The purpose of this study was to evaluate the efficacy of a guided imagery (GI) intervention for stress reduction in pregnant African American women beginning early in the second trimester. This prospective longitudinal study of 72 women used a randomized controlled experimental design with two groups conducted over 12 weeks. The intervention was a CD with 4 professionally recorded tracts designed and sequenced to influence study variables. Participants in both GI and usual care (UC) completed measures and donated 5 cc of blood at baseline, 8 weeks and 12 weeks. Participants also completed a daily stress scale. A mixed-effects linear model tested for differences between groups for self-reported measures of stress, anxiety, and fatigue as well as corticotrophin releasing hormone (CRH), a biologic marker of stress. Significant differences in perceived stress daily scores and at week 8 but not week 12 were found in the GI group compared to UC group. The GI group reported significantly less fatigue and anxiety than the UC group at week 8 but not week 12. There were no significant differences in CRH levels between groups. Results suggest that GI intervention may be effective in reducing perceived stress, anxiety, and fatigue measures among pregnant African American women.
Maternal stress has been associated with pregnancy complications such as hypertension and preeclampsia as well as negative perinatal outcomes such as intrauterine growth restricted (IUGR), low birth weight (LBW) infants, preterm birth (PTB), and neuropsychological developmental delays of affected offspring [
Of equal importance is the major disparity in rates of stress and pregnancy complications between African American women and Caucasian or Hispanic women [
The pathway from maternal stress to negative health outcomes is a multifaceted and complex process which represents the connection between the mind and the body [
Given the mind-body connection between maternal stress and adverse health events, it is possible that a complementary and alternative medicine (CAM) intervention may be effective in reducing self-reported and biologic measures of stress during pregnancy [
GI is a psychophysiological dynamic modality in which a person imagines and experiences an internal reality in the absence of external stimuli [
The use of GI has been reported to reduce self-reported measures of stress, anxiety, and fatigue as well as neuroendocrine measures of stress such as cortisol among nonpregnant participants [
Hospitalized pregnant women with preterm labor were found to benefit from a GI intervention. In Taiwan, this population was found to have lower state anxiety scores but not perceived stress scores compared to control group immediately after 3 sessions and over time until delivery [
Not all studies have reported a reduction in anxiety. For example, the use of imagery during prenatal classes reported no significant differences in anxiety state [
Because imagery links perceptions, emotions, and physiological responses, the selection of appropriate images is of utmost importance [
While the results of these studies are promising, variations in research design, dosage of intervention, and selected study variables limit the applicability of these findings. Additional research addressing these limitations is warranted. Therefore, the purpose of this study was to investigate the effects of a 12-week professionally developed GI intervention on self-reported and biologic measures of perceived stress and the associated symptoms of anxiety and fatigue in pregnant African American women.
This study used a repeated measures, two-group randomized experimental design to examine the effect of GI on perceived stress, and the associated symptoms of anxiety and fatigue, as well as CRH, a neuroendocrine stress measure.
Participants were recruited from 2 academic obstetric clinics affiliated with the Virginia Commonwealth University Health System and the Riverside Regional Health System. Inclusion criteria included being (1) pregnant African American women between 14 and 17 weeks gestation, (2) ≥18 years of age or older, (3) able to read, write, and understand English, and (4) able to verbalize a source of social support. Exclusion criteria included (1) multiple pregnancy, (2) cervical cerclage, (3) current use of oral corticosteroids, (4) uterine or cervical abnormality, (5) dissociative disorders, borderline personalities, and psychotic pathology, (6) medical and/or pregnancy complications, and (7) current use of GI techniques. Criteria were selected to reduce sources of variability between and within subjects to include (1) variables that could potentially influence CRH levels such as ethnicity, development and growth of the placenta, gestational age of the pregnancy, use of glucocorticoids, and medical and/or pregnancy complications potentially associated with infection/inflammation [
A total of 148 potential participants were assessed for eligibility. Of these 148 women, 74 did not meet inclusion criteria, 2 were eligible but declined to participate, and 72 pregnant African American women between 14 and 17 weeks gestation were enrolled after completing the informed consent process (Figure
Flow diagram of enrollment and randomization.
The 12-week intervention consisted of a CD with 4 GI tracts, each 20 minutes in length. The script for each CD was developed and professionally recorded by the author certified in GI. Key components included relaxation, focused breathing, and a variety of multisensory images to promote reduction of stress and anxiety as well as to restore levels of energy. The intervention contained multiple styles of imagery to include (1) feeling state imagery designed to shift the participant’s mood to one of peace and calm, (2) end state imagery suggesting that participants see themselves the way they wish to be, and (3) energetic imagery focusing on restoring normal levels of energy [
The content and order of the CDs were sequenced to focus on the desired outcomes and were based on examples in the professional literature and the Academy for Guided Imagery [
The R-GI group continued usual care with the addition of the GI intervention. Participants were instructed to listen to the CD once a day for 12 weeks in a recommended order for weeks 1–4 and used in any order for weeks 5–12. This plan was based on a study reporting that some women got tired of hearing the same script after three weeks [
Participants in the UC group continued their usual obstetric care. To reduce the possible effect of researcher attention, a member of the researcher team called participants every week to ask if they have completed the NRSS measures for that week.
A self-reported demographic and health history questionnaire was developed for this study to assess participants’ basic demographic information as well as medical, mental, and/or obstetric risk factors and current health behaviors. Estimated gestational age (EGA) was determined by the first day of the last menstrual period and, if possible, confirmed by ultrasound [
This is a 10-item, self-reported measure of global perceived stress that measures the degree to which a respondent appraises one’s life as being stressful during the past month [
This is a unidimensional scale developed for the purpose of quantifying the intensity of stress [
This is a self-report measure consisting of two 20-item subscales measuring state anxiety and trait anxiety [
This is a 9-item self-report measure that assesses the severity of fatigue and interference on daily functioning over the past 24 hours [
Frequency of GI CD tract use, barriers to use, and perceived benefits of the intervention were assessed by participants’ documentation on a daily practice log developed for this study.
Venous blood samples were drawn from a subset of participants using standard venipuncture procedures in the clinic during a prenatal care visit. Blood samples were collected into tubes containing the anticoagulant ethylenediaminetetra-acetic acid (EDTA) and the protease inhibitor aprotinin (0.6 TIU/mL of blood) to prevent degradation of the sample [
Eligible women who consented and agreed to participate were randomized using a computer generated masked assignment list using envelopes to the GI or UC group. At time 1, participants completed the demographic and health history questionnaire and the PSS, STAI, and BFI. Blood was drawn for plasma CRH. Folders containing the daily logs, instructions for completion of the NRSS, and contact information were given to all participants. Participants in the GI group received the CD, CD player, and extra batteries and were individually introduced to GI by a member of the research team.
Approximately 8 weeks (EGA 22–25 weeks) from the initial visit, a member of the research team met with the participants at the prenatal visit. Participants completed the PSS, STAI, and BFI and had blood drawn for plasma CRH. The logs from the previous 8 weeks were collected.
Approximately 4 weeks from time 2 (EGA 26–29 weeks), participants completed the PSS, STAI, and BFI and had blood drawn for plasma CRH. The logs from the previous 4 weeks were collected.
The NRSS was completed daily. The UC group was instructed to complete it in the evening and the GI group before and after listening to the CD. In addition, the GI participants were asked to record any perceived benefits or barriers to using the CD on the daily practice log. The research team telephoned all participants weekly to inquire about their wellbeing and any difficulties in completing the NRSS forms.
Descriptive statistics were used to summarize demographic characteristics. At baseline (Time 1), groups were compared using two-sample
A mixed-effects linear model was used to test for differences between the GI intervention group and the UC group across time for the behavioral variables (PSS, STAI, and BFI) and biologic variable (CRH). Following the intent to treat principle, all participants with postbaseline data were used in the mixed effects linear model. The fixed effects included visit number (baseline (visit 1), visit 2, and visit 3), intervention group (GI and UC), and visit by intervention group interaction. Participant was modeled as a random effect. CRH was positively skewed; thus, a log-transformation served the dual role of normalizing the data and stabilizing the variance.
NRSS scores were obtained daily from each participant. In the GI group, two NRSS sores were obtained (prior to and after listening to the GI intervention) and one NRSS score in UC group was obtained. These daily data were averaged to obtain weekly means (weeks 1–12). Thus, each subject had as many as 12 weekly averages. Two different questions are posed regarding these data. The first is “do the weekly means differ from before to after intervention for the subjects within the GI group?” The mixed-effects linear model used to test this question contained fixed effects for week (1–12), time (pre- or post-CD), and a week by time interaction and a random effect for participant. In order to model the doubly repeated measures (pre/post and across 12 weeks), an unstructured variance structure was used to model the repeated time (pre/post) and a compound symmetric variance structure was used to model the repeated week (1–12). The second question asked of the daily data was “do the post-GI group means differ from the UC group means across the weeks?” The data used to answer this question was the post-GI data and the UC group data. The mixed-effect linear model used to test this question contained fixed effects for week (1–12), intervention group (GI versus UC), and a week by intervention group interaction and a random effect for participant.
Because this was a pilot study, no adjustment for multiplicity was used; the results from this trial should be regarded as exploratory and should be confirmed in a larger trial. JMP and SAS were used for all data analyses.
The study protocol was approved by the Institutional Review Boards (IRB) of Virginia Commonwealth University and Riverside Regional Medical Center.
Participants ages ranged from 18 to 39 years with a mean age of 24.26 years (1.06) (Table
Demographic variables by group (
Variable | GI |
UC |
Total |
|
---|---|---|---|---|
Age | 24.75 (1.06) | 23.78 (0.73) | 24.26 (0.64) | 0.454 |
Height | 5.31 (0.05) | 5.38 (0.04) | 5.35 (0.03) | 0.315 |
Weight | 176.56 (6.20) | 171.49 (7.15) | 174.03 (4.71) | 0.594 |
Number of pregnancies | 2.89 (0.31) | 2.37 (0.30) | 2.63 (0.22) | 0.236 |
Weeks pregnant | 15.53 (0.16) | 15.32 (0.22) | 15.43 (0.13) | 0.446 |
|
||||
% ( |
% ( |
% ( |
||
|
||||
Income | ||||
Less than 15,000 | 72% (26) | 64% (23) | 68% ( |
0.617 |
Between 15,000 and 44,999 | 25% (9) | 28% (10) | 26% ( |
|
More than 45,000 | 3% (1) | 8% (3) | 5% ( |
|
Marital status | ||||
Married/married before | 17% ( |
11% ( |
14% ( |
0.496 |
Single never married | 83% |
89% ( |
86% ( |
|
Employment | ||||
Full-/part-time job | 39% ( |
39% ( |
39% ( |
0.924 |
No Job | 58% ( |
61% ( |
60% ( |
|
Education | ||||
High school or less | 47% ( |
50% ( |
49% ( |
0.896 |
Posthigh school | 36% |
39% ( |
37% ( |
|
Postcollege | 17% ( |
11% ( |
14% ( |
|
Drink alcohol | ||||
No | 86% |
97% ( |
92% ( |
0.199 |
Yes | 14% ( |
3% ( |
8% ( |
|
Smoking | ||||
No | 75% ( |
89% ( |
82% ( |
0.220 |
Yes | 25% ( |
11% ( |
18% ( |
|
Beverages w/caffeine | ||||
No | 22% ( |
36% ( |
29% ( |
0.195 |
Yes | 78% ( |
64% ( |
71% ( |
|
Use stress management | ||||
No | 80% |
89% ( |
85% ( |
0.343 |
Yes | 20% ( |
11% ( |
15% ( |
The mean PSS scores for the GI and UC groups for each time point are presented in Table
Least square means and standard errors from mixed linear model analysis comparing GI and UC groups.
Measurement | GI |
UC |
|
---|---|---|---|
PSS | |||
Baseline | 20.30 (1.41) | 19.23 (1.36) | |
Time 2 | 16.89 (1.42) | 19.99 (1.37) | 0.012 |
Time 3 | 17.11 (1.42) | 17.48 (1.37) | 0.381 |
STAI-S | |||
Baseline | 39.59 (2.32) | 39.35 (2.24) | |
Time 2 | 36.42 (2.38) | 39.10 (2.27) | 0.382 |
Time 3 | 36.38 (2.35) | 34.44 (2.30) | 0.606 |
STAI-T | |||
Baseline | 43.79 (2.32) | 43.14 (2.26) | |
Time 2 | 39.69 (2.36) | 44.45 (2.26) | 0.041 |
Time 3 | 39.11 (2.34) | 39.35 (2.28) | 0.735 |
BFI Total | |||
Baseline | 5.26 (0.43) | 4.55 (0.42) | |
Time 2 | 4.04 (0.44) | 4.59 (0.42) | 0.040 |
Time 3 | 3.60 (0.44) | 3.87 (0.43) | 0.104 |
|
|||
Median (minimum and maximum) |
Median (minimum and maximum) |
||
|
|||
CRH |
|||
Baseline | 0.97 (0.03, 5.75) | 0.85 (0.03, 9.49) | |
Time 2 | 8.80 (1.06, 47.32) | 8.30 (3.35, 85.29) |
|
Time 3 | 36.47 (4.90, 87.17) | 45.14 (2.78, 146.87) |
|
Differences in perceived stress scores over time between the GI and UC group.
The first question of the NRSS variable was “do the weekly means differ from before to after intervention for the subjects within the GI group?” From the mixed model analysis, time (pre- versus post-CD) was statistically significant
Pre- and postintervention numeric rating scale of stress scores over 12 weeks in the GI group.
Differences in UC and GI postintervention numeric rating scale of stress scores over 12 weeks.
The mean STAI scores for the GI and UC groups for each time point are presented in Table
The mean BFI scores for the GI and UC groups for each time point are presented in Table
Differences in Brief Fatigue Inventory Scores over time between the GI and UC group.
Due to laboratory assay problems beyond our control, the plasma of only 31 participants (GI = 16; UC = 15) were used for the CRH analysis. The median (pg/mL) and range for CRH levels are presented in Table
The frequencies of CD tract usage from week 5–12 are presented in Table
Frequencies of GI CD tract usage from week 5 to week 12.
GI CD tract | Week 5 | Week 6 | Week 7 | Week 8 | Week 9 | Week 10 | Week 11 | Week 12 |
---|---|---|---|---|---|---|---|---|
1 | 55 |
64 | 38 | 38 | 59 | 50 | 38 | 42 |
2 |
42 |
45 | 55 | 45 | 44 | 46 | 44 | 39 |
3 |
44 | 40 | 45 | 52 | 34 | 39 | 46 | 34 |
4 |
30 | 34 | 35 | 42 | 38 | 34 | 26 | 53 |
% of participants reporting using CD during week | 84.23% | 90.15% | 85.22% | 87.19% | 86.2% | 83.25% | 75.86% | 82.75% |
One hundred percent of GI group participants wrote in their daily logs at least 1 benefit they had received by listening to the intervention and most participants documented multiple benefits. Comments such as “definitely helped me calm down and not think of anything negative,” “I felt like a weight was lifted,” “it helped me deal with stress better,” “I let the energy come so I could be stress-free,” “ it brings me closer to my baby,” “ease feelings I was having about the baby,” “tiredness and headache went away,” and “learned how to control myself when I am in a tight spot” are just several of many comments describing the perceived benefit of the intervention.
The majority of participants reported no perceived barriers. Of those documented, all barriers pertained to the delivery system (e.g., CD player skipped and batteries ran low) or environmental disruptions that made it difficult to concentrate.
The purpose of this study was to investigate the effects of GI on self-reported and biologic measures of stress and the associated symptoms anxiety and fatigue in pregnant African American women. There was a statistically significant reduction in perceived stress scores at 8 weeks (time 2) but not at 12 weeks (time 3) in the GI compared to UC group. The significant decrease in PSS scores for the GI group at 8 weeks has not been reported in previous studies with pregnant women [
The NRSS weekly mean stress scores were significantly lower in the GI compared to UC group and, in the GI group, the NRSS scores were significantly lower after intervention compared to before intervention. These findings are consistent with prior reported research [
While the GI group’s state anxiety scores trended downward and the UC group’s state anxiety scores remained fairly constant at 8 weeks, there were no significant differences in the state anxiety scores at time 2 or 3. While previous studies [
There was a significant difference in trait anxiety scores with the GI group reporting decrease levels at time 8 but not time 12 compared to UC group. Prior studies [
Fatigue at 8 weeks (time 2) was decreased in the GI group compared to UC group. We believe this is the first published study to document the effects of GI on fatigue in pregnant African American women. This finding is consistent with reports of researchers examining the effects of GI on fatigue in other populations [
As expected, the CRH increased at each measurement. In a normal pregnancy, levels of CRH in maternal blood increase exponentially from weeks 15 to 36 gestation with a significant rise at weeks 26 to 30 [
The intervention was well received by the GI group. The GI participants had never used GI prior to this study and were open to participating and listening to the intervention. As evidenced by the written comments in the daily practice log, the participants found the intervention helpful in feeling less stressed and calmer, relieving symptoms, and soothed pregnancy-related concerns. The positive comments coupled with the fact that only one participant withdrew because the intervention was “too much bother” suggest that CAM interventions are potentially acceptable to African American pregnant women. This is an important finding given the current health disparity in the use of CAM interventions in the United States. Existing data suggest that CAM use is lower among African Americans compared to non-Hispanic Whites and Asian Americans [
There are multiple research implications resulting from this study. An interesting and important area for a future focus is to examine if a booster (s) between 8 and 12 weeks would continue to help improve the stress levels of the group receiving GI as boosters have been found to sustain and/or improve the effectiveness of health behavior interventions [
Our results must be balanced with study limitations. In terms of adherence to the intervention, the majority of participants reported daily use of the GI intervention. However, self-reported measures are dependent on the participant’s accurate reporting. It is possible that participants under- or overreported information or did not complete the form in a timely manner, which are all potential threats to validity. Future studies could address this limitation through the use of expanding mobile technology that can facilitate self-report data collection and validate intervention use. In some instances, participants’ health history information was self-reported and the recruitment site was unable to verify the information. This is a potential limitation when trying to control for confounding factors. For example, fatigue during pregnancy is known to be influenced by coexisting factors such as anemia, excessive weight gain, diet, and physical activity [
This study investigated the effects of a GI intervention on maternal stress and associated symptoms of anxiety and fatigue in pregnant African American women beginning in the second trimester. The findings demonstrated significant effects of the GI intervention on perceived stress, anxiety, and fatigue. Within the GI group, the intervention had an immediate effect as evidenced by significant differences in pre- and postintervention averaged daily scores and the GI group reported significantly lower perceived stress scores compared to UC group at 8 weeks. In addition, the GI group reported significantly less trait anxiety and fatigue compared to UC group at 8 weeks.
This study also provides direction for consideration in future studies. For example, the addition of appropriate booster sessions may strengthen or sustain the effects of the intervention and the use of mobile technology may facilitate self-report data collection and potentially enhance the evaluation of treatment fidelity, which is often missing in CAM mind-body intervention research. In addition, when examining measures of stress and associated symptoms in pregnant women, consideration should be given for the addition of pregnancy anxiety measures.
The results of this study also support the feasibility and acceptability of this intervention. Of those who met the inclusion criteria, 97% of women agreed to participate and, of those, 83% completed the 12-week study. The majority of participants (85%) reported at the beginning of the study that they were not using any stress management technique. Given the negative health outcomes associated with stress during pregnancy, it is imperative that effective, low-cost, and easy to use interventions be available to this population. GI, a CAM mind-body intervention, is a promising intervention to meet that need.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This work is supported by NIH Grant P30 NR011403 (Grap, PI), Center of Excellence for Biobehavioral Approaches to Symptom Management, and UL1TR000058 from the National Institutes of Health’s National Center for Advancing Translational Science and the CCTR Endowment Fund of the Virginia Commonwealth.