The effects, outcomes and acceptability of sterile water injections to relieve low back pain among birthing women in a tertiary hospital in Ghana: a non-randomised study Authors: Awube Menlah, Sue Kruske, Clare Davison, Isabella Garti, Yu Gao, Jonas Awuku-Afari, Nigel Lee Published: 2026 ABSTRACT AbstractBackgroundApproximately 30% of women experience low back pain when in labour. Large placebo-controlled trials conducted in high-resource settings have demonstrated the efficacy of sterile water injections (SWI) to manage back pain in labour. The procedure has also been endorsed by influential international clinical guidelines. However, the effectiveness and acceptability of the procedure amongst women in low-resource settings such as Ghana, where access to any analgesia may be limited, has not been explored. Investigating the clinical utility of SWI, assessing effectiveness and acceptability, is essential for determining its potential benefits for birthing women in this low- and middle-income country (LMIC).MethodsA non-randomised quasi-experimental study was conducted. Women in labour at term with self-reported back pain chose either a treatment with SWI or standard care. Participants in the SWI group received intradermal injections of sterile water, 0.1 mL, four points bordering the lumbo-sacral region (Michaelis’ Rhomboid). Self-reported back pain scores using a Visual Analogue Scale (VAS) were obtained prior to and following treatment up to 120 min after the injections for both groups. The difference in VAS score between groups was analysed using a multilevel repeated-measure analysis. Secondary endpoints included women’s satisfaction with the intervention and maternal and neonatal birth outcomes.ResultsDuring the study period, 73 women were enrolled, 51 choosing SWI and 22 standard care. There was no difference between the two cohorts in self-reported back pain prior to treatment. Women who received SWI reported a significant reduction at 30 min (95% Confidence Interval (CI) -98.29 to -88.60; p < .001) in VAS scores compared to women opting for standard care at all time points (10, 30, 60, 90, and 120 min). The maternal satisfaction survey, conducted with the intervention group, indicated that over 90% of participants were satisfied with their treatment experience and were willing to recommend SWI to others.ConclusionSWI was clinically effective and acceptable to women birthing in Ghana, who reported satisfaction and endorsement. There was a strong association between SWI and a reduction in labour back pain with no adverse effects recorded during the study period.Trial registrationThis study was prospectively registered with the Pan African Clinical Trial Registry with the identification number PACTR202307748580298 on 28th July 2023. Prospectively registered.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12884-026-09174-2.Keywords: Back pain, Labour pain, Sterile water injections, Acceptability Background Women experience considerable variation in the location and course of labour pain [1]. About one-third of all labouring women report low back pain during birth [2]. As a common variant of labour pain, the presence of low back pain can pose specific challenges to both the woman giving birth and the care provider (midwife) as it is often intense, continuous and unpredictable [3]. There are subjective reports from women providing vivid descriptions of their back pain as strong and intense localised back pain [4, 5], prompting women to seek analgesia. A simple, low-resource method for managing labour back pain is the use of sterile water injections (SWI) [6]. Small amounts of sterile water injected into the low back of labouring women can provide low back pain relief almost immediately after the injection. The analgesic effect of sterile water injections is thought to occur through a counter-irritation mechanism, activating pain-modulating pathways consistent with the gate control theory of pain, thereby reducing the perception of labour-related back pain [2]. The analgesic effect can last up to 120 min with options for repeated injections at the request of the woman [7]. A recent practice guideline [7] and multi-centre randomised controlled trials (RCTs) [6, 8] have demonstrated the effectiveness of SWI in relieving low back pain among birthing women across Western settings. Sterile water injections are likely to be more suitable for birthing women who want to avoid pharmacological analgesia or in low-resource settings where pain relief options may be limited or unavailable [3]. In low to middle-income countries (LMIC) like Ghana, limited resources significantly restrict access to pharmacological and non-pharmacological interventions for managing pain during childbirth. This lack of resources can include insufficient equipment, a shortage of trained personnel and a general unavailability of alternative non-pharmacological therapies [9–11]. Women in these regions often lack also access to pharmacological pain relief options, leading to a reliance on some opioids such as meperidine (pethidine), which has been shown to be largely ineffective in labour pain management [12, 13]. However, due to a lack of awareness and training among Ghanaian and LMIC healthcare providers, coupled with the aforementioned resource limitations, SWI and similar interventions are underutilised or not used at all [14, 15]. This gap in knowledge and practice highlights the need to explore the clinical, maternal acceptability, and effectiveness of non-pharmacological pain management options, specifically SWI, among birthing women in Ghana. While SWI is endorsed in other settings with evidence-based practice guidelines as a safe and effective method for managing labour back pain [7, 16], its use remains largely unexplored in the Ghanaian context. Given the limited experience that Ghanaian women have with pain relief options, it was important to assess effectiveness as a measure of acceptability and feasibility. The unique clinical and resource-constrained environments of low- and middle-income countries may benefit from SWI. Therefore, this study aimed to evaluate the clinical outcomes and maternal acceptability of SWI as a novel approach to labour pain management in the Ghanaian setting, addressing a need gap in both local evidence and practice. Methods Design This study utilised a non-randomised approach to evaluate SWI’s clinical utility in Ghanaian maternity care through measures of effectiveness and acceptability as a pain relief method. Large well-designed placebo controlled RCT’s had previously demonstrated the analgesic efficacy of SWI [8]. Furthermore, the procedure has been endorsed in internationally recognised clinical guidelines including the UK’s 2023 National Institute for Health and Care Excellence [16]. Therefore, we considered that subjecting women to randomisation away from an effective and widely available pain relief strategy was unnecessary and unethical. Hence, a non-randomised design allowed participants to choose their preferred pain management option, and facilitated the evaluation of clinical outcomes, effects and acceptability. The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) (Supplementary File 1) [17]. Setting The study was undertaken at a primary national tertiary health facility in Ghana, recognized as the largest public health facility in West Africa. Located in the Greater Accra region, the hospital serves clients from across Ghana. The maternity department is divided into five units, each overseen by a senior obstetric consultant and a chief midwifery officer. On average, there are 9600 births recorded annually [18]. Predominantly Ghanaian women from urban and peri-urban Accra and surrounding regions, reflecting the city’s multiethnic, multilingual mix (English plus widely used local languages such as Twi, Ga, and Ewe). Midwives are the primary caregivers in this facility and attend to uncomplicated births. Commonly in Ghana, labour pain is managed by midwives by providing support and non-pharmacological strategies such as reassurance and massage. In some instances, pharmacological interventions such as pethidine are administered after prescription and approval by medical officers and obstetricians. Participants eligibility Pregnant women who met the eligibility criteria were considered for participation. Inclusion criteria Participants were provisionally eligible to enrol in the study if they were aged at or above 18 years at term (37–41 completed weeks). Final screening for inclusion occurred when a provisionally consented participant presented in labour with low back pain assessed by VAS as ≥ 7. Exclusion criteria Women were excluded if their labour was considered ‘high risk’ which was defined as older primigravida (≥ 40 years), hypertensive disorders (including pre-eclampsia), insulin-dependent diabetes, or a known coagulopathy. Women with a prior caesarean birth were also excluded, as their intrapartum care was provided by physicians rather than midwives. In addition, women were excluded if independent informed consent could not be obtained because a spouse’s consent was culturally required. Finally, women who could not communicate in English or any of the common Ghanaian languages such as Twi, Ga, Hausa, or Ewe and had no interpreter available were also not eligible. Once the participating woman presented in labour, further screening was undertaken to ensure the woman was in the active phase of a spontaneous or induced labour and presented with low back pain scored ≥ 70 mm on a segmented 0–100 mm Visual Analogue scale (VAS). Recruitment and consent of participants Midwives and research assistants received structured training on participant recruitment, administration of the SWI intervention, and standardised data collection procedures. Training was tailored to align with their respective clinical units such as antenatal clinic to ensure consistency and compliance to the study protocol. The recruitment of the pregnant women occurred at the hospital antenatal clinic (ANC), maternity emergency, and maternity wards. These wards had pregnant women at term or in the early stages of labour and in anticipation that a proportion would experience low back pain during labour. Information sharing sessions and advertisements were displayed on the wards to advise potential participants of the study. Eligible participants were identified by registered midwives trained in the trial processes. Once a woman was screened as potentially eligible, a member of the research team, approached her to discuss the study. Information was provided in the ANC for women presenting before labour or on admission to the labour ward during the latent/early-active phase on admission. Study information was presented using a plain-language Participant Information Sheet available in English and local languages as needed. Interpreters and/or a literate independent witness were used where required. Written informed consent was obtained from all participants. Some women were approached during labour but for women experiencing extreme pain or distress at first contact, the discussion was deferred until comfort measures were in place and the woman indicated she was ready to proceed. If readiness was not achieved, recruitment was not pursued. Intervention and procedure Upon presentation to the birth suite with back pain and providing consent women were offered the option of either SWI or standard care. Women who chose SWI received four intradermal injections of 0.1 mL sterile water using a 25-gauge needle [7]. Injections were administered simultaneously by two midwives into the skin bordering the Michaelis rhomboid and followed immediately by two additional simultaneous injections 3 cm below and 1 cm of the first injection (Fig. 1: Anatomical injection sites of Sterile water injection sites). Legend: This figure shows the anatomical sterile water administration sites at the low back, including the first and second injection points. Four intradermal sterile water injection sites are marked (x) over the posterior superior iliac spine (PSIS) region. Two injection points are positioned bilaterally, approximately 1–2 cm medial and inferior to each PSIS. Each cross (x) represents one intradermal injection of 0.1 mL sterile water. The schematic diagram of injection points in Fig. 1 below. Anatomical injection sites of sterile water injection figure (used with permission) The participant was positioned to expose the lower back, and two midwives administered the injections with the woman in a sitting position during a contraction to distract from the discomfort associated with the intervention. Regardless of whether participants chose SWI or not, they all had unrestricted access to standard care consisting of all other forms of natural and medical pain-relieving measures as per the unit’s protocol and according to hospital policy. Outcome measurements Acceptability of medical interventions related to analgesia is a multidimensional construct consisting of both perceived effectiveness (pain relief) and approval as a more global measure of timing, delivery and experience of the procedure [19]. Primary outcome The primary outcome of this study was the change in pain intensity 30 min following SWI or application of standard care. Pain intensity was assessed using VAS in both groups. The primary outcome of acceptability was assessed through the perceived effectiveness after the intervention. Secondary outcomes The main secondary outcome was women’s satisfaction with the intervention, evaluated using a structured postnatal survey administered following birth for only those who received the intervention. Other pain secondary outcomes for the study included the trajectory of pain intensity over time. Pain intensity was recorded using VAS scores at 10, 30, 60, 90 and 120 min after the injection. Additional secondary outcomes were the proportion of women who reported at least a 30% reduction in VAS scores and the proportion who reported at least a 50% reduction in VAS scores at 30 min following SWI administration. Mode of birth was also recorded as a secondary outcome and categorised as spontaneous vaginal birth, instrumental birth (vacuum or forceps), or caesarean section to explore potential associations between labour pain management and labour outcomes. Although SWIs do not directly influence birth mode, improved pain relief may have indirect effects on labour outcomes [8]. Clinical utility was assessed through changes in pain intensity measured using VAS, trajectory of pain over time, and the number of women achieving clinically meaningful pain reduction, alongside consideration of the practicality of administering the intervention within routine clinical care. Finally, local or systemic adverse effects such as localised discomfort, skin irritation, infection or vasovagal symptoms were predefined before data collection. These were assessed using participant self-report and observation by midwives as well as the researchers immediately following the intervention and during follow-up. Data collection We used prospective screening logs and individual case-report forms to capture eligibility, consent and group allocation (SWI vs. standard care). Pain intensity was assessed with a 100 mm VAS (0 = “no pain,” 100 = “worst pain imaginable”). Trained staff provided a brief scripted instruction; women marked current pain on the VAS. A baseline VAS was recorded immediately before SWI (or the parallel time point in controls), and post-intervention VAS ratings were obtained at pre-specified intervals during labour using the same schedule in both groups. In addition to the pain assessments, a structured postnatal survey was administered to women in the intervention group to gather information on their satisfaction with the intervention. The purpose of the survey was to assess the acceptability of SWI amongst participants who had received the injections. Hence, the survey was not provided to women receiving only standard care. The instrument chosen was developed by Lee et al. [6] and adapted for this study. This survey comprised 5 items which was used to assess how satisfied women were with the SWI. The acceptability of SWI among participants was explored through Likert scale and free text questions. The survey tool was reviewed for both face and content validity by all authors and several Ghanaian midwives. There were no adverse effects reported during the study period. Adverse effects were monitored through continuous clinical observation and monitoring as well as self-reports. Routine documentation in the participants’ records of any local or systemic reactions was recorded by attending staff, midwives and researchers. Minimising bias Neither participants nor clinicians (including data assessors) were blinded to SWI, so we minimised performance or detection bias through standardised eligibility, pre-specified outcomes, and uniform data capture [20]. To optimise intervention fidelity and standardise outcome measurement, we used a scripted 100-mm VAS at fixed time points. Co-interventions and protocol deviations were prospectively logged, and case-report forms were completed at the time of data collection. Wherever feasible, the midwife administering the SWI was not the person collecting VAS scores; when separation was not possible, outcome assessors adhered strictly to the standardised script and timing windows. To promote consistency in intervention delivery, midwives received structured training incorporating didactic teaching, an instructional video, and supervised practice. The training included key components relevant to both the intervention delivery (SWI techniques, injection administration), participant recruitment procedures, and standardised data collection processes to ensure consistency in intervention delivery and outcome assessment. Ongoing fidelity was supported through real-time clarification via a moderated messaging channel and periodic file audits with targeted feedback. Sample size We powered the study on a 15% minimally clinically important difference (MCID) in labour back-pain (VAS), with 80% power, α = 0.05 [21]. Although MCIDs for pain vary, literature shows relative MCIDs for acute pain reduction to be about 13% to 15% [22]. We determined the MCID for the study based on the findings of the systematic review by Olsen et al. that suggests an MCID of 15% in pain studies employing multiple measurements anchored on one-dimensional pain scales such as VAS using a mean change approach [22]. A sample size of 32 participants per group along with an anticipated attrition rate of 20%, brought it to a total population of 73 birthing women. Because the primary endpoint was defined as relative VAS change, the sample-size calculation did not require a standard deviation (SD). Data analysis Data were analysed using an intention-to-treat (ITT) approach in IBM SPSS Statistics for Windows, Version 29.0.1.0 (IBM Corp., 2023). A complete case analysis was applied. Demographic and baseline characteristics were summarised using descriptive statistics, with categorical variables presented as frequencies and percentages, and continuous variables summarised as means and standard deviations where normality assumptions were met. Normality was assessed visually using histograms and normal probability plots. The primary and secondary VAS scores were analysed using an adjusted mixed-effects model to compare pain scores between the SWI and standard care group. The mixed-effects model was employed to account for the nested structure of the data, specifying individual participants as random effects. The model was adjusted for parity and onset of labour. We assessed linear and non-linear changes in pain scores over time and evaluated model fit using marginal and conditional R² (random effects). Responses to the postnatal survey regarding the perceived best and worst aspects of the intervention were standardised to a 0–5 Likert scale through linear transformation to ensure consistency across items. This analytic strategy incorporated both continuous and categorical measures while accounting for baseline differences and data structure. Secondary pain outcomes were analysed as clinically meaningful pain reduction, by converting the pre-post changes in VAS scores into binary variables representing ≥ 30% and ≥ 50% reductions in pain. These proportions were compared between groups using Pearson’s Chi-square tests [8]. Secondary labour and birth outcomes were described using standard deviations but were not included in further modelling. Results Recruitment occurred from July 2023 to September 2023. Out of the total 249 eligible women approached, 29 declined participation and 220 provided consent. Low back pain eligibility was then assessed on admission in labour, and 73 (33.18%) women met the specified criteria and formed the study cohort, a proportion comparable to the approximately 30% reported in previous studies, with 51 allocated to the SWI group and 22 to the routine care group. (Fig. 2: Participant flowchart). Legend: This figure shows the recruitment procedures based on eligibility criteria. Rectangular boxes indicate stages of recruitment. Arrows represent participant progression through the study. Participant flowchart Other than spontaneous onset of labour, there were no other significant differences found between the two groups in their baseline clinical characteristics (Table 1: Sociodemographic characteristics of the two groups). Individuals in the intervention group had a higher BMI compared to those in the control group (p = .030). Onset of labour was significantly associated with mode of labour initiation, with a difference observed between spontaneous and induced labour (p = .007). Table 1Sociodemographic characteristics of the two groupsVariableSWI GroupN=51n (%)Control GroupN=22n (%)p-valueMean age (SD)29.7 (6.2)27.2 (5.4)0.114BMI0.030 Normal (18.5–24.9)19.6%40.9% Overweight (25–29.0)39.2%50.0% Obese41.2%9.1%Marital status0.346 Single13 (25.5%)8 (36.4%) Married38 (74.5%)14 (63.6%)Educational status0.241 No formal education2 (3.9%)1 (4.5%) Primary9 (17.6%)8 (36.4%) Secondary16 (31.4%)3 (13.6%) Tertiary24 (47.1%)10 (45.5%)Occupation0.959 Formal employment17 (33.3%)8 (36.4%) Self-employment25 (49.0%)10 (45.5%) Unemployed9 (17.6%)4 (18.2%)Parity0.348 Primiparous42 (82.4%)20 (90.9%) Multiparous9 (17.6%)2 (9.1%)Mean gestation weeks at birth (SD)39.2 (1.2)38.8 (1.2)0.445Onset of labour0.007 Spontaneous51 (100%)19 (86.4%) Induced03 (13.6%)Cervical dilatation on vaginal examMost recent exam0.142 0–4 cm17 (33.3%)5 (22.7%) 5–6 cm26 (51.0%)9 (40.9%) 7–10 cm8 (15.7%)8 (36.4%)Head descent via abdominal palpationaMost recent exam0.17 Less than 1/536 (70.6%)13 (59.1%) 1/5 and above15 (29.4%)9 (40.9%)aAbdominal examination: engagement is considered to have occurred when two-fifths (2/5) of the fetal head are palpable (Norwitz et al., 2002) Sociodemographic characteristics of the two groups aAbdominal examination: engagement is considered to have occurred when two-fifths (2/5) of the fetal head are palpable (Norwitz et al., 2002) Primary outcomes There was no difference between both groups in the pre-injection VAS for back pain, with mean values of 91.19 (SD − 9.03) in the intervention group and 94.95 (SD − 5.15) in the control group (p = .151). The primary outcome was to compare the intervention and control groups based on their VAS scores before and after injection. At the primary outcome time point (30 min), women in the SWI intervention group showed a significantly greater reduction in low back pain on the VAS than those who received standard care (p < .001). Table 2 presents the mean change of both groups at 30 min post injection. Table 2Mean change in VAS pain scores from baseline to 30 min post interventionGroup N Baseline VAS Mean ± SD30 min VAS Mean ± SDMean Change (VAS) ± SDMean Difference (SWI – Control) [95% CI]p-valueSWI Group517.8 ± 1.12.1 ± 1.5−5.7 ± 1.7−5.4 [− 6.3 to − 4.5]< 0.001Control Group227.9 ± 1.27.6 ± 1.0−0.3 ± 0.8-Negative VAS indicates pain reduction. Values represent adjusted means from mixed-effects model controlling for baseline VAS Mean change in VAS pain scores from baseline to 30 min post intervention Negative VAS indicates pain reduction. Values represent adjusted means from mixed-effects model controlling for baseline VAS Secondary outcomes Repeated measures with multilevel mixed modelling Labour onset (spontaneous vs. induced) and parity were adjusted for in the multilevel mixed-effects analysis as they are known confounders that could influence both pain perception and response to interventions during labour [23, 24]. Although BMI was statistically significant when included as a continuous variable, it was not retained in the final model as it did not meet the criteria for clinical relevance for pain [25]. The model’s prediction of pain levels before and after injections showed statistically significant differences across all time points on the VAS score. Across time points, mean VAS scores in the SWI group decreased markedly, with values such as 13.35 (SE 17.48) at 10 min and 2.53 (SE 6.51) at 30 min, while scores in the standard group remained high at the same time points 95.14 (SE 5.49) and 96.18 (SE 5.03), respectively. Similar patterns were observed at later intervals, including 60, 90, and 120 min (Table 3). Table 3Repeated measures of VAS scores over time between SWI intervention and control groups using multilevel mixed modellingTime of VAS scoreSWI Groupn=51mean (SE)Control Groupn=22mean (SE)p-value95% CIPre-injection VAS91.19 (9.03)94.95 (5.15)0.151-8.39–1.29VAS 1013.35 (17.48)95.14 (5.49)0.001-86.42– -76.74VAS 302.53 (6.51)96.18 (5.03)0.001-98.29– -88.60VAS 601.10 (4.42)96.41 (5.02)0.001-100.0– -90.27VAS 902.43 (13.81)96.64 (6.03)0.001-98.88– -88.80VAS 1204.53 (10.59)98.0 (3.86)0.001-98.61– -87.66Adjustment for onset of labour and parity over time Repeated measures of VAS scores over time between SWI intervention and control groups using multilevel mixed modelling Adjustment for onset of labour and parity over time The differences between pre- and post-injection pain levels for women who reported 30% and 50% or more reduction in low back pain across all VAS measurement time points were computed and compared. Results were statistically significant at all the time intervals (Table 4). Table 4Number of women 30% and 50% or greater VAS score reductionTime after injectionSWI Group ≥ 30%N=51n (%)Control Group ≥ 30%N=22n (%)SWI Group ≥ 50%N=51n (%)Control Group ≥ 50%N=22n (%)Missing Data (Combined)*p-value10 min< 0.001Yes51 (100%)047(92.2%)0-30 min< 0.001Yes51 (100%)051 (100%)0-60 min< 0.001Yes49 (96.1%)051 (100%)0SWI 3.9%90 min< 0.001Yes39 (76.5%)039(76.5%)0SWI 21.6%120 min< 0.001Yes35 (68.6%)036(70.6%)0SWI 29.4%; Control 18.2%Missing data: women who were in active labour or had already given birth. Only “Yes” responder data presented Number of women 30% and 50% or greater VAS score reduction Missing data: women who were in active labour or had already given birth. Only “Yes” responder data presented Postnatal survey All participants in the intervention group completed the postnatal survey. Among respondents, satisfaction with SWI was high. Most of the participants selected the highest satisfaction category “very satisfied” 60.8% (n = 31), and the rest indicated they were “satisfied” with SWI intervention. Results in the ‘willingness to reuse’ and ‘recommend to others’ questions were similar. Injection discomfort was identified by some as the least favourable aspect, though this did not appear to affect overall satisfaction of the Likert scale (Table 5). Table 5Maternal satisfactionSurvey QuestionsN = 51n (%)How satisfied or unsatisfied were you with SWI? Very Unsatisfied (1)0 Unsatisfied (2)0 Neither satisfied nor unsatisfied (3)0 Satisfied (4)20 (39.2%) Very satisfied (5)31 (60.8%)Likelihood to use SWI for relief of back pain in labour? Highly unlikely (1)0 Unlikely (2)0 Neither likely nor unlikely (3)1 (2%) Likely (4)25 (49%) Highly likely (5)25 (49%)Will you recommend SWI as a method of back pain relief? Strongly not recommended (1)0 Not recommended (2)0 Neither recommend nor not recommend (3)0 Recommended (4)15 (29.4%) Strongly recommended (5)36 (70.6%)What was the best part (what you liked most) about using SWI? Relief of low back pain41 (80.4%) Reduced low back pain7 (13.7%) Quick Effect3 (5.9%)What was the worst part (what you didn’t like) about using SWI? Stingy pain40 (78.4%) Too many injections7 (13.7%) Pain relief short2 (3.9%) Other pain present2 (3.9%) Maternal satisfaction Labour and birth outcomes The mode of birth did not differ significantly between the intervention and control groups. SVD was the most common mode of birth for both groups, followed by Caesarean Section (CS) largely due to fetal distress. There were no significant differences in maternal or neonatal outcomes observed between the two groups (Table 6). Table 6Labour and birth outcomesMeasured ItemsSWI GroupN=51mean (SD)Control GroupN=22mean (SD)p-valueBirth mode0.385 Spontaneous vaginal delivery (SVD)35 (68.6%)16 (72.7%) Caesarean section (CS)12 (23.5%)4 (18.2%) Vacuum4 (7.8%)0 Missinga02 (9.1%)Duration of active phase of labour (minutes) Mean (SD)544.8 (281.4)481.5 (179.4)0.354Duration of expulsive phase of labour (minutes) Mean (SD)22.61 (21.43)38.60 (20.79)0.006Birth weight0.769 Less than 2.5 kg4 (7.8%)2 (9.1%) 2.5 kg and above47 (92.2%)18 (81.8%) Missinga02 (9.1%)APGAR7.6 (1.42)7.3 (1.26)0.970 Less than 710 (19.6%)4 (18.2%) 7 and higher41 (80.4%)16 (72.7%) Missinga02 (9.1%)Maternal / neonatal complications0.382 Yes8 (15.7%)5 (22.7%) No42 (82.4%)15 (68.2%) Missinga1 (2.0%)2 (9.1%)Breastfeeding initiated within 60 min0.830 Yes32 (62.7%)12 (54.5%) No19 (37.3%)8 (36.3%)aMissing values: Data not available for 2 (9.1%) participants Labour and birth outcomes aMissing values: Data not available for 2 (9.1%) participants Discussion This study provides evidence that SWI is an effective and acceptable non-pharmacological intervention for the management of low back pain during labour, supporting its potential inclusion in maternity care protocols in Ghana and other low- and middle-income countries (LMICs). Women who received SWI experienced significant and clinically meaningful reductions in pain compared with the control group, with analgesic effects occurring immediately and lasting up to two hours. This duration of relief is especially valuable in low-resource settings where continuous or pharmacological pain management options are limited [10, 26]. The reduction in VAS pain scores observed after SWI administration in this study reflect findings from trials conducted in Australia and the United Kingdom, where the proportion of participants reporting pain reductions of ≥ 30% or ≥ 50% were statistically significant following the injections compared to placebo controls [8]. Comparable results were also documented in Sweden, where Mårtensson et al. [27] demonstrated significant relief of labour-related low back pain after SWI. In the trial by Almassinokiani et al. [28], the authors reported that both intradermal and subdermal techniques were more effective than dry needling. In that same study, dry needling was used as the comparator rather than a sham control, which other researchers have opined may have introduced an active analgesic effect in musculoskeletal pain conditions [29, 30], reduced the observed between-group difference, and could lead to an underestimation of the true efficacy of SWI. Our findings indicate that SWI was associated with a rapid reduction in labour-related low back pain, with measurable analgesic effects evident within 30 min of administration. Reductions in pain scores were observed at early post-intervention assessments conducted at 10-minute intervals, supporting a fast onset of analgesic action. This pattern is consistent with findings from studies in high-income settings, which have similarly reported significant pain reduction within the first 10- and 30-minutes following injection [6, 7, 31]. Beyond its early analgesic effect, SWI has practical advantages that are particularly relevant in low-resource maternity care settings. The intervention is inexpensive, requires minimal equipment, and can be administered without specialised training or infrastructure. In contexts such as Ghana, where access to pharmacological labour analgesia remains limited, these characteristics suggest that SWI may be a feasible non-pharmacological option for addressing labour-related back pain when timely pain relief is sought. The intervention was well-accepted by participants, with all women in the SWI group reporting satisfaction; 60% described themselves as “very satisfied,” while the remaining participants reported being “satisfied.” This reflects not only the perceived effectiveness of SWI but also aligns with findings from previous studies in which women expressed high satisfaction with the technique [8, 32–34]. However, in the Ghanaian context, this positive response may carry additional layers of meaning. Labouring women in Ghana often have limited access to effective pain relief, with midwives frequently reporting constraints in offering pharmacological methods [35]. As such, women may experience a heightened sense of relief and gratitude when any form of analgesia is made available, particularly one that provides rapid onset and does not require specialised resources. However, it is also possible that some participants were reluctant to express dissatisfaction with the intervention or study procedures in order to avoid appearing critical of their care providers or assessors, which may have influenced self-reported pain outcomes. No adverse maternal or neonatal outcomes were reported following SWI in this study, consistent with previous reports [26, 33]. Although obesity has been associated with variations in labour and postpartum outcomes [36], BMI did not appear to influence the relationship between SWI and pain reduction in this cohort. Similarly, while induction of labour is often linked to increased pain intensity [23, 37], it did not influence the observed analgesic response to SWI in the present study. Although the intervention was highly acceptable overall, women in the SWI group consistent with findings from earlier studies reported brief but intense discomfort at the injection site. This sensation is well-documented in the literature, where the immediate pain associated with SWI has been cited as a deterrent for some women [38]. The presence of this discomfort highlights the need to prepare women for the transient injection pain and support them through it. Clear counselling, reassurance about the short duration of the discomfort, and empathetic communication from providers may help improve women’s tolerance and overall experience. Addressing the injection discomfort through improved education and supportive care could further enhance uptake and optimise its use in routine labour management. Clinical implications / what this study adds This research is innovative in the Ghanaian context, providing new evidence on the use of SWI and contributing to the growing body of evidence as a non-pharmacological pain management strategy for low back pain during birth. Given the limited local studies on alternative pain management approaches in LMICs, this study fills an important gap in knowledge. In Ghana, access to pharmacological and non-pharmacological pain relief during labour is extremely limited with most common options largely unavailable, and many women report unmanaged pain, fear, and reduced birth satisfaction [39, 40]. Introducing a safe, low-cost, non-pharmacological option such as SWI therefore represents a significant and much-needed contribution to improving intrapartum care. Our study suggests a high degree of effectiveness and acceptability in the Ghanaian healthcare context. Low back pain is a concern for about one-third of birthing women. This study highlights SWI as a safe, low-cost, and easily implementable intervention that can be integrated into routine and standard clinical practice in both primary and tertiary healthcare settings in Ghana. By offering an alternative to non-pharmacological and pharmacological pain management, the study supports efforts to reduce dependence on opioids and other medications, thereby minimising potential side effects and drug-related complications. Strengths/limitations This study has several important strengths. It is, to our knowledge, the first to evaluate the clinical effectiveness through acceptability and outcomes of SWI for labour-related low back pain in a Ghanaian tertiary hospital, thereby addressing a key evidence gap for LMIC settings. The quasi-experimental design was implemented prospectively with pre-specified outcomes, prospective registration, and adherence to STROBE guidance, which strengthens the transparency and rigour of the work. Pain intensity was measured using a validated 100-mm VAS at standardised time points, and outcome assessment followed a scripted protocol to improve consistency. The use of multilevel mixed-effects modelling allowed us to account for repeated measures within individuals and adjust for important confounders such as parity and onset of labour. In addition, the study combined clinical outcomes with women’s reported satisfaction, providing a more holistic assessment of SWI acceptability and practical utility in a busy maternity unit. However, this study has several limitations. The non-randomised, self-selected design introduces potential selection bias, as women opting for SWI may differ systematically in motivation or openness to new interventions, influencing reported pain and satisfaction outcomes. The absence of blinding among participants, clinicians, and most outcome assessors increases the risk of performance and detection bias for subjective measures such as VAS scores and outcome data despite attempts to standardise assessment procedures. It is also possible that some participants were hesitant to report negative perceptions of the intervention to avoid appearing ungrateful or critical of their care providers, which may have influenced self-reported outcomes. Conducting the study in a single tertiary hospital in urban Accra may limit generalisability to smaller or rural facilities with different staffing patterns and pain-management norms. The relatively small control group reduces the precision of effect estimates and limits adjustment for additional confounders. Missing data at later assessment points typically due to advanced labour or birth may have affected longer-term pain trajectories. Finally, satisfaction was assessed only among women receiving SWI, preventing direct comparison with those receiving routine care and limiting interpretation of acceptability outcomes. Conclusion By demonstrating the clinical effectiveness and acceptability as well as potential scalability of SWI, the study provides a transformative approach to pain management in Ghana, particularly in resource-limited settings. These findings support the integration of SWI into national pain management protocols, ultimately improving woman-centred care and broadening the scope of non-pharmacological interventions in the Ghanaian healthcare system. Supplementary Information Supplementary Material 1. Acknowledgements The financial support was provided by an Australian Government Research Training Program (RTP) Scholarship and the Molly Wardaguga Institute as part of the corresponding author’s PhD studies at Charles Darwin University.We gratefully acknowledge all the women who participated in the study and the midwives who supported its implementation. We also thank our research assistants, Destina, Priscilla, Maria, and Gifty, as well as the dedicated research midwives, for their contributions. Special thanks to the Data Safety and Monitoring Board Committee (DSMB) Dr. Naa Barke Armah, Dr Justice Dogbey, Dr Fred Osei and Ms Constance Odonkor for their valuable oversight and guidance. All authors whose works were used have been duly cited. Abbreviations Antenatal Clinic Charles Darwin University CDU Research Data Management Confidence Interval Caesarean Section Clinical Trials Epidural Analgesia Food & Drugs Authority Ghana Human Research Ethics Committee Intention To Treat Korle-Bu Teaching Hospital Low- and Middle-Income countries Minimum Clinically Important Difference National Institute for Care and Health Excellence Pan-African Clinical Trial Registry Randomised Clinical Trials Registered Midwife Relative Risk Sustainable Development Goal Spontaneous vaginal delivery/birth Sterile Water Injection or Injections United Kingdom United States of America Visual Analog Scale World Health Organization Authors’ contributions AM conceived the research idea and led the study. SK, CD, IG, YG, JAA and NL were all principal investigators and contributed to the implementation of the study. YG, AM and NL are responsible for the data analysis plan and undertook the data analysis. AM wrote the first draft of the manuscript with input from SK, CD, IG, YG, JAA and NL. The corresponding author confirms that all individuals listed as authors fulfill the established authorship criteria and that no eligible contributors have been excluded. Funding No funding declared. Data availability The data supporting the results of this study are owned by Charles Darwin University (CDU), Australia and are not publicly available. However, the datasets generated and/or analyzed in this study may be made available from the corresponding author on reasonable request, subject to approval by the university. Declarations Ethics approval and consent to participate Ethical approval for this study was obtained from Charles Darwin University (CDU) Human Research Ethics Committee (CDU-HREC H22055), the Korle-Bu Teaching Hospital (KBTH) Ethics Review Board in Ghana (KBTH-IRB 00170/2022) as well as the Food and Drugs Authority – Ghana (FDA/CT/233) gave approval for this study. Additionally, the study was conducted in accordance with the Declaration of Helsinki and its later amendments. The study was subsequently registered with the Pan-African Clinical Trial Registry (PACTR202307748580298). Written informed consent was obtained from all participants prior to their inclusion in the study. Consent for publication Not applicable. 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 Supplementary Materials Supplementary Material 1. Data Availability Statement The data supporting the results of this study are owned by Charles Darwin University (CDU), Australia and are not publicly available. However, the datasets generated and/or analyzed in this study may be made available from the corresponding author on reasonable request, subject to approval by the university.