The prevalence of communication delays or disorders is increasing and may be ascribed to environmental factors such as unemployment, limited medical resources, lack of educational services, violence, crime and HIV or AIDS (Guralnick 2013). Paediatric HIV or AIDS, for instance, is a challenging condition as it not only influences the well-being of infants but also results in prematurity and low birth weight, and later attention difficulties and speech and language delays (Rossetti 2001; Samuels, Slemming & Balton 2012). South African infants and children are particularly vulnerable because of the high prevalence of predisposing environmental factors such as these (Mayosi & Benatar 2014; Samuels et al. 2012).

The high prevalence of developmental delays or disorders amongst infants in South Africa (Samuels et al. 2012) necessitates selection and implementation of effective screening or developmental surveillance tools to identify at-risk infants as early as possible (Van der Linde et al. 2015). If communication delays remain undetected until primary school years, a child is at greater risk for behavioural problems, academic failure and socio-emotional disturbances (Squires et al. 2009; Yew & O’Kearney 2013). With a direct link between school performance, communication skills and the role that communication plays in general development and emotional and behavioural outcomes, the importance of early identification of communication delays is obvious (Rossetti 2001; Wankoff 2011; Yew & O’Kearney 2013). Early identification of and early intervention for infants at risk can prevent or reduce future developmental difficulties and academic failure whilst improving the quality of life for the infant and family (Samuels et al. 2012). Furthermore, these services can prevent or lessen developmental and communication difficulties (Hawa & Spanoudis 2014), which implies less future financial expenditure for parents with regard to medical costs, transport fees to medical centres and/or speech therapy expenses.

Infants receiving early intervention services, including early detection by means of developmental screening and/or surveillance as first point of access, make greater progress when the whole family is involved (Guralnick 2013). Because parents are usually the first to identify their children’s developmental difficulties, they are considered a good resource by healthcare providers when conducting screening tests (Williams & Holmes 2004). The resource-constrained public healthcare system in developing countries like South Africa (Mayosi & Benatar 2014) requires time-efficient and accurate screening tools to ensure it is practically feasible with low false-positive rates that do not result in over-referral. Parents can be used as a resource in identifying their child’s strengths and weaknesses (Glascoe 2013), and providing important information to professionals. A parent-administered test may therefore be appropriate for the South African context if it is sufficiently accurate and time efficient. Furthermore, selecting a comprehensive screening tool that accurately detects communication delays in addition to other developmental delays may be more suitable than developmental domain–specific screening tools in the South African, resource-constrained public healthcare context.

Early identification of developmental delays, including communication delays, can be facilitated by a variety of valid standardised tools. In South Africa, developmental screening is implemented nationally as part of the Road to Health Booklet (Tarwa & Villiers 2007). However, the Road to Health Booklet has not been validated and its accuracy for developmental screening has been questioned (van der Linde et al. 2015). The Ages and Stages Questionnaire or ASQ (Squires et al. 2009), Denver Developmental Screening Test II (Frankenburg et al. 1992) and the Parents’ Evaluation of Developmental Status or PEDS (Glascoe 1997) are all well validated and standardised screening tools with large bodies of supporting evidence (Macy 2012). All three tools include infants from birth; however, the Denver Developmental Screening Test II is a clinician-administered test, whereas the ASQ and PEDS tools are parent-administered tools (Macy 2012).

The ASQ and the PEDS elicit parental concerns regarding their children’s development and behaviour. In a comparison study conducted in Canada, both the ASQ and PEDS rendered similar outcomes and it was concluded that either one can be selected for implementation (Limbos & Joyce 2011). Taking into consideration the cost of the tools and administration time, the PEDS tools have been deemed more appropriate for use in the South African primary healthcare (PHC) context.

Whilst a recent study evaluated the accuracy of the PEDS and PEDS-Developmental Milestones (PEDS-DM) for developmental delays in the private healthcare sector in South Africa (Silva 2010), the accuracy of the PEDS test detecting communication delays or disorders in infants in the South African PHC context has not yet been established. This study therefore evaluated the accuracy of the PEDS tools in detecting communication delays in infants, aged 6–12 months, in a PHC context in South Africa.

A comparative cross-sectional within-subject design was employed to evaluate the accuracy of the PEDS tools in detecting communication delays using the Rossetti Infant-Toddler Language Scales (RITLS) as a gold standard.

The fail rates of the PEDS, PEDS-DM and PEDS tools were high (47% – 52%). This was to be expected as an at-risk population was used. Several high-risk factors for developmental delay were present in the study population. The majority of participants had one or more risk factor(s) for developmental delays, such as poverty (71%), three or more children in a home (32%) and limited parental education (62%). An estimated 45% of the South African population is poor, whilst 20% live in extreme poverty (Statistics South Africa 2011). Multiple risk factors increase the probability that development will be delayed (Glascoe & Leew 2010) and high-risk children are 24 times more prone to have IQs below 85 than low-risk children (Sameroff et al. 1987).

Specificity and sensitivity values of an accurate screening tool should fall between 70% and 80% (Glascoe 2013). The results in this study demonstrated domain-specific (i.e. expressive language and receptive language) sensitivity scores that were low to very low across the PEDS, PEDS-DM and PEDS tools. Such low sensitivity values may result in a failure to identify a large number of infants who require early communication intervention services. The PEDS tools, on the other hand, did show an accurate sensitivity (71%) and specificity (73%) rating for receptive and expressive language and socio-emotional domains in combination. High sensitivity and specificity for socio-emotional developmental delays indicated that the infant delays in the study sample were accurately identified by means of the PEDS and PEDS tools. Autism spectrum disorders, for example are characterised by such impairments in social interaction, communication and behaviour, which are ostensible before the age of 3 years (Baio 2012). Because the results of this study indicated that PEDS and PEDS tools are able to accurately detect socio–emotional developmental delays in infants, these tools may possibly aid in the early diagnosis of autism spectrum disorders in PHC.

The lack of parental concern regarding their infants’ communication development in the current study population, as illustrated by the fail rate of the PEDS for receptive (3%) and expressive language (3%), were similar to previous research findings. A study performed by Glascoe (2013) revealed that parents of infants, 11 months or younger, do not have many communication-related concerns. However, when there are concerns, it usually pertains to their children’s motor, health, behavioural, self-help and socio-emotional skills (Glascoe 2013). This is possibly because gross motor milestones, such as sitting and crawling, are more observable than infant’s speech sounds and language comprehension (Glascoe 2013).

The low sensitivity and specificity ratings of the screening tools for receptive (22% – 44%) and expressive language (5% – 23%) reported in the current study are likely because of the difficulty to identify communication delays before the age of 12 months (Eadie et al. 2010). It can be expected that parents’ awareness of their child’s communication development might be better at a later stage when the child is older and more communicative (Eadie et al. 2010). It is therefore recommended that future research should evaluate the accuracy of the PEDS tools for communication delays in 2- to 5-year-old children within the South African PHC context. Because the interviews and assessments were not conducted in the home languages of the sample population, it may be deemed a limitation of the current study. Future research should explore the accuracy of translated tools in detecting communication delays in infants and young children. Preventative strategies, such as developmental surveillance and awareness campaigns, should be considered as a way to support underserved communities where the majority of infants are at risk of communication and/or other developmental delays.

The PEDS tools demonstrate limited sensitivity scores for receptive and expressive language domains in young infants, although sensitivity for the socio-emotional domain was high. Obtained values for the PEDS tools did demonstrate a high degree of accuracy when considering a combination of receptive and expressive language and socio-emotional domains with sensitivity and specificity of 71% and 73%, respectively. Future research determining accuracy of the PEDS, PEDS-DM and PEDS tools for children aged 2–5 years in detecting communication delays should be explored.