Child: care, health and development

Current Literature doi:10.1111/j.1365-2214.2009.00987.x

Current Literature cch_987 586..591

Relations among speech, language and reading disorders. Finally, the authors present the evidence for aetiological/
Pennington B. F. & Bishop D. V. M. (2009) Annual Review of genetic overlap in these disorders. They emphasize that SSD,
LI and RD have complex, multi-factorial aetiologies and that
Psychology, 60, 283–306.
the genes involved have probabilistic rather than dichotomous
In this paper, we critically review the evidence for overlap effects which also interact with the child’s environment. They
conclude that LI, RD and SSD are familial, heritable and are
among three developmental disorders, namely speech sound
linked to trait loci in certain chromosomal regions. There is
disorder, language impairment and reading disability, at three some evidence of trait loci which may produce both RD and
levels of analysis: diagnostic, cognitive and etiological. We find SSD phenotypes but less evidence for linkages with LI. However,
that while overlap exists at all three levels, it varies by comor- the need for much more research in this area is underlined, as is
bidity subtype, and the relations among these three disorders the need to understand the nature of influence of environmental
are complex and not fully understood. We evaluate which factors.
Any child presenting to child health and educational profes-
comorbidity models can be rejected or supported as explana-
sionals with an RD, SSD and/or LI should therefore have access
tions for why and how these three disorders overlap and what to comprehensive and detailed assessment of their speech and
new data are needed to better define their relations. language abilities to uncover all areas of need and the probable
implications for educational progress, in particular for literacy
Developmental models of developmental disorders emphasize
skills, should be acknowledged and addressed. Intervention
interactivity between domains over time. In this context explo-
approaches should include work to improve phonological pro-
ration of the neglected issue of co-morbidity could provide
cessing skills for all three groups of children.
insights regarding the genetics, cognitive processes and aetiology
of these disorders. This paper explores co-morbidity with
respect to language impairment (LI), reading disability (RD) Cristina MacKean
and speech sound disorder (SSD). Newcastle University
The evidence for co-morbidity among these disorders is sum-
marized with reference to relative risk (RR). All combinations
occur at above chance levels (RD + LI = 1.9–6.9; RD + SSD =
1.6–2.5; LI + SSD = 2.2–6.9); however, they vary with age as SSD Skeletal surveys in infants with isolated skull fractures.
tend to resolve or respond to treatment in the pre-school or Wood J. N., Christian C. W., Adams C. M. & Rubin D. M.
early school years and RD can only present once literacy instruc-
(2009) Pediatrics, 123, e247–e252.
tion has begun.
Of key interest to practitioners is the ability to determine
DOI: 10.1542/peds.2008-2467.
which children may be at risk of developing later RD. In children
Objective The goal was to describe the utility of skeletal
with LI and/or SSD, an increased risk of later RD is almost
entirely restricted to SSD + LI (RR = 4.6–8.9) whereas the rate of surveys and factors associated with both skeletal survey use and
later RD in SSD without LI is negligible. The relative risk of RD referral to child protective services for infants with skull frac-
in LI without SSD ranges from 1.9 to 6.9 and further longitudi- tures in the absence of significant intracranial injury.
nal research is recommended to consider whether resolved SSD Methods A retrospective chart review was performed for
in some children with LI could explain the variability in inci- infants who were evaluated at a tertiary children’s hospital
dence found between studies.
because of an isolated, non-motor vehicle-related, skull
The evidence for cognitive overlap is explored and placed in
the context of current cognitive models of LI, SSD and RD. The fracture between 1997 and 2006. Logistic regression analyses
authors advance models such that, for all three disorders were used to test for associations of demographic factors,
phonological deficits are a core underlying cause. RD and LI, clinical findings that raised suspicion for abuse (absence of
however, are described as involving multiple deficits; children trauma history, changing history, delay in care, previous child
with RD having deficits in both phonological processing and
protective services involvement and other cutaneous injuries),
rapid serial naming (a difficulty thought to imply slower per-
ceptual or processing abilities), and children with LI having and fracture type (simple vs. complex) with the primary out-
deficits in both phonological processing and ‘grammatical’ lan- comes of skeletal survey use and reports to child protective
guage skills. services.

586 © 2009 Blackwell Publishing Ltd

Current Literature 587

Results Among the 341 infants in the study, 31% had clinical Effectiveness of educational materials designed to change
findings that raised suspicion for abuse and 42% had complex knowledge and behaviors regarding crying and shaken-baby
skull fractures. Skeletal surveys were obtained for 141 infants syndrome in mothers of newborns: a randomized, controlled
(41%) and detected additional fractures for only two (1.4%) of trial.
those 141 infants. Child protective services reports were made Barr R. G., Rivara F. P., Barr M., Cummings P., Taylor J.,
for 52 (15%) of the 341 children. Both infants with positive Lengua L. J. & Meredith-Benitz E. (2009) Pediatrics, 123,
skeletal survey findings had other clinical findings that raised 972–980.
suspicion for abuse, and they were among those reported. With DOI: 10.1542/peds.2008-0908.
controlling for race and age, Medicaid-eligible/uninsured
Background Infant crying is an important precipitant for
infants were more likely than privately insured infants to receive
shaken-infant syndrome.
skeletal surveys and child protective services reports in the pres-
Objective To determine if parent education materials [The
ence of a complex skull fracture or clinical findings that raised
Period of PURPLE Crying (PURPLE)] change maternal knowl-
suspicion for abuse.
edge and behaviour relevant to infant shaking.
Conclusion Skeletal surveys were ordered frequently for
Methods This study was a randomized, controlled trial con-
infants with isolated skull fractures, but they rarely added addi-
ducted in prenatal classes, maternity wards and pediatric prac-
tional information, beyond the history and physical findings, to
tices. There were 1374 mothers of newborns randomly assigned
support a report to child protective services.
to the PURPLE intervention and 1364 mothers to the control
This paper warrants careful reading because it reports striking
group. Primary outcomes were measured by telephone 2
and stark findings from a large well-documented case series months after delivery. These included two knowledge scales
despite some acknowledged weaknesses in the study. The paper about crying and the dangers of shaking; three scales about
presents evidence in support of a change in guidance about child behavioural responses to crying generally and to unsoothable
abuse investigations. The weaknesses in the study are that real crying, and caregiver self-talk in response to unsoothable
life and clinical practice is invariably messier than carefully
crying; and three questions concerning the behaviours of
designed research protocols, and as a result the main findings of
this study – that no occult fractures are found among infants sharing of information with others about crying, walking away
with isolated skull fractures as long as there are no other ‘red if frustrated and the dangers of shaking.
flag’ risk factors for maltreatment – depend on only those cases Results The mean infant crying knowledge score was greater
who actually had a skeletal survey. These were a minority of all in the intervention group (69.5) compared with controls (63.3).
infants with an isolated skull fracture. Although there were clini-
Mean shaking knowledge was greater for intervention subjects
cal features which influenced the decision of whether to carry
out a skeletal survey, there were also socio-economic and insur-
(84.8) compared with controls (83.5). For reported maternal
ance status factors. The authors acknowledge this may have led behavioural responses to crying generally, responses to
to underascertainment of bony injuries among some specific unsoothable crying, and for self-talk responses, mean scores for
socio-economic groups, and overinvestigation of child maltreat- intervention mothers were similar to those for controls. For the
ment among others. behaviours of information sharing, more intervention mothers
However, these cavils aside, these are important findings and
reported sharing information about walking away if frustrated
ought to influence clinical policy and practice. The point of the
paper is that risk factors for maltreatment ought to influence and the dangers of shaking, but there was little difference in
the decision of whether or not to carry out a skeletal survey, sharing information about infant crying. Intervention mothers
rather than the presence of a skull fracture. The authors found also reported increased infant distress.
no evidence that simple or complex fractures were any more or Conclusions Use of the PURPLE education materials seem to
less likely to be associated with other injuries. The paper makes
lead to higher scores in knowledge about early infant crying
the point that unnecessary skeletal surveys can expose infants
to radiation, and parents to distress. The related question,
and the dangers of shaking, and in sharing of information
which this paper is unable to touch on, is which of these behaviours considered to be important for the prevention of
infants require a computerized tomography scan – as the shaking.
amount of radiation from this far exceeds that from plain
radiography. Preventing non-accidental head injury in infants remains a
major challenge in all countries as far as I am aware. This trail
Richard Reading provides evidence about one possible step. The package
described which focuses on infant crying does seem effective in
increasing mothers’ knowledge of crying and how to deal with

© 2009 Blackwell Publishing Ltd, Child: care, health and development, 35, 4, 586–591
588 Current Literature

this. However, behavioural outcomes were less modified by the seats but not booster seat use. Research is needed to identify
intervention, and what participants of the trial know and write methods and messages that will empower centre providers
in their study diaries may not equate with what they actually do
to promote booster seats effectively and reach high-risk
– all the more so as the intervention explains to them what they
should do. There is no evidence that change in knowledge trans-
populations.
lates into change in the desired endpoint of reduced rates of
head injury, even though it is a necessary prerequisite. It is also This is a good example of a research study where an educational
questionable whether mothers are the appropriate target group and behavioural intervention led to an increase in knowledge
for an intervention to reduce non-accidental head injury in but not a change in the end safety practice. As well as informa-
infants given that the majority of adult perpetrators are fathers tion and advice, free booster seats were distributed, and a law
or other male carers. That aside, this is an important trial was enacted in the state mandating safe restraint of children
because provision of information and advice is a practical and under 5 years of age. Despite all this there was little effect of
feasible step that healthcare providers could easily implement. fitting and use of the seats. There are potential weaknesses in a
cluster randomized trial such as this, perhaps the greatest being
contamination of the effect of the intervention to the control
Richard Reading
population, although the study team convincingly argue that
this was not likely on a major scale. They are currently working
Effects of a booster seat education and distribution program on a culturally sensitive intervention among Latino families in
order to understand this disconnection between what ought to
in child care centers on child restraint use among children
happen and what does happen.
aged 4 to 8 years.
Thoreson S., Myers L., Goss C. & DiGuiseppi C. (2009) Richard Reading
Archives of Pediatrics & Adolescent Medicine, 163, 261–267.

Objective To compare child care centre-based booster seat

education and distribution with no intervention when imple-
Pediatric primary care to help prevent child maltreatment:
mented immediately after booster seat legislation.
the Safe Environment for Every Kid (SEEK) model.
Design Cluster randomized controlled trial.
Dubowitz H., Feigelman S., Lane W. & Kim J. (2009)
Setting Thirty-nine urban child care centres.
Pediatrics, 123, 858–864.
Participants A total of 854 parents and 1010 children aged 4 to
DOI: 10.1542/peds.2008-1376.
8 years in vehicles leaving centres.
Intervention We trained 168 staff members at 20 centres to Context Effective strategies for preventing child maltreatment
give parents and children messages promoting booster seats and are needed. Few primary care-based programmes have been
supplied lesson plans, children’s activities, and free booster developed, and most have not been well evaluated.
seats. Objective Our goal was to evaluate the efficacy of the Safe
Main outcome measures Observed booster seat use, ‘good Environment for Every Kid (SEEK) model of pediatric primary
practice’ restraint use and legal restraint use. care in reducing the occurrence of child maltreatment.
Results Parents at intervention centres were more likely to Methods A randomized trial was conducted from June 2002
report receiving restraint information from the centre [adjusted to November 2005 in a university-based resident continuity
odds ratio (AOR), 4.06; 95% confidence interval (CI), 2.48– clinic in Baltimore, Maryland. The study population consisted
6.67], speaking with staff about booster seats (AOR, 3.95; 95% of English-speaking parents of children (0–5 years) brought
CI, 2.26–6.88) and using fit to decide when to move children in for child health supervision. Of the 1118 participants
into seat belts (AOR, 3.39; 95% CI, 1.91–5.99). Groups did not approached, 729 agreed to participate, and 558 of them com-
differ in proportions using booster seats (44% vs. 43%; AOR, pleted the study protocol. Resident continuity clinics were
1.03; 95% CI, 0.62–1.73), good practice (42% vs. 41%; AOR, cluster randomized by day of the week to the model (inter-
1.11; 95% CI, 0.70–1.74) or legal restraints (65% vs. 65%; AOR, vention) or standard care (control) groups. Model care con-
0.79; 95% CI, 0.48–1.31). Results were similar for children aged sisted of (1) residents who received special training; (2) the
4 to 5 and 6 to 8 years. All outcomes were significantly less likely Parent Screening Questionnaire; and (3) a social worker. Risk
among children riding in pickup trucks or with Hispanic or factors for child maltreatment were identified and addressed
black drivers. by the resident physician and/or social worker. Standard
Conclusions The intervention increased parents’ receipt of care involved routine pediatric primary care. A subset of
information from centre staff and knowledge about booster the clinic population was sampled for the evaluation. Child

© 2009 Blackwell Publishing Ltd, Child: care, health and development, 35, 4, 586–591
Current Literature 589

maltreatment was measured in three ways: (1) child protective The clinical effectiveness of different parenting programmes
services reports using state agency data; (2) medical chart for children with conduct problems: a systematic review of
documentation of possible abuse or neglect; and (3) parental randomised controlled trials.
report of harsh punishment via the Parent-Child Conflict Dretzke J., Davenport C., Frew E., Barlow J., Stewart-Brown
Tactics scale. S., Bayliss S., Taylor R. S., Sandercock J. & Hyde C. (2009)
Results Model care resulted in significantly lower rates of child Child and Adolescent Psychiatry and Mental Health, 3, 7.
maltreatment in all the outcome measures: fewer child protec- DOI: 10.1186/1753-2000-3-7.
tive services reports, fewer instances of possible medical neglect
Background Conduct problems are common, disabling and
documented as treatment non-adherence, fewer children with
costly. The prognosis for children with conduct problems is
delayed immunizations and less harsh punishment reported by
poor, with outcomes in adulthood including criminal behav-
parents. One-tailed testing was conducted in accordance with
iour, alcoholism, drug abuse, domestic violence, child abuse
the study hypothesis.
and a range of psychiatric disorders. There has been a rapid
Conclusions The SEEK model of pediatric primary care seems
expansion of group-based parent-training programmes for the
promising as a practical strategy for helping prevent child mal-
treatment of children with conduct problems in a number of
treatment. Replication and additional evaluation of the model
countries over the past 10 years. Existing reviews of parent
are recommended.
training have methodological limitations such as inclusion of
Evidence that clinic-based interventions can reduce levels of non-randomized studies, the absence of investigation for
child maltreatment is always welcome and noteworthy, espe-
heterogeneity prior to meta-analysis or failure to report
cially if the evidence is high-quality such as that from a ran-
domized controlled trial. The programme described here was confidence intervals. The objective of the current study was to
fairly simple, of training doctors and providing them with systematically review randomized controlled trials (RCTs) of
some resources to identify and manage risk factors, of provid- parenting programmes for the treatment of children with
ing screening instruments and resources to increase parent’s conduct problems.
awareness of thesefactors, and providing a dedicated social Methods Standard systematic review methods were followed
worker in the clinic. This appears to be an easily implemented
including duplicate inclusion decisions, data extraction and
intervention and may already be being done in many centres.
One interesting result not mentioned in the abstract is the quality assessment. Twenty electronic databases from the fields
large numbers of mother’s with depressive symptoms who of medicine, psychology, social science and education were
were identified in the intervention clinics – which is worth comprehensively searched for RCTs and systematic reviews to
considering in the light of the proposition by David Finkelhor February 2006. Inclusion criteria were: RCT; of structured,
that one reason for the reduction in US child maltreatment
repeatable parenting programmes; for parents/carers of chil-
reports in the last decade may be the increase in psychomedi-
cation use in the population. dren up to the age of 18 with a conduct problem; and at least
What of the weaknesses in the study? This was a cluster ran- one measure of child behaviour. Meta-analysis and qualitative
domized trial. These work best spread across a wide area and synthesis were used to summarize included studies.
incorporating large numbers of clinics – to avoid the problems Results A total of 57 RCTs were included. Studies were small
of contamination of the intervention, bias in ascertainment of with an average group size of 21. Metaanalyses using both
the outcome measures and low statistical power. This study
parent (SMD1 –0.67; 95% CI: –0.91, –0.42) and independent
suffered all of these, although these would be likely to have
minimized any differences in results. There is a source of bias (SMD –0.44; 95% CI: –0.66, –0.23) reports of outcome
that may have falsely biased the results in the other way showed significant differences favouring the intervention
however. The authors say the analysis was by intention to treat group. There was insufficient evidence to determine the rela-
although this does seem to be strictly true as they only tive effectiveness of different approaches to delivering parent-
included families who completed the protocol in the analysis.
ing programmes.
Some of the outcome measures, such as review of referrals to
child protection services, could have been collected for all Conclusion Parenting programmes are an effective treatment
entrants into the study, which would have increased confidence for children with conduct problems. The relative effectiveness of
in the findings of the study. The authors acknowledge that different parenting programmes requires further research.
these results can only be considered as preliminary, but they do
provide a strong argument for more extensive evaluations of This is an extensive and high-quality review of the evidence
this programme. backing up the value of parenting interventions in the man-
agement of children with conduct problems. It is not the first
Richard Reading 1
Standardised Mean Difference.

© 2009 Blackwell Publishing Ltd, Child: care, health and development, 35, 4, 586–591
590 Current Literature

such review, and there are several others referred to in the • Health policy informed by composite outcome measures that
paper; however, this paper claims to have tighter inclusion and take account of both the U5MR and the distribution of the
evaluation criteria. The paper finds consistent evidence that
burden of mortality across social groups would help to over-
such programmes are effective, and there is little evidence that
any one approach is significantly superior to other approaches.
come this.
This may be to do with the small number of high-quality This is a bit of an old chestnut – and has been discussed exten-
studies and more research is needed to evaluate differences sively with respect to health inequalities within countries: the
between different types of programme – for example, group vs. difficulty of improving overall levels of health without increas-
individual, behaviourally orientated vs. relationship orientated, ing inequalities in health. Many interventions seem to do just
etc. this, particularly those aimed at behavioural change or policy
If the research evidence is so unequivocal about the benefits, which is more likely to benefit the rich than the poor. This
why do they seem so patchy in clinical experience? Certainly, paper makes the case for international health, and the risks
some parents return from these programmes with a trans- of uncritical pursuit of the Millennium Development Goals
formed relationship with their children, but many more find it without considering the impact on possible inequity. As the
of little benefit, or an initial improvement is not sustained. The Millennium Development Goals are largely about the resolving
review gives some insight into this. First, the research studies by problems associated with extreme poverty, this is a very impor-
and large select families who are keen to be involved. Second, we tant issue.
may clinically be seeing the ‘hard core’ of families who are The paper attempts to illustrate this with a hypothetical
perhaps less likely to benefit. Third, we may in fact have too example, which does make the case well but inevitably overlooks
pessimistic a view of the potential of families to change. Fourth, the complexity and subtleties of the real world. The case is made
and most interestingly to me, the measurement of effect has for an equity adjusted measure of the overall mortality, and, by
mainly been carried out with validated measures of behaviour. extension, the other Millennium Development Goals as well. In
But we know little about how these measures translate into other words, if the aim is to reduce the under 5 mortality rate by
quality of life or health utility. In other words, what does a two-thirds, how much would we be prepared to accept a short-
significant change in the Child Behaviour Checklist score trans- fall if this was done equitably across social groups. This inevita-
late into in terms of clinical effect and parents’ or schools’ views bly requires some numerical adjustment which would be the
of the child in question? As ever ‘more research is needed’, but subject of intense debate. Even if this were never agreed, the
this review helps define very clearly what this research should be debate and argument itself would be of immense value.
directed towards.

Richard Reading
Richard Reading

The millennium development goals fail poor children: the Assessment of herd protection against trachoma due to
case for equity-adjusted measures. repeated mass antibiotic distributions: a cluster-randomised
Reidpath D. D., Morel C. M., Mecaskey J. W. & Allotey P. trial.
(2009) PLoS Medicine, 6, e1000062. House J. I., Ayele B., Porco T. C., Zhou Z., Hong K. C., Gebre
DOI: 10.1371/journal.pmed.1000062. T., Ray K. J., Keenan J. D., Stoller N. E. & Whitcher J. P. (2009)
The Lancet, 373, 1111–1118.
Summary points:
Background Trachoma-control programmes distribute oral
• The Millennium Declaration is a statement of principles azithromycin to treat the ocular strains of chlamydia that cause
about the kind of future that world governments seek; a the disease and to control infection. Theoretically, elimination
future that they envisage to be more equitable and more of infection is feasible if untreated individuals receive an
responsive to the socially most vulnerable. indirect protective effect from living in repeatedly treated
• The Millennium Development Goals (MDG4) represent the communities, which is similar to herd protection in vaccine
operational targets by which we may judge their actions. programmes. We assessed indirect protection against trachoma
• The reduction of the under 5-year-old mortality rate (U5MR) with mass azithromycin distributions.
by two-thirds by 2015 is one of the MDG4. Methods In a cluster randomized trial, 24 subkebeles
• The reduction in U5MR can, however, be achieved through a (government-defined units) in Amhara, Ethiopia, were ran-
diversity of policy interventions, some of which could leave domized, with use of a simple random sample, to distribution
the children of the poor worse off. A celebrated MDG4 success four times per year of single-dose oral azithromycin to children
can, thus, be a Millennium Declaration failure. aged 1–10 years (12 subkebeles, 4764 children), or to delayed

© 2009 Blackwell Publishing Ltd, Child: care, health and development, 35, 4, 586–591
Current Literature 591

treatment until after the study (control; 12 subkebeles, 6014 As with immunization, treatment of children can have a wider
children). We compared the prevalence of ocular chlamydial effect on public health in the whole population. Trachoma is a
dreadful blinding disease in which childhood infection leads to
infection in untreated individuals 11 years and older between
chronic inflammation and irritation leading to scarring of the
baseline and 12 months in the treated subkebeles, and at 12 cornea and blindness in adulthood. This disease is now confined
months between the treated and control subkebeles. Healthcare to low income countries, particularly in Africa, but remains one
and laboratory personnel were blinded to study group. Analysis of the leading causes of blindness and impaired vision globally
was intention to treat. The study is registered with clinicaltri- (although not the leading cause which is uncorrected refractive
als.gov, number NCT00322972. error). This complex and extensive trial shows that repeated
antibiotic treatment of children under 10 years of age reduces the
Findings At 12 months, 637 children aged 1–10 years and
incidence of infection across the whole community. This is clearly
561 adults and children aged 11 years and older were analysed a very important study – as the authors point out, children are
in the children-treated group, and 618 and 550, respectively, in more easily accessed for treatment than adults and treating this
the control group. The mean prevalence of infection in chil- group is cheaper than treating the whole population.
dren decreased from 48.4% (95% CI 42.9–53.9) to 3.6% (0.8– This is not the last word on the subject. The study also com-
pared this strategy with a one off treatment of all the population
6.4) after four mass treatments. At 12 months, the mean
which led at 1 year to even lower infection rates, but evidence
prevalence of infection in the untreated age group (ⱖ11 years) from elsewhere would suggest this is unlikely to give any more
was 47% (95% CI 33–57) less than baseline (P = 0.002), and long-lasting protection. An accompanying editorial [Taylor H.
35% (95% CI 1–57) less than that in untreated communities R. (2009) Elimination of blinding trachoma revolves around
(P = 0.04). children. The Lancet, 373, 1061–1063] describes the WHO SAFE
Interpretation Frequent treatment of children, who are a core policy for Trachoma which includes facial cleanliness and envi-
ronmental change to reduce overcrowding, etc., and comments
group for transmission of trachoma, could eventually eliminate
that this may have an even greater impact, although regular
infection from the entire community. Herd protection is offered targeted azithromycin treatment of children does seem an
by repeated mass antibiotic treatments, providing a strategy for important additional strategy.
elimination of a bacterial disease when an effective vaccine is
unavailable. Richard Reading

© 2009 Blackwell Publishing Ltd, Child: care, health and development, 35, 4, 586–591