ESPE Abstracts

Background: Childhood growth is an indicator of health/well-being. Growth monitoring identifies treatable conditions in apparently healthy children and prevents inappropriate referrals. Systematic growth monitoring is not currently a UK priority and growth disorders are frequently diagnosed late.

Objective: Develop and test the accuracy of GrowthMonitor, an app which enables families to measure a child’s height at home as a cost-effective alternative to primary care growth monitoring.

Methods: ‘GrowthMonitor’ calculates height data using augmented reality. The app uses algorithms to calculate height expressed as standard deviations (HSDS) and distance from target height (THSDSDEV) relative to UK population-based height references. Patients were measured using the app in parallel to stadiometer (gold standard) height measurements, as part of routine clinic visit. A subset of parents also used the app to measure their child’s height at home. Precision and reproducibility were assessed with coefficient of variance across 3 consecutive app measurements for each patient. The relationship between the app and stadiometer measurements was assessed with linear regression. The error rate (%) was calculated as [Stadiometer height measurement (cm) – Mean GMA Measurement (cm)]/ Stadiometer height measurement (cm) x 100. Accuracy (%) was then determined as 100% - error rate. Recorded heights were compared with pre-established thresholds that discriminate normal/abnormal growth in ethnicity-matched populations. Predefined thresholds trigger green (normal), amber (monitor) or red (seek medical advice) app alerts. The primary target was to achieve 95% of app growth measurements within +/-0.5SDS of gold standard (stadiometer) measurements.

Results: A total of 88 (46M) patients had three consecutive height measurements made in clinic by the study team using the app. Linear regression showed a significant correlation between the app and the stadiometer height measurements (R20.99; P<0.0001). The average coefficient of variance for height measurements from the app relative to the stadiometer was 0.9%, indicating excellent precision. The average accuracy for the app measurements ±95% CI was 99.2±0.14%. The app provided height measurements within +/- 0.26SDS (95%CI: 0.22-0.30SDS) of the stadiometer. Height measured by parents (n=28) using the app at home also showed a significant correlation (R20.98; P<0.0001) with stadiometer measurements performed in clinic, with a two-tailed paired T-test value of 0.421 suggesting no significant difference in the height measurement methods (app vs. stadiometer).

Conclusion: GrowthMonitor produces accurate, reliable height measurements and can be used by parents/carers in the community to capture serial height measurements and can facilitate early referral/diagnosis of growth disorders.