The aims of the present study were to develop the clinical use of the parent-reported SDQ in preventive child health care by public health nurses, describe the parent-reported SDQ, evaluate empirical cutoff values within the context of the Starting Right™ project in relation to the Swedish, Danish, and UK cutoffs, and evaluate the representativeness of the study sample with regard to parental socioeconomic status.
Our main findings showed that boys had higher total difficulties and impact scores than girls. The differences in means between boys and girls were largest in the case of externalizing symptoms and hyperactivity subscore. However, girls had higher prosocial scores than boys. Our findings consistently indicated that girls had better SDQ scores than boys in the included age group. Moreover, fewer children would be identified as having mental health challenges using the UK cutoff values than using the Scandinavian cutoff values. Applying the 80th percentile cutoffs in the present study, 158 of 665 children were identified as having mental health difficulties.
The mean parent-reported total difficulties scores in the present study were nearly similar to those reported in previous Norwegian [18] and Nordic studies [22]. Even though Norwegian norms and cutoffs are not established [18], the availability of country-specific normative data is of interest in psychosocial research because psychosocial functioning is known to be country- and culture-specific [17]. In the Starting Right™ project, the UK cutoffs were used to guide the child and school health nurses’ interpretation of individual results. As shown in the results, the nurses may overlook more children (approximately 67%) with or at risk of mental health problems using the UK cutoffs than using the Swedish/Danish cutoffs or the “in-study” 80th percentile cutoffs (Table 3).
Even though cutoffs can be difficult to establish, the SDQ total difficulties score correlates with mental health challenges at the full range of scores [24]. Moreover, the use of different concepts related to what SDQ measures are reported in the literature; Sveen et al. [23] used emotional and behavioral disorders whereas Goodman [15] used the concept of “a total difficulties score.” Consequently, an interpretation by a clinician may be either in the direction of using the score and a cutoff to diagnose the child (or hypothesize a diagnosis and refer to specialist health care), or to gain insight into a child’s mental health symptoms as perceived by the parent of the child. Moreover, screening may be the first step in a dialogue and provide valuable insight into children’s mental health. Use of the SDQ may facilitate communication between family/child and the public health nurse, and areas of concern can be identified and discussed. Such knowledge will be important to make decisions regarding the child’s needs and possible support strategies. For clinical use, gender and population-based 80th and 90th percentile cutoffs may help the clinician to focus their efforts on understanding and supporting the children with the most mental health symptoms, which represents a different approach than categorizing the children in terms of psychopathology.
Although Kornør and Heyerdahl [18] did not make the abovementioned distinction clear, they suggested that the parent-reported SDQ should not be used to screen for psychopathology, which would refer to a clinical use of the SDQ for diagnosing disorders. Nevertheless, they emphasized the importance of a low cutoff if the SDQ is to be used in municipal services with a low incidence of mental disorders, which may apply for the current study. However, a recent systematic review concluded that the SDQ demonstrated predictive validity for language and behavioral concerns in preschool-aged children in a community setting [27].
Sveen et al. [23] suggested a Norwegian cutoff score ≥10 to determine psychopathology at age 4 years with satisfactory sensitivity and specificity. However, in their follow-up study of children aged 6 years, they reported many nonpersistent cases and a rather low positive predictive value (9.5%) and a high negative predictive value (99.6%) [28]. Such findings would practically mean that children at 4 years of age with low or unproblematic scores are at low risk of mental health problems 2 years later, while children identified with problems at 4 years of age may not have persistent problems 2 years later. Hence, the clinical implications may be in line with the recommendations made by Kornør and Heyerdahl [18], to not use the parent-reported SDQ as a tool for suggesting or predicting psychiatric disorders. In line with this finding, a Danish study among preschool children reported that the SDQ was useful for screening at a preschool age to identify children at an increased risk of mental health problems. However, the authors emphasized that early screening with the SDQ predictive algorithms cannot stand alone, and repeated assessments of children are needed, especially regarding internalizing mental health problems [29]. Screening tools rely on predictive validity and may imply the risk of false positive and/or false negative cases [1]. Applying the 90th percentile from the current study as a cutoff (≥ 12, see Table 2) for abnormal range in the study population of Sveen et al. [28] would have led to a sensitivity of 54%, which would implicate many cases missed. Our study has a limited contribution for establishing Norwegian norms and cutoffs for diagnostic purposes, whereas the novelty concerns the study of clinical anamnestic considerations. In the current study context, screening is not used to decide who is receiving follow-up from the public health nurses, but to inform the content of the follow-up and strengthen the dialog between the public health nurses and the family. For instance, public health nurses may guide parents about how to relate to children with different difficulties reported by the parents using the SDQ instrument. Our study’s 80th percentile, as well as the study by Sveen et al. [28], indicate that public health nurses should pay attention to all children with a total difficulties score above 9 or 10. Nevertheless, context and age specific cutoffs to guide content of follow-up in primary health care need further investigation.
Moreover, Nilsen et al. [10] reported that internalizing and externalizing mental health problems can be traced to as early as 18 months of age. However, they did not measure mental health problems using the SDQ. In addition, children who show early signs of mental health problems and have mothers who receive the appropriate support can change their trajectory in a healthier direction [10]. A plausible clinical implication would then be to increase or facilitate support to parents of children with the most symptoms. The latter may support a clinical use of the SDQ, not to diagnose or predict diagnosis and problems, but to adjust the effort of the primary child health care providers to individuals at current risk.
In our study, parents overall reported more total difficulties symptoms among boys than girls, particularly driven by externalizing symptoms and the subdomain of hyperactivity. This finding is in line with those of earlier studies. In a Dutch study, boys (aged 4–5 years) scored higher than girls on the hyperactivity and total difficulties domains, and more boys than girls scored in the clinical range of prosocial behavior [17]. Hence, public health nurses should be aware that gender differences could represent the different needs of children.
The mental health of a child may follow certain trajectories but also vary through age. An important factor determining mental health according to the Norwegian TOPP study is personality, and how the environment of the child challenges, or reacts/responds to, the child’s needs [10]. A recent policy statement of the American Academy of Pediatrics also highlighted the importance of personalizing the response to individual children’s needs when facing any adversity [30]. Hence, the SDQ may represent an outcome mostly relevant to the child (and family) from a subjective perspective. Because children are born with different personalities and temperaments, and may face adversity in different environments, they also need different nurturing responses from their environments for their healthy development [7, 10, 31].
Asking children and families about the children’s symptoms, acknowledging and being responsive to the child’s personality and needs, facilitating reduction of family stress, and helping the parents to incorporate core skills, as suggested by Garner and Yogman [30], may represent a feasible and suitable use of the SDQ in primary health care. In such a context, cutoffs may be used as advisory instead of definers and as markers of unhealthy trajectories and/or reflections of adversity. Advisory cutoffs may then motivate efforts in child primary health care on children’s own premises.
Representativeness of the study sample
Our study sample represented all groups of socioeconomic status and immigrant backgrounds; however, it had a minor overrepresentation of parents with higher education and Norwegian background. Nevertheless, 11% of mothers and 10% of fathers had basic school level education (9–10 years of schooling) only.
Representativity in population-based studies may often be a challenge. Public health nurses have also raised concerns about whether all types of families can be included in the Starting Right™ project and if responses will be received from the immigrant population owing to barriers due to language skills or technical issues (smartphone and the secure ID). Our findings indicated that few responses (1%) were received from parents of children in cases where both parents and the child were born outside Norway and had since immigrated to Norway. However, in general, the proportion of children in the population who had immigrated was rather low (4%). For children born in Norway, we had a relatively representative sample including children with and without immigrant backgrounds.
Strength and limitations
The consent rate to this study was 63%, which we consider relatively high. However, it is a limitation that we did not have information about the group not consenting to this research. Data from the included individuals indicated that they represented children with different socioeconomic backgrounds both in terms of education and income, which is a strength of this study. However, a limitation is that our income data included the percentile of income compared with the income of the entire population aged > 16 years, while our study population only included mothers and fathers with a mean age of 35.2 years and 37.6 years, respectively. Responses from mothers were overrepresented because in most cases, the health centers only had the mothers’ phone numbers. Hence, the text message may only have been sent to the mother. Furthermore, despite that the current study corresponds with previous Scandinavian studies concerning identification rates, we cannot tell if the children in need of follow-up are identified. However, the instrument is implemented through ordinary services and all children receive individual follow-up by the public health Nurses.
Implications
Advisory cutoffs of the SDQ, relying on the Scandinavian, but not on the UK, norms may be used to reflect children’s individual and present needs in Norway, and help public health nurses to personalize their care and focus on children and families with the highest needs. Parent-reported SDQ in children aged 4 and 6 years can be representatively collected in municipal health services using an online tool. Furthermore, Norwegian SDQ norms and cutoffs should be further developed.
