Table 1. Summary table of the 16 articles in the review.

PTS-FC: post-traumatic stress following childbirth, AS-FC: acute stress immediately following childbirth

Authors	Purpose	Study design	Study population	Methods	Limitations	Key findings
Campbell et al. (2007) [34]	To investigate the association between support with a doula and the perception by the mother of herself, the infant, and the support of others at 6-8 weeks postpartum	Randomized controlled trial	600 low-risk, nulliparous women	300 with doula support (minimally trained family member or friend), 300 with standard care, and 494 participants completed a second study over the phone of a 42-item questionnaire	There was a significant difference between the groups regarding race, there was no blinding for the research assistant conducting the interviews, recall bias could have affected the responses, and there was a lack of standardized measures of validity and reliability utilized in the study	Doula support by a minimally trained female selected by the mother was an inexpensive way to improve the outcome of both mother and infant.
Mottl-Santiago et al. (2008) [32]	To determine whether hospital-based doula support during labor influenced birth and breastfeeding outcomes	Retrospective cohort study	11,471 full-term women (greater than or equal to 37 weeks gestation) during labor	Tested over the first seven years of the hospital-based doula program, full-term women (greater than or equal to 37 weeks gestation) with and without doula support during labor, regression analyses were used to compare outcomes	Possible experimental group bias utilized a preexisting database, which prevented researchers from collecting certain variables, and limited demographic analysis in regards to distinguishing ethnic group and racial categories	The hospital-based doula support program used during labor had significantly higher breastfeeding rates than those who did not have doula support.
McGrath et al. (2008) [24]	To investigate how nulliparous women with a middle income and a male partner present during labor and birth were perinatally affected by doula support	Randomized controlled trial	420 nulliparous (never given birth), low-risk women in the third trimester	224 women with a doula present from admission through labor and delivery, 196 women without a doula	None noted	The presence during labor of a doula significantly decreased the need for an epidural and the possibility of cesarean delivery, and all women and their partners had positive opinions regarding the presence of a doula during labor and delivery.
Kozhimannil et al. (2014) [30]	To identify women who were able to use doula support and those that did not but desired to use doula support and determine the relationship between the desire of doula services, utilizing doula services, and cesarean delivery with evaluation of non-indicated cesareans	Retrospective cohort study	2,400 women 18-45 who delivered a single infant in the USA from 2011 to 2012	Listening to Mothers III Survey given to women who were a part of Harris Interactive maintained online panels	Self-reported study, Listening to Mothers Survey did not have clinical or diagnostic data due to its self-reporting nature, and a limited sample size	Women with doula support compared to those without doula support and those without support but desired it had lower rates of cesarean delivery overall and non-indicated cesarean delivery.
Everson et al. (2018) [22]	To examine a nationwide sample of the outcomes of both adolescent mothers and neonates who received doula support during birth from 2000 to 2013	Retrospective cohort study	1,892 adolescent mothers (included 1,896 infants, four sets of twins) in the USA between 2000 and 2013 that had doula support present	DONA International birth doula collection form was filled out by the doula and mother; entries included those of adolescent mothers (ages 15-19) between 2000 and 2013	The collection form was not a research design so important covariables were not collected, data entry errors could have occurred as the information was input twice (once by the doula and again by volunteers entering data into the Master file), possible undefined confounding variable of midwifery-led care, and true outcomes need to be gained by a study including a comparison group	The adolescent mothers and their infants in doula-supported births had lower intervention rates and better health outcomes than the national statistics for adolescent births.
Kozhimannil et al. (2013) [16]	To compare the effect on birth outcomes in Medicaid recipients of prenatal education and delivery support from doulas to a national sample to determine cost-saving potential	Retrospective cohort study	279,008 nationwide singleton births on Medicaid, 1,079 singleton births on Medicaid that were doula-supported in Minnesota	Analyzed Medicaid-funded births nationwide and those in Minnesota that were doula-supported using descriptive statistics to determine the effects of doula care and cost deduction associated with decreased cesarean delivery	The doula care only included one state, the nationwide sample and the doula sample were collected during two different periods, the information was collected from discharge reports that could have included under-reporting compilations of doula-supported births, possible selection bias in those Medicaid beneficiaries who chose doula-supported care, and the cost of delivery was estimated	Due to the decreased rate of cesarean deliveries in doula-supported birth and as a result decreased cost of delivery, Medicaid programs should consider covering doula services for birth.
Gruber et al. (2013) [31]	To compare birth outcomes in mothers with and without doula support that are enrolled in the same education program	Prospective cohort study	226 pregnant women enrolled in a healthy moms and healthy babies’ childbirth education class between January 2008 and December 2010	Mothers who attended at least three of the classes were given the choice to work with a doula from the Healthy Beginnings Doula Program, 129 pregnant women without doula support, 97 pregnant women with doula support	Participants self-selected to work with a doula, no information was provided on the mother’s other support systems outside of the doula, and the doula support was not the only service and support provided to the mothers	Mothers receiving doula support had better birth outcomes when compared to those without doula support in regard to low birth weight, complications experienced during birth and delivery, and breastfeeding initiation.
Falconi et al. (2022) [18]	To determine when doula care initiation and with which clinical providers will have the best effect on expectant mothers and determine whether there is a difference in outcomes with a doula based on health status and race/ethnicity	Retrospective cohort study	596 women (298 pairs of women that were matched on age, hospital type, state, socioeconomic status, and race/ethnicity where one received doula support and the other did not)	Collected Medicaid claims from multiple states in the USA between January 1, 2014, and December 31, 2020, to compare the birth outcomes of those women that did and did not have doula support	The extent of doula care was not evaluated, the population of women could not be generalized to the broader US population, and the concordance of race between the doula and the mother was not addressed in its possible influence on health outcomes	Women receiving doula support had lower rates of cesarean delivery and postpartum depression/anxiety diagnosis compared to those without doula support
Hans et al. (2018) [25]	To investigate the effect of doula’s performing home visits on newborn care practices, birth outcomes, and postpartum health of both mother and infant	Randomized controlled trial	312 young pregnant women	Women randomly assigned to doula home services (experimental) or case management services (control), interviews created a baseline then at 37 weeks, three weeks postpartum, and three months postpartum, the participants completed the interview again	The small sample population included only four home visit doula programs in one state, high-risk adolescent women were excluded, and the information was provided by the mother so information on administrative records was not available	Compared to the case management group, the doula home visit group was more likely to attend prenatal education courses, less likely to utilize epidural or other pain medication in labor, and had a higher likelihood of initiating breastfeeding with their infant
Darwin et al. (2017) [28]	To evaluate trained volunteer doula services for disadvantaged women in England	Program evaluation	137 women who received a volunteer doula service before December 2012	Questionnaires (n=136) were done with the help of a researcher or interpreter or were self-completed, individual or group interviews (n=12) were used also	Low response rate of the questionnaire (21.7% of women who used the service) and recipients of the survey could not be contacted from many years prior	Women with the volunteer doula service reported benefits in emotional health and well-being. Most women also viewed the volunteers as a constant help, focused on the mother throughout labor and delivery. An important finding was that the benefits of the volunteer did not depend on the timing of the help, whether it was prior, during, or after birth.
Paterno et al. (2012) [23]	To evaluate a student-nurse doula program specifically in interventions, labor analgesia, and cesarean birth	Secondary analysis	648 records of mothers in the Birth Companions Program	T-tests, chi-squared statistics, and logistic regression models	The data depended on the birth companion’s documentation; the length of labor could not be used as a variable	Epidural use and cesarean birth occurrence decreased with an increase in the number of interventions used by birth companions.
Thomas et al. (2017) [27]	To evaluate the New York City Department of Health and Mental Hygiene’s By My Side Birth Support Program (BMS)	Program evaluation	489 mothers’ data collected by doulas who took part in the By My Side program from 2010 to 2015	BMS analyzed the demographic characteristics, birth outcomes, and follow-up interviews of the women they served from 2010 to 2015	Self-selection bias	The BMS program provided benefits to disadvantaged communities. BMS participants had lower rates of preterm births and low birth weight infants. In follow-up interviews, program participants stated that they would recommend the program to others.
Nommsen-Rivers et al. (2009) [26]	To examine associations between doula care, early breastfeeding outcomes, and breastfeeding duration	Prospective cohort study	Low-income, full gestation primipara receiving doula care (n=44) or standard care (n=97)	The authors collected data from hospital records on birth outcomes and feeding data, follow-up interviews were conducted for lactogenesis onset and breastfeeding behavior	Selection bias, certain results were gained from maternal recall, lack of population generalizability, and the Doula Care Project had a narrow range of eligibility criteria	Doula care resulted in a shorter stage 2 of labor and the onset of lactogenesis within 72 hours postpartum. Doula care patients were 89% more likely to start breastfeeding by six weeks when compared to standard care patients.
Mosley et al. (2021) [29]	To evaluate birth outcomes of a community of refugee women in Georgia	Program evaluation	9,136 hospital clinical records of mothers taking part in the Embrace Refugee Birth Support program from 2016 to 2018	They used bivariate tests to compare the Embrace participants and the comparison group’s descriptive statistics, they used chi-squared tests for categorical predictors and outcomes and T-tests for continuous predictors and outcomes, they also used multivariate analyses	No data on the refugee status of women in the comparison group	The Embrace participants had a 48% lower chance of labor induction when compared to the comparison group. Embrace participants were also less likely to have cesarean deliveries.
Campbell et al. (2006) [33]	To compare labor outcomes between women with doula support and women with no additional support	Randomized controlled trial	600 nulliparous (never given birth) women	They had a control group who were women who did not have doula support and an experimental group who had doula support, the outcome measures were the length of labor, type of delivery, type and timing of analgesia/anesthesia, and Apgar scores	No statistical significance for the type and timing of analgesia/anesthesia and the type of delivery	The length of labor was statistically shorter in the doula support group. The doula group also had higher Apgar scores.
Rousseau et al. (2021) [19]	To determine the risk of developing PTS-FC in women who had prenatal anxiety and if doula support had any effect on PTS-FC development	Longitudinal cohort study	149 low-risk nulliparous pregnant women	The women were split into a doula group and a non-doula group, the mothers self-reported via questionnaires during the last trimester (State-Trait Anxiety Inventory-trait questionnaire), about 48 hours after delivery (Stanford Acute Stress Reaction Questionnaire), and one month postpartum (Posttraumatic Stress Disorder Checklist- Specific Version), the analysis was run on the group to determine PTS-FC development and then compared the doula and non-doula groups to determine the effect of doula support	Analysis of the doula support was secondary instead of primary; doula care itself was not analyzed; it was not a randomized controlled design, which could influence results based on mothers’ feelings toward doulas; small doula group; no clinically diagnostic interviews of the mothers to determine prenatal trait anxiety, AS-FC, and PTS-FC; possible lack of generalizability based on exclusion criteria; and lack of investigation of other vulnerability factors such as state anxiety that could contribute to PTS-FC or AS-FC development	Prenatal trait anxiety and AS-FC are significant risk factors for PTS-FC. Doula support was found to be a possible mitigating factor for PTS-FC development.