The state of infant massage use in neonatal intensive care units

TL;DR

This study examines how infant massage is used in NICUs and finds that while it is common, practices vary widely among therapists.

Contribution

The study highlights the need for standardized protocols to ensure safe and effective infant massage in NICUs.

Findings

90% of respondents use infant massage, with 71% considering it standard care.

Only 53% follow a specific protocol, and practices vary in techniques and frequency.

Safety concerns include infant stress, but adverse events are rare.

Abstract

To characterize current infant massage practices in neonatal intensive care units (NICUs) and identify variability in approaches among neonatal therapists.

A cross-sectional survey was distributed to NICU-based occupational therapists, physical therapists, and speech language pathologists. The survey inquired about massage use, training, protocols, techniques, and safety concerns. Descriptive statistics were used for analysis.

Among 101 respondents from 32 states, 90 (90%) used infant massage, with 64 (71%) considering it standard care. Infant massage was most often administered by occupational therapists (77, 76%), physical therapists (70, 69%), and parents (46, 46%). Despite high training rates (87, 97%), only 48 (53%) followed a specific protocol. Techniques, frequency, and use of emollients varied widely. Safety concerns included infant stress and physiological instability, though adverse events were rare.

Despite widespread use of infant massage in NICUs, variability in findings underscores the need for standardization to ensure safe, effective delivery of massage.

Full text

Introduction

From birth, preterm infants often require intensive, prolonged medical care in the neonatal intensive care unit (NICU). During this time, high risk of infant death and comorbidities creates a barrier to natural infant sensory-motor experiences and parent-infant bonding opportunities. Infant massage in preterm infants has been studied for more than 4 decades [1], with the first Cochrane Review published in 2000 [2]. More recently, infant massage has emerged as a promising intervention in NICUs as early as 32 weeks postmenstrual age (PMA) [3, 4] with benefits reported for both infants and parents [5, 6]. Numerous studies support its safety and efficacy when administered by skilled professionals or caregivers following appropriate training [5].

Research on infant massage has reported a range of infant physiological benefits such as increased oxygen saturation levels, enhanced vagal tone, improved gastric motility, and reduced pain, all of which contribute to shorter hospital stays and better post-discharge outcomes [7, 8]. Infant massage is also associated with improved anthropometric measures (e.g., weight, length, and head circumference) and improved oral feeding skills [3, 9]. Beyond these reported infant benefits, infant massage may also benefit parents who administer it to their infants by increasing opportunities for enhanced maternal-infant bonding and alleviating maternal mental health symptoms [6, 10].

There are specific protocols for infant massage, and some that were developed for different populations have been applied or adapted for infants in the NICU [9, 11–17]. However, the absence of standardized, evidence-based protocols for infant massage in the NICU limits its widespread adoption. Several reviews published between 2012 and 2025 [4, 5, 7, 18–20] have documented significant variability in infant massage protocols, including differences in duration, frequency, pressure, use of oils, and who administers the intervention. This heterogeneity in study design, intervention protocols, and implementation fidelity has significantly impeded the translation of infant massage research into standardized, evidence-based clinical practice that ensures both safety and efficacy.

The objectives of this study were to (1) characterize the current state of infant massage in clinical practice in NICUs across the United States from the perspective of neonatal therapists, and (2) identify variability in massage approaches across institutions and practitioners. Findings from this study may inform current research needs as well as the development of standardized protocols to optimize outcomes for hospitalized preterm infants.

Subjects and methods

Study design and participants

This cross-sectional study was approved by the Institutional Review Board at the University of Southern California (USC) with a waiver of written informed consent. All study procedures were conducted in accordance with relevant institutional guidelines and regulations and with the ethical principles outlined in the Declaration of Helsinki. Participants were provided with general information about study participation prior to survey initiation. Eligible participants included occupational therapists (OTs), physical therapists (PTs), and speech language pathologists (SLPs) currently practicing in the NICU. These disciplines were selected due to their central role in delivering sensory-based interventions, including infant massage, within neonatal care settings.

Survey development

Survey prompts were developed by two neonatal therapy experts (DM, RP) following a comprehensive literature review and identification of gaps in existing knowledge [4]. The initial survey draft underwent iterative refinement by the full study team to improve clarity, reduce redundancy, and ensure relevance. Next, to enhance content validity, the survey was reviewed in three rounds by a total of 12 neonatal therapy (PT, OT, SLP) professionals. Each round included feedback from 3 to 5 therapists, with revisions incorporated after each round. By the third round, no new feedback was received, indicating consensus on survey content and structure.

The final survey was programmed in REDCap (Research Electronic Data Capture) [21], a secure, web-based application designed for data collection. Branching logic was used to tailor questions based on participants’ reported use of infant massage in their clinical practice.

Recruitment and dissemination

Survey dissemination occurred between January and March 2025. A multi-pronged recruitment strategy was employed to maximize reach and diversity of respondents. Recruitment channels included professional social media platforms (e.g., Facebook groups, Instagram), email outreach to NICU therapy networks and colleagues, newsletters targeting neonatal therapists, and announcements at national neonatal therapy conferences. Participation was voluntary, and no incentives were provided. The survey link directed participants to the REDCap platform, where they could review the study information and proceed to the survey.

Survey responses were collected anonymously and stored securely within the REDCap system. Data were exported to StataMP16 for analysis [22]. Descriptive statistics were used to summarize participant responses. Open-ended responses were systematically reviewed and analyzed to identify recurring themes and patterns. Similar responses were combined and summarized to capture the essence of participants’ perspectives. Representative quotes were selected to illustrate common themes and provide insight into the range and depth of responses.

Results

Participant characteristics

A total of 101 responses were received from OTs, PTs, and SLPs currently practicing in the NICU. Survey responses represented 32 states in the U.S, as well as one respondent from each of the following: India, United Arab Emirates, Peru, and Guatemala. NICU levels spanned II-IV, with the majority working in Level III (59, 58%) and Level IV (37, 37%) units. Nine participants were from the same hospital as another participant (3 sets of 2 were from the same hospital, and one set of 3 participants were from the same hospital). The mean number of NICU beds reported by participants were 51.1 (31.4) beds (range 6-150), and therapists had a average of 12.3 (8.7) years (range 0.25-39) of NICU experience. Sixty four (63%) of respondents were Certified Neonatal Therapists (CNTs) [23] (Table 1).Table 1. Participant and hospital characteristics.Mean (SD) or n^a^ (%)NICU leveln = 101 Level II5 (5%) Level III59 (58%) Level IV37 (37%)Total number of bedsn = 10051.1 ± 31.4Years of clinical experience in NICUn = 10012.3 ± 8.7Certified Neonatal Therapist (CNT)n = 101 Yes64 (63%) No37 (37%)Use of specific program or sensory-based interventions in the NICUn = 101 Yes49 (49%) No52 (51%)Name of sensory-based intervention or program usedn = 47 Supporting and Enhancing Neonatal Intensive Care Unit Sensory Experiences (SENSE)45 (96%) I-Rainbow2 (4%) The ladder approach1 (2%) Protocol developed locally2 (4%) Newborn Individualized Developmental Care and Assessment Program (NIDCAP)1 (2%)Massage used as a developmental interventionn = 100 Yes90 (90%) No10 (10%)Infant massage is considered ‘standard of care’ developmental interventionn = 90 Yes64 (71%) No26 (29%)Infant massage administered byn = 101 Occupational therapist77 (76%) Physical therapist70 (69%) Parent46 (46%) Nurse25 (25%) Speech-language pathologist16 (16%) Developmental interventionist3 (3%) Child Life Specialist1 (1%) Music therapist1 (1%) Other4 (4%)Received formal or informal infant massage programn = 90 Yes87 (97%) No3 (3%)Name of training programn = 101 Neonatal touch and massage certification69 (63%) International Loving Touch Foundation14 (13%) Peer training14 (13%) International association of infant massage7 (6%) Field method4 (4%) Hospital to Home: Optimizing the Infant’s Environment (H-HOPE); Massage+30,10,5; Auditory, Tactile, Visual, and Vestibular stimulation (ATVV)3 (3%) Infant Massage USA, Certified Infant Massage Educator1 (1%) LiddleKidz or Certified Infant Massage Teacher (CIMT)1 (1%) Post-graduate fellowship training1 (1%) Shantala Foundation1 (1%) M-technique1 (1%)Use of a particular massage protocoln = 90 Yes48 (53%) No42 (47%)Name of massage protocol usedn = 64 International Loving Touch Foundation19 (21%) Neonatal Touch and Massage13 (14%) Listening Touch6 (7%) Massage+30,10,56 (7%) Infant Massage USA5 (6%) Eclectic mix of protocols5 (6%) International Association of Infant Massage3 (3%) Field Method0 (0%) M-technique1 (1%) Yakson0 (0%) Unstandardized1 (1%) None of the above5 (6%)Participant’s self-reported comfort administering infant massage (scale of 1–10 with 1 not comfortable to 10 completely comfortablen = 988.43 ± 2.25(not comfortable) 11 (1%)24 (4%)31 (1%)42 (2%)54 (4%)64 (4%)74 (4%)814 (14%)917 (17%)(completely comfortable) 1047 (48%)^a^n reflects number of respondents for each question.

Use of infant massage in the NICU

Ninety (90%) of therapists respondents reported that infant massage was used as a developmental intervention in their NICU. Among these, 64 (71%) of therapists described infant massage as a standard of care practice. Therapists reported that infant massage was most commonly administered by OTs (77, 76%), PTs (70, 69%), and parents (46, 46%), with some units also involving nurses, developmental specialists, and massage therapists. Respondents who wrote in responses for “Other” indicated that trained volunteers, music therapists, nurses who underwent Neonatal Touch & Massage Certification (NTMC), and nurse practitioners also performed infant massage. Therapists reported a high level of comfort providing infant massage, with a mean comfort rating of 8.4 on a 10-point scale (1 = not comfortable, 10 = completely comfortable) (Table 1).

Training and protocols

Among 90 respondents who reported using infant massage, 87 (97%) received formal or informal training. Fourteen (13%) were trained by peer therapists. Approximately 48 (53%) respondents indicated that their unit followed a specific massage protocol, while others reported using an eclectic mix of techniques. Although most participants received formal training through a specific infant massage program, use of the same training program’s protocol was not reported as frequently. For example, of the 64 respondents who reported use of a particular infant massage protocol, the most frequently cited training programs were the Neonatal Touch and Massage Certification (NTMC) (69, 68%) and International Loving Touch Foundation (ILTF) (14, 14%), yet only 13 (14%) and 19 (21%) respondents reported use of the NTMC and ILTF protocols, respectively. See Table 1.

Massage techniques and environment

Table 2 provides reported characteristics for massage interventions. Massage was most often performed in an open crib (86, 96%), followed by the infant’s incubator (73, 81%) and in the lap (68, 76%). A kinesthetic, or movement, component was added after tactile stimulation in 48 (54%) of cases. In open responses, therapists described the kinesthetic components as joint compressions and/or approximations, range of motion, and facilitating movement towards midline (e.g., hand to mouth). Infant massage was integrated into multimodal interventions among 47 (53%) of respondents, often in combination with skin-to-skin care, music therapy, or other sensory-based interventions. R7rding pressure and speed, most respondents described using medium pressure (67, 74%) and slow strokes (77, 86%), with slow stroke durations descriptions ranging from 1-20 seconds, or 6 seconds on average. Only 2 respondents (2%) described infant pressure as “deep” and no therapists reported use of fast strokes. Respondents described a range of estimated infant massage frequencies from 1x/week (5, 6%) to more than 4x/day (1, 1%); however, most commonly reported infant massage frequencies were 3x/week (24, 29%) and 4x/week (17, 20%). Respondents estimated that parents administered infant massage approximately 37% ± 25% (range 0-100) of the time.Table 2. Massage intervention characteristics.Mean (SD) or n^a^ (%)Use of kinesthetic component following infant massagen = 89  Yes48 (54%)  No41 (46%)Type of kinesthetic componentn = 43  Facilitated movement/guided movement/stretching14 (33%)  Joint compression or weight bearing11 (26%)  Developmental positioning, pelvic tilts, or flexion movements10 (23%)  Hand to mouth4 (9%)  Myofascial techniques4 (9%)  Midline orientation3 (7%)  Rocking/vibration3 (7%)  Grasp2 (5%)  Dynamic or static touch1 (2%)  Trigger reflexes largely into flexion1 (2%)  Neurodevelopmental Treatment (NDT) intervention1 (2%)Massage is part of a multimodal intervention (e.g., kangaroo care, with music)n = 89  Yes47 (53%)  No42 (47%)Use of additional sensory inputs during massagen = 43  Auditory (singing, speaking, music, humming)24 (56%)  Skin-to-skin19 (44%)  Vestibular/rocking/movement5 (12%)  Gustatory/Olfactory (parent smell, tastes)4 (9%)  Holding2 (5%)  Swaddle bathing2 (5%)  Visual1 (2%)  Containment/proprioception1 (2%)  Pacifier1 (2%)  Proprioception1 (2%)Location of infant massagen = 90  Open crib86 (96%)  Inside the incubator73 (81%)  On the caregiver/therapist’s lap68 (76%)  During Kangaroo Care21 (23%)  On the floor17 (19%)  Other5 (6%)Describe the typical speed of the strokes during massage in strokes per secondn = 90 Slow77 (86%)  Estimated strokes per secondRange 1-20; 6.3 (4.2) Medium13 (14%)  Estimated strokes per secondRange 1-9; 4.7 (2.5) Fast0 (0%)Describe the typical amount of pressure used during massagen = 90  Light21 (23%)  Medium67 (74%)  Deep2 (2%)Use of medium (e.g., emollient, oil) when providing massagen = 101  Oil45 (45%)  Lotion23 (23%)  Aquaphor or Vaseline18 (19%)  Massage over a swaddle or clothing with no emollient23 (23%)  Massage over skin but without use of emollient25 (25%)  Other or unstandardized5 (5%)Specific oil usedn = 45  Calendula oil19 (42%)  Grapeseed oil13 (29%)  Combination oil5 (11%)  Coconut oil4 (9%)  Sunflower oil3 (7%)  Almond oil1 (2%)  Safflower oil1 (2%)  Jojoba oil1 (2%)  Parent brings in oil1 (2%) Medium-chain-triglyceride oil1 (2%)Use of gloves during massage?n = 90  Yes32 (36%)  No19 (21%)  Only in certain circumstances39 (43%)If yes, is there a unit policy for use of gloves during massage?n = 32  Yes31 (97%)  No1 (3%)Estimated massage frequency (times per week received)n = 844.41 ± 3.6  1 time per week5 (6%)  2 times per week11 (13%)  3 times per week24 (29%)  4 times per week17 (20%)  5 times per week14 (17%)  6 times per week1 (1%)  Daily4 (5%)  1-2x Daily7 (8%)  More than 4 times daily1 (1%)Estimated percentage of massage conducted by parentsn = 84(Range 0-99)37.6 ± 24.9^a^n reflects number of responses for each question.

Use of mediums and gloves

The use of emollients varied. Forty-eight (48%) respondents either massaged infants over clothing without an emollient or massaged directly on infant skin without using an emollient. Those who used a medium directly on infant skin during infant massage most commonly used oils (45, 45%). Open responses regarding the types of oil used for infant massage revealed a wide variety of preferences with calendula oil emerging as the most frequently reported (19, 42%). In open text responses, some respondents reported difficulty getting use of oil mediums approved in their NICUs due to fragile infant skin and unit policies. Gloves were always used during infant massage among 32 (36%) of respondents, and gloves were used only in certain circumstances when isolation, infection, or exposure precautions were in place among 39 (43%) of respondents. Open responses indicated that parents were encouraged to massage their infants without gloves to promote skin-to-skin contact and positive touch experiences.

Goals of infant massage

Responses for indications and appropriate populations for massage are reported in Table 3. Respondents (n = 101) identified a wide range of goals for infant massage in the NICU. The most frequently endorsed infant-centered goals included providing positive sensory experiences (e.g., tactile, kinesthetic, proprioceptive input) (n = 88, 87%), promoting relaxation and/or sleep (86, 85%), and enhancing state regulation and/or facilitating calm alert states (85, 84%). Respondents also endorsed the following parent-centered goals: encouraging parent-infant interaction (84, 83%) and improving parent competence and responsiveness (84, 83%).Table 3. Infant massage indications and appropriate populations.n^a^ (%)Infant-Centered goals for massagen = 101  Provide positive sensory experiences88 (87%)  Promote sleep and relaxation86 (85%)  Facilitate calm alert state/Enhance state regulation85 (84%)  Reduce pain85 (84%)  Improve infant tolerance to handling84 (83%)  Improve respiratory function84 (83%)  Improve skin texture83 (82%)  Address withdrawal symptoms82 (81%)  Improve feeding outcomes79 (78%)  Weight gain73 (72%)  Improve gastrointestinal issues73 (72%)  Control edema57 (56%)  Ease of diapering57 (56%)  Improve tone and movement53 (52%)  Improve bone density45 (45%)  Improve bilirubin levels34 (34%)  *Other (desensitization)1 (1%)Parent-centered goals for massagen = 101  Facilitate parent-infant interaction84 (83%)  Improving parent competence and responsiveness84 (83%)  Increase parent engagement74 (73%)Infant conditions and clinical presentations that indicate massagen = 101  Infants with challenges with state regulation82 (81%)  Late preterm (33-36 weeks BGA)78 (77%)  Hypertonia76 (75%)  Neurological conditions (e.g., hypoxic ischemic encephalopathy)62 (61%)  Hypotonia61 (60%)  Infants with occupational therapy orders60 (59%)  Infants with physical therapy orders51 (50%)  Congenital anomalies48 (48%)  Respiratory conditions45 (45%)  Very low birth weight/very preterm ( < 32 weeks BGA)31 (31%)  Infants with surgical conditions29 (29%)  Cardiac conditions28 (28%)  Extremely low birth weight/very preterm ( < 28 weeks BGA)23 (23%)  Other12 (12%)PMA of infant when massage is typically initiatedn = 90Range 27–26 weeks PMA32.2 (1.4)Massage is appropriate at the following PMAs 23 weeks PMAn = 88  Never84 (95%)  Sometimes4 (5%) 24 weeks PMAn = 88  Never82 (94%)  Sometimes5 (6%) 25 weeks PMAn = 87  Never83 (95%)  Sometimes5 (6%) 26 weeks PMAn = 88  Never78 (89%)  Sometimes10 (11%) 27 weeks PMAn = 87  Never75 (86%)  Sometimes12 (14%) 28 weeks PMAn = 89  Never61 (69%)  Sometimes23 (26%)  About half of the time5 (6%) 29 weeks PMAn = 92  Never54 (61%)  Sometimes27 (30%)  About half of the time8 (9%) 30 weeks PMAn = 92  Never41 (45%)  Sometimes32 (35%)  About half of the time12 (13%)  Frequently7 (7%) 31 weeks PMAn = 93  Never35 (38%)  Sometimes33 (35%)  About half of the time14 (15%)  Frequently11 (12%) 32 weeks PMAn = 97  Never5 (5%)  Sometimes25 (25%)  About half of the time26 (27%)  Frequently29 (30%)  Always10 (10%) 33 weeks PMAn = 99  Never2 (2%)  Sometimes17 (17%)  About half of the time23 (23%)  Frequently46 (46%)  Always11 (11%) 34 weeks PMAn = 97  Sometimes7 (7%)  About half of the time11 (11%)  Frequently61 (61%)  Always18 (18%) 35 weeks PMAn = 97  Sometimes5 (5%)  About half of the time6 (6%)  Frequently63 (65%) Always23 (24%) 36 weeks PMAn = 96  Sometimes2 (2%)  About half of the time6 (6%)  Frequently59 (61%)  Always29 (30%) 37 weeks PMAn = 93  Sometimes2 (2%)  About half of the time3 (3%)  Frequently59 (63%)  Always29 (31%) 38 weeks PMAn = 96  Sometimes2 (2%)  About half of the time1 (1%)  Frequently61 (64%)  Always32 (33%) 39 weeks PMAn = 96  Sometimes2 (2%)  About half of the time1 (1%)  Frequently61 (64%)  Always32 (33%) 40 weeks PMAn = 96  Sometimes2 (2%)  About half of the time1 (1%)  Frequently58 (60%)  Always35 (36%)40+ weeks PMAn = 96  Sometimes2 (2%)  About half of the time2 (2%)  Frequently56 (58%)  Always36 (38%)Massage is appropriate in the following circumstances Within 72 hours for ELBW infantn = 93  Never87 (91%)  Sometimes6 (6%) Chest tuben = 89  Never27 (30%)  Sometimes38 (43%)  About half of the time8 (9%)  Frequently14 (16%)  Always2 (2%) ECMOn = 87  Never59 (68%)  Sometimes23 (26%)  About half of the time3 (3%)  Frequently2 (2%) Endotracheal intubationn = 92  Never13 (14%)  Sometimes40 (43%)  About half of the time10 (11%)  Frequently25 (27%)  Always4 (4%) Fragile bonesn = 92  Never31 (34%)  Sometimes41 (45%)  About half of the time10 (11%)  Frequently10 (11%)  Always0 (0%) During hypothermian = 91  Never72 (79%)  Sometimes13 (14%)  About half of the time4 (4%)  Frequently2 (2%) Non-invasive ventilationn = 95  Never10 (11%)  Sometimes27 (28%)  About half of the time16 (17%)  Frequently33 (34%)  Always9 (10%)Neonatal abstinence syndrome or neonatal opioid withdrawal syndromen = 97  Never2 (9%)  Sometimes3 (3%)  About half of the time4 (4%)  Frequently52 (54%)  Always36 (37%) Open woundsn = 92  Never57 (62%)  Sometimes31 (33%)  About half of the time2 (2%)  Frequently1 (1%) Sepsisn = 93  Never75 (81%)  Sometimes11 (12%)  About half of the time3 (3%)  Frequently4 (4%)Infant body parts avoided during massagen = 65  Fragile bone, limb at risk of deep vein thrombosis; skin infection, rash, open wound16 (25%)  Abdomen/belly12 (18%)  Genitals7 (11%)  Limbs with intravenous or catheter lines (PIV, PICC)8 (12%)  Anterior chest6 (9%)  Belly button/umbilical area4 (6%)  Head4 (6%)  Face4 (6%)  Spine2 (3%)  A painful limb1 (2%)Safety issues observed during massagen = 101  Behavioral motor stress signs62 (61%)  Infant physiological instability36 (36%)  Respiratory changes14 (14%)  Temperature instability13 (13%)  Lines concerns (dislodgment, pulling)9 (9%)  Allergic response to medium6 (6%)  Skin safety4 (4%)  Other (movement of nasal prongs)1 (1%)^a^n reflects number of responses for each question.

Infant eligibility and timing

The average PMA reported for initiation of infant massage was 32.2 (1.4) weeks (range 27–36). Infant conditions and clinical presentations that indicated massage as part of the therapy plan of care were most commonly state regulation dysfunction (82, 81%), and infants born at late preterm gestation (78, 77%). The most endorsed conditions for initiating infant massage “frequently” or “always” included Neonatal Opioid Withdrawal Syndrome (NOWS) (87, 90%). Open text responses to “Other” populations (12, 12%) discussed use of infant massage as part of an early mobilization protocol for infants undergoing extracorporeal membrane oxygenation (ECMO) and continuous renal replacement therapy (CRRT). In contrast, the majority of respondents identified ECMO (59, 66%) as one of several clinical conditions under which infant massage is never initiated.

Few respondents endorsed initiating massage before 32 weeks PMA, however, 4 (4%) of respondents reported that massage was “sometimes” appropriate as early as 23–24 weeks PMA. Open responses mentioned considerations of birth gestational age (BGA), medical readiness, and physiological stability to determine the appropriateness of massage, with some presenting guidelines discouraging massage for infants <1500 g or those with active infections, sepsis, or unhealed surgical sites.

Safety concerns

When asked about safety issues observed during massage, respondents (n = 101) most frequently reported infant stress (62, 61%) and physiological instability (e.g., changes in heart rate or oxygen saturation) (36, 36%), with fewer respondents concerned about allergic reactions to massage mediums (6, 6%) or skin safety (4, 4%). Open-ended responses about safety concerns discussed the importance of monitoring stress cues, ensuring thermoregulation, and adjusting interventions based on infant tolerance or behavior. Additional concerns were noted about potential safety issues with insufficient caregiver training and pressure inconsistencies.

Narrative description of infant massage

Finally, respondents were invited to describe infant massage practices in their unit using the following free-text prompt: “Describe the massage you do in your unit. (What is the environmental set-up, position of the infant, where are your hands, where do you start, how many strokes, what body parts are included, how much pressure is applied?)” Of the 101 individuals surveyed, 78 (77%) provided detailed responses. As noted in previous multiple-choice responses, open-ended responses revealed considerable variability in infant massage practices across neonatal units; however, common themes emerging in environmental set-up, infant positioning, techniques, and responsiveness to infant cues are described below.

Regarding environmental setup, infant massage was typically performed in low-stimulation environments, including incubators, open cribs, or on the therapist’s or parent’s lap. Dim lighting, warm blankets, and quiet surroundings were frequently used to promote infant comfort. Massage sessions were often integrated into care routines, such as before or after feeding or diapering. Regarding positioning, infants were usually positioned in prone, side-lying, or supine positions depending on their medical stability and tolerance. Respondents described use of semi-swaddling to contain the infant, with only the body part being massaged exposed. Sessions typically began with static touch and progressed to dynamic strokes of commonly massaged areas: back, extremities, and face. Abdominal massage, such as the “I Love You” technique, was used to support gastrointestinal function. Some therapists reported use of scar massage or manual edema mobilization when clinically indicated. Number of strokes per body part ranged from 3 to 10, though many respondents emphasized that stroke count was determined by infant cues rather than a fixed protocol. Pressure was described as light to moderate, often guided by visual cues such as slight skin blanching or wrinkling. In general, respondents reported that infant massage was paused or discontinued if the infant demonstrated stress cues.

Discussion

This study offers an in-depth description of infant massage practices in NICUs, revealing its widespread use, substantial variability in technique, and shared safety concerns. Survey responses demonstrate that neonatal therapists not only recognize the potential benefits of infant massage but also advocate for its integration into clinical practice. Additionally, the data suggest that these clinicians are well-informed regarding the evidence base for infant massage, supporting its indications and therapeutic benefits.

While over 70% of respondents reported that massage was considered standard of care practice in their NICU, a central finding of this study was the significant variability in infant massage techniques and protocols described—an observation that mirrors the heterogeneity documented in prior reviews [5, 7, 18, 20]. Across the reviewed literature, infant massage protocols described share several core characteristics, including the use of tactile and kinesthetic stimulation, moderate pressure techniques, and structured session durations typically ranging from 10 to 20 minutes. However, reported protocols differ in stroke speed, weekly frequency, and use of emollients, with no universally accepted standards [5, 7, 18, 20]. A 2015 meta-analysis [20] and more recent reviews by Alvarez et al. [7] and Pados et al. [5] highlight similar inconsistencies, particularly in how “moderate pressure” is defined and applied, despite its use in infant massage being associated with improved weight gain and physiological outcomes.

Estimated frequency of infant massage in the NICU varied considerably in our study, from 1x/week to 4x/day, with 3-4x/week most commonly reported. These variables and mostly low frequencies of massage perhaps stand in the greatest contrast to one of the most researched infant massage protocols, the Field Method [16, 17], which employs infant massage 3 times per day. The Field Method has been associated with positive outcomes [16, 17, 24], but it is unclear if infants receive the same degree of positive benefits when massage is delivered only 2–3x/week as reported by our survey respondents. Two small clinical trials published within the last 2 years examined two different infant massage protocols employed daily over a 2-week period in preterm infants >30 weeks PMA with opposing results. A study by Weerakul et al. [25] demonstrated increased weight gain and reduced hospital stays in the massage group—consistent with earlier studies employing infant massage at a frequency of 3 times daily. In contrast, a trial by Alinejad-Naeini et al., which evaluated the M-Technique massage protocol [9], did not find significant differences in infant weight gain between intervention and control groups; however, the intervention group demonstrated more favorable behavioral response scores than the control group—findings that may have important implications for parent-infant interaction [26].

Use of ILTF and NTMC protocols was most frequently reported, yet, despite even higher reported rates of formal training in these protocols, many therapists appear to opt for eclectic approaches that draw from multiple sources or personal experience. This variability may be attributed to the fact that, although ILTF and NTMC protocols are accessible to therapists who have completed training programs, continued access to parent education materials, program resources, and updated protocols often requires re-certification or annual membership fees. These additional requirements may discourage therapists or institutions from committing to a specific program or protocol [11, 12].

Survey responses indicated that use of emollients or mediums varied considerably. When mediums were not used, open-text responses revealed that it was typically because unit policies prohibited emollients in the NICU due to concerns about the fragility of preterm infants’ skin. Topical emollients are more commonly used in low- to middle-income countries due to cultural norms that support oil massage for newborns and because they offer a low-cost, accessible way to reduce infection risk and support skin barrier function [27]. Evidence from clinical trials in these regions has demonstrated benefits such as improved skin integrity, enhanced growth, and reduced neonatal mortality, yet, according to the 2023 WHO recommendations, topical emollient use is conditionally recommended due to “low certainty” evidence on type, dose, timing, and duration [28]. The report notes that sunflower or coconut oil may be appropriate with gentle application to protect skin integrity and that use of emollients should be guided by clinical judgment and in collaboration with parents [28].

Although emollient therapy is used globally to support skin integrity and reduce infection rates, in the US, neonatal therapists primarily use a medium during infant massage to promote comfort and relaxation. An early study by Field et al. [29] found that typically developing 1-month old infants massaged with oil showed fewer limb movements, fewer stress behaviors, less gaze aversion, and greater physiological regulation, including lower salivary cortisol levels during massage than those massaged without oil. A follow-up study by Field et al. in 2016 [30] found that full-term, healthy newborns who received daily massage with lotion slept longer and experienced fewer night wakings over a one-month period compared to those massaged without lotion. Therefore, more research is needed to identify the safest emollient for use on preterm infant skin in the context of infant massage application.

Respondents to our survey felt that infant massage’s application should be approached with caution, particularly for infants under 32 weeks PMA, those undergoing ECMO, or those with chest tubes, sepsis, or open wounds. These findings reflect high levels of clinical decision making and align with prior studies that have emphasized the need for careful assessment of infant medical stability before initiating massage [31]. Inappropriate infant PMA and medical conditions described in our results are aligned with recent clinical trials of preterm infant massage which excluded infants <28 weeks PMA and <30 weeks PMA, infants with congenital anomalies (e.g., congenital heart disease, congestive heart failure, lung or airway malformations), requiring endotracheal intubation or chest tubes, exhibiting unstable vital signs (including persistent bradycardia or tachycardia), diagnosed with convulsions, septic shock, or grade III/IV intracranial hemorrhage, or experiencing severe respiratory distress. While some therapists reported performing infant massage on infants with high-acuity conditions, such as those receiving ECMO, these practices were not widely endorsed and highlight the need for clearer contraindication guidelines.

Safety concerns were recognized by most respondents, with 61% having observed infant behavioral stress signs and 36% observing physiological instability during infant massage. These concerns were generally managed through responsive clinical decision making, such as adjusting pressure or speed, pausing, or discontinuing massage. Systematic reviews have identified adverse effects such as changes in heart rate, oxygen saturation, and skin integrity when infant massage is applied without appropriate clinical oversight [31], but overall, conclude that adverse events during infant massage are extremely rare [7]. However, safety issues or adverse events are not consistently reported in published research.

Although strategies for managing adverse events in the NICU are well-documented in the literature [32, 33], most studies focus on medication-related safety concerns, and no existing reviews specifically address adverse events associated with infant massage. Gao et al. investigated the effectiveness and safety of combining sucrose with massage, music, non-nutritive sucking, and gentle human touch to manage repeated procedural pain in preterm infants. As a secondary outcome, they monitored adverse events such as vomiting, abdominal distension, coughing, choking, oral infection, sustained tachycardia, bradycardia, tachypnea, dyspnea, and hyperglycemia before, during, and after procedures. Among the 71 infants studied, vomiting was the only adverse event reported, occurring in one infant from the routine care group and one from the intervention group [34]. Given the limited reporting of adverse events in systematic reviews and our survey findings, neonatal therapists appear aware of potential risks but generally view the benefits of infant massage as outweighing those risks. Authors of a systematic review published in 2017 recognized the need for adaptive, cue-based approaches that account for infant medical status and individual tolerance during infant massage [7], which aligns with how neonatal therapists approach infant massage based on open-text responses in our survey.

Limitations

This study has limitations. The survey format and means of recruitment may have led to underrepresentation of certain groups, particularly those in Level II NICUs or units without dedicated neonatal therapists. Most respondents were therapists who regularly perform infant massage and who may have a positive bias toward its benefits, potentially skewing the data. Additionally, while other disciplines involved in infant massage were mentioned, this survey included only neonatal therapists so lacks representation from non-therapy providers who may also administer massage. Duplicate responses from the same hospital were identified and addressed, but this may still influence the generalizability of findings. Further, despite our multi-pronged recruitment approach to increase our reach, we acknowledge that our response rate of 101 therapist respondents represents only a small fraction (<1%) of the estimated 4232 neonatal therapists in the US [35]. Additionally, we acknowledge that because respondents represented various areas of the US and outside of the US, there may be cultural differences in massage characteristics reported and this is an important area to consider for future research.

Conclusion

Infant massage is a widely used and valued intervention in the NICU, yet its implementation is marked by considerable variability and clinical caution. Therapists adapt their approaches based on infant cues and medical status, suggesting that flexibility is both common and necessary in this complex environment. While safety concerns are not uncommon, they appear to be managed with thoughtful clinical judgment. Based on these results, the majority of neonatal therapists are aware of the indications and benefits of infant massage based on available evidence. These findings highlight the need not only for clearer definitions in infant massage techniques and more transparent and reproducible infant massage protocols in research, but also for standardized, evidence-based protocols and training to ensure safe and effective infant massage practices in neonatal care.