Exploring the Influence of Early Developmental Intervention on Preterm Infants’ Cognitive and Physical Development: A Meta-Analysis

3.3. Risk of Bias Assessment

All 15 trials reported appropriate randomisation procedures, with 7 specifying parallel or stratified RCT designs, indicating strong methodological consistency. 3 studies described adequate blinding, yielding higher scores (5/5). The remainder lacked feasible double blinding due to the nature of developmental interventions. All studies accounted for participant follow-up or stated reasons for attrition Table 4. Overall, median Jadad score = 3 (moderate quality), indicating low-to-moderate risk of bias across the evidence base.

4. DISCUSSION

4.1. Summary of Main Findings

The present meta-analysis consolidates recent randomised controlled evidence on the impact of early developmental interventions on cognitive and physical (weight gain) outcomes in preterm infants. The pooled findings reveal that structured, multisensory, and parent-mediated interventions enhance neurocognitive performance and growth trajectories more than standard neonatal care. These results reaffirm the biological and psychosocial rationale that early, enriched environments during the critical window of brain plasticity can alter developmental pathways toward more adaptive outcomes, directly addressing the study’s aim to quantify their overall effectiveness across both domains.

4.2. Interpretation in Context

The earlier mentioned positive impact corresponds with an increasing amount of neonatal research showing that early stimulation promotes the maturation of the nervous system and the organisation of behaviours, particularly in parent-child interactions. Recent contributions by Liang et al. [19] and Birberg Thornberg et al. [10] demonstrate that family-centred developmental care and early collaborative interventions lead to measurable changes in the Bayley Scales of Infant Development (BSID-III) scores, confirming the earlier work of Nordhov et al. [5] and Kynø et al. [17], who reported similar cognitive benefits that persisted until the child turned five. The mechanisms resonating in the benefits of early child imitative stimulation and movements are multifactorial, and the integration of tactile, visual, and auditory stimuli facilitates synaptogenesis and myelination in the cortical areas of attention and memory. At the same time, Gilkison [24] explains that positive parent–infant emotional exchanges support socio-emotional regulation and self-regulatory behaviours. Crucially, the homogeneity across studies in the cognitive synthesis suggests a stable and reproducible effect of developmental care despite variations in country, gestational age, and program format. This consistency contrasts with earlier meta-analyses of Spittle et al. [25], where heterogeneity was moderate, reflecting methodological variability in older trials. One plausible explanation is the refinement of intervention design over the last decade; current programs integrate standardised protocols, trained facilitators, and validated outcome tools, thus reducing inter-study noise. This reinforces the conclusion that cognitive advantages are statistically and clinically meaningful when developmental programs are structured and caregiver-engaged.

The results obtained are also consistent with other studies, particularly within neuroscience, which suggest that some structural components of the brain are amenable to modification if worked on during the early stages of development. Lean [26] noted increased cortical thickness and improved white matter integrity in preterm infants who, during their NICU stay, participated in parent-infant interaction programmes. These neuroanatomical changes support the view that the majority of behavioural advances are a result of real neurodevelopmental change rather than simply the passive effects of the assessment. Similarly, Baumann et al. [27] found that infants who participated in early development programmes based on physiotherapy exhibited increased functional connectivity in the regions of the brain engaged in the integration of sensory and motor activities, thus showing improved cognitive function and underlying neural plasticity. That said, the gain in mental functioning captured and quantified in this analysis was positive yet not substantial. This echoes Stephenson’s [28] findings, which suggest that preterm-born children have a far wider range of development than their term-born peers. The absence of longitudinal studies and inconsistent follow-up data after infancy, however, limits the claim that these developmental programmes having long-lasting effects. Such compensatory interventions may provide some relief but are not a long-term solution.

4.3. Mechanistic Explanations

Outcomes of extreme preterm weight results, for example, meta-analyses and developmental programmes appear to be the most heterogeneous of the outcome measures, while others are more elaborate and stricter. For example, reports on weight gains and other measurements of growth variables. On the other hand, Maguire et al. and Fan et al. [15, 16] reported weak and not statistically significant negative growth. Such differences might be due to variations in the intensity and duration of the interventions. Furthermore, supportive, nurturing and prolonged periods of exposure to rest metabolic and growth hormonal stimulators increase stimulation.

The effect seems to align with the model proposed by Pirazzoli [29], which states that additional touch and movement hyperstimulation enhances vagal tone with resultant improvements in gastric motility, caloric absorption, and secretion of insulin-like growth factor. Controlled massage and kangaroo care, in particular, have been shown to improve weight gain, thermoregulation, and overall stability in the postnatal epoch, especially in infants less than 32 weeks of gestation [30]. These findings, thus, support the view that the benefits of developmental practices are not only numerous but are also essential for optimising neural and metabolic change in the post-hospital home environment.

4.4. Limitations

The lack of homogeneity demonstrates the presence of differing levels of intervention effectiveness. Caregiver education with sustained home involvement (e.g., family integrated care, early collaborative intervention) appears to be more effective than short, centre-based approaches. This is consistent with Spratt’s [31] work regarding emotionally supportive intervention parent sessions, where the parent control sessions are predicted to result in greater weight and length gains than the therapist-alone model due to the supportive caregiving behaviours reinforced in the sessions. Further context moderators, such as socio-economic status, maternal education, and healthcare system configuration, particularly matter. Pitchik et al.’s [32] work is illustrative of the greater and more consistent gains coming from high-income countries relative to low-resource countries with insufficient post-discharge follow-up, implying dilution of the intervention effects. Importantly, while short-term growth is the primary target of the immobilisation interventions, the long-term impact is more uncertain. Haschke et al. [33] observed the weight and length trajectory converge to the control group by the age of 2 years corrected age, implying supportive gains would not be maintained without ongoing intervention. This underscores the need to integrate developmental programmes within multi-faceted care approaches across the early childhood years.

The symmetrically shaped funnel plot indicates slight publication bias towards the cognitive outcomes, which validates the thorough nature of the studies about reporting. On the other hand, the weight-gain funnel plot shows an asymmetry, which suggests the presence of small-study effects or the selective publication of favourable outcomes. This asymmetry is characteristic of neonatal interventions, in which smaller, single-centre studies dominate. This suggests that although the effects have a strong direction, convincing evidence stems from the necessity of future large-scale multicentre RCTs with standardised outcome reporting.

Despite the strengths above, there lies a series of limitations. The analysis of weight gain demonstrated a substantial Degree of heterogeneity, which stemmed from differences in intervention type, parental training, and Unit of Care protocols (Lee et al. [18]; Namprom et al. [13]). Most of the studies had small sample sizes. The duration of the follow-up (3–24 months) in comparison with the data available was relatively short, resulting in the data perhaps being compressed, inflating the short-term effects and underestimating the long-term persistence (Liang et al. [19]; Yu et al. [23]). Blinding was often not possible, increasing the risk of performance bias, and the different methods of anthropometry (absolute gain vs. z-scores) complicated the standardisation of the data. The funnel plot demonstrates asymmetry, suggesting a potential publication bias due to small single-centre studies with positive outcomes. Finally, several studies failed to control for the mother’s socio-economic, nutritional, and mental health, which limits causality. The wide 95% prediction interval quantitatively underscores between‐study variability and moderates the generalisability of results. Policymakers should prioritise multi-center, adequately powered RCTs with standardised measurement frameworks to strengthen neonatal intervention guidelines and inform equitable global implementation.

The strength of the meta-analysis is its rigorous focus on randomised controlled trials (RCTs). They constitute the foundation of any rigorous methodological framework and play a critical role in minimising selection bias. For example, robust randomisation and transparent reporting of trials supported the internal validity of pooled cognitive and growth outcomes, particularly those by Birberg Thornberg et al. [10] and Silveira et al. [12]. Trials such as Van Hus et al. [11] and Altunalan et al. [9], which are well-characterised, added greater comparability through the standardised implementation of the Bayley Scales of Infant Development (BSID-II/III). Maguire et al. [15] and Fan et al. [16] enriched the longitudinal scope by evaluating developmental care given post-discharge. Their use of both fixed and random effects models increased the analytical generalisability of the data by reducing variability in studies on the intervention’s cross-gestational age, duration, and intensity. PRISMA 2020 and Cochrane protocols concerning meta-analysis and systematic reviews on new studies (2021-2025) ensured the integration of new high-quality studies reflecting current neonatal practices to parental involvement, and a structured developmental framework by Youn et al. [22] and Samsudin et al. [14] targeted abstraction and reproducibility. Indeed, the addition of outcomes, as well as developmental somatic outcomes, on top of cognitive outcomes provides multiple angles on neurodevelopment, underscoring the relevance of this work to clinician translators in neonatology and early years’ policymakers.