- What it is: An AI system that analyses growth trajectories across 5 dimensions — not just a chart that shows where the dot falls
- 5 analysis dimensions: WHO centile trajectory · Growth velocity · Three-metric interplay · Weight-for-length · Head circumference trajectory
- Data source: WHO Child Growth Standards — constructed to show how children should grow, not just how they do
- Key insight types: Trajectory stability vs centile crossing · Velocity below expected range · Weight-length divergence · Proportionate vs disproportionate smallness · Head growth trajectory
- Core principle: Centile position is not a quality score — trajectory is what matters. A baby on the 9th centile following a stable trajectory is growing well; a baby on the 25th centile with a falling trajectory warrants investigation
- What AI cannot do: Diagnose growth disorders — any growth concern requires clinical assessment by a paediatrician with physical examination, dietary history, and medical evaluation
What Is an AI Growth Tracker — Beyond the Chart
A growth chart plots a point. An AI growth tracker analyses a trajectory. This distinction — between a point-in-time position and a longitudinal pattern — is the entire value of AI in growth monitoring, because the clinical significance of any single growth measurement is almost entirely determined by the context of the measurements that preceded it.
A baby on the 9th centile for weight who has been on the 9th centile since their 6-week check is growing appropriately — they are constitutionally small. A baby on the 9th centile for weight who was on the 50th centile four months ago has experienced significant downward centile crossing — a trajectory that warrants clinical investigation regardless of the absolute centile they have reached. Both babies are on the 9th centile. Only the trajectory tells you which clinical picture you are looking at.
| Capability | Growth Chart App | AI Growth Tracker |
|---|---|---|
| Plots measurements on WHO chart | ✓ | ✓ Plus annotated trajectory analysis |
| Shows current centile position | ✓ | ✓ Plus whether trajectory is stable, rising, or falling |
| Growth velocity calculation | ✗ | ✓ Per measurement pair · vs age-expected velocity range |
| Three-metric interplay analysis | ✗ | ✓ Weight + length + head circumference simultaneously |
| Weight-for-length calculation | ✗ | ✓ Distinguishes proportionate smallness from underweight |
| Early faltering detection | ✗ | ✓ Velocity deceleration before centile crossing is visible |
| Head circumference trajectory | Sometimes — chart only | ✓ Separate trajectory analysis · Flags macro/microcephaly concern |
| Well-baby visit summary | ✗ | ✓ All measurements · Trajectories · Velocities · Flags |
| Concern flag calibration | ✗ | ✓ Two-centile-line crossing · Velocity below range · Weight-for-length threshold |
The Five AI Growth Analysis Dimensions — What the AI Actually Looks At
Dimension 1 — WHO Centile Trajectory Analysis
WHO Child Growth Standards centile charts were constructed from a multinational study of children raised under optimal conditions across six countries, specifically designed to show how children should grow rather than describing how children in a given population happen to grow. This distinction matters: WHO charts provide the most appropriate universal reference for infant and toddler growth assessment.
AI centile trajectory analysis tracks the direction of growth across every logged measurement pair — not just the current centile position. The AI identifies whether the growth trajectory is following a stable centile channel (the ideal pattern for most babies), crossing centiles upward (which may reflect catch-up growth or constitutional trajectory adjustment), or crossing centiles downward (which ranges from normal constitutional adjustment in the first 6–12 months to a clinically significant pattern that warrants investigation). The AI applies the clinical significance thresholds correctly: crossing one centile space is common and usually normal; crossing two or more major centile lines downward is the threshold that triggers a 'raise at your next visit' flag.
📊 WHO Centile Positions — What Each Line Means
AI concern flag threshold: downward crossing of 2 or more major centile lines (e.g. 50th → 9th, or 75th → 25th) across successive measurements. A baby on the 9th centile following their channel since birth is growing appropriately. The centile position is not the concern — trajectory direction is.
Dimension 2 — Growth Velocity Calculation
Growth velocity is the rate of growth between two successive measurements: grams per week for weight, centimetres per month for length and head circumference. It is calculated from each pair of logged measurements and compared against age-appropriate expected velocity ranges — the ranges within which weight gain, linear growth, and head growth are considered to be proceeding normally for a baby of that age.
Growth velocity is a more sensitive early indicator of growth change than centile position, because centile position can appear stable for several months while velocity has been declining — the centile position smooths the trajectory in a way that masks early deceleration. AI growth velocity analysis detects this early signal, flagging when velocity has dropped below the lower expected range for the baby's age before any centile crossing has become visible on the chart. This earlier detection is one of the most clinically valuable features of AI growth tracking.
| Age | Expected Weight Gain | Expected Length Gain | Expected Head Growth |
|---|---|---|---|
| 0–3 months | 150–200 g/week (peak growth) | 3–3.5 cm/month | ~2 cm/month |
| 3–6 months | 100–150 g/week | 2–2.5 cm/month | ~1 cm/month |
| 6–9 months | 70–90 g/week | 1.5–2 cm/month | ~0.5 cm/month |
| 9–12 months | 60–80 g/week | 1–1.5 cm/month | ~0.5 cm/month |
| 12–18 months | 40–60 g/week | 1–1.2 cm/month | ~0.3 cm/month |
| 18–24 months | 30–45 g/week | 0.8–1 cm/month | ~0.2 cm/month |
| 2–3 years | 25–35 g/week | 0.7–0.9 cm/month | ~0.1 cm/month |
Dimension 3 — Three-Metric Interplay Analysis
The three primary growth measurements are not independent variables — they are interrelated in ways that carry specific clinical meaning. A baby's weight centile position means something very different depending on their length centile and head circumference centile. An AI growth tracker analyses all three metrics simultaneously and generates insights based on their interrelationship, which is qualitatively different from reporting three separate centile positions.
Pattern 1 — Proportionate Growth Across All Three
All three measurements (weight, length, head circumference) on similar centile positions, following stable channels. This is the ideal growth pattern regardless of the absolute centile — a baby consistently on the 9th centile for all three metrics is growing proportionately and appropriately. AI assessment: no concern; continue monitoring at scheduled well-baby visits. The AI notes this pattern explicitly in the growth summary — 'proportionate growth across all three metrics, trajectory stable' — which is the reassurance parents and paediatricians most need to hear confirmed with data.
Pattern 2 — Weight Falling, Length Stable
Weight trajectory is declining or velocity is below expected range while length trajectory remains on channel. This pattern — weight faltering with preserved linear growth — is the classic early growth faltering signature and the highest priority AI flag in growth tracking. It indicates that the body's first response to inadequate nutrition has occurred: weight gain is the first thing to suffer when caloric intake is insufficient, because the body prioritises linear growth and head growth. This pattern warrants nutritional assessment — feeding review, diet history, breastfeeding support if applicable — before it progresses to length faltering, which is a later and more serious stage. AI flags this pattern with specific metric data and a clear 'contact your healthcare provider' prompt.
Pattern 3 — All Three Metrics Declining
Weight, length, and head circumference trajectories all crossing centiles downward simultaneously. This pattern — three-metric decline — suggests a systemic cause rather than nutritional insufficiency alone, because adequate nutrition typically preserves linear and head growth even when weight is affected. Three-metric decline warrants prompt medical assessment: possible causes include metabolic disorders, chronic illness, endocrine conditions, or other systemic factors that affect growth across all dimensions simultaneously. AI flags this pattern as a clinical priority — the most urgent growth flag it generates — and recommends contact with your healthcare provider before the next scheduled well-baby visit.
Pattern 4 — Large Head, Average Body
Head circumference significantly above the centile for weight and length — for example, weight and length on the 25th centile with head circumference on the 75th or 91st centile. This pattern (macrocephaly relative to body) is common in infancy and is usually benign — familial macrocephaly is the most common cause (check parental head circumferences). However, it requires clinical assessment to exclude raised intracranial pressure (which presents with additional symptoms: bulging fontanelle, sunset sign, irritability, vomiting, growth acceleration) and benign communicating hydrocephalus. AI flags significant head-body centile discordance with a 'worth measuring parent head circumferences and raising at your next visit' prompt — not alarming language, but a specific and clinically appropriate signal.
Dimension 4 — Weight-for-Length Analysis
Weight-for-length is the WHO-standardised comparison of a baby's weight against their length — it answers the question 'is this baby's weight appropriate for how tall they are?' rather than 'is this baby's weight appropriate for their age?'. It is particularly important for distinguishing two very different clinical situations that may present with similar weight centile positions.
Consider two babies, both on the 9th centile for weight at 6 months. Baby A is also on the 9th centile for length — they are proportionately small. Their weight-for-length is normal; they are constitutionally small and growing appropriately. Baby B is on the 50th centile for length but on the 9th centile for weight — their weight is significantly below what is expected for their height. Their weight-for-length is low; this is a genuine weight-relative-to-length concern that warrants nutritional assessment. Both babies have the same weight centile. Only weight-for-length tells you which clinical picture you are looking at.
WHO defines wasting (acute malnutrition) as weight-for-length below -2 SD (approximately the 2nd centile for weight-for-length). AI growth trackers calculate weight-for-length at each measurement pair and generate flags when it falls below the 2nd centile — one of the clearest growth concern signals in the WHO framework.
Dimension 5 — Head Circumference Trajectory
🧠 Head Circumference — Expected Growth by Age
avg 34–35 cm
avg ~40 cm
avg ~43 cm
avg ~46 cm
avg ~48 cm
WHO trajectory. Growth velocity. Three-metric interplay. AI analysis that reads what the numbers mean together.
Lunara analyses your baby's growth across all five dimensions — centile trajectory, growth velocity, three-metric interplay, weight-for-length, and head circumference trajectory — generating insights calibrated to WHO Child Growth Standards and your baby's exact age and sex.
How AI Detects Growth Faltering Early — Before the Chart Shows It
Growth faltering — previously called 'failure to thrive' — is a pattern of inadequate growth that can have nutritional, medical, or psychosocial causes. Its earliest clinical sign is a deceleration in weight gain velocity, which precedes any visible downward centile crossing by weeks to months. An AI growth tracker that calculates growth velocity at every measurement pair can detect this early velocity deceleration signal before the standard chart review at a scheduled well-baby visit would show anything of concern.
Stage 1 — Velocity deceleration: Weight gain velocity drops below the lower expected range for the baby's age across 2+ consecutive measurement intervals. The centile trajectory may still appear stable because the most recent centile position has not yet shifted visibly. AI detects this stage and generates a 'monitor closely, raise at next visit' flag with velocity data. This is the earliest possible detection point — 4–8 weeks before standard chart review would identify the pattern.
Stage 2 — Weight-length divergence: Weight trajectory begins to cross centiles downward while length trajectory remains stable. Weight-for-length begins to fall. This is the established early growth faltering signature — weight is being lost relative to linear growth because the body is prioritising linear growth over weight gain under caloric restriction. AI flags this as a higher-priority concern: 'raise at your next well-baby visit or contact your healthcare provider if the next visit is more than 2 weeks away.'
Stage 3 — Centile crossing: Weight has now crossed two or more major centile lines. This is the stage at which standard chart review at a scheduled well-baby visit would typically identify the pattern. AI would have flagged this at Stages 1 and 2 — weeks earlier — giving the family and clinician time to investigate and intervene before the growth deficit has deepened.
Understanding Centile Positions — What Parents Most Often Misread
The most persistent misunderstanding in baby growth monitoring is that a low centile position is a problem. It is not. The centile system is a description of the normal distribution — by design, 25% of all well-nourished, healthy babies will be below the 25th centile, and 9% will be below the 9th centile. These babies are not growing poorly. They are simply on the smaller end of a normal distribution.
What Centile Position Actually Tells You
A centile position tells you how your baby compares to the distribution of well-nourished babies of the same age and sex — nothing more. A baby on the 9th centile for weight is lighter than 91% of babies at their age. A baby on the 75th centile is heavier than 75%. These are descriptions, not quality scores. A baby on the 9th centile who is following a stable trajectory is growing normally. A baby on the 75th centile whose trajectory has fallen from the 91st centile across three measurements may be showing a growth concern. Centile position without trajectory direction is incomplete information — the AI's trajectory analysis provides the context that centile position alone cannot.
Genetic Target Height and Mid-Parental Height
A child's constitutional size is primarily determined by their genetic inheritance — their mid-parental height (the average of both parents' heights, adjusted for sex) is the strongest single predictor of their adult height. A child who is on the 9th centile for length with parents both below average height is likely following their constitutional growth trajectory correctly. A child on the 9th centile for length with parents both significantly above average height may have a linear growth concern. AI growth trackers that allow parental height entry can calibrate the expected centile channel against the genetic target, providing much more accurate trajectory interpretation. Without parental height data, the AI uses the WHO population-level reference; with it, it can personalise the expected centile range.
Birthweight and Constitutional Adjustment
Birthweight is influenced by maternal factors (pre-pregnancy BMI, gestational nutrition, gestational diabetes, placental function, smoking) and gestational age — none of which determine genetic growth potential. A baby born large because of gestational diabetes may drift from the 91st centile at birth to the 50th centile by 6 months as they grow toward their genetic size. A baby born small because of placental insufficiency may rise from the 9th centile at birth to the 50th centile by 12 months in the same way. AI growth trackers can note birth context and distinguish expected constitutional adjustment in the first 12 months from trajectory changes that are not explained by constitutional adjustment — an analytically important distinction that a chart review alone cannot easily make.
Solid Food Introduction and Weight Patterns
The introduction of solid foods at 6 months is associated with a common and normal growth pattern: milk intake may reduce slightly as solids increase; the caloric density of the diet may temporarily change; and the weight trajectory may briefly decelerate or level before re-establishing. AI growth trackers that are integrated with feeding tracking can contextualise a brief weight velocity deceleration at 6–7 months against the solid food introduction timeline, reducing false-positive growth concern flags during this transition. This is one of the cross-domain analysis benefits that an integrated AI baby tracker provides over a standalone growth tracking tool — the feeding data adds context that makes the growth data more accurately interpretable.
How to Use an AI Growth Tracker — 6 Strategies for Maximum Accuracy
Log Clinical Measurements at Every Well-Baby Visit
The AI growth trajectory is only as accurate as the data it has — and the most accurate data comes from clinical-grade measurements taken by trained healthcare professionals using calibrated equipment. Log all three measurements (weight, length, head circumference) at every well-baby visit immediately after the appointment. If the health record or personal child health record (Red Book in the UK) shows the measurements, log them exactly as recorded — do not estimate or round. The trajectory analysis is built from these data points; a single inaccurate entry can distort the trajectory calculation for subsequent measurements. If you notice a measurement that seems significantly different from the previous one, note it and ask your healthcare provider at the next visit whether it should be re-measured.
Home Measurements — Use With Context
Home weight scales and tape measures vary considerably in accuracy — consumer baby scales typically have an accuracy of ±50–100 grams, which represents approximately one-third of a week's expected weight gain for a 3-month-old. This margin of error makes home measurements useful for tracking broad trends but not reliable enough for precise velocity calculation. If logging home measurements, flag them as 'home measurement' in the app and understand that the AI will give appropriate weight to clinical vs home measurement accuracy in its trajectory analysis. The most valuable use of home measurements is tracking the direction of weight change (trending up, stable, or declining) rather than calculating precise velocity from short-interval home weigh-ins, which are prone to measurement artefact.
Generate a Growth Summary Before Every Well-Baby Visit
Generate the AI growth summary 2–3 days before each well-baby visit and review it yourself before the appointment. The summary will show: all three measurements at every logged visit with dates; current centile positions across all three metrics; trajectory direction; growth velocity calculations for the most recent measurement intervals; weight-for-length analysis; head circumference trajectory; and any AI-flagged concerns. Bring this summary to the appointment and use it as the basis for any growth discussions — 'the app shows weight velocity has dropped from 120 grams per week to 60 grams per week over the last two measurement intervals; is that within normal range for her age?' gives your paediatrician a specific, data-backed question that can be assessed and answered in the clinical encounter rather than a vague 'is she growing normally?'
Enter Both Parents' Heights for Personalised Calibration
Entering both parents' heights allows the AI to calculate the genetic target height range for your baby and use it to personalise the expected growth centile channel. This significantly improves trajectory interpretation: a baby whose length is on the 9th centile when their genetic target suggests the 25th–50th centile is a different clinical picture from a baby whose length is on the 9th centile when their genetic target suggests the 9th centile. Without parental height data, the AI uses WHO population-level references; with it, it can flag when a baby's length trajectory appears below their genetic target — an early signal of linear growth concern that standard centile chart review cannot provide. This is one of the highest-value data inputs in AI growth tracking.
Act on Velocity Flags Promptly — These Are the Earliest Signals
When AI flags a growth velocity concern — weight gain velocity below the expected range for the baby's age — treat it as an early signal that warrants attention, not a confirmed diagnosis of growth faltering. The appropriate response: contact your healthcare provider for a feeding review or additional weight check between scheduled well-baby visits if the AI has flagged the concern across two or more measurement intervals. Do not wait for the next scheduled appointment if it is more than 3 weeks away — growth velocity flags represent the earliest detection window available, and the earlier feeding support or dietary review is provided, the less growth deficit is accumulated before intervention. A brief additional weighing appointment and feeding review is low-cost; the cost of delayed response to genuine early growth faltering is much higher.
Connect Growth Data to Feeding and Sleep Tracking
The most complete picture of a baby's health comes from connecting growth data to feeding and sleep tracking — the three pillars of baby health monitoring that are most interdependent. Inadequate feeding drives growth faltering; growth faltering affects sleep (overtired and malnourished babies have disrupted sleep architecture); sleep deprivation affects feeding (an overtired baby feeds less effectively). An integrated AI baby tracker that connects growth velocity data to feeding volume trends and sleep pattern data can identify cross-domain patterns — for example, declining weight velocity in the context of reducing feeding volumes and increasing night waking frequency — that a standalone growth tracker cannot detect. Use growth tracking as part of integrated baby tracking rather than in isolation for the richest analytical picture.
Common Mistakes With AI Growth Trackers
❌ Focusing on Centile Position Rather Than Trajectory
The problem: The most common misuse of growth tracking — AI or otherwise — is treating the centile position as the key number to watch. Parents who see their baby on the 9th centile and worry, or on the 75th centile and feel reassured, are using an incomplete metric. A baby on the 9th centile with a stable trajectory is growing well; a baby on the 75th centile with a falling trajectory has a growth concern. The AI's trajectory analysis is the key output — not the centile number — but parents habituated to centile-centric thinking may not focus on it.
What to do instead:- Read the AI trajectory summary first — 'trajectory stable / crossing upward / crossing downward' — before looking at the centile number
- If the AI reports a stable trajectory, the current centile position requires no further interpretation regardless of its absolute value
- If the AI reports centile crossing, focus on the direction, the number of centile lines crossed, and the velocity data — these are the clinically meaningful signals
❌ Logging Only Weight and Ignoring Length and Head Circumference
The problem: Weight is the easiest measurement to obtain at home and the one most parents focus on. Length and head circumference require more effort — a tape measure, a flat surface, a cooperative baby — and are often not logged between well-baby visits. This partial logging prevents three-metric interplay analysis and weight-for-length calculation, both of which depend on all three measurements being logged simultaneously. An AI growth tracker with only weight data can calculate weight velocity and weight centile trajectory — valuable insights — but cannot distinguish proportionate smallness from weight-relative-to-length concern, which is the most important clinical distinction in growth assessment.
What to do instead:- Log all three measurements at every well-baby visit — length and head circumference are both taken routinely and recorded in the health record
- If you weigh at home between visits, also measure length and head circumference with a cloth tape measure — accuracy is lower but direction of change is still informative
- At minimum, log clinical measurements (all three) immediately after every well-baby visit
❌ Comparing Your Baby's Centile to Other Babies
The problem: Baby centile comparison between families — common in parenting groups and social media — generates anxiety that has no clinical basis. A baby on the 25th centile whose parent is comparing them to babies on the 75th centile in a parenting group is not growing poorly — they are on the smaller end of a normal distribution, as 25% of all healthy babies are. Centile comparison between babies ignores the most important growth factor (trajectory) and produces only the least informative growth metric (absolute centile position) in a social context that encourages anxiety rather than accurate interpretation.
What to do instead:- Focus exclusively on your baby's own trajectory — the only comparison that matters is your baby's current measurements against their own historical pattern
- If other parents or well-meaning relatives express concern about your baby's size, point them to the trajectory: 'the paediatrician and the app both show a stable trajectory — she's consistently on her centile channel, which is what matters'
- Contact your healthcare provider if you have a genuine concern — not a parenting group, whose members cannot access your baby's trajectory data
❌ Not Logging the Birth Measurements
The problem: The birth measurements — birthweight, birth length, birth head circumference — are the anchor points from which the entire growth trajectory is calculated. Without them, the AI cannot establish the starting centile position or identify constitutional centile adjustment in the first 12 months (which is only interpretable against the birthweight centile). Many parents begin logging growth measurements only after discovering the app at 2–3 months, missing the critical first-trimester data. The birth measurements are in the hospital records and the personal child health record (Red Book) — they can be entered retrospectively and doing so significantly improves the accuracy of all subsequent trajectory analysis.
What to do instead:- Log birth measurements retrospectively if you did not log them at the time — check the hospital discharge summary or Red Book
- Include the 1–2 week weight check (if available) and 6–8 week well-baby visit measurements as additional early data points
- The more early data points the AI has, the more stable and accurate the trajectory estimation for all subsequent measurements
Frequently Asked Questions — AI Growth Tracker
An AI growth tracker for babies is a baby tracking app that uses artificial intelligence to analyse logged growth measurements — weight, length/height, and head circumference — across five analytical dimensions: WHO centile trajectory, growth velocity, three-metric interplay, weight-for-length, and head circumference trajectory. Unlike a growth chart app (which plots a measurement and shows a centile), an AI growth tracker analyses the trajectory of growth across multiple measurements, detects early velocity deceleration before centile crossing thresholds are reached, reads the three-metric interplay to distinguish constitutional smallness from genuine growth concern, and generates structured summaries for well-baby visits. It is a decision-support tool, not a diagnostic tool — any growth concern identified by AI requires clinical assessment.
A comprehensive AI growth tracker analyses five dimensions: (1) WHO centile trajectory — the direction of growth across multiple measurements (stable, rising, or falling) against WHO Child Growth Standards; (2) Growth velocity — the rate of growth between measurements vs age-appropriate expected velocity ranges; (3) Three-metric interplay — the relationship between weight, length, and head circumference simultaneously; (4) Weight-for-length — whether the baby's weight is appropriate for their height (distinguishing constitutional smallness from underweight); (5) Head circumference trajectory — separate trajectory analysis flagging macro and microcephaly concern thresholds. The AI's output: trajectory direction, velocity comparison, pattern classification (proportionate growth, weight-length divergence, three-metric decline, head-body discordance), and well-baby visit summaries.
WHO Child Growth Standards centile charts were constructed from a multinational study of children raised under optimal conditions across six countries — designed to show how children should grow, not just how children in a given population happen to grow. This makes them the most appropriate universal reference for infant and toddler growth assessment. AI growth trackers use WHO charts because they provide the most valid and internationally consistent growth reference for the 0–5 year age range. The centile lines (2nd, 9th, 25th, 50th, 75th, 91st, 98th) divide the population into groups — a baby on the 25th centile is heavier than 25% and lighter than 75% of well-nourished babies of the same age and sex. Position alone is not a quality score; trajectory direction is the clinically meaningful analysis.
Centile crossing means the growth trajectory has shifted from one centile channel to another. Modest crossing of one centile space is common and often normal — particularly in the first 6–12 months as babies adjust toward their genetic size. Clinically concerning centile crossing: crossing two or more major centile lines (2nd, 9th, 25th, 50th, 75th, 91st, 98th) downward in weight; a falling weight trajectory while length remains stable; or downward crossing in all three metrics simultaneously. AI applies these clinical thresholds precisely — flagging patterns that warrant clinical discussion while not generating unnecessary alarm about normal constitutional adjustment that crosses one centile space during the first year.
AI detects growth faltering through three mechanisms: (1) Growth velocity calculation — detecting when weight gain rate drops below the age-appropriate expected range across 2+ measurement intervals, before centile crossing is visible on the chart; (2) Three-metric interplay — detecting when weight growth decelerates while length growth continues on trajectory (weight-length divergence is the earliest established growth faltering signature); (3) Trajectory analysis across all data points — detecting deceleration from the established trajectory even when the centile position appears stable. Together these allow AI to flag a developing growth concern 4–8 weeks before it would become apparent on standard chart review at a scheduled well-baby visit.
The three-metric interplay is the analysis of weight, length/height, and head circumference simultaneously — because the clinical significance of any single measurement changes depending on the other two. Four key patterns: (1) All three proportionate and stable — ideal; (2) Weight falling, length stable — early growth faltering signature; nutritional assessment warranted; (3) All three declining — systemic cause suggested; prompt medical assessment warranted; (4) Significant head-body centile discordance — familial variation most likely, but clinical assessment needed to exclude other causes. AI generates pattern classification and flags based on all three simultaneously, which is qualitatively different from reporting three separate centile numbers independently.
No — a low centile position alone is not a sign of a growth problem. By definition, 9% of all well-nourished healthy babies are below the 9th centile and 25% are below the 25th centile — these babies are not growing poorly, they are simply on the smaller end of the normal distribution. The clinically meaningful question is whether the baby is following their established growth trajectory consistently. A baby on the 9th centile since birth who is still on the 9th centile at 12 months is growing well. A baby who was on the 50th centile and has dropped to the 9th centile is showing significant centile crossing that warrants investigation — regardless of the absolute position reached. AI growth tracking focuses on trajectory direction, not centile position alone.
Growth velocity is the rate of growth between two measurements: grams per week for weight, centimetres per month for length and head circumference. AI calculates it from each logged measurement pair and compares against age-appropriate expected velocity ranges. Velocity matters because it is more sensitive than centile position: a baby's centile may appear stable for months while velocity has been declining below expected range — the centile position smooths the trajectory in a way that masks early deceleration. Expected weight gain velocity: 0–3 months: 150–200 g/week; 3–6 months: 100–150 g/week; 6–12 months: 70–90 g/week; 12–24 months: 40–60 g/week. Below-range velocity across 2+ measurement intervals triggers an AI flag.
Log all three primary measurements at every well-baby visit: weight (grams for infants; kilograms from toddlerhood), length/height (centimetres — lying length under 2 years; standing height from 2 years), and head circumference (centimetres — measured at widest circumference above eyebrows and ears). Log clinical measurements taken by your healthcare provider as the primary data source — these use calibrated equipment and standardised technique. Home measurements can be logged between visits for broad trend tracking but have higher measurement error. Also log birth measurements retrospectively if not done at the time — they are the trajectory anchor point and can be found in hospital records or the personal child health record.
Standard well-baby visits provide growth measurement at birth, 1–2 weeks, 6–8 weeks, 3 months, 4 months, 6 months, 9 months, 12 months, 15 months, 18 months, and 24 months (varies by country) — this is sufficient for AI trajectory analysis in a baby growing well on a stable centile channel. For babies with an AI-flagged growth concern, more frequent measurements every 2–4 weeks allow faster assessment of whether the trajectory is responding to intervention. Avoid more frequent than weekly weighing at home — day-to-day weight variation from hydration, feeding timing, and bowel status makes very frequent measurements difficult to interpret and potentially anxiety-producing without clinical context to ground the reading.
Weight-for-length compares a baby's weight against their length — 'is this baby's weight appropriate for how tall they are?' Two babies on the 9th centile for weight have very different clinical pictures: Baby A on the 9th centile for both weight and length is proportionately small (constitutional, weight-for-length normal). Baby B on the 9th centile for weight but 50th centile for length is underweight relative to their linear growth (weight-for-length low, nutritional assessment warranted). WHO defines wasting (acute malnutrition) as weight-for-length below -2 SD (approximately 2nd centile for weight-for-length). AI calculates weight-for-length at each measurement pair to distinguish constitutional smallness from genuine nutritional concern — a distinction the centile charts alone cannot provide.
Growth concerns worth discussing promptly with your paediatrician: downward centile crossing of two or more major centile lines in weight; weight trajectory flattening (very low or zero weight gain over 2–3 weeks) under 12 months; weight-for-length below the 2nd centile; head circumference crossing two or more centile lines in either direction; three-metric decline (all three falling simultaneously); any return to weight loss after 2 weeks or failure to regain birthweight by 3 weeks. Contact your healthcare provider before the next scheduled visit if you observe unexpected weight loss, very poor feeding, or any strong concern — do not wait. AI growth flags are a prompt to seek clinical input, not a reason for self-management.
An AI growth tracker generates a structured summary for well-baby visits: all three measurements at every logged visit with dates, current centile positions across all metrics, trajectory direction, growth velocity calculations for recent intervals, weight-for-length, head circumference trajectory, and any AI-flagged concerns. A parent who arrives with this summary gives the paediatrician longitudinal growth context across the full period since birth — not just the measurements taken at this visit and the previous one. Generate the summary 2–3 days before each visit, review it yourself, write down specific data-prompted questions, and bring it as the basis for growth discussions rather than asking 'is she growing normally?' which cannot be meaningfully answered without the trajectory data.
A growth chart app plots measurements on a WHO centile chart and shows where the dot falls — that is its complete function. An AI growth tracker does that and: analyses the trajectory across multiple measurements; calculates growth velocity vs age-expected ranges; performs three-metric interplay analysis; calculates weight-for-length; detects early growth faltering before centile crossing thresholds are reached; generates well-baby visit summaries; and flags concerns with clinical context and 'raise at your next visit' prompts. Growth chart app output: 'your baby is on the 25th centile.' AI growth tracker output: 'weight trajectory has crossed from the 50th to the 25th centile across three measurements while length remains stable — weight-for-length below expected; worth discussing at your next visit with this data.'
Lunara's AI growth tracker analyses logged weight, length/height, and head circumference measurements against WHO Child Growth Standards across five dimensions: centile trajectory analysis, growth velocity calculation (vs age-specific expected ranges), three-metric interplay pattern classification, weight-for-length analysis, and head circumference trajectory. Parental heights can be entered for genetic target calibration. Concern flags are calibrated to clinically meaningful thresholds — two-centile-line crossing, below-range velocity, weight-for-length below 2nd centile — and presented with clinical context and 'raise at your next visit' prompts. Both parents log on one shared profile. Well-baby visit summaries generated automatically. All AI growth science clinically reviewed. Free to start.
The Bottom Line on AI Growth Trackers
Growth monitoring has always depended on longitudinal data — a single measurement tells you almost nothing, while a trajectory tells you almost everything. The challenge is that meaningful trajectory analysis across multiple measurements, across three interrelated metrics, with growth velocity calculation and weight-for-length computation, is not something a 15-minute well-baby visit can do comprehensively from two data points. An AI growth tracker does it continuously, from birth, from every measurement the family logs.
The clinical value is concentrated in three places: early growth faltering detection — identifying velocity deceleration weeks before centile crossing is visible on the standard chart; the three-metric interplay — distinguishing constitutional smallness from genuine weight-relative-to-length concern; and the well-baby visit summary — giving paediatricians longitudinal growth context that transforms the quality of the clinical encounter. Used consistently — with all three measurements logged at every well-baby visit, parental heights entered for genetic calibration, and the AI summary brought to every appointment — an AI growth tracker provides the continuous growth surveillance that well-baby visits alone cannot offer, and the early warning system that gives the families who need intervention the maximum time to access it.
AI Growth Tracker — Quick Reference
- 0–3 months: 150–200 g/week weight · 3–3.5 cm/month length · 2 cm/month head
- 3–6 months: 100–150 g/week · 2–2.5 cm/month · 1 cm/month head
- 6–12 months: 70–90 g/week · 1–2 cm/month · 0.5 cm/month head
- 12–24 months: 40–60 g/week · 0.8–1.2 cm/month · 0.2–0.3 cm/month head
- AI flags: below-range velocity across 2+ measurement intervals → raise at next visit
- Downward centile crossing of 2+ major lines in weight → raise at next visit
- Weight falling while length is stable → nutritional assessment warranted
- Three-metric decline (all three falling) → prompt medical assessment
- Weight-for-length below 2nd centile → nutritional review urgently
- Head circumference crossing 2+ centile lines (either direction) → clinical assessment
- All three proportionate and stable → ideal; no concern
- Weight falling, length stable → early growth faltering; nutritional assessment
- All three declining → systemic cause; prompt medical assessment
- Head significantly above body centile → familial macrocephaly likely; clinical assessment
- Head significantly below body centile → familial microcephaly likely; clinical assessment
- Log all three measurements (weight + length + head) at every well-baby visit
- Enter birth measurements retrospectively — the trajectory anchor point
- Enter parental heights for genetic target calibration
- Generate AI summary 2–3 days before every well-baby visit
- Act on AI velocity flags — contact provider before next visit if more than 2–3 weeks away
Five dimensions. One analysis. AI growth intelligence that reads trajectory, not just position.
WHO centile trajectory · Growth velocity · Three-metric interplay · Weight-for-length · Head circumference trajectory — all analysed from your logged data, calibrated to WHO Child Growth Standards and your baby's exact age and sex. Both parents on one profile. Free to start.