Drs Wilson and Samiei are co-founders of iED, a company which specialises in creating medical apps. There are no further patents, products in development or marketed products to declare. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: GS EP JW BW HS NC WJ. Performed the experiments: GS. Analyzed the data: GS. Contributed reagents/materials/analysis tools: GS. Wrote the paper: GS WJ. Affiliation with the commercial company iED who designed the app: BW HS.
Advancements in knowledge of obesity aetiology and mobile phone technology have created the opportunity to develop an electronic tool to predict an infant’s risk of childhood obesity. The study aims were to develop and validate equations for the prediction of childhood obesity and integrate them into a mobile phone application (App).
Anthropometry and childhood obesity risk data were obtained for 1868 UK-born White or South Asian infants in the Born in Bradford cohort. Logistic regression was used to develop prediction equations (at 6±1.5, 9±1.5 and 12±1.5 months) for risk of childhood obesity (BMI at 2 years >91st centile and weight gain from 0–2 years >1 centile band) incorporating sex, birth weight, and weight gain as predictors. The discrimination accuracy of the equations was assessed by the area under the curve (AUC); internal validity by comparing area under the curve to those obtained in bootstrapped samples; and external validity by applying the equations to an external sample. An App was built to incorporate six final equations (two at each age, one of which included maternal BMI). The equations had good discrimination (AUCs 86–91%), with the addition of maternal BMI marginally improving prediction. The AUCs in the bootstrapped and external validation samples were similar to those obtained in the development sample. The App is user-friendly, requires a minimum amount of information, and provides a risk assessment of low, medium, or high accompanied by advice and website links to government recommendations.
Prediction equations for risk of childhood obesity have been developed and incorporated into a novel App, thereby providing proof of concept that childhood obesity prediction research can be integrated with advancements in technology.
Childhood obesity is one of the most daunting global public health threats
The aetiology of childhood obesity is complex
The advancing knowledge of risk factors for childhood obesity and in mobile phone technology has created the opportunity to develop an electronic tool that predicts during infancy an individual’s risk of becoming obese. In the UK, growth monitoring in infancy is part of routine National Health Service (NHS) care
The present study aims to (1) develop prediction equations which can be used during infancy for the early identification of risk of childhood obesity, (2) validate the prediction equations internally using statistical methods and externally in a different population, and (3) integrate the equations into a novel user-friendly App.
The sample comprised 1868 participants in the Born in Bradford (BiB) birth cohort study (
Weight and length at six, 12, and 24 months of age were measured by trained study workers as part of the BiB 1000 assessment schedule. Weight in kilograms (kg) was assessed using Seca baby scales and length in centimetres (cm) using a standard neonatometer (both from Harlow Health Care, London UK). These data were supplemented by infant weight and length measurements collected by health visitors as part of routine NHS care. At the beginning of BiB, a measurement protocol/standard was produced and all health workers received training in anthropometry
Data on childhood obesity risk factors were obtained from a number of sources. Maternal height (cm), ethnicity (
Logistic regression was used to develop the prediction equations. Anthropometric data were converted to age- and sex-adjusted z-scores by comparison to the UK90 reference
To ensure that the App could be utilised in children over a wide range of ages in infancy, we developed three series of equations, the first to be used at 6±1.5 months (equation 1), the second at 9±1.5 months (equation 2), and the third at 12±1.5 months (equation 3). The App would therefore be able to assess risk for childhood obesity in infants aged 4.5 to 13.5 months. Sample selection was based on complete covariate data, birth weight and weight/length data at age two years (±2 months) in addition to weight data in at least one of the three age periods when assessment would take place. The number of infants in each prediction equation differed slightly, but was always greater than 700 (
Equation 1 sample | Equation 2 sample | Equation 3 sample | |
N = 1022 | N = 1528 | N = 731 | |
Source of data | |||
BiB 1000 data | 310 (30.3) | 281 (18.4) | 287 (39.3) |
Routine NHS data | 712 (69.7) | 1247 (81.6) | 444 (60.7) |
Baby’s sex | |||
Boys | 538 (52.6) | 785 (51.4) | 368 (50.3) |
Girls | 484 (47.4) | 743 (48.6) | 363 (49.7) |
Birthweight z-score, mean (SD) | |||
Boys | −0.57 (1.2) | −0.56 (1.2) | −0.58 (1.2) |
Girls | −0.56 (1.2) | −0.52 (1.2) | −0.58 (1.2) |
Weight change z-score, mean (SD) |
|||
Boys | 0.07 (0.9) | 0.06 (1.0) | 0.02 (0.9) |
Girls | −0.08 (1.0) | −0.06 (1.1) | −0.02 (1.1) |
Infant BMI z-score at 2 years, mean (SD) | |||
Boys | −0.03 (1.1) | −0.03 (1.1) | −0.08 (1.1) |
Girls | −0.08 (1.2) | −0.03 (1.2) | −0.01 (1.1) |
Ethnicity | |||
White British | 492 (48.1) | 654 (42.8) | 331 (45.3) |
South Asian | 530 (51.9) | 874 (57.2) | 400 (54.7) |
Maternal BMI, mean (SD) | 25.9 (5.6) | 26.0 (5.6) | 25.8 (5.5) |
Mother’s education |
|||
<5 GCSE equivalent | 190 (18.6) | 330 (21.6) | 142 (19.4) |
≥5 GCSE equivalent | 310 (30.3) | 472 (30.9) | 222 (30.4) |
‘A’ level equivalent | 172 (16.8) | 233 (15.3) | 124 (17.0) |
Degree level equivalent | 291 (28.5) | 406 (26.6) | 195 (26.7) |
Other | 59 (5.8) | 87 (5.7) | 48 (6.6) |
Smoked during pregnancy |
136 (13.3) | 201 (13.2) | 97 (13.3) |
Gestational diabetes |
76 (7.4) | 113 (7.4) | 58 (7.9) |
Gestational age (<37 weeks) |
47 (4.6) | 75 (4.9) | 33 (4.5) |
Obesity and rapid weight gain at 2 years |
83 (8.1) | 121 (7.9) | 61 (8.3) |
Weight change z-score from birth to 6, 9 or 12 months in equation 1, 2 and 3 samples respectively.
GSCE = General Certificate of Secondary Education; A-level = Advanced level.
Dichotomised
Infant BMI >91st centile and growth from birth to 2 years of age ≥1 centile band.
Values are numbers (percentages) unless stated otherwise.
All potential predictors were entered into backward stepwise multivariable models that retained predictors with a p-value <0.05. These models tested possible interactions of sex by birth weight, sex by conditional infant weight z-score gain, ethnicity by birth weight, and ethnicity by conditional infant weight z-score gain. Individual risk prediction scores were calculated using the coefficients (where α is the constant and β1 to βk is a vector of predictors) from each of the three final prediction equations:
Sensitivity, specificity, and positive and negative predictive values (PPV and NPV, respectively) were calculated at risk score distribution cut-off points of 10%, 20% and 30% and area under the curves (AUCs) for the final logistic regression models were obtained to quantify the overall discrimination of the equations.
All analyses were conducted using Stata/SE v12
Internal validity was assessed using bootstrapping methods
External validity was assessed by applying the equations to an external sample and calculating the AUCs. We used data from the Children in Focus (CiF) subsample of the Avon Longitudinal Study of Parents and Children (ALSPAC)
Equation 2 sample | Equation 3 sample | |
N = 880 | N = 867 | |
Baby’s sex | ||
Boys | 481 (54.7) | 470 (54.2) |
Girls | 399 (45.3) | 397 (45.8) |
Birthweight z-score, mean (SD) | ||
Boys | −0.10 (1.08) | −0.09 (1.06) |
Girls | −0.01 (1.01) | −0.02 (1.01) |
Weight change z-score, mean (SD) |
||
Boys | 0.01 (0.98) | 0.30 (0.96) |
Girls | −0.01 (0.94) | −0.35 (0.92) |
Infant BMI z-score at 2 years | ||
Boys | 0.18 (0.98) | 0.17 (0.98) |
Girls | 0.25 (0.93) | 0.26 (0.93) |
Ethnicity | ||
White | 865 (98.3) | 852 (98.3) |
Other | 15 (1.7) | 15 (1.7) |
Maternal BMI, mean (SD) | 23.4 (4.0) | 23.4 (4.1) |
Mother’s education |
||
<5 GCSE equivalent | 92 (10.5) | 91 (10.5) |
≥ GCSE equivalent | 323 (36.7) | 318 (36.7) |
‘A’ level equivalent | 240 (27.3) | 236 (27.2) |
Degree level equivalent | 132 (15.0) | 130 (15.0) |
Other | 93 (10.6) | 92 (10.6) |
Smoked during pregnancy |
150 (17.1) | 146 (16.8) |
Gestational diabetes |
3 (0.3) | 2 (0.2) |
Gestational age (<37 weeks) |
36 (4.1) | 35 (4.0) |
Obesity and rapid weight gain at 2 years |
84 (9.6) | 84 (9.7) |
Weight change z-score from birth to 9 or 12 months in equation 2 and 3 samples respectively.
GSCE = General Certificate of Secondary Education; A-level = Advanced level.
Dichotomised
Infant BMI >91st centile and growth from birth to 2 years of age ≥1 centile band.
Values are numbers (percentages) unless stated otherwise. Equation 1 could not be validated due to insufficient numbers.
The prevalence of childhood obesity in the BiB 1000 and ALSPAC samples respectively was 8.1% in equation 1 (insufficient data in the ALSPAC sample), 7.9% and 9.6% in equation 2, and 8.3% and 9.7% in equation 3.
Equation 1 | Equation 2 | Equation 3 | |
Birthweight z-score | 2.09 (1.59, 2.75) | 1.67 (1.36, 2.05) | 2.28 (1.64, 3.12) |
Weight change z-score | 4.45 (3.28, 6.04) | 4.48 (3.52, 5.72) | 8.80 (5.45, 14.21) |
Maternal BMI | 1.05 (1.00, 1.09) | 1.05 (1.01, 1.09) | |
South Asian ethnicity |
1.80 (1.05, 3.11) | ||
Gestational age (<37 weeks) |
0.26 (0.07, 0.96) |
Reference categories:
White British;
Gestational age ≥37 weeks.
Values are odds ratios with 95% confidence intervals.
The final multivariable bootstrap model for the equation 1 sample demonstrated statistical significance of birthweight z-score, weight gain z-score and maternal BMI, but not gestational age and ethnicity. However, the AUC (95% CI) for the bootstrapped model was the same as for the development model (85.8% (81.6–90.0%)). The bootstrapped models for equations 2 and 3 retained the same variables as the development models, with no change in AUCs.
Due to insufficient numbers equation 1 could not be validated using ALSPAC data. Equations 2 and 3 had AUCs (95% CI) of 85.0% (81.0–89.1%) and 88.6% (85.2–92.0%) respectively when applied to the ALSPAC sample.
As birthweight z-score and weight gain z-score were significant predictors of childhood obesity in the development and validation models, they were selected as covariates in the sex-adjusted prediction model for the App. Discrimination accuracy of the risk scores for predicting childhood obesity was excellent, equation 1: AUC 85.3% (95% CI 81.0–89.6%), equation 2: AUC 85.7% (82.4–89.0%) and equation 3: AUC 91.1% (87.9–94.4%). The coefficients used to derive the childhood obesity risk scores from the final multivariable regression model are presented in
Coefficient values | ||||
Variable | Equation 1 | Equation 2 | Equation 3 | |
Α | Constant | −3.718 | −3.542 | −3.937 |
β1 | Female sex | 0.488 | 0.288 | 0.234 |
β2 | Birthweight z-score | 0.599 | 0.551 | 0.824 |
β3 | Weight z-score change |
1.501 | 1.508 | 2.174 |
Weight change z-score from birth to 6, 9 or 12 months in equation 1, 2 and 3 samples respectively.
Probability childhood obesity = 1/(1+ e -[α+β1+ β2+ β3]).
Proportion of the populationabove risk score threshold (%) | Risk scorethreshold | Sensitivity %(95% CI) | Specificity %(95% CI) | PPV %(95% CI) | NPV %(95% CI) |
Equation 1 | |||||
30 | 0.0731 | 78.3 (67.9, 86.6) | 74.2 (71.3, 77.0) | 21.2 (16.7, 26.2) | 97.5 (96.1, 98.5) |
20 | 0.1155 | 69.9 (58.8, 79.5) | 84.3 (81.9, 86.6) | 28.3 (22.2, 35.0) | 96.9 (95.5, 98.0) |
10 | 0.2072 | 50.6 (39.4, 61.8) | 93.6 (91.9, 95.1) | 41.2 (31.5, 51.4) | 95.5 (94.0, 96.8) |
Equation 2 | |||||
30 | 0.0662 | 77.7 (69.2, 84.8) | 73.9 (71.5, 76.2) | 20.4 (16.8, 24.4) | 97.5 (96.3, 98.3) |
20 | 0.1104 | 68.6 (59.5, 76.7) | 84.1 (82.1, 86.0) | 27.0 (22.1, 32.4) | 96.9 (95.8, 97.8) |
10 | 0.2082 | 53.7 (44.4, 62.8) | 93.5 (92.1, 94.8) | 41.7 (33.8, 49.8) | 95.9 (94.7, 96.9) |
Equation 3 | |||||
30 | 0.0609 | 86.9 (75.8, 94.2) | 75.1 (71.6, 78.3) | 24.1 (18.6, 30.3) | 98.4 (96.9, 99.3) |
20 | 0.1065 | 77.0 (64.5, 86.8) | 85.1 (82.1, 87.7) | 32.0 (24.5, 40.2) | 97.6 (96.0, 98.7) |
10 | 0.2391 | 65.6 (52.3, 77.3) | 94.9 (93.0, 96.5) | 54.1 (42.1, 65.7) | 96.8 (95.2, 98.0) |
PPV = positive predictive value; NPV = negative predictive value.
The coefficients used to derive the risk scores for a further three models which additionally included maternal BMI are presented in
Coefficient values | ||||
Variable | Equation 1 | Equation 2 | Equation 3 | |
α | Constant | −4.920 | −4.745 | −4.625 |
β1 | Female sex | 0.493 | 0.255 | 0.230 |
β2 | Birthweight z-score | 0.577 | 0.505 | 0.798 |
β3 | Weight change z-score |
1.494 | 1.501 | 2.149 |
β4 | Maternal BMI | 0.044 | 0.046 | 0.026 |
Weight change z-score from birth to 6, 9 or 12 months in equation 1, 2 and 3 samples respectively.
Proportion of the populationabove risk score threshold (%) | Risk scorethreshold | Sensitivity %(95% CI) | Specificity %(95% CI) | PPV %(95% CI) | NPV %(95% CI) |
Equation 1 | |||||
30 | 0.0696 | 81.9 (72.0, 89.5) | 74.5 (71.6, 77.3) | 22.1 (17.6, 27.2) | 97.9 (96.6, 98.8) |
20 | 0.1156 | 71.1 (60.1, 80.5) | 84.5 (82.0, 86.7) | 28.8 (22.7, 35.5) | 97.1 (95.7, 98.1) |
10 | 0.2183 | 50.6 (39.4, 61.8) | 93.5 (91.7, 95.0) | 40.8 (31.2, 50.9) | 95.5 (94.0, 96.8) |
Equation 2 | |||||
30 | 0.0646 | 76.9 (68.3, 84.0) | 74.0 (71.6, 76.3) | 20.3 (16.7, 24.2) | 97.4 (96.2, 98.3) |
20 | 0.1076 | 69.4 (60.4, 77.5) | 84.2 (82.2, 86.1) | 27.5 (22.5, 32.8) | 97.0 (95.9, 97.9) |
10 | 0.2042 | 52.1 (42.8, 61.2) | 93.7 (92.3, 94.9) | 41.4 (33.5, 49.7) | 95.8 (94.6, 96.8) |
Equation 3 | |||||
30 | 0.0612 | 88.5 (77.8, 95.3) | 75.2 (71.8, 78.5) | 24.5 (19.0, 30.8) | 98.6 (97.2, 99.4) |
20 | 0.1051 | 77.0 (64.5, 86.8) | 85.2 (82.3, 87.8) | 32.2 (24.7, 40.4) | 97.6 (96.0, 98.7) |
10 | 0.2404 | 67.2 (54.0, 78.7) | 95.1 (93.2, 96.6) | 55.4 (43.4, 67.0) | 97.0 (95.3, 98.1) |
PPV = positive predictive value; NPV = negative predictive value.
The App can be used on all Apple devices (iPhone, iPad and iPod Touch) and is free to download from the App store.
Childhood obesity is a major public health threat in the UK
There is extensive literature on the early life risk factors for obesity
A recent meta-analysis of 10 birth cohort studies including 47,661 participants reported that a one unit increase in weight z-score change between birth and one year of age conferred a two-fold increase in risk for childhood obesity at ages between six and 14 years
A paper-based tool to predict an infant’s risk of childhood obesity has been proposed
We present the development of a practical mobile phone application that can be used during a wide range of ages (4.5 to 13.5 months) in infancy when growth monitoring is part of routine health care
The lack of a hard outcome in adolescence or adulthood is perhaps the greatest limitation of the present study, because even though obesity risk tracks across the life course
As a next step, qualitative work is needed to understand how this tool will be received by health care practitioners. As children in the Born in Bradford cohort grow up there will be an opportunity to update our prediction equations using later life health outcomes. Alternatively, a similar approach to that used in the Druet et al
In conclusion, we have developed data driven prediction equations for risk of childhood obesity and incorporated them into a mobile phone application, thereby providing proof of concept that childhood obesity prediction research can be integrated with advancements in technology to deliver a clinically relevant tool to practitioners.
Born in Bradford: We are grateful to all the families who took part in this study, to the midwives for their help in recruiting them, the paediatricians and health visitors and to the Born in Bradford team which included interviewers, data managers, laboratory staff, clerical workers, research scientists, volunteers and managers.
ALSPAC: We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The UK Medical Research Council and the Wellcome Trust (Grant ref: 092731) and the University of Bristol provide core support for ALSPAC.
We would also like to thank Simon Faulkner from NHS Airedale, Bradford and Leeds for providing us with the Child Health Data, and Lesley Fairley from Born in Bradford for appraising the manuscript.