Assessing developmental delay

An 18 month old boy was referred to the local child development centre because he was not walking independently. He walked around by holding on to furniture; he had first crawled at age 11 months.

We wanted to use an evidence based approach to guide our assessment and management, so before seeing him, we considered three issues. Firstly, we wanted to know whether, in an 18 month old child referred to a community paediatrician because he cannot walk independently, the most likely outcome is cerebral palsy, a primary muscle disorder, another neurological abnormality, or simply delayed motor development. Secondly, we wanted to know how many children with Duchenne muscular dystrophy present with delayed walking at 18 months. Thirdly, we wanted to know whether a creatine kinase level is a worthwhile screening test if there are no abnormal signs.

Search

To determine the most likely outcome, we used a Medline database to obtain evidence (at http://biomed.niss.ac.uk). We used a filter to select articles dealing with prognosis (from wwwlib.jr2.ox.ac.uk/caspfew/filters/). We found additional papers by using the exploded form of “prognosis” (“exp prognosis”) instead of the text word. We combined this filter with “developmental disabilities” and “motor skills” or “locomotion” in the exploded form and “limiting to infants under 2 years.” To assess prevalence of Duchenne muscular dystrophy we used a simple search for articles dealing with early diagnosis. To determine whether screening is worth while we used data obtained from the two searches described above.

Determining likely underlying cause

We identified 45 papers. After studying the abstracts, we obtained the five most promising papers from the library. The article by Chaplais et al,¹ was the most relevant prospective, population based study. The others contained information irrelevant to our question. We appraised the article using a template provided by Sackett et al.²

Appraisal of study by Chaplais et al¹

Sample—Was a defined, representative sample of patients assembled at a common point in the course of their disease? General practitioners and health visitors screened 84% of infants in a community population. This is usually considered a sufficient proportion to provide an estimate of prevalence.

Follow up—Was the follow up of patients sufficiently long and complete? Follow up was for at least 30 months. A longer follow up would have been reassuring, as pathological reasons for late walking might subsequently become obvious.

Blinding—Were objective outcome criteria applied in a blind fashion? Assessments were not blind; a paediatrician assessed referred children at home.

Adjustments—If subgroups with different prognoses were identified, was there adjustment for important prognostic factors? Bottom shuffling and a family history of late walking were both found in half of the idiopathic late walkers.

Conclusion—We had minor concerns about the length of follow up and the diagnostic methods. We felt that the population was similar to ours and the outcomes were applicable to our patient. Although this was the largest study found in our search, less common conditions—for example, muscular dystrophy—may be underrepresented (see below).

Importance of results of study by Chaplais et al¹

Under 10% of children presenting with delay in walking had neurological problems (cerebral palsy or minor neurological abnormality in 5.8% (95% confidence interval 2.9% to 8.7%)); 11% had delay in other areas, but limited follow up did not allow precise developmental diagnosis. Without abnormal examination findings or delay in other areas, we were confident that the child was unlikely to have a substantial underlying abnormality.

Prevalence

We found 123 potential articles. Studying the abstracts, we selected 14 articles to review.^3–16 The paper by Gardner-Medwin et al seemed to be the basis for the practice of screening boys who are not walking at 18 months.¹⁴ They described when normal controls and patients with Duchenne muscular dystrophy first walk. The methods were poorly described, but half of their patients with Duchenne muscular dystrophy walked at 18 months compared with 97% of normal controls. Clearly, half of patients with Duchenne muscular dystrophy (those who walk before 18 months) are missed using this screening method.

Screening

Bradley et al gave the incidence of Duchenne muscular dystrophy as 1 in 3802 male births in Wales.⁶ If half of these cases present with delayed walking at 18 months, this equates to 1 in 7604 male births. If 3% of the male population are not walking at 18 months, this equates to 228 in 7604 male births. Therefore in our child (not walking at 18 months), the risk of having Duchenne muscular dystrophy is about 1 in 228.

We would need to screen 228 children presenting in this way, to detect one case of Duchenne muscular dystrophy. This “number needed to screen” allows comparison of the cost of testing 228 children with not detecting one child with Duchenne muscular dystrophy at age 18 months. Testing is relatively cheap in most countries (about £10 in the United Kingdom) but causes discomfort to children and causes parental anxiety regarding the diagnosis. An undetected child will present later with deteriorating muscular dystrophy. Parents might miss the opportunity of genetic advice or antenatal testing. Clearly this is a complex decision.

Conclusion

This case shows how an evidence based approach can find, analyse, and apply evidence to medical problems. Our management was guided by evidence not available from textbooks. Decisions remained personal to the doctor and the family but were better informed. In this case a normal screening test excluded muscular dystrophy, and the child subsequently walked at age 20 months.

Figure.

Using an evidence based approach to assessing developmental delay can be useful