Reading abilities in school-aged preterm children: a review and meta-analysis

Abstract

Aim: Children born preterm (at ≤32wks) are at risk of developing deficits in reading ability. This meta-analysis aims to determine whether or not school-aged preterm children perform worse than those born at term in single-word reading (decoding) and reading comprehension.

Method: Electronic databases were searched for studies published between 2000 and 2013, which assessed decoding or reading comprehension performance in English-speaking preterm and term-born children aged between 6 years and 13 years, and born after 1990. Standardized mean differences in decoding and reading comprehension scores were calculated.

Results: Nine studies were suitable for analysis of decoding, and five for analysis of reading comprehension. Random-effects meta-analyses showed that children born preterm had significantly lower scores (reported as Cohen's d values [d] with 95% confidence intervals [CIs]) than those born at term for decoding (d=-0.42, 95% CI -0.57 to -0.27, p<0.001) and reading comprehension (d=-0.57, 95% CI -0.68 to -0.46, p<0.001). Meta-regressions showed that lower gestational age was associated with larger differences in decoding (Q[1]=5.92, p=0.02) and reading comprehension (Q[1]=4.69, p=0.03) between preterm and term groups. Differences between groups increased with age for reading comprehension (Q[1]=5.10, p=0.02) and, although not significant, there was also a trend for increased group differences for decoding (Q[1]=3.44, p=0.06).

Interpretation: Preterm children perform worse than peers born at term on decoding and reading comprehension. These findings suggest that preterm children should receive more ongoing monitoring for reading difficulties throughout their education.

Figure 1

Flow diagram for all stages of article selection.

Figure 2

Forest plots of the effect sizes and heterogeneity statistics for decoding and reading comprehension meta-analyses. Negative effect sizes indicate poorer performance on reading tasks for the preterm sample in comparison with the term-born sample. CI, confidence interval; T, term-born; PT, preterm.

Figure 3

Meta-regression (slope with corresponding 95% CIs) of decoding effect sizes with (a) gestational age of the preterm sample and (b) age at assessment. Mean gestational age of the preterm sample was significantly associated with decoding. Age at assessment was not significantly associated with decoding.

Figure 4

Meta-regression (slope with corresponding 95% CIs) of reading comprehension effect sizes with (a) gestational age of the preterm sample and (b) age at assessment. Mean gestational age of the preterm sample was significantly associated with reading comprehension. Age at assessment was significantly associated with reading comprehension.