Background:

Psychopathology in children can be conceptualized as a normal development that has gone awry. That is, some conducts which are expectable at a certain age could turn to be inappropriate and pathological if they persist. When some traits, conducts or signs are very conspicuous and they are frequently present together, they are conventionally called syndromes.

Studies registering children’s observed conducts by the parents have been very useful to identify groups of symptoms, and several scales have been designed to elicit psychopathology such as the Children’s Behavior Questionnaire (CBQ), Conner’s scales, and the Child Behavior Checklist, CBCL-P. With the exception of the CBQ, the other two instruments, although frequently used as screening instruments in several studies, are too long or too specific to be systematically employed at the general practice services and in the community. More recently, Goodman designed the Strengths and Difficulties Questionnaire (SDQ), which is a 25-item instrument showing an acceptable predictive validity for three groups of disorders: conduct, emotional, hyperactivity and inattentive. In Mexico, our epidemiological work on children’s mental health started at the end of the eighties using the Report Questionnaire for Children (RQC) which is a 10-item screening instrument developed at the end of the seventies for a WHO collaborative research with the aim of extending psychiatric services to primary care settings. In our population, the instrument showed good efficiency with a positive predictive value (PPV) of 76% and a negative predictive value (NPV) of 99%, and it has been useful in detecting mental health problems both in the general population, as well as in primary care services. However, the need to identify what kind of disorders are they and estimating their prevalence remains.

The Brief Screening and Diagnostic Questionnaire (CBTD) was built based on previous experience using the RQC. Seventeen items which explored symptoms frequently reported as motives for seeking attention at the out-patient mental health services were added to the original 10 questions of the RQC. Most of them are items included in the CBCL-P, which explore hyperactivity, impulsivity, attention deficit, sadness, inhibition, oppositional and antisocial behaviors, and eating behaviors associated with low or high weight. The aim was to include cardinal symptoms that could lead to identify probable specific syndromes and disorders, based on the parent’s report.

The reliability of the instrument was measured using the Kuder- Richarson coefficient (KR-20), obtaining a 0.81 value. Based on responses obtained in a general population sample of 1686 children aged 4 to 16 years in Mexico City, the score at the 90th percentile, five symptoms, was established to define probable caseness. Also, using logistic regression analysis, the association between the cardinal symptoms for different disorders -as defined in the DSMIV and ICD-10 diagnostic criteria- and the rest of the items from the questionnaire was studied in order to obtain symptom profiles or syndromes signaling probable psychiatric disorders.

The main objective of the present study was to evaluate the concurrent validity and the efficiency of the diagnostic algorithms of the CBTD, as compared with the psychiatric diagnoses of children attended at two out-patient mental health services in Mexico City.

Method:

A random sample of consecutive new out-patients aged 4 to 16 years was obtained. The CBTD was administered to the accompanying parent before the consultation. Clinical evaluation was done independently and blind to these results; the psychiatrists emitted diagnoses following the ICD-10 criteria. Diagnostic reliability between this initial evaluation and further diagnosis of hyperactivity and attention deficit disorder, depressive disorder, oppositional and conduct disorder and anxiety disorders, established at the different clinics of the children’s psychiatric hospital showed good agreement with Kappa values ranging from 0.60 to 0.83.

Concurrent validity between the diagnostic algorithms of the CBTD and the psychiatric diagnoses was measured using Kappa and Yule statistics. Efficiency measures: sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were also obtained.

Results:

A total of 530 children were included in the study; 73% were male and 27% female; mean-age was 9.35 (s.d. 3.12) year old.

Only eight patients reported less than five symptoms in the CBTD. The 4/5 cut-off point showed a sensitivity of 98.7% and PPV of 99.8%. However, as very few children were below the cut-off point, specificity resulted in 50% and NPV was 12.5%.

Also, as the children attended the psychiatric services, they showed a highly symptomatic profile (median= 11 symptoms), concurrent validity analyses were first carried out in a sub-sample including only those patients with three CBTD syndromes at the most (n= 102). Diagnostic algorithms for attention deficit and hyperactivity, depression, and conduct disorders showed fair agreement with the corresponding psychiatric diagnoses: Yule statistic range from 0.43 to 0.55. As it could be expected, sensitivity (range: 71% to 84%) and NPV (range: 85% to 97%) were higher for the most general algorithms, while specificity and PPV were higher for the most stringent definitions.

Analyses including the whole sample showed a sensitivity ranging from 54% to 95%, and NPV from 70% to 98% for the different diagnostic algorithms, and thus indicating a high efficiency of this brief instrument.

Conclusion:

The CBTD seems to be a good and efficient screening instrument, useful for the detection of the most frequent psychiatric disorders in childhood and early adolescence. Results suggest that it should be tested and incorporated as a tool at primary health services for the systematic surveillance of mental health during childhood and adolescence.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )