seen in outpatient general pediatrics, endocrinology and gastroenterology clinics, with documentation occurring in only 48.0% of the visits. However, they used obesity-related terms to define diagnoses and did not also consider ICD codes. This may have overestimated the true extent of obesity underdiagnosis in that population. In a nationally-representative sample (2005–7) of US children ages 2–18 years, Patel et al. [8] reported that only 18.0% of preventive care visits with a BMI ≥ 95th percentile had a documented obesity diagnosis. A more recent study found that only 21.5% of children ages 1–6 years were properly documented as obese despite utilizing ICD-9 codes and natural language processing (NLP) on provider notes to identify documentation [23]. However, this study’s focus was restricted to severe obesity (e.g., BMI ≥ 99th percentile) which is less common in the general pediatric population.
The present study is not directly comparable to the aforementioned studies not only because the interest was primarily children born preterm but also because the sample focused on early childhood and included children <24 months of age. One reason few studies have included this population is that there is no universally accepted definition of obesity for children <24 months of age, and clinicians are not generally encouraged to diagnose children during this period [11]. However, the perinatal period and infancy, particularly the “first 1000 days”, have been identified as a critical window for the development and prevention of childhood obesity, especially among those born preterm [31, 32]. A recent study of extremely preterm children found that many children with obesity in adolescence had elevated weight gain during the first 2 years of life [31]. These findings not only highlight the importance of the first “1000 days” but solidify the need to begin monitoring and diagnosing children <24 months of age. This a unique period of time when children have frequent contact with their healthcare providers during well-child visits. Furthermore, weight-management interventions are more likely to be effective among younger children, which makes an early diagnosis of obesity crucial [32, 33].
It is important that all clinicians realize that they can play a key role in identifying and diagnosing obesity, even if their specialization may not be regarded as a priority setting for obesity diagnoses. Although Primary Care was a common setting for undiagnosed obesity, subspecialties were as well, indicating that there is room for progress in regards to obesity screening and documentation [11]. However, the burden of obesity screening cannot be placed solely on Primary Care considering that some families, especially those with children born preterm, may have challenges in maintaining frequent Primary Care visits schedule alongside potentially frequent subspecialist follow-up care for chronic conditions [34]. Furthermore, prior research has shown that children born preterm are just as likely as other children to lack a medical home [35]. Thus, some children born preterm may receive some subspecialty care but lack primary care (or vice versa). As such, it becomes all clinicians’ role to identify and document obesity when applicable.
To the authors’ knowledge, no prior studies have examined obesity diagnosis rates in early childhood among children born preterm compared to those at term. Reasons why children born preterm are more likely to be undiagnosed are not clear, but may include factors such as competing health priorities, time constraints during patient encounters, and lack of published clinical guidelines regarding obesity screening in children born preterm [36, 37]. Clinicians may overlook obesity in children born preterm due to an early focus on survivorship and treatment of what may be considered more serious chronic conditions of prematurity like neurodevelopmental delay or chronic lung disease. Furthermore, clinicians may perceive obesity as uncommon in children born preterm because traditionally they have focused on the problem of postnatal growth failure in preterm infants, which can persist into early childhood [38, 39]. Because there is little consensus on the ideal postnatal growth patterns of preterm infants, clinicians have minimal guidance on screening and managing obesity in children born preterm [37].
Documentation of an obesity diagnosis in the medical record is the first step in managing pediatric obesity in early childhood. Children who are diagnosed with obesity are more likely to have an intervention plan in place and to be referred for specialized treatment [40]. This study found that despite ongoing attention to childhood obesity, many opportunities for obesity diagnoses in pediatric clinical care settings are being missed. Children born preterm, a special population at risk for obesity [31], were more likely to have their obesity go undiagnosed compared to children born at term. While past research has shown that overweight and obesity risk was restricted to late preterm children, recent studies have shown an emergence of overweight and obesity in early preterm children as well [31, 41]. Thus, it is imperative that clinicians are actively monitoring growth acceleration and obesity development among all children born preterm. It is unlikely that parents will present it to the clinician as a concern given that they tend to view their preterm children as underweight even when they are maintaining a healthy or excess weight status [42, 43].
Strengths and limitations
This study offers several strengths. First, is the use of a racially and socioeconomically diverse longitudinal sample of children from a large U.S. academic children’s hospital that includes more than 40,000 children and 700,000 measurements that span across infancy into early childhood. Children included in the sample received inpatient and/or outpatient care, providing greater diversity of participants and their clinical care patterns. Secondly, children’s EMR data was linked to their birth certificates to examine prenatal and sociodemographic risk factors based on their established relationship to preterm delivery and obesity (e.g., maternal BMI, maternal smoking history) [31, 32].
One limitation of the study is that anthropometric data were collected during routine pediatric clinical care, not specifically for research purposes and therefore, could have been subject to measurement error. However, NCH has implemented age-specific procedures for measuring and recording anthropometrics during clinical encounters that specify appropriate positioning of the patient and equipment use. Furthermore, past studies have verified that anthropometrics recorded during pediatric clinical care are of suitable quality for many research purposes [44, 45]. Second, children may have received some of their care outside of NCH, including primary care. Therefore, some children may have had a diagnosis noted in a medical record external to NCH, and this information may or may not have been available to NCH clinicians (e.g., via parent report or transmission of records). Nevertheless, documentation of obesity elsewhere would not alleviate the need for obesity screening and documentation in NCH records as well. Third, it is possible that a child’s BMI could have fluctuated between the obese and non-obese range from the time of the initial obese measurement to the time of diagnosis, which could have overestimated the time from an obese measurement to diagnosis. Last, this was a single-site study, so findings may not be generalizable to other populations. For instance, there are vast black-white racial differences in the sample when comparing children born preterm to those born at term. Differences across groups are expected given the vast racial disparities in preterm birth in the US. However, because there are no prior studies of similar purpose and size, it is unclear if the racial differences in our sample are typical for large pediatric academic centers or specific to the Nationwide Children’s Hospital system.
Conclusion
In conclusion, preterm birth was associated with an increased risk of undiagnosed obesity in early childhood. These findings highlight the need to enhance obesity screening and management guidelines, particularly in the preterm population. Additional research is also needed to clarify and understand the mechanisms by which obesity in children born preterm may be more likely to go undiagnosed.
Supplementary Material
Acknowledgements
We thank the Nationwide Children’s Hospital Electronic Data Warehouse, Research Information Solutions and Innovation (RISI), and the Ohio Department of Health for its support services. SAK, MAK, ROF, KB study concept and design; TTI, SAK, and KB acquisition of data; TTI, RL, SAK, MAK, ROF, JR analysis and interpretation of data; TTI and RR wrote the paper; TTI, RL, RR, MAK, ROF JR, KMB, SAK revised the manuscript for important intellectual content; SAK had primary responsibility for the final content. The authors have no conflicts to disclose. The study sponsors did not have a role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation or approval of the manuscript; and decision to submit the manuscript for publication. This study includes data provided by the Ohio Department of Health which should not be considered an endorsement of this study or its conclusions. All authors have read and approved the final version of the manuscript.
Funding
This study was supported by grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD)/ National Institutes of Health (1R03HD084927-01A1).
Footnotes
Conflict of interest The authors declare no competing interests.
Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41366-021-00834-1.
References
- 1.Martin JA, Hamilton BE, Michelle JK, Osterman MHS Births in the United States, 2017. Hyattsville, MD: National Center for Health Statistics, U.S. Department of Health and Human Services; 2017. 1–8 p. [Google Scholar]
- 2.Stoll BJ, Hansen NI, Bell EF, Shankaran S, Laptook AR, Walsh MC, et al. Neonatal outcomes of extremely preterm infants From the NICHD neonatal research network. Pediatrics. 2010;126:443–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sipola-Leppänen M, Vääräsmäki M, Tikanmäki M, Matinolli H-M, Miettola S, Hovi P, et al. Cardiometabolic risk factors in young adults who were born preterm. Am J Epidemiol. 2015;181:861–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wang G, Johnson S, Gong Y, Polk S, Divall S, Radovick S, et al. Weight gain in infancy and overweight or obesity in childhood across the gestational spectrum: a prospective birth cohort study. Sci Rep. 2016;6:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.O’campo P, Burke JG, Culhane J, Elo IT, Eyster J, Holzman C, et al. Neighborhood deprivation and preterm birth among non-Hispanic Black and White women in eight geographic areas in the United States. Am J Epidemiol. 2007;167:155–63. [DOI] [PubMed] [Google Scholar]
- 6.Kaufman JS, Dole N, Savitz DA, Herring AH. Modeling community-level effects on preterm birth. Ann Epidemiol. 2003;13:377–84. [DOI] [PubMed] [Google Scholar]
- 7.Martoccio T, Brophy-Herb H, Miller A, Peterson K, Contreras D, Horodynski M, et al. Family Risk and Body Mass Index in Children Born Preterm or With Low Birth Weight. Society for Research in Child Development; March 21; Baltimore, Maryland: 2019. [Google Scholar]
- 8.Patel AI, Madsen KA, Maselli JH, Cabana MD, Stafford RS, Hersh AL. Underdiagnosis of pediatric obesity during outpatient preventive care visits. Academic Pediatrics. 2010;10:405–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Higgins A, McCarville M, Kurowski J, McEwen S, Tanz RR. Diagnosis and screening of overweight and obese children in a resident continuity clinic. Global pediatric health. 2014;1:2333794X14559396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lydecker JA, Grilo CM. The missed diagnosis and misdiagnosis of pediatric obesity. Psychother Psychosom. 2017;86:173–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Barlow SE. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics. 2007;120(Suppl 4):S164–92. [DOI] [PubMed] [Google Scholar]
- 12.Benson L, Baer HJ, Kaelber DC. Trends in the diagnosis of overweight and obesity in children and adolescents: 1999–2007. Pediatrics. 2009;123:e153–e8. [DOI] [PubMed] [Google Scholar]
- 13.Kuo DZ, Lyle RE, Casey PH, Stille CJ. Care system redesign for preterm children after discharge from the NICU. Pediatrics. 2017;139:e20162969. [DOI] [PubMed] [Google Scholar]
- 14.Yi-Frazier JP, Larison C, Neff JM, Grow HM, Liu LL. Obesity in pediatric specialty clinics: an underestimated comorbidity. Clin Pediatr. 2012;51:1056–62. [DOI] [PubMed] [Google Scholar]
- 15.Centers for Disease Control and Prevention. A SAS Program for the 2000 CDC Growth Charts (ages 0 to <20 years). 2000.
- 16.World Health Organization. WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards: Growth velocity based on weight, length and head circumference: Methods and development. Geneva: World Health Organization; 2009. [Google Scholar]
- 17.Bustamante R, Earles A, Murphy JD, Bryant AK, Patterson OV, Gawron AJ, et al. Ascertainment of aspirin exposure using structured and unstructured large-scale electronic health record data. Medical Care. 2019;57:e60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Greenland S Basic problems in interaction assessment. Environ Health Perspect. 1993;101:59–66. suppl 4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Marshall SW. Power for tests of interaction: effect of raising the type I error rate. Epidemiol Perspect Innov. 2007;4:4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Rossen LM, Schoendorf KC. Measuring health disparities: trends in racial– ethnic and socioeconomic disparities in obesity among 2-to 18-year old youth in the United States, 2001–2010. Ann Epidemiol. 2012;22:698–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Isong IA, Rao SR, Bind M-A, Avendaño M, Kawachi I, Richmond TK. Racial and ethnic disparities in early childhood obesity. Pediatrics. 2018;141:e20170865. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Riley MR, Bass NM, Rosenthal P, Merriman RB. Underdiagnosis of pediatric obesity and underscreening for fatty liver disease and metabolic syndrome by pediatricians and pediatric subspecialists. J Pediatr. 2005;147:839–42. [DOI] [PubMed] [Google Scholar]
- 23.Brady CC, Thaker VV, Lingren T, Woo JG, Kennebeck SS, Namjou-Khales B, et al. Suboptimal clinical documentation in young children with severe obesity at tertiary care centers. Int J Pediatics 2016;2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chen AY, Escarce JJ. Peer reviewed: family structure and childhood obesity, early childhood longitudinal study—kindergarten cohort. Prevent Chronic Dis. 2010;7(3):A50. [PMC free article] [PubMed] [Google Scholar]
- 25.Grow HMG, Cook AJ, Arterburn DE, Saelens BE, Drewnowski A, Lozano P. Child obesity associated with social disadvantage of children’s neighborhoods. Soc Sci Med. 2010;71:584–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Hales CM, Carroll MD, Fryar CD, Ogden CL. Prevalence of obesity among adults and youth: United States, 2015–2016. NCHS data brief, no 288 Hyattsville, MD: National Center for Health Statistics; 2017. [PubMed] [Google Scholar]
- 27.Ogden CL, Fakhouri TH, Carroll MD, Hales CM, Fryar CD, Li X, et al. Prevalence of obesity among adults, by household income and education—United States, 2011–2014. MMWR Morbidity Mortality Weekly Report. 2017;66:1369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ruiz M, Goldblatt P, Morrison J, Porta D, Forastiere F, Hryhorczuk D, et al. Impact of low maternal education on early childhood overweight and obesity in Europe. Paediatr Perinat Epidemiol. 2016;30:274–84. [DOI] [PubMed] [Google Scholar]
- 29.Dorsey KB, Wells C, Krumholz HM, Concato JC. Diagnosis, evaluation, and treatment of childhood obesity in pediatric practice. Arch Pediatr Adolesc Med. 2005;159:632–8. [DOI] [PubMed] [Google Scholar]
- 30.O’Brien SH, Holubkov R, Reis EC. Identification, evaluation, and management of obesity in an academic primary care center. Pediatrics. 2004;114:e154–e9. [DOI] [PubMed] [Google Scholar]
- 31.Wood CT, Linthavong O, Perrin EM, Leviton A, Allred EN, Kuban KCK, et al. Antecedents of obesity among children born extremely preterm. Pediatrics. 2018;142:e20180519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Baidal JAW, Locks LM, Cheng ER, Blake-Lamb TL, Perkins ME, Taveras EM. Risk factors for childhood obesity in the first 1,000 days: a systematic review. Am J Prevent Med. 2016;50:761–79. [DOI] [PubMed] [Google Scholar]
- 33.Zolotarjova J, Ten Velde G, Vreugdenhil A. Effects of multi-disciplinary interventions on weight loss and health outcomes in children and adolescents with morbid obesity. Obesity Rev. 2018;19:931–46. [DOI] [PubMed] [Google Scholar]
- 34.Coker TR, Chung PJ, Cowgill BO, Chen L, Rodriguez MA. Low-income parents’ views on the redesign of well-child care. Pediatrics. 2009;124:194–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Boone KM, Nelin MA, Chisolm DJ, Keim SA. Gaps and factors related to receipt of care within a medical home for toddlers born preterm. J Pediatr. 2019;207:161–8.e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Klein JD, Sesselberg TS, Johnson MS, O’Connor KG, Cook S, Coon M, et al. Adoption of body mass index guidelines for screening and counseling in pediatric practice. Pediatrics. 2010;125:265–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Villar J, Giuliani F, Barros F, Roggero P, Zarco IAC, Rego MAS, et al. Monitoring the postnatal growth of preterm infants: a paradigm change. Pediatrics. 2018;141:e20172467. [DOI] [PubMed] [Google Scholar]
- 38.Moyses HE, Johnson MJ, Leaf AA, Cornelius VR. Early parenteral nutrition and growth outcomes in preterm infants: a systematic review and meta-analysis. Am J Clin Nutr. 2013;97:816–26. [DOI] [PubMed] [Google Scholar]
- 39.Cooke RJ, Ainsworth SB, Fenton AC. Postnatal growth retardation: a universal problem in preterm infants. Arch Dis Childhood Fetal Neonatal Ed. 2004;89:428–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Walsh CO, Milliren CE, Feldman HA, Taveras EM. Factors affecting subspecialty referrals by pediatric primary care providers for children with obesity-related comorbidities. Clin Pediatr. 2013;52:777–85. [DOI] [PubMed] [Google Scholar]
- 41.Vohr BR, Heyne R, Bann CM, Das A, Higgins RD, Hintz SR. Extreme preterm infant rates of overweight and obesity at school age in the SUPPORT neuroimaging and neurodevelopmental outcomes cohort. J Pediatr. 2018;200:132–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Saleska JL, Sheppard K, Turner AN, Boone KM, Keim SA. Parental perceptions and behaviors regarding child weight status among toddlers born preterm. Am J Perinatol. 2020;37:525–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Murphy LT, Skinner AC, Check J, Warner DD, Perrin EM. Parental perceptions of weight status in preterm compared with term infants. Am J Perinatol. 2016;33:1371–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Oza-Frank R, Hade EM, Conrey EJ. Inter-rater reliability of Ohio school-based overweight and obesity surveillance data. J Acad Nutr Diet. 2012;112:1410–4. [DOI] [PubMed] [Google Scholar]
- 45.Smith N, Coleman KJ, Lawrence JM, Quinn VP, Getahun D, Reynolds K, et al. Body weight and height data in electronic medical records of children. Int J Pediatr Obes. 2010;5:237–42. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.