Article Text

Parent-reported prevalence and persistence of 19 common child health conditions
  1. Tracy Liu1,2,
  2. Raghu Lingam1,3,
  3. Kate Lycett1,2,
  4. Fiona K Mensah1,2,
  5. Joshua Muller1,
  6. Harriet Hiscock1,2,4,
  7. Md Hamidul Huque1,
  8. Melissa Wake1,2,5
  1. 1Murdoch Children’s Research Institute, Parkville, Victoria, Australia
  2. 2Department of Paediatrics, The University of Melbourne, Parkville, Victoria, Australia
  3. 3Institute of Health and Society, Newcastle University, Newcastle Upon Tyne, UK
  4. 4Health Services Research Unit, Royal Children’s Hospital, Parkville, Victoria, Australia
  5. 5Department of Paediatrics and Liggins Institute, The University of Auckland, Auckland, New Zealand
  1. Correspondence to Professor Melissa Wake, Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, VIC 3052, Australia; melissa.wake{at}


Objective To estimate prevalence and persistence of 19 common paediatric conditions from infancy to 14–15 years.

Design Population-based prospective cohort study.

Setting Australia.

Participants Parallel cohorts assessed biennially from 2004 to 2014 from ages 0–1 and 4–5 years to 10–11 and 14–15 years, respectively, in the Longitudinal Study of Australian Children.

Main outcome measures 19 health conditions: 17 parent-reported, 2 (overweight/obesity, obesity) directly assessed. Two general measures: health status, special health care needs. Analysis: (1) prevalence estimated in 2-year age-bands and (2) persistence rates calculated at each subsequent time point for each condition among affected children.

Results 10 090 children participated in Wave 1 and 6717 in all waves. From age 2, more than 60% of children were experiencing at least one health condition at any age. Distinct prevalence patterns by age-bands comprised eight conditions that steadily rose (overweight/obesity, obesity, injury, anxiety/depression, frequent headaches, abdominal pain, autism spectrum disorder, attention-deficit hyperactivity disorder). Six conditions fell with age (eczema, sleep problems, day-wetting, soiling, constipation, recurrent tonsillitis), three remained stable (asthma, diabetes, epilepsy) and two peaked in mid-childhood (dental decay, recurrent ear infections). Conditions were more likely to persist if present for 2 years; persistence was especially high for obesity beyond 6–7 (91.3%–95.1% persisting at 14–15).

Conclusions Beyond infancy, most Australian children are experiencing at least one ongoing health condition at any given time. This study’s age-specific estimates of prevalence and persistence should assist families and clinicians to plan care. Conditions showing little resolution (obesity, asthma, attention-deficit hyperactivity disorder) require long-term planning and management.

  • Epidemiology
  • Comm Child Health
  • Adolescent Health

Statistics from

What is already known on this topic

  • Children’s health conditions impose substantial physical, social and economic burden both individually and societally.

  • Chronic and/or recurrent conditions now dominate paediatric practice, superseding acute issues.

  • Comprehensive population evidence is lacking on the prognosis and natural history of such conditions.

What this study adds

  • Most Australian children experience ongoing health conditions at every age beyond infancy.

  • Our longitudinal age-specific graphs could help families and clinicians to plan care for numerous conditions.

  • Obesity, asthma and attention-deficit hyperactivity disorder will require innovative strategies to prevent persistence into adulthood.


As practising child health clinicians, we are asked to give advice to children and their parents about the prevalence, short-term and long-term prognosis of an array of health conditions. However, most conditions lack the empirical information to do so with any level of accuracy. As well as the  direct value of prevalence and persistence data to clinicians and families, this information would enable policy-makers to better target health services to children according to age and condition.

Large cross-sectional studies have documented the prevalence of many childhood conditions.1 2 In the Australian National Health Survey, asthma was the most frequent long-term childhood condition, followed by allergy, refractive vision errors, chronic sinusitis, eczema and mental health issues.1 Overweight/obesity and dental decay are highly prevalent,3–5 while diabetes, epilepsy, attention-deficit hyperactivity disorder (ADHD), anxiety, depression and autism spectrum disorder (ASD) have been identified as key chronic conditions due to their burden.6 Many adolescents are also presenting for headaches and abdominal pain.7 However, this evidence is piecemeal and its value is limited by aggregation into wide age-bands that may not capture variation across childhood.

Even less information is available supporting accurate prognosis for these conditions, which likely varies by condition and age at which present. Large longitudinal studies have reported persistence and resolution of some conditions such as asthma, eczema, overweight/obesity and ADHD.8–13 However, few have followed up children frequently through childhood and into adolescence,8 with most studies including only one follow-up9–12 or large intervals between time-points.13 There has been very little data spanning childhood for conditions like recurrent ear infections, constipation and soiling, day-wetting, recurrent tonsillitis, sleep problems and injury.

This study tracks 19 paediatric physical and mental health conditions longitudinally, in two cohorts recruited within a single population, over a decade. Our research uses parallel, population-based Australian cohorts comprising the Longitudinal Study of Australian Children (LSAC) to answer the important clinical question that paediatricians often face: ‘How common is it, doctor, and (when) will it get better?’ The study aims to estimate (1) the prevalence of 19 common conditions and (2) their persistence rates in 2-year age-bands from infancy to 14–15 years.


Design and setting

Data were drawn from the first six waves of LSAC, a population-based study following two representative samples of Australian children from early childhood. LSAC is conducted jointly by the Australian Department of Social Services, the Australian Institute of Family Studies and the Australian Bureau of Statistics.

Procedure and participants

The sampling design and methods are detailed elsewhere.14 In summary, in a two-stage national clustered sampling design, postcodes were stratified by state then randomly selected to represent urban/rural distributions. Children aged 0–1 years (Birth (B) cohort) and 4–5 years (Kindergarten (K) cohort) were randomly selected within these postcodes using the database of Medicare, Australia’s universal health care system into which ~98% of children nationwide are enrolled by 1 year of age.15 Response rates to the initial mailed invitation in 2004 were 57.2% for the B cohort (5107 infants) and 50.4% for the K cohort (4983 children 4–5 years old).

Data have been collected biennially for six waves from 2004 to 2014, to ages 10–11 years (B cohort) and 14–15 years (K cohort). Non-participants at each wave after recruitment were invited to return at subsequent waves. At each wave, a trained interviewer conducted a 90 minute home visit, including interviewer-administered, computer-assisted questions and direct child assessments.


Table 1 details the repeated health measures available in LSAC, their response formats and the waves in which they were measured. Parents reported whether their child had any of 17 ongoing health conditions, explained as ‘exists for some period of time (weeks, months, years) or re-occurs regularly’: eczema, asthma, recurrent ear infections, ADHD, recurrent tonsillitis, anxiety/depression, ASD, diabetes, epilepsy, injury, sleep problems, recurrent abdominal pain, dental decay, day-wetting, constipation, soiling and frequent headaches.

Table 1

Measures table detailing specific questions for each health condition or problem

Figure 1

Prevalence of conditions and health status measures in 2-year age-bands. (A) Chronic conditions; (B) acute/recurrent and developmental conditions; (C) general health measures. Note the Y-axis scale for graphs changes between 0-35% and 0-5% according to condition prevalence. Survey weights were applied to the multiply imputed data. Technical papers detailing the estimation of weights are available from the Longitudinal Study of Australian Children website. ADHD, attention-deficit hyperactivity disorder.

The interviewer measured each child’s height (from 2 years) and weight at every wave. Weight was measured in light clothing to the nearest 50 g using standardised glass bathroom scales (eg, Salter Australia, Code 79985). Height was measured to the nearest 0.1 cm using a portable rigid stadiometer (Invicta, Code IPO955). Children’s body mass index (BMI, kg/m2) was classified using the International Obesity Task Force age-specific and sex-specific criteria at each half-year age interval.16 This yielded two further conditions—overweight/obesity and obesity.

Parents also reported two general indicators of their child’s health. Global health was measured as a single item, dichotomised as ‘very good/excellent’ versus ‘good/fair/poor’, as was the two-item short form of the parent-reported Children with Special Health Care Needs screener.17

Sociodemographic data were collected for comparison of those retained with those not retained across waves. Neighbourhood disadvantage was derived from postcodes using the Australian census-derived Socio-Economic Index For Areas (SEIFA; national mean 1000, SD 100; higher scores indicate less disadvantage).18

Statistical methods

Statistical analyses were conducted using Stata V.14.1 (StataCorp, Texas, USA, 2016). All waves used the confidentialised LSAC Wave 6 data release (December 2015), accessed under the Murdoch Children’s Research Institute institutional license.

All conditions were dichotomised (yes/no) for analysis. Prevalence estimates with 95% CIs at each wave were then calculated. Cohort-specific and wave-specific cross-sectional weights were applied to compensate for differences between the sample and the target national population arising from non-response and sample attrition. Technical papers detailing the estimation of sample and population weights for each wave are available on the LSAC website.19 Children in categories with higher non-response and attrition rates (eg, mothers who had not completed high school, first language not English) received higher weighting.19

Persistence was estimated for the cohort of children with a particular condition at any given age, by calculating the percentage that still had the condition at every subsequent wave (ie, those who still had it 2, 4, 6, 8 and 10 years later). For obesity, persistence rates included whether they continued to be classified as overweight or obese, in order to capture those who did not return to normal BMI status. B and K cohort data were averaged for ages 4–5, 6–7 and 8–9, where available. Analyses excluded diabetes (only affected <0.5% of children and unlikely to resolve), ASD, anxiety/depression and epilepsy (each measured for only three waves), and injury and dental decay (recorded as more acute events). Weighting and survey methods were not applied in the estimation of persistence as these were conducted within the subsamples of children with a condition present at a given age.

Both the prevalence and the persistence analyses were repeated using multiple imputation to account for non-response across the six waves.20 The prevalence and persistence rates estimated using the multiply imputed data are presented in the results to follow. Specifically, multivariate normal imputation models were employed using all the health conditions across the six waves and demographic (SEIFA, maternal education and Aboriginal/Torres Strait Islander status) and other factors that are known to be predictors of missingness, to derive the imputed values. Due to the large number of imputed variables, the number of imputations in our analysis was limited to the computational capacity of Stata IC (10 imputations for B cohort and 5 for K cohort). However, often 5–10 imputations are considered adequate.21 The imputed datasets were first rounded using adaptive rounding and then combined using Rubin’s rule.22 These estimates differed only slightly from those estimated using the non-imputed data (available on request from Dr Liu).


Sample characteristics

See the supplement for participant flow (online supplementary eFigure 1). Table 2 details baseline characteristics. Proportions of males (51.0%) and females were similar, the proportion identifying as Aboriginal/Torres Strait Islander was 4.1% and the mean neighbourhood disadvantage index was 1005 (SD 78).

Table 2

Baseline characteristics of the sample at Wave 1 in 2004 when B cohort children were aged 0–1 and K cohort children were aged 4–5 years

Compared to children with complete data, more children with missing data identified as Aboriginal/Torres Strait Islander (7.7% vs 2.3%) and had a mother who did not complete high school (50.3% vs 31.1%). On average, they resided in slightly more disadvantaged areas (SEIFA mean 996 vs 1009), and a slightly higher proportion had good/fair/poor (rather than very good/excellent) general health at baseline (14.6% vs 11.8%).

Estimated prevalence of common conditions (Aim 1)

At 0–1 year, 35.3% (n=5107, B cohort only) of infants experienced at least one ongoing condition. This rose steeply, with more than 60% of children 2–15 years old experiencing at least one condition (2–3, 62.9% (n=4606, B); 4–5, 60.7% (n=9369, B and K); 6–7, 60.3% (n=8706, B and K); 8–9, 71.0% (n=8416, B and K); 10–11, 72.1% (n=7933, B and K); 12–13, 68.2% (n=3956, K); and 14–15, 69.1% (n=3537, K); non-imputed values).

Figure 1 presents the estimated imputed prevalence of the reported conditions at each wave and online supplementary eTable 1 provides the underlying estimates with 95% CI. The most common condition was dental decay, reaching a peak of 32.3% at age 8–9 years. The prevalence of overweight/obesity, recent injury and special health care needs each exceeded 15% in multiple age-bands. The least common were diabetes and epilepsy (<1% from age 6–7 to 14–15). 

Four distinct prevalence patterns emerged. Eight conditions rose with age: recent injury (rising steeply from 6.7% of infants to 26.2% at 14–15), overweight/obesity, obesity, frequent headaches, recurrent abdominal pain, ADHD, anxiety/depression and ASD. The two general indicators, special health care needs and poorer general health, also rose from infancy to 14–15 (6.3% to 20.1% and 13.2% to 18.8%, respectively). Six conditions fell with age: day-wetting, soiling, constipation, recurrent tonsillitis, sleep problems and eczema, whose rates nearly halved from 2-3 (17.7%) to 14–15 (9.1%). Three conditions (asthma, diabetes, epilepsy) were relatively stable with age, while two conditions (dental decay, recurrent ear infections) peaked then fell after mid-childhood.

Persistence rates among children experiencing the condition at each age (Aim 2)

Imputed persistence estimates similarly displayed four broad patterns as demonstrated in figure 2 (online supplementary eTable 1 provides the underlying exact estimates).

Figure 2

Persistence of conditions and health status from infancy to 14–15 years. (A) Chronic conditions and general health indicators; (B) acute/recurrent conditions. The point prevalence estimates (%) for each 2-year age-band, from B or K or both cohorts as applicable, are provided at the top of each line for reference. ADHD, attention-deficit hyperactivity disorder.

Asthma and eczema showed high persistence with slow resolution. On average, 70.9% of children with asthma and 56.4% with eczema at any given wave still had the condition 2 years (one wave) later. A further quarter of these children who had these conditions at 4–5 resolved by mid-adolescence (asthma 27.9%, eczema 24.8%)

Overweight/obesity, obesity and ADHD displayed high persistence with plateauing resolution, as did the special health care needs indicator. Overweight/obesity persisted after 2 years much more often in adolescents (78.3% for those overweight/obese at 10–11) than younger children (57.9% for those overweight/obese at 2–3). Beyond age 6–7, very few obese children returned to normal BMI (7.6%, averaged across waves).

The remaining eight conditions showed low persistence rates at all ages within 2 years. Rates thereafter remained relatively stable (low persistence with plateauing resolution) for moderate-to-large sleeping problems and day-wetting, as well as good/fair/poor health status. Others (constipation, recurrent tonsillitis, soiling, recurrent ear infections, recurrent abdominal pain, frequent headaches) continued to steadily resolve after the following wave (low persistence and continuing resolution).


Principal findings

This population-based study followed a large, nationally-derived sample from infancy to mid-teen years, reporting longitudinal prevalence and persistence rates for multiple health conditions within the same non-clinical cohort. It provides age-specific answers to the question, ‘How common is it, doctor, and (when) will it get better?’ To our knowledge, such information has not been available previously.

From 2 years onwards, more than 60% of children were experiencing at least one health condition at any age, rising to over two-thirds from age 8–9 years. This suggests that experiencing ongoing health conditions may be a ‘normal’ part of childhood. Individual conditions differed widely in prevalence and persistence by 2-year age-bands, showing the importance of not aggregating data into broad age-groups. Resolution over the subsequent 2 years also varied widely. Conditions that persisted across 2 years were much less likely to resolve thereafter.

Strengths and limitations

This study was enabled by our large population-based prospective cohorts, frequent follow-up and relatively high retention rates. The resulting longitudinal prevalence graphs provide unique insight into changing prevalence rates of different health conditions throughout childhood within a single sample. The narrow age-bands provide data relevant to the individual child and family presenting to the primary clinician. Unlike persistence estimates drawn from condition-specific cohorts, our estimates included children who developed the condition post-baseline and at each age-band, giving a more accurate picture of conditions over time. The two cohorts, 4 years apart in age, enabled data collection over a wider age span (from infancy to adolescence) in a shorter period of time than would a single cohort. The cross-cohort similarity in same-age prevalence patterns provides confidence in their robustness. We used repeated measures for multiple conditions, including many that have not been otherwise studied in such large samples over time. Through multiple imputation, we accounted for the 33% of children who did not participate at all waves.

LSAC included conditions for their relevance to child health and wellbeing on expert advice, but this was somewhat arbitrary. Some topical conditions were omitted from persistence analyses because questions only entered LSAC in later waves (eg, anxiety/depression, ASD); nonetheless, these important mental health conditions were included in prevalence analyses. Other conditions were not collected at all, and food allergy and visual refractive problems were excluded due to notable variation in question design across waves. We could not estimate how care provided may have influenced persistence or, where conditions persisted, their severity and/or impact.

As in most large multi-purpose longitudinal studies, conditions other than overweight and obesity were parent-reported. Parental report is just one perspective of the prevalence of these common conditions; despite likely imperfect reliability, it is an important perspective since it drives healthcare use.23 However, relying on clinical presentation also has disadvantages: (a) it is probably impossible to collect accurately within the same population at a national level for multiple conditions simultaneously, (b) many of these conditions do not have firm diagnostic criteria and (c) not all children who genuinely have these conditions present to primary clinicians, which would lead to underestimating true prevalence through disregarding self-care and care in different sectors, for example in allied health. Probably the only way to fully understand prevalence and persistence is in studies that combine both perspectives, but no such study exists to our knowledge.

Interpretation in light of other research

ASD showed a higher-than-expected prevalence (up to 4% and 3% in B and K cohorts, respectively) compared with a systematic review where prevalence of ASD was found to be less than 1%.24 This may reflect any of several possible explanations: rising ASD prevalence estimates worldwide; inexact diagnostic processes and definitions that may vary from country to country; ASD’s known reverse social gradient,25 generating a spuriously high prevalence in long-running longitudinal studies such as LSAC, which tend to experience high attrition among the more disadvantaged participants; and/or Australia’s special Medicare-funded services for children with diagnosed ASD,26 which could potentially inflate diagnoses in an environment scarce in resources for children with special needs.

The high prevalence of obesity, asthma and ADHD into adolescence aligns with the Lancet Commission on Adolescent Health’s 2016 findings of a predominance of non-communicable disease in ‘high income’ countries.27 Over 90% of children who were obese at 6–7 did not resolve to normal weight by 14–15. This supports other studies in which only 5–20% of children with overweight/obesity resolved to normal weight after 3 years.10 Our study design enabled us to delineate fine-grained patterns in overweight/obesity resolution: resolution to normal weight beyond 2 years from any given age is unlikely, and the proportion who resolve in that 2-year ‘window of opportunity’ dwindles markedly with age. The other condition showing both high prevalence and persistence was asthma, reinforcing existing studies followed up to age 10–12.8 28 In contrast to overweight/obesity and most other conditions, both asthma and eczema have a gradual tendency towards resolution continuing to at least age 14–15. Though less prevalent, the persistence of ADHD was high, consistent with two previous studies showing persistence rates of 66% and 78% at 4 and 11 years post-baseline, respectively.12 13 Like obesity, persistence rates plateaued if ADHD had not resolved within 2–4 years of its presence at any given age.

These appear to be the first epidemiological data on persistence patterns for several conditions, some of which show high resolution rates. These include soiling, recurrent ear infections, recurrent tonsillitis and day-wetting. This could help guide informed decision-making in clinical management.29

Implications for clinicians and policy-makers

Paediatric conditions show unique prevalence and persistence patterns across childhood. These data could reassure parents that, although the majority of children are affected by at least one health condition at any given age, many resolve within a 2-year window. However, innovative intervention models are required to prevent and manage highly prevalent and persistent conditions like asthma and obesity, which have long-term implications for adult health.

This information should help policy-makers target services to chronic paediatric conditions based on age. It is also a first step towards a clinical tool to facilitate discussions between clinicians, parents and children that could provide answers, informed by strong evidence, to the frequently encountered questions: ‘How common is it, doctor, and (when) will it get better?’


This study brings new depth and breadth to understanding prevalence and persistence of paediatric conditions within the same population from infancy to mid-adolescence. Further research could determine the interplay between conditions and how they cluster within and across ages. Extending previous cross-sectional work in this same cohort, health determinants and outcomes could be examined, including healthcare use and burden throughout childhood.30–34


We would like to thank Professor John Carlin and Dr Jon Quach for their advice and guidance regarding aspects of the statistical analysis. No compensation was received for such contributions. We thank all families for taking part in the Longitudinal Study of Australian Children.


View Abstract


  • Contributors As guarantor, MW had final responsibility for the decision to submit for publication and takes overall responsibility for all aspects of this study and this manuscript. TL, KL, HH, RL and MW determined the aims of the study. TL conducted the analysis of the data with advice from FKM, JM and MHH. TL, KL, RL and MW were involved in the interpretation of data. Authors drafted (TL, KL, RL, MW) or critically reviewed the article for important intellectual content (HH, FKM, JM, MHH), and all have given final approval of the version to be published.

  • Funding This paper uses unit record data from Growing Up in Australia, the Longitudinal Study of Australian Children. The study is conducted in partnership between the Department of Social Services, the Australian Institute of Family Studies and the Australian Bureau of Statistics. The findings and views reported in this paper are those of the authors and should not be attributed to any of these three agencies. The following authors were supported by the Australian National Health and Medical Research Council: MW, Senior Research Fellowship 1046518; FKM, Career Development Fellowship 1111160; HH, Career Development Award 607351. MW was also supported by Cure Kids New Zealand. Research at the Murdoch Children’s Research Institute is supported by the Victorian Government’s Operational Infrastructure Support Program.

  • Competing interests None declared.

  • Patient consent Details have been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.

  • Ethics approval The Australian Institute of Family Studies Ethics Committee approved each wave and parents provided written consent to participate.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The integrity of the LSAC dataset rests with the Growing Up in Australia study, which makes the data available to researchers under license. TL and MW had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.