Systematic review of Doppler for detecting intrapartum fetal heart abnormalities and measuring perinatal mortality in low‐ and middle‐income countries

Abstract Background Using Doppler to improve detection of intrapartum fetal heart rate (FHR) abnormalities coupled with appropriate, timely intrapartum care in low‐ and middle‐income countries (LMIC) can save lives. Objective To review studies using Doppler to improve detection of intrapartum FHR abnormalities and intrapartum care quality in LMIC health facilities. Search strategy PubMed, Web of Science, Embase, Global Health, and Scopus were searched from inception to October 2018 by combining terms for Doppler, perinatal outcomes, and FHR monitoring. Selection criteria Selected studies compared Doppler and Pinard stethoscope for detecting/monitoring intrapartum FHR, or described provider and maternal preferences for FHR monitoring in LMIC settings. Data collection and analysis Two team members independently screened and collected data. Risk of bias was assessed by Cochrane EPOC criteria. Results Eleven studies from eight countries were included. Doppler was superior at detecting abnormal intrapartum FHR as compared with Pinard stethoscope, but was not associated with improved perinatal outcomes. Using Doppler on admission helped to accurately measure perinatal deaths occurring after facility admission. Conclusion Studies and program learning are needed to translate improved detection of FHR abnormalities to improved case management in LMICs. Doppler should be used to calculate a facility indicator of intrapartum care quality. PROSPERO registration: CRD42019121924.

Interruption of placental blood flow during labor can result in fetal heart rate (FHR) acceleration, deceleration, bradycardia (<120 beats per minute) and/or tachycardia (>160 bpm). Such FHR abnormalities have been associated with low Apgar score, intrapartum stillbirth, and neonatal death. 5,6 Early detection of FHR abnormalities, linked to timely and appropriate obstetric case management practices, can potentially reduce adverse perinatal outcomes.
A 2017 Cochrane review found that continuous monitoring of FHR by using cardiotocography-the standard of care in highincome countries-was associated with increased numbers of cesarean and assisted deliveries, without a corresponding decrease in adverse newborn outcomes. 7 This may have contributed to the WHO's recommendation to use intermittent FHR monitoring. 8,9 That guidance, however, contains no recommendation of which device (Pinard stethoscope or Doppler) should be used for auscultation 9 ; as a result, many studies have examined the effectiveness of Doppler for intrapartum FHR monitoring in LMIC settings.
The utility of Doppler in the intrapartum care setting is not limited to the diagnosis of fetal heart abnormalities. The importance of an indicator that can be used to track intrapartum deaths in health facilities was noted in a call to action in the Lancet in 2007. 10 Subsequent studies have used Doppler to confirm timing of fetal demise in order to measure stillbirths and neonatal deaths that occur after admission to the health facility.
Maternal preference may increasingly influence which method is used for FHR monitoring in LMIC settings. 9 Some laboring women have noted that hearing the fetal heartbeat amplified by Doppler is a positive experience, and others have reported that the Pinard fetoscope causes discomfort. 11,12 To our knowledge, maternal preferences for the method of FHR monitoring in the LMIC health facility setting have not been systematically described.
The aim of the present systematic review was, therefore, to determine (1) whether Doppler for intrapartum FHR monitoring is associated with a decrease in adverse perinatal outcomes; (2) whether Doppler can be effectively used to calculate a facilitybased indicator of perinatal mortality; and (3) whether women and healthcare providers express a preference for Doppler over Pinard stethoscope for intrapartum FHR monitoring in LMIC settings.

| Search strategy and search terms
The review was registered with PROSPERO (reference CRD42019121924) and followed guidelines detailed in the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. 13 The following databases were searched from inception up until October 31, 2018: PubMed, Web of Science, Embase, Global Health, and Scopus.
The following search terms were used: (Doppler OR fetoscope OR Pinard) AND (newborn OR labor OR labour OR delivery OR perinatal OR intrapartum OR stillbirth OR still birth OR fetal OR foetal OR fetus OR neonatal OR "intermittent fetal heart rate monitoring" OR "fetal heart"). Searches were limited to English and had no date restriction. Both American and UK English spelling was considered in the search terms.
Records retrieved through the searches were imported into Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia) and duplicates were removed automatically. Additional studies were identified by using backward searches (snowballing) of references in relevant articles.

| Inclusion criteria
For inclusion, the studies must have been conducted in a LMIC, assessed an intervention that included Doppler in the intrapartum (not pregnancy) period, have been conducted in a health facility or with health facility staff, have tested use of Doppler to improve the detection of FHR abnormalities to inform intrapartum interventions, address maternal or healthcare provider preference for tools of FHR monitoring during the intrapartum period, or have tested the validity or application of an indicator in which Doppler is used to assess timing of fetal demise. Systematic reviews, case reports, abstracts, and unpublished reports were excluded.

| Data collection and analysis
Titles and abstracts were screened on the basis of the inclusion and exclusion criteria. At this stage, the abstract was perused to assess fit to the given criteria. Studies were selected for inclusion by two researchers (MP, BK), working independently. Disagreements between the two authors were resolved by discussion and review by a third researcher (SW).
After screening, full text versions of eligible studies were examined. Data were extracted by using a pre-defined data extraction form.
Abstracted data included study setting and design, study outcome measures, key findings, summary of limitations, type and characteristics of the intervention, outcome measures, and effect of the intervention on the outcome measures. Qualitative data were described by using textual narrative synthesis, as recommended for systematic reviews. Risk of bias and quality of evidence were assessed by using the Cochrane Effective Practice and Organisation of Care (EPOC) criteria. 14

| Search results and included studies
The initial search yielded 1464 records. After de-duplication, 1463 articles remained. Of these, 1446 articles did not meet the inclusion criteria and the remaining 19 studies were reviewed in full. Of these, 11 studies from Tanzania, Uganda, South Africa, India, Pakistan, Democratic Republic of Congo, Kenya, and Zimbabwe met the inclusion criteria and were included in the review (Fig. 1).
Of the 11 included studies, all but one 15

| FHR abnormalities and adverse perinatal outcomes
Six studies addressed the effectiveness of Doppler versus Pinard stethoscope for the detection of abnormal FHR during intermittent or continuous FHR monitoring in the intrapartum period (Table 1). 15

| Findings on detection of abnormal FHR
All but one study 17

| Findings on adverse perinatal outcomes
Adverse perinatal outcomes were defined as intrapartum stillbirth, neonatal death within 24 hours, neonatal seizures, hypoxic ischemic encephalopathy, bag and mask ventilation, or admission to the neonatal intensive care unit (NICU). Two studies documented a reduction in perinatal adverse events associated with intermittent Doppler monitoring of intrapartum FHR as compared with intermittent monitoring with the Pinard fetoscope. 15,18 In the oldest study, Mahomed et al. 15 reported a reduction of perinatal mortality in the arm using Doppler for intermittent monitoring, with neonatal death rates of 0.6% in the Doppler arm as compared with 2%-3% in the two Pinard arms. No statistical data were presented to demonstrate the significance of the finding.
In a more recent study in Tanzania, among newborns with abnormal intrapartum FHR who were delivered vaginally, lower rates of adverse outcomes (composite of fresh stillbirth, perinatal death, and NICU admission) were seen in the Doppler than in the Pinard arm (16.3% vs 45.3%, P=0.021). 18 In the same study, however, there was no decline in adverse perinatal outcomes when all newborns in the study were considered. In the other four studies, no difference in adverse perinatal outcomes was seen between Doppler and Pinard fetoscope for FHR monitoring (Table 1). [15][16][17]19

| Findings on clinical management associated with abnormal FHR
Multiple studies looked at intrapartum clinical management procedures that would be expected to increase after detection of abnormal FHR and might be associated with a reduction in perinatal mortality.
These measures included cesarean delivery, [15][16][17][18]20 shortening the length of time from abnormal FHR detection to delivery, [15][16][17][18]20 vacuum delivery, NICU admission, and intrauterine resuscitation. 20 Two studies showed a higher rate of cesarean delivery with use of Doppler. In a randomized controlled trial (RCT) in Zimbabwe, the relative risk of cesarean after Doppler monitoring as compared with routine monitoring with Pinard was 1.6 (95% CI, 1.2-2.0). 15 In an observational study in Tanzania, cesarean rates were 5.4% for women with continuous Doppler monitoring, as compared with 2.6% for those with intermittent Pinard monitoring (P<0.001). 19 Other studies in Uganda 16 and Tanzania 17,18 showed no difference in cesarean rates between Doppler and Pinard groups.
In another RCT in Tanzania, an increase in risk of intrauterine resuscitation was observed for women continuously monitored with Doppler as compared with those intermittently monitored with Pinard (RR, 2.07; 95% CI, 1.4-2.9); as described above, there was no difference in adverse perinatal outcomes between the two arms. 20 In Tanzania, two RCTs of intermittent monitoring with Doppler versus intermittent monitoring with Pinard did not find a difference in time from abnormal FHR detection to delivery between the two arms. 17,18 In Zimbabwe, there was no difference in mean duration of labor among the four study groups. 15 The observational study in Tanzania found that continuous FHR monitoring with Doppler was associated with a shorter time from last FHR assessment to delivery (median 45 minutes post-vs 60 minutes pre-intervention, P<0.001). 19 The RCT in Uganda did not report any measure of time associated with clinical management of the women. 16

| Risk of bias and quality of evidence
For the six studies, risk of bias and quality of evidence were assessed by Cochrane EPOC criteria. 14 The most pervasive risk in all of the RCTs was the lack of blinding regarding the device that the participants and study staff used (

| Doppler as a tool for improving measurement of facility perinatal death
Two studies assessed the feasibility and validity of measurements of perinatal mortality in health facilities based on using Doppler to verify the presence or absence of an FHR on admission to labor and delivery services 21,22 (Table 3). A multi-country study was conducted to determine the level of potentially preventable perinatal deaths occurring in study facilities and to describe the feasibility of the measure.  Rates of perinatal mortality occurring in-facility ranged from 4.2% (regional hospital); 1.5%-2.7% (district hospitals); and 0.3%-0.5% (health centers); Use of Doppler on admission and recording the FHR in the register produced a more specific measure as compared with crude perinatal death rate, which included macerated stillbirth and was thus less reflective of quality of intrapartum care

| Healthcare provider and maternal preferences for Doppler versus Pinard stethoscope
Abbreviations: DRC, Democratic Republic of Congo; FHR, fetal heart rate.
T A B L E 4 Studies related to healthcare provider or maternal preference for Doppler vs Pinard. adequate education on Doppler for healthcare providers when introducing the device into pre-service and/or professional training.

Study
All three studies had notable limitations that lessened the generalizability of results. The Tanzanian RCT was conducted with relatively few midwives from one health facility, and reflected device use based on random assignment rather than on provider preference. 11 The South African study, which compared maternal preferences among Doppler, Pinard and cardiotocography, did not test FHR monitoring throughout labor, but rather at a single point during the first stage of labor. 23 In addition, the authors did not address the potential effects of being in active labor while giving feedback, nor did they describe the information that they provided to participants about the efficacy of the devices for FHR monitoring. Lastly, the study did not provide statistics to test significance of the findings.
The qualitative study from Tanzania, which assessed women's perceptions on Doppler for continuous monitoring of FHR during labor, reflected views from women who attended services at one facility and included only women who had healthy newborns. 11 Interviews were conducted before discharge from the facility, which might have affected the women's openness to answer questions honestly.

| DISCUSSION
An estimated 1 million neonatal deaths and half of all maternal deaths might be prevented with higher quality maternal and newborn care. 25 Lack of intrapartum monitoring of FHR according to standards contributes to persistently high levels of perinatal and neonatal death in LMIC. 2,26 Although assessment of the fetus at the time of admission to labor and delivery services is supposed to be routine, 27

| Doppler and perinatal mortality
Except in one instance, 19 none of the reviewed studies reported a reduction of perinatal mortality associated with use of Doppler for FHR monitoring as compared with Pinard fetoscope. This finding echoes that of a broader systematic review of intrapartum fetal surveillance in LMIC. 29 In multiple studies where Doppler was used for FHR monitoring, [15][16][17][18]20  Continuous monitoring of FHR has been associated with an increase in cesarean delivery, which may not benefit the mother. 8 Given WHO guidance cautioning about potential overuse of cesarean in LMIC, 31 any quality improvement work that introduces Doppler, particularly continuous monitoring, should also monitor potential overuse of this intervention.

| Doppler to improve measurement of facility perinatal mortality
The WHO has called for a metric for perinatal mortality occurring after admission to a health facility that can be used to monitor quality of intrapartum care. 11,32 In two studies in five countries, Doppler was used to detect FHR among women on admission, allowing for verification of whether fetal deaths occurred before or after facility admission. This information is useful to calculate an indicator of perinatal mortality that occurs in a health facility (i.e., the mother was admitted to the facility with a documented FHR and was discharged with a stillborn or deceased newborn). It can be presumed that many of these cases represent poor quality of care.
Both of the studies concluded that such a facility perinatal mortality indicator is a feasible and useful measurement 21,22 ; one study also noted the feasibility of integrating the indicator into the national health information system. 22 Despite the small number of studies, the findings support increased use of Doppler to accurately measure preventable perinatal death (intrapartum stillbirth and early neonatal death) occurring after admission to labor and delivery services in LMIC health facilities.
Further studies might address the feasibility of integrating the indicator into health information management systems, provider acceptance of the indicator, costs associated with scaling up Doppler use, and national policy-makers' understanding of the need for the indicator.

| Healthcare provider and maternal preference for Doppler as a means of FHR monitoring
The WHO considers maternal and healthcare provider preferences to be key elements for a positive childbirth experience, 9 in addition to the importance of the woman having informed choices regarding interventions in labor. 27 A strong maternal or healthcare provider preference for Doppler over Pinard may be sufficient to justify integrating the device into LMIC intrapartum care protocols. Three studies addressed healthcare provider and maternal preference for Doppler as compared with other devices for monitoring FHR. All three had substantial limitations regarding generalizability that restricts their utility in drawing programmatic or policy conclusions. The current evidence on maternal and provider preferences should be bolstered with studies that have greater generalizability and include the perspectives of women who experienced deliveries with fetal distress.

| Limitations
The review has some limitations. First, the findings rely on the quality of included studies. All studies that examined adverse perinatal outcomes were designed with perinatal outcomes as a secondary outcome measure, and hence had relatively low power to detect these differences. Second, two studies indicated that, although FHR monitoring protocols were properly followed due to study oversight, there were delays in proper case management, impacting perinatal death rates. 16,17 Third, the review did not include a meta-analysis owing to dissimilarity of interventions and outcome measures among the studies. Last, none of the included studies addressed the feasibility of scaling-up use of Doppler, which would require an assessment of infrastructure-related needs such as power, ultrasound gel, and maintenance, and which will ultimately be an important consideration in Doppler scale-up In LMIC.

| CONCLUSIONS
On the basis of the reviewed studies, it is reasonable to conclude that Doppler may be a better diagnostic tool than Pinard fetoscope for monitoring FHR in the LMIC facility setting. In all but a few cases, the studies that assessed interim measures of clinical management (i.e., cesarean delivery, intrauterine resuscitation, and time from detection of abnormal FHR to delivery) showed that these interventions were the same in the

CONFLICTS OF INTEREST
The authors have no conflicts of interest.