How Do Neurocognitive Tests Relate to Reported Child Difficulties at 6 Years of Age in Benin?

Abstract

Objective

The aim of this study was to examine the relationship between children’s difficulties perceived by parents in the Ten Questions questionnaire (TQ) and children’s assessments by the Kaufman Assessment Battery for Children, Second Edition (KABC-II) and the Bruininks-Oseretsky Test of Motor Proficiency, Second Edition (BOT-2).

Methods

The study was carried out in the district of Allada, a semi-rural area of Benin, involving a cohort of 562 6-year-old children. A response was considered positive if the parent reported a difficulty for their child compared to other children.

Results

The proportion of TQ-reported difficulties was 34.2%. More difficulties were reported by parents when their child had lower scores on the KABC-II (p < 0.001) and on the BOT-2 (p < 0.01). Greater family wealth, higher maternal education, and better child school attendance were associated with higher KABC-II Mental Processing Index (MPI) scores, higher BOT-2 scores, and fewer reported difficulties on the TQ. The likelihood of parents reporting difficulties was ˃60% when KABC-II MPI scores were low. When considering the KABC-II as the gold standard to define mild delays (mean minus 1 SD), the TQ demonstrated a sensitivity of 51.0% and specificity of 69.4%, which increased to 58% and 72%, respectively, when children attended school.

Conclusion

Objective assessments of children’s abilities were associated with parents’ perceived difficulties. In a low-resource rural setting in sub-Saharan Africa, the TQ is a useful tool for screening neurodevelopmental difficulties, particularly for children attending school.

INTRODUCTION

It is estimated that ˃200 million children in low- and middle-income countries (LMICs) fail to reach their full cognitive developmental potential (Black et al., 2017; Grantham-McGregor et al., 2007). To improve children’s development, a clear understanding of the risk factors and mechanisms involved is necessary. For this, tools to evaluate child neurodevelopment are needed in LMICs.

Easy-to-use screening tools are very useful in LMICs and serve as a first step toward more in-depth evaluations of children’s performances or abilities. For instance, the Ten Questions questionnaire (TQ) was designed to screen for disabilities in children aged 2–9 years. It is a rapid and low-cost screening tool for identifying children’s disabilities in low-income countries. However, little is known about how children’s performance relates to the child’s difficulties reported by caregivers. This information would be crucial for better interpreting test results and efficiently guiding families to more advanced care.

In Benin, we conducted a large birth cohort study enrolling ˃1,000 pregnant women before 28 weeks of gestation (Gonzalez et al., 2014) and following their offspring with detailed long-term neurodevelopmental assessments at 1 years and 6 years of age to identify risk factors for poor child neurodevelopment. Since 2011, we used validated tests, including the Mullen Scales of Early Learning (MSEL) and the Kaufman Assessment Battery for Children, 2nd edition (KABC-II), for which we first checked construct validity (Bodeau-Livinec et al., 2019) and demonstrated their predictive validity between 1 year and 6 years (Boivin et al., 2021). Each of these tests has been validated in children in sub-Saharan Africa (SSA) studies (Chernoff et al., 2018). When using these tests, there is a need to provide adequate and meaningful feedback to parents, especially in a context with a high prevalence of risk factors for poor child neurodevelopment. Developmental difficulties in this context are sometimes perceived by parents but are rarely assessed by neurocognitive tests that allow for an accurate diagnosis and adequate early remediation. Parental perceptions seem to be more consistent in determining mental health problems than neurocognitive developmental issues (Black et al., 2017). Awareness of developmental difficulties is low, and traditional explanatory models are widespread among parents, as well as a preference for traditional treatment options in cases of neurocognitive dysfunction (Abera et al., 2015; Adebowale & Ogunlesi, 1999; Choudhry et al., 2016). Some studies have shown that traditional beliefs and cultural differences may influence parental perceptions (Poduska, 2000), especially in Africa (Adebowale & Ogunlesi, 1999). Therefore, a better understanding of what is meaningful to parents, based on their perceptions and the child’s performance, is needed.

In our Beninese cohort, we showed that the MSEL was highly correlated with the TQ in infants, at 1 year of age, but this correlation was not known in school-aged children. Therefore, the main objective of this study is to investigate the relationship between the neurocognitive performance of children at the age of 6 years and difficulties reported by caregivers. Our goal was to provide tools to assess children’s motor and cognitive development and to screen for poor child development in a semi-rural setting in a French-speaking, low-income African country. We hypothesized that the difficulties in cognitive and motor development perceived by parents in their 6-year-old children collected with the TQ would be associated with children’s performance assessed with two neurocognitive tests, the KABC-II and the Bruininks-Oseretsky Test-2 (BOT-2) (Boivin et al., 2007). More specifically, we hypothesize that overall, KABC-II performance will be most strongly related to the TQ domains related to cognitive function (understanding, learning, speech, language, intellectual ability). In addition, we hypothesize that BOT-2 will be most strongly related to the TQ domains related to motor function (motor ability and movement).

METHODS

Study design and data collection procedure

This research is carried out as part of a longitudinal evaluation of a cohort of children recruited in Benin, West Africa. The study population consisted of children (N = 963) born to mothers who were included in a cohort from 2009 to 2010 during the second trimester of their pregnancy and followed in a trial that compared two intermittent preventive treatments for malaria (Sulfadoxine-Pyrymethamine and Mefloquine) (Gonzalez et al., 2014). Pregnant women were rigorously monitored until delivery, as were their offspring from birth to 1 year. Of these, 747 1-year-old children were assessed at the psychomotor level between 2011 and 2013 (Mireku et al., 2015), with the main objective being to study the impact of maternal anemia during pregnancy and its effect on child development at 1 year of age. At 6 years of age, children were invited to the clinic for assessment. At the clinic, 580 of these children were assessed in a quiet room alone with the examiner (Garrison et al., 2022). Only 562 children could be assessed with all tests on the same day and were retained for the study. Assessment included the KABC-II and the Bruininks-Oseretsky Test of Motor Proficiency, Second Edition (BOT-2), the Test of Variables of Attention (TOVA), the Strengths and Difficulties Questionnaire (SDQ), and the TQ. These tests were administered to assess the child’s cognitive development, motor development, attention deficits and hyperactivity, behavior, and parents’ perceived difficulties, respectively. Although the tests were provided in French to the evaluators, the instructions were given to the child and mother aloud in a local Benin language (Fon). The evaluators were psychologists and nurses trained to administer these different tests. In case of fever or illness, children were tested for malaria and anemia and treated if necessary. They were invited for psychological assessment only in the absence of symptoms.

Kaufman assessment battery for children, second edition

The KABC-II provides a comprehensive assessment of neurocognitive and psychological abilities in children up to 13 years of age, according to the French version, and has been adapted and validated in malaria studies involving African children in Uganda (Boivin et al., 2007) and in Benin during a pilot study (Koura et al., 2013). It comprises five cognitive domains: sequential processing, simultaneous processing, learning, planning, and knowledge (Kaufman A, 2004). Apart from the Mental Process Index (MPI), which provides an overall level of executive functioning, the KABC-II also produces a Non-Verbal Index (NVI). This French version is exclusively published and distributed by Éditions du Centre de Psychologie Appliquée (ECPA par Pearson) with the permission of AGS Publishing, Pearson Products Inc., with standards developed based on the French population. To ensure the quality of the assessments, a monthly independent evaluation of each interviewer was carried out (Garrison et al., 2022). Mild and moderate delays were defined as the mean from the population minus 1 and 2 Standard Deviations, respectively.

Bruininks-Oseretsky test of motor proficiency, second edition

The BOT-2 is a standardized test that provides a comprehensive measure of gross and fine motor skills in children (Bruininks, 2005). BOT-2 total scores (0–72) are age-standardized to US norms and applied as such to children in Benin.

Ten questions questionnaire

The TQ is a short, easy-to-use questionnaire that asks the caregiver to compare the child with other children of the same age regarding developmental difficulties or disabilities the child may have (Durkin et al., 1994). For example it was asked to the caregiver: “Compared with other children of his or her age, does the child appear in any way mentally backward, dull, or slow?”. It is a tool for screening children for neurological impairments and disabilities. This screening questionnaire, validated in other studies, is commonly used in LMICs (Kakooza-Mwesige et al., 2014) and has been validated for the children in this cohort at 1 year of age (Koura et al., 2013). The TQ is administered to children aged 2–9 years and is an effective screening tool for identifying cognitive, motor, sensory, and seizure disorders in children in low-resource environments (Mung'ala-Odera et al., 2004). It takes little time to administer and is typically done in a face-to-face interview. It has been used in numerous studies in several countries and provides information on children’s neuropsychological development.

In this study, the child’s caregiver was asked to compare the child with other children of the same age and list difficulties in terms of vision, hearing, behavioral difficulties, and cognitive and motor development. The TQ was used and translated into French and Fon in this cohort when children were 1 year of age (Koura et al., 2013). The same translation was used for the present study at 6 years of age. We selected specific questions from the TQ that were closely related to the KABC-II and the BOT-2. Screening with the TQ is considered positive if at least one difficulty is reported in any of the ten questions.

Socio-economic status

In our study, two variables were used to assess family socio-economic status: the family’s wealth and the mother’s level of education (Mung'ala-Odera et al., 2004) at the time of child assessment. The family wealth scale was described in another study on this same cohort of children (Koura et al., 2013) and was assessed using scoring items that included a checklist of family material possessions (radio, television, bicycle, motorcycle, and car), cow ownership, and access to electricity. Maternal education was categorized based on whether mothers had received some schooling, at least at the primary school level.

Statistical analysis

The statistical analysis included a description of the socio-demographic characteristics of the children and their mothers. Various child measures were summarized with means and standard deviations for continuous variables, and percentages for categorical variables. Maternal education (yes/no), child sex (boy/girl), and child education at the time of assessment (yes/no) were binary variables. Bivariate analyses were performed between some of these variables, which are known risk factors for poor child neurocognitive development, and a subset of the TQ. Associations between KABC-II and BOT-2 composite scores and selected subsets of the TQ were examined. Relevant questions from the TQ were selected for each test. Based on the similarities found between certain TQ questions and the neurocognitive and psychomotor assessment tests used, we investigated potential relationships. As a result, six of the TQ questions were studied in relation to the KABC-II, and two in relation to the BOT-2, to analyze possible associations. TQ questions 1 (Motor ability) and 5 (Movement), relating to motor difficulties, as well as TQ questions 4 (Understanding), 7 (Learning), 8 (Speech), 9 (Different language), 10 (Intellectual ability), and 12 (Hearing), relating to neurocognitive development, were studied for associations between BOT-2, KABC-II and TQ. Means, standard deviations, and predictive values of a TQ subset associated with KABC-II and BOT-2 composite scores were calculated. Finally, these analyses were computed in relation to at least one TQ difficulty reported. Scatterplots were created to illustrate the probability of a positive screening score according to the KABC-II scores. Logistic regression was used to predict parents’ responses to the TQ based on the KABC-II scores to explore whether low KABC-II scores were predicted by parents’ answers. All statistical analyses were performed using SAS 9.4 software, and statistical tests were two-tailed. A p-value <0.05 was deemed significant.

Ethical and data protection procedures

The EXPLORE study protocol was approved by the local Research Ethics Committee Committee of Ethical Research of the Applied Biomedical Sciences Institute (CER-ISBA) in Benin (#87). Written informed consent was obtained at the clinic or thumbprints were provided for consent if the caregiver could not read or write. Treatment for malaria, helminths infections and anemia was given if relevant. Recommendations regarding lead exposure was given.

RESULTS

General description

Of the 963 live singleton births to mothers participating in the clinical trial, 562 children were evaluated with data available for the KABC-II, BOT-2 and TQ, and are included in our analyses. The mean age was 6.5 years (standard-deviation: 0.47, age range: [4.0–8.7]). A total of 293 boys (52%) and 269 girls (48%) underwent comprehensive assessments with all relevant tests. The characteristics of children by sex, as assessed by the KABC-II and BOT-2 composite scores, are presented in Table 1. Boys presented higher score of KABC-II sequential processing scores, delayed recall scores, and BOT-2 standardized scores compared with girls.

Socioeconomic status and child neurodevelopment (Tables 2 and 3)

A large proportion of mothers had no schooling (64%). Additionally, one-third of children were not attending school. As expected, greater family wealth, higher maternal education, and better child’s school attendance were statistically associated with higher scores on the KABC-II, the BOT-II, and fewer positive responses on the TQ. Maternal marital status was not associated with KBAC-II, BOT-II, or TQ. Gender was not associated with TQ (p-value: 0.45).

Association of KABC-II and BOT-II derived scores with the TQ (Tables 4 and 5)

Some TQ scores were associated with KABC-II (Table 4), and others with BOT-2 scores (Table 5). Overall, 34.2% of parents reported at least one difficulty for their child (positive TQ). The mean KABC-II scores, including the standardized Mental Processing Index (MPI), the global MPI, the standardized sequential processing score, the standardized learning score, the standardized delayed recall score, and the standardized non-verbal index, were lower when the TQ was positive, compared to when parents did not report any difficulties for their child (p < 0.001). These associations were particularly strong for the questions related to the child’s speech (question 8) and intellectual ability (question 10). The mean BOT-II scores were associated with the question related to motor ability (question 1, p = 0.002) and with the overall TQ (p = 01), as shown in Table 4.

Probability to report a positive TQ according to KABC-II scores, sensitivity and specificity (Figs. 1 and 2, Table 6)

The probability of parents reporting a difficulty was higher than 50% when KABC-II scores were lower (Fig. 1). The probability was higher among children attending school compared with those not attending school (Fig. 2), but it was very close between girls and boys. When considering the KABC-II as the gold standard to define mild delays (mean from the population minus 1 SD), the probability of a positive report was 51.1%, reaching 57.6% when children were attending school. The sensitivity, specificity, positive predictive, and negative predictive values were 51.0%, 69.4%, 26.3%, and 86.9%, respectively. When considering the KABC-II as the gold standard to define moderate delays (mean from the population minus 2 SD), the probability was 80.0%. However, only 10 children were included in the moderate delay category.

Sensitivity analyses

One child was 4 years of age, and 3 were 5 years of age. In addition, 13 and 21 children were 7 and 8 years of age, respectively. When excluding children below 5.5 and children above 7 years of age, results were unchanged. When stratifying on gender, the same associations were found between TQ and KABC-II derived scores. The only difference was in girls where the association between TQ and NVI did not reach significance (p-value = 0.12).

DISCUSSION

The aim of this study was to examine the relationship between parents’ perceived difficulties, as reported in the TQ and children’s performance on the KABC-II and the BOT-2. We observed a strong association between the parents’ reports and the neurocognitive performance-based assessments of the children in this cohort. To our knowledge, this is the first study to investigate the relationship between parents’ reports and actual assessments of children’s performance in such a large study in SSA. We found that 34.2% of parents reported at least one difficulty on the TQ. This is consistent with a study conducted in Uganda where 27% of children screened positive (Kakooza-Mwesige et al., 2014).

Associations on motor questions

There was a significant association with the BOT-2 and questions 1 (Motor ability) and 5 (Movement) of the TQ. These results are comparable to those of a study conducted in Benin on this same cohort of 1-year-old children, which showed the association of these two questions with the motor functions of the MSEL (Koura et al., 2013). However, the associations for motor functions appeared weaker than those for cognitive outcomes. These difficulties may be harder for parents to detect unless they are more pronounced.

Socio-demographic conditions

Boys scored higher than girls on some cognitive tasks. Studies have shown that, apart from sex, children’s family wealth, environment, and education were significantly correlated with the raw, standardized measure of global cognitive performance on the KABC-II and MPI (Bodeau-Livinec et al., 2019; Boivin et al., 2021). Family wealth was also associated with higher scores on the receptive language scale (Koura et al., 2013).

Difficulties reported

Analysis of the TQ items, combined with the KABC-II and BOT-2, revealed that the TQ was effective in detecting difficulties in children. Significant associations were found between TQ-reported difficulties and neurocognitive test scores. The TQ screen is considered positive if any of the questions is answered affirmatively. When children are not attending school, many parents have fewer benchmarks to gage their children’s neurocognitive development (Mung'ala-Odera et al., 2004). Indeed, the TQ asked parents to compare their child with others of the same age. These children share similar socio-economic and cultural conditions, meaning the children being compared are likely to have similar risk factors for poor development. Parents have more reference points for comparison when children are in school. The probability of TQ-reported difficulties increases with very low KABC-II and BOT-2 scores, and exceeds 50%. These results may be influenced by socio-cultural and economic factors, including the high number of children not attending school, limited access to external influences (such as media), and parents’ lack of awareness about developmental and neurocognitive disabilities. However, the cutoff used in our population was the mean minus one standard deviation, which is not considered indicative of developmental delay, but may include some children with difficulties. Therefore, these results are consistent with what might be expected.

Usefulness of TQ screen

The validity of the TQ has been extensively studied in Africa and around the world, including countries such as Bangladesh, Jamaica, Pakistan, Saudi Arabia, Ghana, South Africa, and Kenya (Mung'ala-Odera et al., 2004) for children aged 2–9 years. A previous study showed the validity of the TQ in the children in this cohort when they were 1 year old. The results revealed a high number of false-positive results (Koura et al., 2013). However, the TQ was significantly associated with the MSEL at 1 year of age. Compared to results in the same children at 1 year of age, the TQ sensitivity was lower at 6 years (81.5% at 1 year versus 51.0% at 6 years), but the specificity was higher (30.1% at 1 year versus 69.4% at 6 years). With the TQ, caregivers are asked to compare their children with other children. In a setting with high prevalence of risk factors for developmental problems, it may be difficult for caregivers to detect problems for their children. This may explain the differences in sensitivity and specificity between 1 year and 6 years of age. These results, with less false-positives, are particularly important in this low-resource setting.

In a study of children aged 6–9 years conducted in Kenya, the TQ was found to be sensitive to visual and hearing impairments but not sensitive to cognitive disorders (Bitta et al., 2021), and could even underestimate child development. Another study in Bangladesh found that a good validity of the TQ to screen disabilities in 10–16 year old children in relation to the Wechsler Intelligence Scale-Revised (WISC-R) assessments by psychologists as a reference standard (Hasan et al., 2023).

By combining the scores from the KABC-II and the BOT-2 with the TQ, the results showed that parents whose children had low KABC-II or BOT-2 reported more difficulties in their children on the TQ. This association highlights the value of the tests in detecting difficulties in children. The TQ can be used to identify developmental delays, which can serve as an indicator of disability. These results are consistent with studies conducted in Uganda, Bangladesh, Pakistan, and Jamaica, which also used the TQ to screen for cognitive impairments in children aged 2–9 years (Durkin et al., 1994). It can be inferred that the TQ-positive children identified in this study represent minimum estimates, as the TQ reliably detects only moderate and severe disabilities, not mild conditions (Mung'ala-Odera et al., 2004).

The standardized scoring of such “western-based” tests necessitates the use of norms from children in American (KABC in English) or European (KABC in French) countries to adjust the scale performance for age. This is because country-specific norms are typically not available for the LMICs in which the KABC is used. Although the resulting standardized scores for African children on the KABC are typically below that of the normative means devised for those measures, the factor structure and construct validity of the test itself remains intact (Bangirana P, 2009). Boivin and Giordani (2009) have suggested that this is because tests such as the KABC can provide a reasonable and valid characterization of neuropsychological brain/behavior integrity in African children because of the more universal features of the foundational neurocognitive skills and abilities derived from this foundational brain/behavior “omnibus” (Boivin & Giordani, 2009).

Children identified by the TQ as being at risk of disability could be referred for evaluation using more in-depth diagnostic measures (MSEL, KABC-2, BOT-2, etc.) to benefit from specific psychiatric, neurological, neurocognitive, or neurodevelopmental treatments (Lorencz EE, 2013). Especially, the abilities related to language and overall learning disability were highly correlated to TQ, and may be of great importance for parents, and should be detected and treated. In this population, where resources are scarce and access to care is limited, these tests are essential and easy to use.

Limitations

Although the results of this study are encouraging and demonstrate a significant association between the KABC-II, BOT-2, and TQ, there are limitations to be addressed. It is important to note that the sample of children evaluated is not nationally representative, as it consists only of parents and children living in a semi-rural region. This population faces high illiteracy rates, and socio-economic and cultural conditions differ from those in developed countries.

A total of 747 children were assessed at 12 months of age, but only 562 were assessed at 6 years of age. This attrition is attributed to loss to follow-up and high mortality rates. Children lost to follow-up tended to come from slightly more disadvantaged families. Although this may slightly underestimate the proportion of positive TQ results, it is unlikely that the observed associations were significantly affected.

We used standardized scores for both the KABC-II and BOT-2. The scores in our study were relatively low and should be interpreted with caution, as no local standardization exists for Benin. Using Western norms may not be optimal, especially regarding age standardization (Bodeau-Livinec et al., 2019). However, because our study compared children within the same population, all of whom were of same age, this limitation is minimized. Furthermore, we determined the cutoff for mild cognitive deficits according to the mean and standard deviation within our population.

Despite having a large sample size, only 10 children (1.8%) fell into the category for moderate delay (KABC-II MPI mean minus two standard deviations). This aligns with expectations, but sensitivity and specificity could not be calculated for this group due to the small number of children in this category.

Children were also assessed the TOVA and the SDQ. These tests were not included in the analyses as they measure other dimensions. These two other dimensions, including an objective assessment (TOVA) of children’s abilities and one reported by parents (SDQ) were also associated (Zoumenou et al., 2025).

Finally, this study is large, with trained interviewers and standardized tests. A qualitative survey could complement this work by providing deeper insights into parents’ perceptions of their child’s development.

CONCLUSION

In this study, the TQ was significantly associated with both motor development and neurocognitive assessments. As a screening tool, the TQ could help identify children who would benefit from early intervention, ultimately contributing to the development of the countries concerned. The TQ is easy to use, requires few resources, and could be integrated into routine primary care for children in areas at risk for poor cognitive development.

FUNDING

Financial Support Data collection costs were supported by funding from the Fondation de France, France (00100075) to F.B-L.

Conflict of interest

None declared.

Author contributions

Romeo Zoumenou (Investigation, Writing—original draft), Jaqueline Wendland (Conceptualization, Funding acquisition, Supervision, Writing—review & editing), Victoria Jacobsen (Formal analysis, Writing—review & editing), Michael Boivin (Conceptualization, Investigation, Writing—review & editing), Nathalie Costet (Conceptualization, Formal analysis, Methodology, Writing—review & editing), and Florence Bodeau-Livinec (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Software, Supervision, Validation, Writing—review & editing)

References