Growth&Development week 10

winniesmith2's version from 2018-01-13 16:11

Section 1

Question Answer
How do you assess the growth of a child one needs to compare his/ her height (or some other measure) against a comparative/ reference population
Two types of growth chartsreferences and standard
ReferencesCross sectional data, distance, more variation, limited inclusion criteria, descriptive growth 'as is' . Uk 1990 reference
Standard longitudinal data,distance and velocity, less variation, strict inclusion criteria. prescriptive growth ' as it ought to be '. WHO 20006 standard.
What did the UK90 do references combine data on 37,000 individuals from 17 distinct cross-sectional surveys (1984-1992) representative of England, Scotland, and Wales. For weight and stature. - cant make judgement on good or bad just compare individual to this population.
Why did the WHO think new growth charts were needed?-differences in weight gain between breast fed and formula fed infants. Healthy breast fed infants show very similar growth patterns around the world. Decided to produce charts that set breast feeding as the norm for infant feeding.
Who charts- development 15 yr programme of planning, data collection and analysis. New study of growth on breast fed infants of non smoking non deprived mothers in 6 countries (use, norway, india, ghana,brazil, oman) birth to 5 years. Very similar growth patterns in all 6 centres. Charts are a description of optimal rather than average growth. suitable for all children worldwide
Issue with WHO chartsIgnores genetics, say 'growth in length the same in all centres-one chart valid for all children worldwide' . Probably not true.
Uk- WHO charts, do we use it?Used from 2 weeks to 4 years. After 4 years they use UK90 reference. AND uk90 for birth.
Plotting in the first 2 weeks Gap in chart data. good-clear difference between the reference and standard. Babies will loose chunk of birth weight within the first 2 weeks.
Assassing neonatal weight loss; how many regain the weight <80% will have regained this by 2 weeks of age
How many babies lose more than 10% at 2 weeks.fewer than 5%. Baby needs careful assessment for feeding problems and unrecognised illness
What is a normal rate of weight gain?Usually tracks within one centile. Acute illness-> weight loss and weight centile fall. Less than 2% of infants will shows a sustained drop through two or more weight centile spaces on the new WHO charts.
When should a single measurement trigger assessment no single threshold. Further assess; all children with measurements below 0.4th centile. Child with height above 99.6th centile plus other concerns. Look up BMI centile.
When is BMI usedafter the first 2 years of age.
for infants and toddlers, read off the weight and height centiles from the growth chat and blot these against each other, read off the corresponding BMI centile from the slanting lines.
How do you transfer preterm to infancy sectionBorn 6 weeks preterm, plot on preterm section of chart until 42 weeks. The plot on infancy section using gestational correction (draw back 6 weeks).
Length and height changes at 2 years child measures standing up, spine squashed, so appear shorter. UK/WHO charts shift down slightly for this.
Adult height predictionPlot most recent height, find corresponding centile on adult scale, 4 out of 5 children will be within -/+ 6cm of this value as an adult
What has changed in BMI charts due to pediatric obesity epidemicmeans that BMI charts now include extreme centiles.
Some tools assess puberty but why no velocity charts?because you need longitudinal data. WHO have published them but the UK don't use them.
Why are there separate centiles depending on birthweight those who are born smaller tend to have higher velocity.
What are Thrive lines5th velocity lines, adding velocity to growth charts. Not used/never caught on, very difficult to use.
Different charts can give different growth diagnoses......because you are comparing against different populations. -chart you use impacts assessment.

Section 2

Question Answer
equationsdon't need to know them, just the principals
why switch from measurement to z scoresthey are standardised for age and sex
50th percentile =z score of 0
BMIintended to standardise index from weight for height, (to get rid of fact taller people are generally heavier). Weight independent of height in the population. Next meant to be used to assess weight of an individual- only used as it is a good as a quick assessment.
Duality of size and growthyou know, weight as first time point, weight at second time point and weight gain inbetween, if you know any 2 of these 3 variables then you can work out the 3rd.
With only two data points our model of human growth represents a constant, continuous process
With more data point our model becomes more detailed saltation and stasis
what is a growth model?
and normally all growth models are?
a mathematical representation of the real process
regression equations- move from individual points to smooth curve
what is an error and what are the types of errordistance between data point and curve
model error (model is incorrect as doesn't go through all points). and measurement error (wasn't measure properly, if so would be on line).

don't want line to pass through every data point exactly as that suggests there is no measurement error- but always is. Just want it to represent the biology of whats actually going on.
Residual (errors) should be serially independent. explainthe error values should be random. page 40. if all positive model predicting too high, negative too low..
Finding the curve; structural modelsequations developed by growth experts to describe growth of some dimension over an age range.
Always imposes structural form as specific functional form, such as for adolescence it always says theres going to be an age of take off and , AGS, peak height velocity and the deceleration- it always imposes that into data.
-Postulate that the curve as a basic functional form
-Usually have few parameters with some biological interpretation
-Usually tend to an upper asymptote (final size)
Finding the curve; non-structural modelNot developed by growth experts for specific process, just a way of smoothing data
-Do not postulate that the curve has a basic functional form .
-Usually have a large number of parameters with no biological interpretation
-Do not tend to an upper asymptote (final size)
-Usually unstable at the extremities of the data range E.g., polynomials and splines

Recent badges