Supporting Women in Labor: Analysis of Different Types of Caregivers

Patricia Rosen, CNM, MSN


J Midwifery Womens Health. 2004;49(1) 

In This Article


The search returned 284 articles. Because 276 did not meet the selection criteria, 8 articles remained. Reasons for exclusion included publication before 1980; retrospective or nonrandomized design; no analysis by intention to treat or high attrition rates in primary analysis and no reasons given; no clear explanation of who provided support; intermittent versus continuous support or no clear explanation of level of continuity; no relevance to labor support; publication in a language other than English or only as an abstract; and being a commentary or opinion piece.

The eight published reports provided information about labor support by 1) unfamiliar, untrained lay women; 2) unfamiliar, trained lay women; 3) female relatives; 4) nurses; and 5) monitrices, or lay midwives acting solely as labor support persons. No studies meeting the selection criteria reported on labor support by fathers, lesbian partners, certified nurse-midwives, certified midwives, or physicians. Furthermore, no studies were found that addressed specifically the issue of who does labor support best.

Untrained Lay Women

Two randomized trials[11,12] compared outcomes of women attended by untrained lay women with outcomes of women who did not have these caregivers present during labor. In both studies, unfamiliar women provided emotional support (either friendly conversation[11] or explanation and encouragement[12]) and physical support in the form of rubbing the woman's back and holding her hands. Dependent variables assessed were duration of labor, use of oxytocin, use of analgesia, mode of delivery, and newborn health. One study[11] also assessed maternal-infant bonding behavior during the first 22.5 minutes postpartum.

Effects were greater on women living alone.[12] This type of support does not appear to influence use of analgesia.[11,12] Support by untrained lay women may improve maternal-infant bonding behavior[11] and newborn health,[11,12] but more research is needed to validate these findings ( Table 1 ).

Doulas are supportive companions trained to provide continuous physical, emotional, and informational support to a mother and her family during and immediately after giving birth. Simkin and O'Hara


have noted that the prevailing model of doula care in North America, in which a woman establishes a relationship with a doula prenatally and receives support early in labor, has not been investigated. Two randomized trials


have compared outcomes of women attended by trained doulas with outcomes of women who did not have doulas present during labor. Dependent variables assessed were duration of labor, use of oxytocin, use of epidural analgesia, mode of delivery, maternal satisfaction with the experience, newborn outcomes, and rates of breastfeeding at 1 month (

Table 1

). These trials suggest that continuous support by an unfamiliar but trained lay woman shortens the duration of labor by 1 to 2 hours. Maternal feelings of control over the birth experience and rates of breastfeeding at 1 month may increase with support by a trained lay woman, but one study


does not permit firm conclusions. The findings related to use of medications, mode of delivery, or newborn health demonstrate no consistency and no conclusions can yet be drawn.

Female Relatives

One randomized trial[5] compared outcomes of women attended by a female relative with outcomes of women who did not have a female relative present during labor. Dependent variables included use of oxytocin, use of analgesia, amniotomy to augment labor, and mode of delivery. The results of the trial suggest that continuous support by a female relative may decrease augmentation with oxytocin or amniotomy, lessen the use of analgesia, and increase the rate of unassisted vaginal births. However, because this was only one trial with a small sample size, no firm conclusions can be made ( Table 1 ).


Two randomized trials[15,16] compared outcomes of women attended by nurses trained to provide specialized support with outcomes of women who had usual nursing support. In one study,[15] nurses were trained to include in their care emotional support, physical comfort, instruction for relaxation and coping techniques, partner support, and regular communication with the health care team. In the other study,[16] nurses were trained by a professional labor nurse and doula to include in their care emotional support, physical comfort, information/advice, and advocacy. These aspects of support are similar to those provided by trained doulas. Dependent variables included duration of labor support, use of oxytocin, use of epidural analgesia, continuous electronic fetal monitoring (EFM), perineal trauma, mode of delivery, newborn health, women's perceptions of control during childbirth, postpartum complications and length of stay, postpartum depression, and preferences for labor support ( Table 2 ).

These two studies suggest that continuous support by intrapartum nurses decreases the use of oxytocin and continuous EFM, but has no significant effect on duration of labor, use of epidural analgesia, likelihood of perineal trauma, mode of delivery, newborn health, women's perceptions of control during childbirth, postpartum complications, or postpartum depression. The findings of the larger, multisite trial indicate that most women prefer continuous versus intermittent support by a nurse in childbirth.[16] The authors concluded that in hospitals characterized by high rates of routine interventions, continuous labor support by nurses does not affect the likelihood of cesarean birth or other medical or psychosocial outcomes of labor and birth.


Details of the only randomized trial[17] evaluating support by lay midwives or lay midwives-in-training are summarized in Table 3 . These midwives were self-employed birth attendants, or "monitrices," and did not manage the care of the women they supported. The study examined the effect of this type of caregiver on duration of labor, use of oxytocin, analgesia/anesthesia, and stirrups, mode of delivery, and perineal trauma.

Continuous support by a monitrice appears to decrease the use of analgesia and/or anesthesia, decrease the use of stirrups, increase the likelihood of an intact perineum, and increase the use of oxytocin. Support by a monitrice does not appear to influence mode of delivery. The authors concluded that type of prenatal education, anxiety (trait or state), and commitment to unmedicated labor had little impact on outcomes. The important predictors of outcomes were continuous professional support and expectations of control. Because this was only one trial with a small sample size, these conclusions must be viewed with caution. Furthermore, the results of this study cannot be generalized to practicing midwives.


Chalmers and Wolmer[18] assert that the most impressive, consistent, and methodologically sound results on labor support come from research on untrained lay women. The earliest studies on labor support took place in Guatemala[11,12] and used the term "doula" to describe an untrained female volunteer who supported women by rubbing their backs and holding their hands, talking with them, and being a friendly presence. In these early studies, women customarily labored in a crowded and unfamiliar environment where hospital policy prohibited family members, friends, or continuous nurse caretakers.[18] According to the investigators,[11] these circumstances may have significantly increased maternal anxiety and exaggerated the effect of a supportive companion. The authors concluded that this type of labor support might be useful for low-income, single, young mothers who lack support from family and may have no formal or strong cultural preparation for childbirth.[11]

The emergence of such positive outcomes despite adverse circumstances has several potential explanations. First, the lay women were not part of a hospital hierarchy and so may have been seen as an ally without other interests. Next, these women came from the same community as the laboring women and may have been able to communicate easily and to relate in terms of shared values. Finally, the lay women were told repeatedly to concentrate on comfort, reassurance, and praise. In situations such as this, it is possible that some of the effects of "labor support" result from the hospital's allocation of a person to stay one-to-one with a woman throughout labor, which may convey a message of concern for and value of the woman as an individual.[2] In addition, the mere presence of an observer may influence a provider's behavior and decrease early interventions; likewise, staff may have focused more on the control group, increasing the number of interventions to compensate for their not having a companion.[5]

After publication of the dramatic findings in the first labor support studies in Guatemala, investigators undertook assessment of the effect of labor support in the current labor and birth environment of the United States and on populations of privately insured, middle-class women giving birth in private hospitals. The randomized trial by Kennell et al.[4] affirmed that young, disadvantaged, nulliparous women in crowded units with limited privacy and opportunities for support benefit from the presence of a doula. However, in this study, women were confined to bed as soon as possible after admission and labored in 12-bed wards, where they were allowed no visitors and were surrounded by unfamiliar staff who often did not speak their language. Furthermore, "if, in the judgment of nursing and/or medical staff, the patient was unable to deal with her pain, as evidenced by vocalization, restlessness, or lack of cooperation,"[4] pain medication was used as chemical restraint. As Richards[19] has noted, given these conditions, it is not surprising that a doula could act as "a buffer against the worst excesses of institutionalized obstetrics."

Studies of women from higher socioeconomic backgrounds have not demonstrated such dramatic effects. Privately insured women tend to be better educated, to have taken childbirth education classes, and to be more aware of their childbirth options than the young, less advantaged women in earlier studies.[20] Langer et al.[14] found higher rates of exclusive breastfeeding at 1 month in the supported group, but the intervention included a postpartum visit during which the doula told the mother about the benefits of breastfeeding and how to solve problems she might encounter.

Continuous labor support may have greater benefits for certain groups of women. Women with a higher level of self-esteem and/or a greater internal locus of control may tend to seek out information and support (e.g., taking childbirth education classes or hiring a doula); however, women who are young, unmarried, and of low socioeconomic status, who are less likely to attend childbirth education classes and who have poor social support, may have more anxiety and, therefore, may benefit the most from labor support.[21]

Appropriate care for high-risk women may spill over to the care offered low-risk women in tertiary care centers, leading to less than optimal outcomes for the latter.[22] Thus, the effect of the doulas in the trial by Langer et al.[14] may have been diminished by the strict routine hospital procedures (80% epidural rate and almost universal use of oxytocin). In both RCTs, doulas were hired and trained by the project, and in one trial[14] the doulas were retired nurses. These women may have been restricted by a primary allegiance to the medical establishment or to the study, rather than to the laboring woman. In the case of the retired nurses, they also may have been desensitized to laboring women, making them less effective in their role as a support person. Finally, it is unclear how the timing and duration of support may affect outcomes: almost 80% of participants in the trial by Langer et al.[14] were admitted after 4 cm of dilatation. Hodnett and Osborn[17] suggested that additional support during early labor may enable women to labor longer or more effectively at home; thus, women admitted to this study later in labor may not have had time to benefit from the presence of a labor support person.

Only one RCT has examined the effectiveness of a female relative labor companion on childbirth outcomes.[5] On the basis of the results of this trial, it appears that female relatives may represent an excellent option for women desiring continuous support during labor. Participants were mainly black, young, single students in an unfamiliar, overcrowded hospital environment with limited privacy, restriction of visitors and companions, and multiple caregivers. In addition, the ratio of staff to women was 1:4. Significantly more mothers in the support group had spontaneous vaginal births, less intrapartum analgesia, less oxytocin, and fewer amniotomies to augment labor.[5] The investigators note that findings cannot be generalized due to the small sample size and the setting in Botswana. However, the findings in this study are consistent with those of studies of untrained lay women in Guatemala.[11,12]

Neither RCT conducted on nursing labor support[15,16] found significant differences in maternal or infant outcomes. In fact, the only differences between outcomes for women in the supported and non-supported groups were modest decreases in the use of oxytocin and continuous EFM. Several possible explanations exist for this lack of effect. The authors of the larger, multicenter RCT[16] defined neither "usual care" nor "support," making it difficult to distinguish between treatment in the experimental and control groups. The study did not control for the presence of another support person, such as a partner or friend. Furthermore, unlike most other RCTs of labor support, this study included multiparous women, women delivering before term, and women with twin pregnancies, all of whom may respond differently to labor support compared with low-risk primigravidas at term. Moreover, the trial randomized women as long as second stage was "not imminent."[16] As mentioned previously, additional support during early labor may enable women to labor longer or more effectively at home[17] and thereby avoid interventions. Finally, hospitals eligible for the larger RCT of nurse support had a cesarean birth rate of at least 15% and a 24-hour epidural service.[16] Settings that use highly technological and medically interventionist approaches show muted effects of labor support.

The other RCT[15] of nurse support had methodological flaws as well, including possible contamination (partners were present in 98.6% of the experimental group and 97% of control group) and baseline differences in the two groups (compared with the experimental group, 10% more women in the control group attended prenatal classes).[15] The authors suggest that suboptimum support by the one-to-one nurses may explain the lack of results, even though these nurses were specifically trained in labor support, whereas many intrapartum nurses are not. Perhaps the simplest explanation for the lack of results in this study is that it had insufficient statistical power to answer even the primary question about cesarean rate.[23]

These RCTs[24,25] suggest the possibility that health professionals may not be the best persons to provide labor support. In general, nurses may be hindered not only by lack of time and ineffective staffing models but also by limited educational preparation and by organizational culture.[6] In a medicolegal environment that rewards technical proficiency,[26] many nurses have not been socialized into a supportive labor and birth role.[27] Furthermore, nurses are part of the hospital hierarchy and may be constrained by policies of the organization or norms of the subculture. The use of technology, intervention, and adherence to institutional policies and procedures encourages intermittent presence and does not allow the woman the opportunity to cope with her pain instinctually or independently.[28] Thus, the control that nurses (and other caregivers) assume throughout the childbirth process may be a barrier to providing supportive care.[28]

Few studies have considered the effect on childbirth outcomes of labor support by a midwife or nurse-midwife. Where providers use active management of labor, favorable outcomes, particularly a low cesarean birth rate, have been related to the beneficial effects of continuous one-to-one labor support by a midwife.[29,30] However, no RCT has investigated childbirth outcomes when the midwife is both managing labor and giving continuous support.

The majority of randomized trials included in this review suffer from multiple methodological problems, including bias due to the inability to blind health professionals, the lasting effects of care before randomization, the challenge of ensuring the intervention is applied appropriately (e.g., continuous versus intermittent support, randomization in early versus late stages of labor, and level of support in the early postpartum period), variations in the definitions of training and support, and the possibility of contamination (e.g., presence of additional support persons or increased support of women in the control group). Furthermore, some studies are small[5,31] or have limited statistical power.[14,15] In all studies, obstetric care seems to have been provided by physicians in hospital settings, so results cannot be generalized to out-of-hospital births or births attended by midwives. In addition, no research has been conducted in the US with untrained lay women or female relatives as support persons.

The impact of labor support on rates of interventions is difficult to assess. As noted above, in settings where technological and interventionist approaches are common, the effects of labor support may be muted. The authors of one metanalysis argued that, in settings with high rates of interventions and extensive use of technology, the effect of labor support may be less significant; therefore, to avoid caregiver-related biases, explicit criteria on the indications for major obstetric interventions must be included in design of trials.[21] The authors of another metanalysis[1] concluded the opposite: the presence of a female labor companion may be particularly useful in situations where the intervention rates are high. Clearly, the relationship between support and intervention rates requires further examination.[10]

Today, partners/husbands attend the majority of births in the United States[32] and, along with nurses, provide the majority of labor support.[33] Although in the past, women have rated their intimate partners' presence during labor and birth as important and helpful,[34] the first national US survey of women's childbearing experiences[33] reported that women rank partners or husbands after doulas, midwives, and other family members in terms of the quality of supportive care. The partner's effect on the course of labor and obstetric outcomes has not been assessed adequately[18,35] and requires further investigation.

Ideally, future research into the effectiveness of various types of labor support persons would include both in- and out-of-hospital settings and obstetric care by both physicians and midwives. Support should begin early in labor and should extend at least 1 hour into the postpartum period, because that is a critical time for the initiation of maternal-infant bonding and breastfeeding. The independent variable should be continuous support by: a partner, friend, or relative of the woman's choosing; an untrained lay woman; a lay woman with standardized training; or a person of the woman's choosing and either an untrained lay woman or a trained lay woman. There should be strict criteria about the type and amount of support offered, and none of the support persons should be affiliated with the birth site. Comparisons between support and no-support groups would include outcomes of labor and birth, breastfeeding, and financial cost. In addition, maternal perceptions of and satisfaction with the birth experience, maternal and paternal self-esteem, role adjustment, and infant attachment could be examined between groups.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.