Discussion
This study found ADHD symptoms prevalence of 6.3% (95% CI; 3.72-10.33) among children visiting a busy paediatric accident and emergency unit of a tertiary care private “not-for profit” teaching hospital. Although seven children were described by parents as having ADHD (hyperactivity-impulsivity and inattention) symptoms they had no observed functional impairment hence they did not meet the diagnostic criteria for inclusion as ADHD. Further, the low response rate from teachers made it difficult to determine whether or not there was academic dysfunction attributable to the condition thus limiting our ability to estimate actual prevalence of ADHD. We unfortunately received only six reports back from teachers despite reminders. This meant we had to rely exclusively on parental reports based on observed home behavior. But parents were also expected to recall the history over the past six months which may suffer from recall bias. A combination of these factors could have contributed to underestimate of the true prevalence. The estimate is nevertheless comparable to prevalence of 5.3% and 6% in neighboring Congo and 8% in Nigeria suggesting error may be marginal [6,7]. The other studies were carried out in schools hence may not be comparable in terms of study population characteristics. Regardless of the setting, it is evident that symptoms of ADHD are prevalent enough in our population to warrant concern. In a technical review by Green et al., prevalence of ADHD in the community ranged from 4-12% compared to 2%-5% in the paediatric clinics suggesting similarity in burden in the two populations. They however observed that prevalence in paediatric clinics varied widely in the few studies available for analysis [9].
Magnitude of prevalence of ADHD is influenced by the criteria used. This type of variation is not unusual as illustrated by Wolraich et al. who encountered a similar inaccuracy in diagnosis when 4323 children were evaluated for ADHD in 10 schools in Tennessee [19]. They found a prevalence of 16% when ADHD diagnosis was based on symptoms alone compared to 6.8% when both symptoms and functional impairment was used as per diagnostic criteria requirement. In review of global prevalence of ADHD, Polanczyk et al. attributed variability to methodological differences [20]. The American Academy of Paediatrics recommends behavioral interventions for children who do not meet the full diagnostic criteria for ADHD although evidence in support of the practice is weak [16].
Unlike some other investigators, this study did not observe any gender difference between children with and without symptoms of ADHD. This could be explained by the small sample size filing to detect a true difference if it indeed existed. It could also have been caused by some unidentified seasonal occurrence like preferential referral of girls over the study period. The National Survey of Children’s Health reported a male to female prevalence ratio of 2.5:1 with clinic based populations showing 10:1 [21]. Spencer et al. attributed the gender difference to boys presenting with disruptive behaviour being referred as compared to girls with inattentive behaviour [22].
We found some association, albeit weak, between past injury, especially burns, and ADHD despite the low power of the study. Whereas a larger sample size is needed to examine this further, Tai et al. prospectively looked at “injury-proneness” of children aged six to eighteen years and found children with ADHD to have a 2-5 fold increase in risk of injury [23]. Additionally, this study found the predominant type of injury to be burns. The findings concurred with those of Fritz et al. [12].
A striking observation from our study was the up to 20-fold increase in risk of repeating classes in children with symptoms of ADHD as a manifestation of poor academic performance. This phenomenon should increase index of suspicion for ADHD among health professionals. In a study of a class of 700 by Breslau et al. on impact of early behavior disturbances on academic achievement, students with attention problems were found to be inefficient learners which limited their ability to acquire basic skills necessary for higher education [11].
Unlike other studies, ADHD in our study was not associated with oppositional defiant disorder, anxiety, depression and conduct disorders in this study [17]. This may be attributed to the fact that our study was not powered to detect such an association if it indeed exists.
VAS would clearly not be recommended for ADHD screening in view of the low sensitivity found in this study as many with the condition would be missed out. However, the high specificity and high positive likelihood ratio argue a case for its use in already suspected diagnosis from say, history of poor school performance or injury and suggestive symptoms of ADHD. Testing positive in such patients would then suggest strong need for referral to a psychiatrist for further assessment. Utility of other behavioral scales such as Conner’s Questionnaires and Strength and Difficulties Questionnaire as alternatives to DSM IV need to be further investigated in subsequent studies.