Figure 2 near here. Flow chart showing the evolution of the project
Legend to Figure 2. A Flow chart which describes the timelines of the various interventions undertaken up to the completion of this manuscript. The project meanwhile continues in both hospitals.
Consent and monitoring (see flow chart Figure 3)
Figure 3 near here. Flow chart showing consent and comments for this initiative
Over 15 months from 31 July 2017 until 24 October 2018, 474 women (157 from CH Rennie and 317 from CB Dunbar Hospitals) admitted in the first or second stages of labour and without obstetric complications were approached. 461 gave their informed consent to participate in the FHR monitoring (97%) and 13 declined.
CH Rennie hospital data
At CH Rennie Hospital, data were collected from 21 March 2018 until 24 October 2018. 157 mothers were approached for consent. Three declined [case number 6 (too much pain); case number 54 (no reason); and case number 75 (no reason). In the 3 women who declined, only fetal monitoring as part of the partograph, continued to be undertaken.
Three (case numbers CHR 4, 10 and 153) had consented but, after monitoring for 12, 8 and 36 contractions respectively, stopped monitoring because of pain and tiredness.
Fetal death was identified on admission during training in the use of the sonicaid in one mother (case CHR 50 in Table 4: see below).
CB Dunbar hospital data
At CB Dunbar Hospital, data were collected from 31 July 2017 until 23 October 2018. 317 mothers were approached for consent. 307 gave their informed consent and 10 declined. In 4 of these cases, where the mother said her pain was too much to allow her to do it, with the woman’s permission, the obstetric clinician and/or midwife continued to undertake the monitoring with the sonicaid. In one of these latter 4 cases (CBD 251) a student midwife undertook all of the monitoring and identified a change in FHR (see Table 4). In the other 6 cases, no FHR monitoring was undertaken except as part of the partograph.
Out of the 307 mothers who consented, there were 7 cases where the mothers initially declined but then changed their minds and undertook monitoring for the rest of their labour. In one of these cases, the mother shared the monitoring with her midwife and obstetric clinician.
In an additional 7 cases, the mother consented but subsequently stopped monitoring (1 through tiredness, 1 because of pain, 4 with no reasons given and 1 because she was worried about delivering on time).
In an additional 20 cases where the mother consented but then stopped monitoring, a midwife or obstetric clinician took over (19 cases) until delivery or shared the monitoring with the mother (1 case). Of these 20 cases, 15 gave pain and 5 gave tiredness or weakness as the reasons.
425 of 461 (92%) participating mothers were able to complete the monitoring themselves (151 from CH Rennie Hospital and 274 from CB Dunbar Hospital).
Maternal age
Maternal age was available for 416 of 461 participants. 48 (12%) were aged under 18 years; 22 mothers aged 17; 17 aged 16; 5 aged 15; 3 aged 14; and one aged 13 years.
Maternal experiences (including evidence regarding the consequences of the lack of any pain control, which was not available during labour in any of the public hospitals in Liberia).
400 participants provided written or verbal (transcribed) comments on their experiences of the monitoring.
387 mothers found listening to their unborn baby a positive experience expressing one or more of the following words or phrases: alright, not bad, good, fine, helpful, loved or liked it, happy, comfortable, gives me joy, or other positive comments such as “Thank you”. A selection of these comments is shown in Table 3, including some from mothers who identified changes in FHR. The complete set of comments from all participants are available in the supplementary information (additional files).
13 participants reported only negative comments: 5 reported weakness or tiredness (including feeling nauseated in one case), 6 reported pain (which in 3 interfered with the monitoring), 1 said it was not easy and 1 said it was bad.
Table 3 (near here) Selected maternal comments from mothers at both Hospitals (additional maternal comments following changes in fetal heart rate are reported in Tables 4 and 6 below).
Out of the 387 providing positive comments, 86 also reported how much they were affected by pain or severe pain. This pain interfered with their ability to undertake the monitoring in 52 of the 86 (60%).
Within the 387 with positive comments, 44 women also reported discomfort, 5 reported tiredness and weakness and 4 reported difficulties applying the sonicaid.
Three mothers said that they would hope that monitoring would be available during their future pregnancies. 6 said they would return to the hospitals to deliver as a result of the project. An additional 6 mothers said they would hope that monitoring will continue to be available so that other mothers can benefit and 5 said they would encourage other mothers to participate. 9 participants said monitoring helped to cope with labour pain.
Technical and administrative problems identified
As the forms were to be completed by busy midwives and obstetric clinicians, not dedicated researchers, forms were designed to record the most relevant information for the purposes of the study objectives rather than detailed information which, although useful, was not necessary for the purposes of this initiative and could have taken the health workers away from their clinical work.
The initial design of the tick sheet documenting each contraction monitored did not always allow enough space to record every contraction and obtaining extension sheets was sometimes a logistic problem. To minimise the workload of the scarce midwifery workforce, forms were also re-designed to record clinical data that were appropriate but not excessive given time restraints.
Due to a communication problem, 57 mothers at CB Dunbar Hospital (between 19 Feb 2018 and 8 May 2018) incorrectly monitored their FHR every 30 minutes (similar to the partograph). However, unlike the partograph, monitoring was always undertaken for approximately 60 seconds immediately following the nearest contraction to each 30-minute window. We did not identify any clinical differences or differences in satisfaction between the 57 women who monitored the fetal heart every 30 minutes and the women who monitored after every contraction.
Birth/delivery data
In 461 participants, there were 33 caesarean sections (7.2%) including 14 with FHR changes. There were 20 vacuum deliveries (4.3%) including 8 with FHR changes, (the latter included a mother with a stillborn baby identified during the training in the use of the sonicaid, confirmed by ultrasound scan).
Clinical information and outcomes for participants where changes in FHR were identified.
Table 4 near here. Clinical information and outcomes where FHR changes were identified by monitoring.
Table 4 describes the clinical information and outcomes relating to identified FHR changes. Changes in FHR were reported in 28 of 461 participants (6.1%,) which in two cases were not confirmed by the attending midwife, giving 26 confirmed cases (5.6%). In 23 of the 26 confirmed cases, the FHR decreased and in 3 the FHR increased. Two changes related to unrecognized obstetric complications, with one mother found to have Bandl’s ring with obstructed labour and the other, cord prolapse.
All 26 neonates with changes identified in FHR survived, including 13 requiring resuscitation at birth. Twelve neonates were admitted to the neonatal unit. One neonate developed birth asphyxia/HIE (case CHR 99) but the immediate clinical outcomes of the 25 other neonates were good and all 26 were discharged home apparently well.
One of the 26 changes in FHR was identified by a midwife who had taken-over monitoring from a mother who became too tired to continue. In a second case, the mother had declined to undertake monitoring herself but had consented to the monitoring being undertaken by a student midwife.
In 17 of the 26 with confirmed FHR changes plus 1 intrauterine fetal death detected on admission , there was accompanying meconium-stained liquor.
13 of the 26 (50%) neonates with prior FHR changes had low Apgar scores and needed resuscitation. There were no deaths following resuscitation. One baby (case CHR 99) had convulsions managed with Phenobarbital, recovered and was feeding normally at discharge home aged 7 days. None of the other 26 neonates developed birth asphyxia (also known as Hypoxic Ischaemic encephalopathy-HIE). However, long-term infant follow-up was not undertaken.
8 of the 26 neonates with confirmed FHR changes plus 1 with intrauterine fetal death (30%) were born by vacuum, 14 by Caesarean Section (CS) (54%), and 6 (23%) by vaginal delivery. In one case (CBD 272), CS followed a failed vacuum delivery.
Clinical information and outcomes in 3 newborn infants needing resuscitation at birth where mothers had declined to participate in FHR monitoring.
TABLE 5 (near here). Clinical information and outcomes in newborn infants needing resuscitation at birth where mothers had declined to participate in FHR monitoring.
In one such case, the baby was born with Apgar scores of 2 at 1 minute and 3 at 5 minutes and, despite resuscitation, died of HIE in the neonatal unit aged 2 days. In 2 other cases, there were low Apgar scores at 1 and 5 minutes (4 and 6; and 2 and 6) and the babies needed resuscitation. Both were admitted to the neonatal unit. One responded well to resuscitation with no evidence of HIE and was discharged home well. The other died aged 3 days from birth asphyxia/HIE.
Clinical information and outcomes in 8 neonates needing resuscitation where no FHR changes had been identified
TABLE 6 (near here). Clinical information and outcomes in newborn infants needing resuscitation at birth where monitoring had not identified any FHR changes.
In addition to the neonates requiring resuscitation where FHR changes had been detected, 8 other neonates without detected changes in the FHR required resuscitation (Table 6). One (case CBD 238) was preterm/low-birth weight and died at birth. None of the remaining 7 developed birth asphyxia/HIE.
In one neonate (Case number CBD 224), it was unclear who had undertaken the monitoring and for how long. A vacuum delivery was undertaken for failure to push, Apgar scores were 5 and 8 and the baby required 10 minutes of bag and mask ventilation. He was discharged home aged 7 days well. No evidence of birth asphyxia/HIE was evident on clinical assessment.
Three cases were born following vacuum delivery and 5 cases by vaginal delivery, including in one mother where an episiotomy was undertaken to expedite delivery (case CBD 240).
Three of the neonates had been monitored in utero only every 30 minutes in a temporary deviation to the protocol because of a communication problem with one of the trainee obstetric clinicians but none developed birth asphyxia/HIE.
Costs associated with collecting data
The costs of the project were low. The fetal doppler monitors (Sonicaids: 12 in total) were USD 40 each. Rechargeable AA batteries were used. Additional costs included paper and printing for the consent, data collection and monitoring forms including the internet costs of scanning and sending them to MCAI for analysis and KY jelly (or locally available clear hair gel) for interfacing the ultrasound probe with the abdomen: commercial ultrasound gel was too expensive.
Missing data
Because of problems with the completion of medical records and the work pressure on the health workers involved, it was sometimes difficult to fill the gaps of any missing information, such as birth weights, retrospectively. Every effort was made by the management committee to minimise missing data, especially regarding maternal and neonatal outcomes.