Triangulating evidence, expertise, and maternal experience to design a gestational age-specific stimulation protocol for preterm infants in indian neonatal intensive care units Authors: Abishek J R, Vadivelan Kanniappan, Manisha Srinivasan, Karthikeyan Saravanan Published: 2026 ABSTRACT AbstractBackgroundEarly stimulation is increasingly recognised as essential for optimising neurodevelopmental outcomes in preterm infants admitted to Neonatal Intensive Care Units (NICUs). However, most existing stimulation protocols are derived from controlled research settings and lack guidance in gestational age (GA) -specific timing, caregiver role allocation, and cultural applicability, limiting their translation into routine practice in resource-constrained settings like India.AimTo develop a GA-specific early stimulation protocol for preterm infants using structured triangulation of trial evidence, therapist-informed practice patterns and maternal experience in Indian NICU settings.MethodsA sequential exploratory mixed-method design was employed. First, a systematic review of randomised controlled trials (2010–2023) mapped stimulation modalities and GA characteristics. Second, a semi-structured interview was conducted with 16 NICU physiotherapists across eight Indian states to capture feasibility-driven clinical expertise. Third, in-depth interviews with 12 mothers of preterm infants explored maternal readiness, emotional acceptance and culturally embedded caregiving practices. Data were integrated using joint display analysis and a fixed decision algorithm.ResultsTriangulation classified modalities into four categories: five with strong convergence, five requiring conditional implementation, two restricted to specialist delivery and two excluded from the protocol. These findings were synthesised into a novel, colour-coded week x modality matrix, operationalising developmental sequencing, modality layering and transitions from therapist-led to parent-led delivery across GA.ConclusionThis study presents the foundational development of a contextually grounded GA-specific early stimulation protocol that maps twelve modalities across 26–40 weeks with colour-coded transitions from therapist-led to parent-led care. By integrating evidence, expertise and lived experience, the framework offers a transparent and culturally informed basis for subsequent feasibility testing, validation and implementation.Keywords: Infant, premature; Neonatal intensive care units; Early stimulation; India; Mixed methods research; Developmental care Introduction Preterm birth remains a leading cause of neonatal morbidity and mortality worldwide, with the burden disproportionately concentrated in low and middle-income countries (LMIC) where health systems face significant resource constraints and high rates of neonatal complications [1–3]. India, an LMIC, is the single largest contributor to global preterm birth, with approximately 3.5 million preterm births per year [4, 5]. In India, preterm infants account for a large share of adverse neonatal outcomes and are at high risk of long-term neurodevelopmental impairment, including motor, cognitive and socio-emotional difficulties [6–8]. In these settings, preterm infants are at heightened risk of adverse neurodevelopmental outcomes due to both biological vulnerability and environmental factors, including limited access to developmentally supportive care and overstretched Neonatal Intensive Care Units (NICUs) [1–3]. Early sensory and motor experiences are critical for brain development in preterm infants, and insufficient or inappropriate stimulation in the NICU can negatively impact autonomic regulation, feeding, and long-term cognitive and socio-emotional outcomes [9–11]. While interventions such as kangaroo mother care, oromotor stimulation and multisensory approaches have demonstrated benefits, most protocols are not tailored to the infant’s gestational age (GA) or the progressive maturation of neurodevelopmental subsystems [9, 10, 12]. Existing guidelines are often derived from high-income countries and may not account for the realities of LMICs, such as limited staffing, parental access restrictions and cultural beliefs influencing caregiving practices [2, 3, 12, 13]. As a result, clinicians in LMIC frequently rely on fragmented evidence and local expertise, leading to inconsistent timing, intensity and caregiver involvement in stimulation practices 41. There is a need in India for a structured, GA-specific stimulation protocol that is biologically appropriate yet feasible within resource-constrained NICUs and that deliberately integrates parents as care partners [14–16]. This study aims to adapt a mixed-methods triangulation approach that combines global evidence of preterm neurodevelopment, Indian NICU practice patterns, and maternal perspectives to inform protocols that are both evidence-based and culturally and operationally adapted to Indian NICUs. The goal was to create a week-by-week, developmentally targeted protocol that not only delineates what interventions are suitable for which gestational age group, but also when, how, and by whom they ought to be given. The research aimed to frame these interventions within the realities of care provision in low- and middle-income country (LMIC) NICU environments, specifically in India. To guide the triangulation process, the study was structured around the following research questions: (1) Which stimulation modalities are supported by evidence at specific gestational ages? (2) Which modalities are feasible within NICU workflows? (3) Which modalities are acceptable to mothers in the NICU context? Methods a) Study design and rationale This study used a sequential exploratory mixed-methods design to develop a context-appropriate multimodal early stimulation protocol for preterm infants admitted to Indian NICUs by integrating both qualitative and structured review components through a formal triangulation process and joint display analysis. Anchored in an implementation science framework that targeted the “efficacy to effectiveness gap” between tightly controlled Randomized Controlled Trial (RCTs) and real-world NICU practice. Reporting followed GRAMMS (Good Reporting of a Mixed-Methods Study) guidelines with PRISMA applied to the systematic review and COREQ to the qualitative components. The core analytic output was a triangulated synthesis that generated meta-inferences to directly inform protocol development. Five sequential phases were employed, with integration at the analytical level (shown in Fig. 1). Sequential mixed-method design informing protocol development Systematic review of RCTs to map evidence on early stimulation modalities and timing by gestational age and implementation parameters. Systematic review of RCTs to map evidence on early stimulation modalities and timing by gestational age and implementation parameters. Qualitative interviews with NICU physiotherapists to capture feasibility and practice-informed clinical expertise. Qualitative interviews with NICU physiotherapists to capture feasibility and practice-informed clinical expertise. Qualitative interviews with mothers of preterm infants to understand maternal readiness, emotional acceptance and cultural practices. Qualitative interviews with mothers of preterm infants to understand maternal readiness, emotional acceptance and cultural practices. Triangulation through joint display analysis and rule-based decision logic. Triangulation through joint display analysis and rule-based decision logic. Development of a gestational age–specific Early Stimulation Protocol. Development of a gestational age–specific Early Stimulation Protocol. b) Rationale for integrating multiple data streams Although clinical evidence hierarchies prioritise RCTs, Neonatal RCTs on early stimulation are often conducted under highly controlled conditions, with fixed intervention durations that do not vary by gestation age at recruitment, physiological stability or length of NICU stay. This limits transferability across the broad 26–40 week gestational spectrum of Indian NICUs. To address this, three complementary evidence streams were integrated, each serving a distinct, non-interchangeable purpose: i.RCTs provided biological plausibility, safety, and dose/timing parameters under controlled conditions. RCTs provided biological plausibility, safety, and dose/timing parameters under controlled conditions. ii.Therapist expertise functioned as a safety and feasibility filter, clarifying which modalities could be delivered reliably within staffing, space and monitoring constraints of NICU. Therapist expertise functioned as a safety and feasibility filter, clarifying which modalities could be delivered reliably within staffing, space and monitoring constraints of NICU. iii.Maternal experience served as an acceptability and readiness indicator, identifying when and how caregiver-led delivery was appropriate and culturally congruent. Maternal experience served as an acceptability and readiness indicator, identifying when and how caregiver-led delivery was appropriate and culturally congruent. All three streams were therefore afforded equal consideration during triangulation. Final protocol decisions followed a tiered hierarchy: Infant safety > clinical evidence > contextual feasibility > maternal readiness. c) Data streams Data stream 1 - systematic review: mapping of intervention characteristics Electronic searches (PubMed, Scopus and ProQuest), identified RCTs from 2010 to 2023 reporting stimulation-based interventions for preterm infants (26–40 weeks GA), including gestational age and the respective implementation parameters. The review was intentionally restricted to recent RCTs to capture contemporary clinical relevance and established safety oversight. Studies without explicit timing by gestational age or conducted outside the NICU were excluded. Three independent reviewers (AJR, MS, KS) extracted data (intervention modality, gestational age at initiation and cessation, session duration and frequency, and delivery agent) in Microsoft Excel version 21, and conflicts were resolved by a fourth reviewer (VK). Risk of bias was assessed using the Cochrane ROB 2, and only trials that are was rated as low risk or some concerns were included. Of the 669 records, 59 RCTs met the inclusion and quality criteria. To reduce spurious inclusion, a modality was mapped only if reported in ≥ 2 instances. Data were synthesised into a gestation-age x modality matrix, providing an evidentiary foundation for triangulation. Data stream 2 – therapist interviews: clinical feasibility and implementation patterns In-depth, semi-structured interviews (phenomenological approach) were conducted with 16 physiotherapists practising in NICUs from eight Indian states (Tamil Nadu, Maharashtra, Haryana, Telangana, Chandigarh, Himachal Pradesh, Uttar Pradesh and Bihar), recruited via purposive and snowball sampling to maximise geographic and institutional diversity. Eligibility required: Minimum two years of NICU-specific experience. Physiotherapists were intentionally selected as key informants because they are formally responsible for prescribing, progressing or withholding developmental and therapeutic stimulation in Indian NICUs and routinely guide or supervise nursing staff in modality-specific delivery. Semi-structured telephonic interviews (45 to 90 min, English) explored intervention sequencing, therapist-only modalities, parent training strategies and decision thresholds for infant readiness and their barriers and facilitators in practice. Interviews were recorded, transcribed verbatim, and analysed with Braun and Clarke’s six-phase reflexive thematic analysis. Coding was conducted manually in Excel version 21 by two independent researchers, with themes refined iteratively. Themes included “routine practices by GA,” “barriers to parent involvement,” “infant stress cue observation,” and “therapist discretion versus protocol-driven care.” Data stream 3 – maternal interviews: readiness and acceptability Maternal readiness was inferred from the convergence of affective (comfort – emotional and support), cognitive (awareness), and behavioural (task initiation) indicators across interview narratives, forming a dynamic readiness rather than technical skill acquisition. Consistent with the synactive theory of development, readiness is not static but evolves through reciprocal interactions between the infant’s behavioural cues and the caregiver’s adaptive response in the NICU. Mothers were purposively selected if they had at least one week of NICU exposure, were actively involved in infant care, and had no reported psychological illness or substance abuse. Mothers were recruited during the infants’ NICU stay. Mothers who experienced neonatal loss or acute psychological distress were not included as the study focused on caregiving readiness and their emotional transitions during ongoing NICU admission. Twelve mothers (22–37 years, whose infant was born between 28 and 36 weeks of gestation) were recruited. At the time of participation, infants’ age ranged from 29 to 38 weeks GA, capturing experience across early and later phases of NICU hospitalisation. Educational background varied with participants having completed high school (8%), undergraduate (75%), and postgraduate education (16%). Most mothers were primiparous (67%) and included both singleton and twin births (25%). The geographical distribution of participants varied with urban (25%), semi-urban (58.3%) and rural (16.6%). All participating mothers had been actively involved in NICU caregiving for at least one week prior to the interview, ensuring experiential familiarity with routine care. Face-to-face, semi-structured interviews were conducted to understand instinctive versus guided caregiving behaviours, perceived therapist-dependent modalities, bonding practices, perceived risk, emotional responses and confidence in handling. Interviews were recorded, transcribed verbatim and analysed using phenomenological thematic analysis by two independent researchers. Codes were synthesised into the maternal readiness matrix by modality and gestational age. Cultural mediators identified informed caregiver role allocation, timing of parent integration and modality delivery within the protocol. Readiness thresholds were operationalised through the convergence of maternal narratives in caregiving behaviours and expressed comfort or hesitation towards specific modalities. Data saturation of qualitative interviews Sample size for both qualitative data streams was determined by thematic saturation, with recruitment continuing until no new codes emerged across two consecutive interviews. Saturation was reached with sixteen therapist interviews and twelve maternal interviews. d) Triangulation and joint display analysis Following the independent analysis of each data stream, the fourth phase involved a four-stage process (raw matrices → convergence coding → Decision algorithm → Final Joint Displays) where convergence/divergence across data types are made explicit and then translated into actionable decisions. Stage 1: Construction of raw triangulation matrices A raw triangulation matrix was constructed for each modality, displaying unaltered data excerpts from all three sources. This ensured analytic transparency and preserved the integrity of each data stream. Stage 2: Interpretive convergence coding Each modality-gestation week pairing was assessed using predefined operational definitions: Convergence: At least two of three data sources aligned on the safety, timing and delivery feasibility of a modality. Convergence: At least two of three data sources aligned on the safety, timing and delivery feasibility of a modality. Conditional convergence: Agreement existed, but one source indicated the need for supervision or staged parental training. Conditional convergence: Agreement existed, but one source indicated the need for supervision or staged parental training. Divergence: Contradiction or insufficient data across sources. Divergent items were discussed until consensus, if unresolved, they were classified as specialist-only or excluded. Divergence: Contradiction or insufficient data across sources. Divergent items were discussed until consensus, if unresolved, they were classified as specialist-only or excluded. Stage 3: Application of triangulation decision algorithm Final protocol decisions were guided by a tiered priority framework, with explicit precedence given to: Infant safety > Clinical evidence > Contextual Feasibility > Maternal Readiness. Safety screening based on gestational age, physiological stability and infant tolerance was conducted before applying any decision rule. For each modality-week pairing, the following IF-THEN decision logic was applied. Rule 1: Convergent modality IF RCT evidence supported the modality AND therapists rated it as routine an feasible within NICU workflows AND mothers demonstrated readiness or instinctive acceptance IF RCT evidence supported the modality AND therapists rated it as routine an feasible within NICU workflows AND mothers demonstrated readiness or instinctive acceptance THEN the modality was included as parent-led (E.g., Kangaroo Mother Care, auditory stimulation) THEN the modality was included as parent-led (E.g., Kangaroo Mother Care, auditory stimulation) Rule 2: Conditional modality (therapist-led with parental observation/ specialist-only) IF RCT evidence supported the modality BUT therapists identified technical complexity or safety concerns OR mothers expressed fear, uncertainty or lack of awareness IF RCT evidence supported the modality BUT therapists identified technical complexity or safety concerns OR mothers expressed fear, uncertainty or lack of awareness THEN the modality was included as therapist-led, with structured parental observation if gradual skill transfer is intended or without parental observation it is specialist-only (E.g., Reflex Locomotive Therapy, Kinesthetic stimulation) THEN the modality was included as therapist-led, with structured parental observation if gradual skill transfer is intended or without parental observation it is specialist-only (E.g., Reflex Locomotive Therapy, Kinesthetic stimulation) Rule 3: Expert-driven modality (specialist-only) IF RCT evidence was absent or limited AND mothers expressed fear, uncertainty or lack of awareness BUT the therapists consensus identified the modality as widely practised and safe in Indian NICUs IF RCT evidence was absent or limited AND mothers expressed fear, uncertainty or lack of awareness BUT the therapists consensus identified the modality as widely practised and safe in Indian NICUs THEN the modality was included as a specialist-only (E.g., Respiratory facilitation techniques, nesting) THEN the modality was included as a specialist-only (E.g., Respiratory facilitation techniques, nesting) Rule 4: Exclusion IF physiological instability is anticipated OR gestational age or safety thresholds were not met IF physiological instability is anticipated OR gestational age or safety thresholds were not met THEN the modality was excluded (E.g., high intensity tactile inputs or thoracic squeezing) THEN the modality was excluded (E.g., high intensity tactile inputs or thoracic squeezing) This algorithm ensured that clinical evidence-informed inclusion, while therapist expertise safeguarded safety and feasibility and maternal readiness governed caregiver-led delivery, thereby supporting contextually valid and developmentally sensitive protocol construction. Stage 4: Joint display synthesis Based on coding and decision rules, two analytic display were developed, listing each intervention modality, its evidence-based timing and delivery structure, therapist feasibility, maternal acceptability, triangulated judgment, and final protocol decision, which is displayed in Table 1. Table 1Joint display matrix – triangulated of data streamsModalitySystematic reviewTherapist practiceMaternal readinessTriangulated judgementFinal protocol inclusionNesting/ PositioningSupported for physiologic stabilityNICU practice. Nesting might be difficult for mothers to perform in NICU. Handling extreme preterm infant requires skillsHave not practiced turning or nestingConditional Convergence26 W onwards; Skill transfer from 28 W; Parent led by 32 WKMCStrong evidence from 26 weeksWidely encouraged. therapist supervisionStrong emotional bond, confidenceStrong convergence26 W onwards parent ledTactileSupported strongly – From 26 W by massageMassage is advised post 31 W, as the skin has reached maturity. Till then, Yakson Touch is advised as an alternative.Willing to perform. Performs in daily care, but is unaware of the sequence of the procedure. Afraid of injuring the babyConditional ConvergenceTouch from 26 W onwards, Progresses to massage from 31 W. Gradual skill transfer to mother from 32 WTactile KinestheticStrongly supported from 26 WSupported. But the risk of IVH was mentioned in the early weeksInstinctively practised. Needs trainingConditional ConvergenceFrom 26 W onwards. Gradual transition to parents from 32 weeksReflex LocomotionLimited evidence28–37 WRare. Therapist-onlyUnaware. Clinical domainDivergent, but specialist-approvedTherapist-only28–37 WOromotor stimulationStrong evidence26 W onwardsRecognised, but was advised not to start before 28 W, as feeding reflexes do not matureInterested in improving the feeding, but are unaware.Conditional ConvergenceTherapist-led, from 28 W, gradual transition to mother-led from 35 WRespiratory PNFNo RCT dataUsed by experts for respiratory enhancement and to reduce distressUnaware.Clinical domainDivergent – Clinician onlyTherapist-only 30–40 WAuditorySupported by RCTFrom 26 WStrongly advised. Can be added during KMCUniversally practicedStrong convergenceMother led26 W onwardsOlfactionIndicated form 28 WEncouraged using breast milk and mother-scented cloths.Motivated to performStrong convergenceParent-led28 W onwardsVisual InteractionSafely implementedfrom 34 W. Limited evidence.Face regards encouragedContrast cards are to be used with cautionHighly motivated when asked to make eye contact with an infantStrongconvergenceParent-led34 W onwardsGustatorySafely implemented from 34 WTo be done with breast milk. Can be done with gauze, milk-dipped finger or syringe.Highly interestedStrong convergenceParent-led34 W onwards with initial therapist guidance.VestibularImplemented form 34 W onwardsGentle rocking, slow-paced movements, completely cradled in the mother’s arms, is preferredMothers feared rocking but were interested in performingConditional ConvergenceTherapist-led at 34 W, with gradual skill transfer to parent-led rocking*W – Weeks; IVH – Intraventricular Hemorrhage; PNF – Proprioceptive Neuromuscular Facilitation Joint display matrix – triangulated of data streams Limited evidence 28–37 W Therapist-only 28–37 W Strong evidence 26 W onwards Unaware. Clinical domain Supported by RCT From 26 W Mother led 26 W onwards Parent-led 28 W onwards Safely implemented from 34 W. Limited evidence. Face regards encouraged Contrast cards are to be used with caution Strong convergence Parent-led 34 W onwards Parent-led 34 W onwards with initial therapist guidance. *W – Weeks; IVH – Intraventricular Hemorrhage; PNF – Proprioceptive Neuromuscular Facilitation A GA x modality matrix was created to visualise developmental sequencing from GA of 26 to 40, which is shown in Table 2. Table 2GA-specific layered protocol for preterm infants admittedin Neonatal Intensive Care Unit GA-specific layered protocol for preterm infants admittedin Neonatal Intensive Care Unit e) Resolution of disagreements Complete divergence occurred for respiratory PNF (absent RCT evidence, therapist-only practice) and Reflex Locomotor therapy (had RCT evidence, therapist questioned the need and feasibility with complete maternal unfamiliarity). Disagreements were resolved through consensus meetings guided by the predefined hierarchy (infant safety > clinical evidence > feasibility > readiness). For example, tactile-kinaesthetic inputs were initially contested: therapists endorsed feasibility, but with the risk of intraventricular haemorrhage at early weeks, mothers feared injury, while RCT supported the applicability. The team resolved this by classifying modality to be therapist-led till 31 weeks, then gradual progression from therapist from 32 to 34 weeks, then to be carried out by the mother. f) Data integration team and reflexivity The three-member analytic team (AJR, MS, KS) reviewed all divergent codes through a consensus meeting and maintained an audit log detailing justification for inclusion/exclusion. Two researchers (AJR, MS) performed initial convergence coding. A third researcher (KS), independent of primary data collection or initial coding, reviewed all matrices to reduce confirmation bias. Manual coding and Excel-based joint display construction were used to maintain full visibility of data transformation across phases and to prevent abstraction through automated software processes, thereby enhancing transparency, traceability and confirmability of protocol development. g) Trustworthiness and reflexivity The study was led by a neonatally trained physiotherapist with prior clinical and qualitative research experience. This professional background positioned the researcher as an “insider”, offering contextual understanding of NICU routines, terminologies and clinical decision-making, while also posing a potential risk of profession-specific interpretive bias. To address this, reflexive field notes were maintained throughout data collection and analysis to document the responses. Responses related to physiotherapy-led care, therapist authority and perceived feasibility of parent-delivered interventions were explicitly bracketed during coding. Investigator triangulation was used to further mitigate bias, with the co-researcher independently viewing transcripts. Therapists’ responses were reviewed by AJR and KS, while maternal responses were reviewed by AJR and MS. Discrepancies in coding of these responses were resolved by reflexive discussion, ensuring that interpretations remained grounded in participants’ accounts rather than professional expectations. None of the investigators holds a clinical role in the participating therapists’ institutes, minimising the power imbalance and role conflict. Participants were informed that the study aimed to understand practice experiences rather than evaluate individual performance. h) Operationalisation of Als’ synactive theory Als’ synactive theory conceptualises neonatal hierarchical development under the interaction of five subsystems: autonomic (physiological stability), motor (tone and posture), state (alertness, sleep-wake state), attention (engagement) and self-regulatory (balancing between the subsystems), each reflecting the infant’s developmental organisation with advancing gestational age. This theory was applied to guide the week-by-week sequencing of the modalities by aligning their introduction and progression within subsystem maturation [17, 18]. Results Independent analyses of the systematic review, therapist interviews and maternal interviews were synthesised through analytic joint-display triangulation. The matrices served as analytic engines that produced meta-inferences connecting efficacy evidence to contextual feasibility. Each modality’s inclusion was determined by cross-stream comparison, documented convergence/conditional convergence/ divergence and final consensus judgement. a) Patterns of convergence and divergence Five modalities demonstrated strong convergence across evidence, feasibility and readiness: kangaroo care, auditory, olfactory, gustatory and visual stimulation. Five modalities showed conditional convergence: vestibular, tactile, nesting/positioning, oromotor and kinesthetic stimulation. Kinesthetic stimulation till 31 weeks of gestation was restricted to the therapist in view of intraventricular haemorrhage. While oromotor stimulation was similarly staged, with therapist-led delivery and gradual skill transfer to mothers because of the intra-oral component. Divergence arose mainly for specialised techniques where maternal involvement was inappropriate: Respiratory proprioceptive neuromuscular facilitation and reflex locomotive therapy. Two modalities, thoracic squeeze and vibration-percussion were removed entirely. Therapists described these as indication-specific procedures rather than developmental care and therefore unsuitable for inclusion as protocolized modalities. Table 1 functions as a joint display by aligning findings from the three different data streams into a single, comparable matrix for each stimulation modality. This side-by-side arrangement makes visible where data converges, conditionally converges, or diverges and links these patterns directly to the final protocol decision. b) Gestation week-specific evolution – theoretical anchoring The week-by-week mapping revealed a progressive layering of stimulation aligned with physiological maturation (Als’ synactive theory). Protocol decisions between 26 and 30 weeks were aligned with the autonomic subsystem, prioritising physiological stability (containment, nesting, skin-to-skin care etc.,), and auditory is introduced to regulate the attention subsystem. From 31 to 34 weeks, the protocol reflected emerging motor subsystems maturation by introducing controlled tactile-kinesthetic input. Between 35 and 40 weeks, interventions targeting state and attention subsystems (visual interaction, gustatory, vestibular and multisensory inputs) were progressively layered (developmentally staged addition of stimulation modalities) to support arousal-regulation, feeding readiness and social engagement. Infant-environment interaction was operationalised through graded modality introduction, observation phases and staged role transitions from therapist-led to parent-led delivery. Self-regulation was inferred from infant tolerance and physiologic stability. Stimulation modalities are delayed, restricted or withdrawn at any signs of stress or instability. Thus environmental input remained responsive to the infant’s regulatory capacity not solely on gestational age. c) Cultural mediators of contextual components Triangulated qualitative synthesis identified sociocultural construct shaping caregiver participation. Maternal readiness emerged as a family-integrated construct with mothers reporting greater confidence when supported by extended family members, particularly grandmothers with prior experience caring for preterm infants. Father’s early involvement boosted mother’s confidence, especially during post-cesarean recovery. Therapists reported routine implementation of kangaroo father care, supporting flexible caregiver role allocation within the protocol, reflecting the joint family caregiving structures. Postoperative pain, restricted NICU access, distance to NICU from postnatal wards, difficulty sustaining caregiving postures, and fear of infant fragility delayed early maternal engagement, particularly among first-time mothers. The perception of NICU as a sterile, restrictive and alien environment further limited spontaneous caregiver-infant interaction informing a graded transition model, with initial observation-only and therapist-led exposure preceding active parent-led caregiving. An adaptation window of approximately three days, during which gradual exposure to caregiving tasks facilitated transition to readiness, was recommended. Protective cultural practices (e.g., of spiritual ornaments and applying sacred ashes to the forehead) were suspended in the NICU for hygiene but reinstated post-discharge. This emphasizes the need for anticipatory caregiver education to mediate this transition and to prevent injury associated with forceful traditional massage practice. When clinical behaviours were reframed as culturally familiar acts such as talking, singing a lullaby, gentle rocking while being cradled in hands or maintaining eye contact with their baby, acceptance was notably increased. d) Synthesis of maternal and clinical narratives Interpretive analysis revealed that mothers initially perceived infants as belonging to the medical team rather than to themselves. Mothers described a liminal phase of motherhood marked by disbelief and emotional suspension. Although mothers expressed a strong desire to hold and comfort their infants, touch was frequently deferred due to the fear of fragility. Clinical expertise placed professionals as gatekeepers of interaction and independent caregiving. Post-cesarean pain further caused a drift between intention and action. Delayed participation in caregiving was often accompanied by guilt. Fathers or grandmothers often assumed caregiving roles, and mothers described it as a sense of relief with shared responsibilities. Therapist narratives highlighted parallel tensions between relational care and institutional efficiency. Mothers’ repeated questions and emotional expressions were sometimes perceived as time-consuming, which limited opportunities for dialogue and positioned therapists as primary decision-makers regarding infant handling and interaction. When therapists explicitly framed routine parental behaviours such as speaking, singing or holding as therapeutic interventions, these actions became clinically legitimised, maternal participation. This dynamic transformation of caregiver identity from spectator to co-therapist to primary caregiver was a key psychosocial output of the triangulation process. Power asymmetry inherent to NICU hierarchies, therapists as gatekeepers of handling permissions, shaped maternal hesitancy and the observed progression from spectator to co-therapist. e) Overall integration outcome The final integrated product was the GA-specific layered protocol, which is shown in Table 2. This table represents the synthesis of all data streams, translating convergence into an operational developmental sequence of NICU implementation. It aligns with ALS’ synactive theory, beginning with autonomic stability and progressing towards multisensory integration. stimulations advance in layering and caregiver participation, shifting from therapist-led to containment and to parent-led multimodal engagement. Table 2 represents a second-level joint display that translated the earlier triangulation into a gestational age by modality matrix. By mapping each intervention across weeks and indicating when it becomes active, this display integrated trial-based timing with therapist judgement and maternal readiness into a single sequence. Discussion This study developed a gestational age-specific, Indian context-adapted early stimulation protocol for preterm infants in an LMIC NICU, integrating parents as a team of providers, using a mixed-method triangulation approach. The protocol visually maps the transition of stimulation responsibilities from healthcare providers to parents across the gestational journey, aiming to optimise neurodevelopmental outcomes while accommodating local resource constraints and cultural practices. Stimulation techniques, when given as cultural care practices, may improve maternal adherence. A key insight from the triangulation was the natural occurrence of multisensory layering, particularly during KMC, where mothers unknowingly combined voice, touch, scent, and sometimes movement. Although neither the literature nor therapists explicitly structured these combinations, the presence of such layering in maternal narratives suggests an opportunity to formalise instinctive caregiving into a scaffolded multisensory progression from 31 weeks onward. This supports the theoretical model of graded sensory integration, wherein sensory modalities are introduced in overlapping, developmentally sequenced layers [19–21]. Prior research has established the benefits of early stimulation and nurturing care for preterm infants, especially in LMICs where risks for neurodevelopment are high. Meta-analyses and systematic reviews confirm that parenting and stimulation interventions can yield small-to-moderate improvements in cognitive, language and motor outcomes in vulnerable populations. However, most existing protocols lack GA-specificity and are not explicitly adapted to resource-constrained setups, resulting in limited scalability and transferability [22, 23]. The current study addresses these gaps by providing a structured, adaptable protocol grounded in both evidence and Indian context-based realities. The current protocol is grounded in Als’ synactive theory, similar to the Newborn Individualized Developmental care and Assessment Program (NIDCAP), which emphasizes fully individualized, cue-based care derived from continuous behavioural observation [24, 25]. In contrast, the present protocol adopts a structured, GA-specific approach that provides a baseline for stimulation, while incorporating safety-based cue monitoring rather than continuous individual assessment. This distinction reflects differences in implementation context. NIDCAP requires intensive training, dedicated personnel, and system-level reorganization, which may not be feasible in many Indian NICUs characterized by workforce constraints and high patient load [26–28]. A structured protocol was therefore chosen as a pragmatic approach to translate core neurodevelopmental principles into routine practice, ensuring consistency across providers while retaining flexibility through staged progression and safety thresholds. This design aims to establish baseline standards of developmentally appropriate care that can be feasibly implemented and slowly can be progressed as individualised care in resource-constrained settings [29, 30]. Compared with SENSE, which is a manualized parent-implemented sensory exposure program that delivers a GA-specific, structured, dose-specified, multisensory package in high-resource NICUs and assumes consistent parental presence, specialised staff and close monitoring [31–33], The current protocol is designed for resource-constrained Indian NICU care pathways with a much lower infant-nurse ratio, a shortage of therapist and frequent task shifting, where comprehensive sensory bundles are harder to sustain [34–37]. The protocol is intentionally modular rather than a fixed bundle, so content and intensity can be adapted to GA, clinical stability and family context while maintaining core practice elements: responsive care and caregiver-infant interaction that are common, effective early childhood interventions in LMICs [10, 38–40]. SENSE gains strength from a highly standardised package that facilitates fidelity and multi-site replication, but this could limit transferability to resource-constrained settings. Whereas the current protocols are more flexible, but, this could make cross-site fidelity and component testing more challenging. The strength of this study is the use of triangulation of data streams stemmed from evidene, expertise and experience, which has ensured the protocol is both scientifically robust and contextually relevant for Indian NICUs. The visual mapping of stimulation responsibilities across GA, with a gradual transition from therapist-led to parent-led by skill transfer, helps parents gain confidence and autonomy over the child, while the proximity of the mother enhances the outcomes of the preterm infant. This increasing involvement of mothers reduces the burden placed on healthcare professionals [22, 41]. Limitations, implementation roadmap and future validation phases While triangulation enhanced contextual validity, interpretive synthesis relied on manual coding, introducing potential researcher subjectivity despite audit trails. The systematic review was limited to English-language publications, which may have excluded relevant studies published in other languages and could have overlooked additional modalities or earlier gestational age initiation timelines. As such, the maternal sample represents caregivers who were emotionally and physically able to participate in caregiving and does not include mothers who experienced neonatal loss or prolonged NICU separation or acute psychological distress. So, the findings may not fully reflect the experiences of the most vulnerable caregivers. Similarly, expert inputs were restricted to therapists, which may not fully capture the multidisciplinary reality of NICU care delivery. The representation of therapists to geographical diversity was underrepresented, which may limit the direct transferability. Other professionals, such as neonatologists, neonatal nurses, etc., may have different feasibility Judgements regarding modality implementation. To directly address this limitation, the next phase will involve multidisciplinary protocol validation using a Delphi consensus process, including neonatologists, neonatal nurses, physiotherapists, occupational therapists, neonatal therapists, lactation specialists, paediatricians, and developmental paediatricians. This step is designed to refine modality timing, role allocation, safety thresholds and developmental appropriateness while ensuring alignment with bedside care practices. Although the protocol is developed within the Indian NICU context, a large and diverse Low- and Middle-Income Country (LMIC), direct transferability to other LMIC settings cannot be assumed, particularly in regions with differing health system structures and resource constraints. Contextual adaptation and local validation would be required prior to implementation in other settings. The Present study reports the protocol development and triangulation phase of a larger, staged Multimodal Early Stimulation development program. Formal protocol validation, feasibility testing and implementation evaluation were intentionally positioned as subsequent phases to avoid premature efficacy claims. Following expert consensus, protocol will undergo pilot usability and feasibility testing in NICU settings to evaluate implementation practicality, staff acceptability, caregiver participation and safety monitoring across varied NICU contexts and varied maternal characteristics. Therapist training frameworks, caregiver competency indicators and fidelity measures will be developed during this phase. Only after feasibility and fidelity are established, preliminary efficacy testing will be undertaken. This phased roadmap ensures that the protocol transitions from the evidence-informed design to safe, contextually grounded clinical implementation. Conclusion This study developed a gestational age-specific early stimulation protocol for preterm infants admitted to NICU through triangulation of trial evidence, therapist expertise and maternal experience. Using an explicit decision algorithm, the protocol translated evidence into NICU practice that is relevant to the Indian context, with Als’ synactive theory guiding the developmental sequencing and promoting caregiver role transitions. While the protocol supports safe and maternal-inclusive care, it represents the foundational development phase of translational intervention, warranting validation, feasibility testing, and implementation evaluation. For clinicians, Table 2 can be used as a structured reference to guide decisions on which stimulation modalities to introduce, when to initiate them, and how to plan the transition from therapist-led to parent-led care. For researchers, the protocol represents a development-phase framework that requires further evaluation of feasibility, fidelity, and preliminary efficacy before broader implementation can be recommended. Acknowledgements Authors would like to extend heartfelt thanks to all the participants for spending their valuable time for the interview sessions. Author contributions AJR: Conceptualization, Study design, Data curation, conducting interviews, screening articles, preparing transcripts, coding and reflexive memos, Triangulating, visualising the results, drafting of results, writing the first draft of manuscript, revision of manuscriptKV: Study design, data analysis, supervision, resources. Methodological guidance, screening, coding, proof reading, revision of manuscriptMS: Data curation, conducting interviews, screening articles, preparing transcripts, coding and reflexive memos, Triangulating, visualising the results, drafting of results, revision of manuscriptKS: Data curation, screening articles, preparing transcripts, coding and reflexive memos, Triangulating, visualising the results, drafting of results, revision of manuscriptAll authors contributed to the revision of the protocol and read and approved the final manuscript. Funding statements Open access funding provided by SRM Institute of Science and Technology for SRMIST – Medical & Health Sciences. tOpen access funding provided by SRM Institute of Science and Technology for SRMIST – Medical & Health Sciences, Kattankulathur, India.The project is funded by Researcher Starter Grant – “Growth & Advancement Towards Excellence”, SRM Institute of Science and Technology, Kattankulathur, India (CFFP001/NOV25). Data availability The datasets used and/or analysed during the current study are available from the PI on reasonable request. Declarations Ethics approval and consent to participate The study was conducted in accordance with Helsinki’s declaration. The study was approved by the Institutional Ethics Committee of SRM Medical College Hospital and Research and Centre, Kattankulathur. ( Reg. No: ECR/8972/INST/TN/2013/RR-19). Confidentiality was strictly maintained, with all data anonymised before analysis. Informed consent was obtained from all participants before participation, and they had the right to withdraw from the study at any stage without consequences. Participants were recruited to the study following their voluntary willingness to participate. As all the participants were above 18 years, consent was obtained only from the participants. Consent for publication All the participants gave consent to publish. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analysed during the current study are available from the PI on reasonable request.