Effect of Oxytocin Massage and Slow-Stroke Back Massage on Breast Milk Production and Reduction of Risk Factors Associated with Stunting: A Randomized Controlled Trial

2.1. Study Design

This study used a Randomized Controlled Trial (RCT) with pre- and post-test control groups. It included postpartum mothers who delivered vaginally or via Caesarean section and who roomed in. Mothers were excluded if they had complications, were in high-care units, or had babies with congenital disabilities. The dependent variables were breast milk production (daily urination, infant not fussy or sleeping well, babies' weight), and babies' length, while the independent variables included oxytocin massage and slow-stroke back massage (SSBM). Control variables were maternal and infant characteristics.

2.2. The Participants

The population was randomized into three groups: two intervention groups and one control group. Although this sample was smaller than the ideal 246, the finite population correction (FPC) indicates that 70 participants would suffice. FPC is used when the population is small (under 10,000) and sampling is without replacement. It reduces the required sample size, since a small sample can represent a small population. Thus, the adjusted sample size was about 70–75. The final sample size, using FPC, was about 75.

2.3. Data Collection

Seventy-five postpartum mothers were divided into three groups. The first group received only oxytocin massage, administered by trained staff. The second group received both oxytocin massage and the SSBM intervention, a structured support program also delivered by trained personnel. The control group received routine hospital care without additional interventions. Breast milk production was evaluated using three indicators: at least six newborn urinations per day, the baby not being fussy or excessively sleepy, and the baby maintaining normal weight and body length. Early stunting detection was conducted using the length-for-age z-score (<-2 SD), calculated from two separate measurements taken over two months.

Breast milk production was assessed by counting daily urinations. At least 6 wet diapers meant adequate production. Mothers reported infants' behavioral states as either Satisfied (calm or sleepy) or Fussy (crying or restless) after breastfeeding. Weight and height were measured using standard techniques. A UNICEF-recommended wooden board and calibrated tape were used. Data collection took about 20–25 minutes per participant, including interviews and physical measurements.

Group one received oxytocin massage intervention. The procedure began by assessing the respondent's condition and positioning her sitting on the edge of the bed with a towel placed under her breasts. The nurse applied baby oil to both palms, then massaged the respondent's back from the base of the neck to both breasts. The massage was performed along both sides of the spine, moving downward (from the neck to the shoulder blades) in small circular motions using the thumbs and applying gentle pressure. This movement was repeated 5–10 times. Finally, the nurse cleaned the respondent's back with warm and cold water, alternating between them, using a washcloth. The materials used in this procedure included a towel, washcloth, basin, and massage oil [11].

Group Two: Combination of Oxytocin Massage and SSBM. The procedure began by assessing the respondents' condition and positioning them sitting on the edge of the bed. A towel was placed under their breasts, and the nurse applied baby oil to their palms. The nurse first performed a rolling massage, followed by SSBM. In the first movement, the nurse held the respondents' upper shoulders with both hands, placed both thumbs just below the base of the skull, and made small circular motions across the upper neck. In the second movement, one hand was placed at the base of the skull and slowly, gently stroked downward along the spine to the waist; the other hand followed as the first returned to its original position. In the third side movement, both hands were placed on each respondent's neck just below the ear, stroking downward across the collarbone while the thumbs glided over the shoulder blades three times. In the fourth movement, both thumbs were placed on each side of the back, starting at the shoulder and moving down to the waist with gentle pressure. In the final side movement, both hands were placed on each respondent's neck, then stroked long, continuous strokes down across the shoulders and the entire spine. This movement was repeated several times, with the entire procedure lasting 15–20 minutes. Once completed, the nurse wiped off any remaining baby oil using alternating warm and cool wet washcloths and helped the patient return to a comfortable position. Materials used in this combined procedure included gloves, pillows, towels, washcloths, a basin, massage oil, a breast milk container, and relaxation music [14].

Group Three: Routine Hospital Care. The third group served as a control group, receiving only routine hospital care. This care generally included standard postpartum hygiene care, such as monitoring uterine involution and breast care for milk expression.

The first assessment occurred at 1 month, followed by a second at 2 months. Data collection involved home visits to assess criteria, including at least six urinations per day, a non-fussy or sleeping behavioral state, and normal physical growth indicators, such as weight and length. Post-collection, data completeness was verified, and the data were then recoded and categorized.

2.4. Data Analysis

Analysis was performed with STATA 13. Statistical analysis was conducted at two levels. First, within-group changes in breast milk adequacy (a binary outcome) across three time points (birth, 1 month, 2 months) were examined using McNemar’s Exact Binomial test within each group. Within-group changes in continuous growth outcomes (body length and length-for-age Z-score) were analyzed using the paired t-test for normally distributed data and the Wilcoxon signed-rank test for non-normal distributions. Second, and critically, between-group comparisons were conducted to directly assess intervention effects. Differences in breast milk adequacy rates across all three groups at each time point were compared using the Chi-square test or Fisher’s Exact test as appropriate, with Cramer’s V reported as the effect size measure. For continuous outcomes (Z-scores and body length), a one-way ANOVA with Bonferroni post hoc pairwise comparisons was used to compare groups at each time point; partial eta-squared (η²) was reported as the effect size. All analyses report point estimates with 95% confidence intervals (95% CI). A significance threshold of α = 0.05 was applied. To address baseline imbalances in family income and infant urination frequency, ANCOVA (for continuous outcomes) and logistic regression (for binary outcomes) were conducted as sensitivity analyses, with the imbalanced variables entered as covariates. Results from both unadjusted and adjusted analyses are reported where applicable.

2.5. Ethical Considerations

The study received ethical approval from the ethics committee at Zainoel Abidin General Hospital, Aceh (Ref No. 267/ETIK-RSUZA/2025). Permission was granted by hospital management. Verbal informed consent was obtained from participants, including explanations of their rights and the study's purpose. Participants were told they could withdraw at any time. The research adhered to ethical standards and received hospital authorization. The stages of this research use the CONSORT pattern as presented in Fig. (1) below:

3.1. Characteristics of Study Participants

A total of 75 postpartum mother including their infants, participated in the study. Table 1 presents the baseline characteristics of postpartum mothers and their infants across three study arms: Intervention Group 2, receiving combined SSBM and Oxytocin Massage (SSBM+OM); Intervention Group 1, receiving Oxytocin Massage alone (OM); and the control group. Each group consisted of 25 mother–infant pairs. Overall, most maternal and infant characteristics were comparable across groups, indicating adequate baseline balance before intervention, with only a few variables showing statistically significant differences.

Regarding maternal characteristics, maternal age distribution did not differ significantly among groups (p = 0.486). The majority of mothers in all groups were aged 21–34 years, representing the typical reproductive age range, while a smaller proportion were ≥35 years. Very few mothers were younger than 20 years, suggesting minimal influence of adolescent maternal age on study outcomes.

Parity distribution was similar across groups (p = 0.356), with most mothers being multigravida in all arms. This consistency suggests that prior childbirth experience was unlikely to affect breastfeeding outcomes differentially across groups. Educational levels showed no significant variation (p = 0.630), with the majority of mothers having secondary education, and smaller groups with primary or tertiary education. Similarly, maternal occupation was evenly distributed across groups (p = 0.846), with homemakers comprising the largest share, followed by entrepreneurs, teachers, and contract workers. Only a few healthcare workers were present, reducing the likelihood that professional bias would affect breastfeeding practices.

Antenatal care visits were mostly regular (≥6 visits) across all groups, with no notable differences (p = 0.727). Prior breastfeeding experience was also similar (p = 0.433), with the majority of mothers having previous experience, indicating a comparable level of initial breastfeeding readiness. Likewise, husband-related factors, such as education level (p = 0.630) and occupation (p = 0.456), showed no significant differences among the groups, suggesting similar household support environments.

However, family income differed significantly between groups (p = 0.032). A larger proportion of mothers in Intervention Group 1 had incomes below IDR 3,165,000 than in the other two groups. In contrast, higher-income households were more common in Intervention Group 2 and the control group. This imbalance suggests potential socioeconomic variation that may influence maternal nutrition, stress levels, or breastfeeding support, and therefore warrants consideration in interpreting outcomes or adjusting subsequent analyses.

Infant characteristics were consistent across groups: sex distribution was similar (p = 0.850), with a slightly higher proportion of male infants overall. Gestational age was also comparable (p = 0.551), with most infants born at term (37–40 weeks), indicating similar neonatal maturity. Birth weight categories showed no significant difference (p = 0.584), with most infants having normal birth weight (2.500–4.000 g). A small proportion had low birth weight, and very few weighed over 4.000 g. Birth length distribution was also consistent (p = 0.688), with most infants measuring 48–52 cm at birth. Initial postpartum infant nutrition was similar (p = 0.730), with exclusive breastfeeding being predominant across all groups. Minimal formula supplementation was observed, suggesting baseline feeding practices were largely uniform before any intervention effects.

In contrast, infant urination frequency varied significantly among groups (p = 0.000). A notably higher proportion of infants in Intervention Group 2 had urination frequency ≥6 times/day compared to Group 1 and the control group. Conversely, low urination frequency (<6 times/day) was more prevalent in the control group. Since urination frequency indicates hydration status and indirectly reflects adequacy of milk intake, this baseline difference may suggest early disparities in breastfeeding effectiveness. It should be considered when analyzing intervention outcomes. Breastfeeding satisfaction showed no statistically significant difference (p = 0.111), though satisfaction was slightly lower in Intervention Group 2. Still, most mothers across groups reported satisfactory breastfeeding experiences at baseline.

In summary, Table 1 shows that maternal and infant characteristics are generally balanced across study groups, supporting their comparability at baseline. Only family income and infant urination frequency differed significantly, suggesting potential socioeconomic and early feeding differences that may warrant adjustment or discussion when evaluating intervention effects. Overall, the similarity of most baseline variables enhances the internal validity of future comparisons between intervention and control groups.

Table 2 shows the within-group changes in breast milk adequacy from birth to two months across three study groups. Breast milk adequacy was evaluated based on clinical criteria: an infant urinating six or more times daily, absence of persistent fussiness, and proper weight and length growth. Each group consisted of 25 mother-infant pairs. These findings are directly relevant to assessing the effectiveness of massage interventions in boosting breast milk production, which could help prevent early stunting.

In Intervention Group 2, which received both SSBM and OM, 68% of infants met the adequacy criteria at birth. Then, it increased to 76% at one month, though the difference was not statistically significant (p = 0.774). From one to two months, the proportion reaching adequacy rose to 100%, showing a significant change (p = 0.031). Overall, from birth to 2 months, breast milk adequacy increased significantly from 68% to 100% (p = 0.007). These results indicate that the combined massage intervention has a cumulative effect, substantially enhancing breast milk production and ensuring sufficient intake for all infants by two months.

In Intervention Group 1, which only received Oxytocin Massage, 68% of infants had sufficient breast milk at birth and at 1 month, with no significant early change (p = 1.000). However, over the period between one and two months, adequacy increased to 96%, which is a statistically significant increase (p = 0.015). Overall, from birth to two months, the proportion improved significantly from 68% to 96% (p = 0.039). These results suggest that Oxytocin Massage alone can improve breast milk adequacy. However, its effect is somewhat weaker than that of the combined intervention, as not all infants achieved full adequacy.

In the control group, breast milk adequacy was significantly lower and showed little change over time. At birth, only 32% of infants received adequate breast milk, and this increased slightly to 44% at one month (p = 0.581), but then declined to 40% at two months (p = 1.000). Overall, from birth to two months, the change was not statistically significant (32% to 40%; p = 0.790). These results emphasize the limited natural improvement in breast milk adequacy without intervention and highlight the risk of inadequate intake and early growth delays.

Overall, Table 2 shows that both massage-based interventions enhance breast milk adequacy, with the combined use of SSBM and Oxytocin Massage being the most effective. Significant improvements between one and two months, and between birth and two months, in both groups support the hypothesis that massage can increase milk production. These gains in breast milk sufficiency are especially relevant for preventing early stunting, as proper nutrition during the first two months is essential for supporting healthy growth and minimizing growth delays.

Table 3 shows within-group changes in infant body length and length-for-age Z-scores from birth to two months across the three study groups: Intervention Group 2, receiving the combination of SSBM and Oxytocin Massage (SSBM+OM); Intervention Group 1, receiving OM only; and the control group. These indicators are directly relevant to the study's focus, as linear growth and Z-scores serve as early markers of stunting and as indicators of the effectiveness of interventions to improve breast milk intake.

In Intervention Group 2 (SSBM + OM), infants demonstrated substantial, statistically significant increases in body length at all time points. The mean body length increased from 46.52 cm at birth to 51.68 cm at one month, showing a 5.16 cm increase (p = 0.000). From one to two months, it grew further to 56.56 cm, with an increase of 4.88 cm (p = 0.000). Overall, infants grew by 10.04 cm in the first two months (p = 0.001). Z-scores, which reflect growth relative to age-specific standards, also improved from −1.61 at birth to −0.70 at two months (increase of +0.92; p = 0.000), showing both absolute linear growth and better growth adequacy. This pattern suggests that the combined SSBM and OM intervention could promote early catch-up growth by increasing breast milk intake, supporting adequate nutrition, and preventing early stunting.

In Intervention Group 1 (OM only), infants showed significant increases in body length: from 46.24 cm at birth to 50.64 cm at one month (+4.40 cm; p = 0.000), and from 50.64 cm to 55.84 cm between one and two months (+5.20 cm; p = 0.000), making a total of 9.60 cm over the two months (p = 0.000). Z-score variations showed a minor decrease in the first month (−1.72 to −1.90; p = 0.347), suggesting that initial linear growth did not quickly translate into better age-specific growth adequacy. A significant improvement was seen in the second month, with Z-scores increasing from −1.90 to −1.05 (a gain of 0.86; p = 0.000). Overall, there was a 0.68 increase from birth to two months (p = 0.016). These results suggest that OM alone can promote infant growth and result in a steady improvement in Z-scores, although this effect appears somewhat delayed relative to the combined intervention.

In the control group, the increases in body length were smaller but statistically significant, with gains of 3.32 cm in the first month (from 47.36 to 50.68 cm) and 3.12 cm in the second month (from 50.68 to 53.80 cm), totalling a 6.44 cm increase (p = 0.000). However, Z-scores revealed a concerning trend: a significant drop in the first month (−0.82 to −1.88; p = 0.002), a modest recovery in the second month (−1.88 to −1.41; gain +0.47; p = 0.004), and an overall non-significant decline from birth to two months (−0.82 to −1.41; −0.59; p = 0.056). This finding indicates that, without intervention, infants' growth falls below age-specific standards, underscoring the risk of early stunting.

Overall, the data in Table 3 indicate that massage interventions, particularly the combined use of SSBM and OM, notably enhance linear growth and increase length-for-age Z-scores during the first two months of life. These results support the hypothesis that enhancing breast milk production through these methods can promote early growth and help prevent stunting. Compared with OM alone, combination therapy appears to achieve more rapid improvements in both absolute and relative growth, highlighting the potential of integrated massage approaches in postnatal care.

Table 4 presents between-group comparisons for all primary outcomes, adjusted for baseline covariates (family income and infant urination frequency), which showed significant imbalance at randomisation (Table 1). All cells report adjusted means with 95% confidence intervals (CIs) for continuous outcomes and Odds Ratios (ORs) with 95% CIs for the binary outcome. For breast milk adequacy, logistic regression adjusted for the two covariates showed a significant group difference at birth (adjusted p = 0.041; OR = 4.21, 95% CI: 1.47–12.06) that strengthened considerably at two months (adjusted p < 0.001; OR = 28.50, 95% CI: 7.82–103.9), indicating that after controlling for baseline imbalances, mothers in Group 2 were 28.5 times more likely to have adequate milk production than the control group at two months. At two months, all mothers in Group 2 (100%) had adequate milk, compared with 96% in Group 1 and 40% in the control group.

For infant body length, ANCOVA confirmed no adjusted difference at baseline (reference row). Significant adjusted between-group differences emerged at one month (adjusted p = 0.021, η²p = 0.13; Group 2 adj. mean: 51.71 cm, 95% CI: 51.05–52.37) and two months (adjusted p < 0.001, η²p = 0.28; Group 2: 56.61 cm, 95% CI: 55.99–57.23), with Group 2 consistently recording the highest adjusted mean length. The adjusted gain from birth to two months was greatest in Group 2 (10.09 cm, 95% CI: 9.55–10.63) versus Group 1 (9.54 cm, 95% CI: 8.96–10.12) and control (6.48 cm, 95% CI: 5.92–7.04) (p < 0.001, η²p = 0.46), indicating a large effect. For length-for-age Z-scores, ANCOVA revealed significant adjusted between-group differences at one month (adjusted p = 0.018, η²p = 0.13) and two months (adjusted p < 0.001, η²p = 0.23; Group 2: -0.68, 95% CI: -0.96 to -0.40). The adjusted gain in Z-score from birth to two months was substantially greater in Group 2 (+0.93, 95% CI: +0.69 to +1.17) and Group 1 (+0.65, 95% CI: +0.37 to +0.93) compared with the control group (-0.57, 95% CI: -0.79 to -0.35) (p < 0.001, η²p = 0.54), a very large effect.

The non-overlapping 95% CIs between Group 2 and the control group for all primary outcomes at two months provide strong evidence that the observed differences are not attributable to chance or to baseline imbalances controlled for in the ANCOVA model. Taken together, these ANCOVA-adjusted findings with 95% CIs confirm that the combined SSBM and oxytocin massage intervention produced significantly superior and clinically meaningful improvements in both breast milk adequacy and infant growth compared with routine care.