| 1. |
- Holmgren, Anton, et al.
(författare)
-
Estimating secular changes in longitudinal growth patterns underlying adult height with the QEPS model: the Grow Up Gothenburg cohorts
- 2018
-
Ingår i: Pediatric Research. - : Springer Science and Business Media LLC. - 0031-3998 .- 1530-0447. ; 84:1, s. 41-49
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Over the past 150 years, humans have become taller, and puberty has begun earlier. It is unclear if these changes are continuing in Sweden, and how longitudinal growth patterns are involved. We aimed to evaluate the underlying changes in growth patterns from birth to adulthood by QEPS estimates in two Swedish cohorts born in 1974 and 1990. METHODS: Growth characteristics of the longitudinal 1974 and 1990-birth cohorts (n = 4181) were compared using the QEPS model together with adult heights. RESULTS: There was more rapid fetal/infancy growth in girls/boys born in 1990 compared to 1974, as shown by a faster Etimescale and they were heavier at birth. The laterborn were taller also in childhood as shown by a higher Q-function. Girls born in 1990 had earlier and more pronounced growth during puberty than girls born in 1974. Individuals in the 1990 cohort attained greater adult heights than those in the 1974 cohort; 6 mm taller for females and 10 mm for males. CONCLUSION: A positive change in adult height was attributed to more growth during childhood in both sexes and during puberty for girls. The QEPS model proved to be effective detecting small changes of growth patterns, between two longitudinal growth cohorts born only 16 years apart.
|
|
| 2. |
- Kriström, Berit, et al.
(författare)
-
The first-year growth response to growth hormone treatment predicts the long-term prepubertal growth response in children.
- 2009
-
Ingår i: BMC medical informatics and decision making. - : Springer Science and Business Media LLC. - 1472-6947. ; 9, s. 1-
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Pretreatment auxological variables, such as birth size and parental heights, are important predictors of the growth response to GH treatment. For children with missing pretreatment data, published prediction models cannot be used. The objective was to construct and validate a prediction model for children with missing background data based on the observed first-year growth response to GH. The accuracy and reliability of the model should be comparable with our previously published prediction model relying on pretreatment data. The design used was mathematical curve fitting on observed growth response data from children treated with a GH dose of 33 microg/kg/d. METHODS: Growth response data from 162 prepubertal children born at term were used to construct the model; the group comprised of 19% girls, 80% GH-deficient and 23% born SGA. For validation, data from 205 other children fulfilling the same inclusion and treatment criteria as the model group were used. The model was also tested on data from children born prematurely, children from other continents and children receiving a GH dose of 67 microg/kg/d. RESULTS: The GH response curve was similar for all children, but with an individual amplitude. The curve SD score depends on an individual factor combining the effect of dose and growth, the 'Response Score', and time on treatment, making prediction possible when the first-year growth response is known. The prediction interval (+/- 2 SD res) was +/- 0.34 SDS for the second treatment year growth response, corresponding to +/- 1.2 cm for a 3-year-old child and +/- 1.8 cm for a 7-year-old child. For the 1-4-year prediction, the SD res was 0.13 SDS/year and for the 1-7-year prediction it was 0.57 SDS (i.e. < 0.1 SDS/year). CONCLUSION: The model based on the observed first-year growth response on GH is valid worldwide for the prediction of up to 7 years of prepubertal growth in children with GHD/ISS, born AGA/SGA and born preterm/term, and can be used as an aid in medical decision making.
|
|
| 3. |
- Shmoish, M., et al.
(författare)
-
Prediction of Adult Height by Machine Learning Technique
- 2021
-
Ingår i: Journal of Clinical Endocrinology & Metabolism. - : The Endocrine Society. - 0021-972X .- 1945-7197. ; 106:7
-
Tidskriftsartikel (refereegranskat)abstract
- Context: Prediction of AH is frequently undertaken in the clinical setting. The commonly used methods are based on the assessment of skeletal maturation. Predictive algorithms generated by machine learning, which can already automatically drive cars and recognize spoken language, are the keys to unlocking data that can precisely inform the pediatrician for real-time decision making. Objective: To use machine learning (ML) to predict adult height (AH) based on growth measurements until age 6 years. Methods: Growth data from 1596 subjects (798 boys) aged 0-20 years from the longitudinal GrowUp 1974 Gothenburg cohort were utilized to train multiple ML regressors. Of these, 100 were used for model comparison, the rest was used for 5-fold cross-validation. The winning model, random forest (RF), was first validated on 684 additional subjects from the 1974 cohort. It was additionally validated using 1890 subjects from the GrowUp 1990 Gothenburg cohort and 145 subjects from the Edinburgh Longitudinal Growth Study cohort. Results: RF with 51 regression trees produced the most accurate predictions. The best predicting features were sex and height at age 3.4-6.0 years. Observed and predicted AHs were 173.98.9 cm and 173.9 +/- 7.7 cm, respectively, with prediction average error of -0.4 +/- 4.0 cm. Validation of prediction for 684 GrowUp 1974 children showed prediction accuracy r=0.87 between predicted and observed AH (R-2 = 0.75). When validated on the 1990 Gothenburg and Edinburgh cohorts (completely unseen by the learned RF model), the prediction accuracy was r = 0.88 in both cases (R-2 = 0.77). AH in short children was overpredicted and AH in tall children was underpredicted. Prediction absolute error correlated negatively with AH (P < .0001). Conclusion: We show successful, validated ML of AH using growth measurements before age 6 years. The most important features for prediction were sex, and height at age 3.4-6.0. Prediction errors result in over- or underestimates of AH for short and tall subjects, respectively. Prediction by ML can be generalized to other cohorts.
|
|
