A prospective, two-group, quasi-experimental, pretest/posttest design was utilized.

A convenience sample of 40 participants, 20 aquatic exercisers, and 20 nonexercisers was recruited from a metropolitan area. Based on prior exercise research experience, in order to reduce study attrition, participants were allowed to self-select either the control group (nonexercise) or experimental group (aquatic exercise intervention). As recommended by the Aquatic Exercise Association, class size was limited to 20 participants per instructor to provide a safe environment (Sova, 2000). The sample size is adequate to detect moderate to large effects between groups (.9 power) (Cohen 1998).

Because complaints of back pain and leg cramps are often expressed during the late stages of pregnancy (Oga et al., 1995), participants were at least 19 weeks gestation. Additional inclusion criteria were required by the recreational center, included age greater than 18 years, health care provider permission, and height of 5 ft or greater due to the pool depth of 4 ft at its shallowest point. In accordance with the ACOG parameters for safe exercise practices during pregnancy (ACOG, 2002), participants with medical complications, such as multiple pregnancies, history of preterm birth, placenta previa or gestational hypertension, or both, were excluded from the study.

Evaluation instruments were selected to measure the major domains of the adapted Pender’s HPM: Individual Characteristics, Behavioral Outcomes, and Health Outcome (Figure 1). Demographic data were collected to identify characteristics of the sample individual characteristics and experiences (Table 1). The psychological individual characteristic of body image was measured by the Pregnancy Body Shape Questionnaire (PBSQ) (Fox & Yamaguchi, 1997). The PBSQ is a 34-item summated, 6-point Likert scale adapted from the Body Shape Questionnaire (Cooper, Taylor, Cooper, & Fairburn, 1987). Higher scores indicate a negative body image attitude. Internal consistency (Cronbach’s alpha) has been assessed at .86 for the revised scale.

In order to evaluate the effects of aquatic exercise in promoting healthy behaviors, the HPLP was used. The HPLP, developed by Walker, Sechrist, and Pender (1987), is a 52-item summated scale, which contains six subscales (e.g., self-actualization, health responsibility, physical activity, nutrition, interpersonal relationships, and stress management). Subscale reliabilities ranged from .70 to .90 with high internal consistency estimated for the total instrument (Cronbach’s α= .92). Higher scores indicate more participation in health-promoting activities. Originally tested using a convenience sample of adults (N= 952), the instrument has been utilized with pregnant women (Telleen, 1993). In a preliminary study of effects of exercise on maternal discomfort and health-promoting behaviors (N= 15), subscale reliabilities ranged from .74 to .89, with an overall scale reliability of .93 (Smith, 2001).

Mobility was evaluated using the Timed Get Up and Go Test, which was originally designed to measure mobility in older adults (Podsiadlo & Richardson, 1991). Reliability has been estimated from .92 to .96 in samples of older adults and pregnant women (Podsiadlo & Richardson; Smith 2001). Prior to completing the test, a demonstration was given and starting procedures were discussed with each participant. Each participant was asked to sit in a standard chair without arms (approximately 46 cm seat height). On the verbal command “go,” the participant was asked to stand up and walk 3 m (a red line was marked on the floor), turn, walk back to the chair, and sit down again. The test was timed with a digital watch to prevent timing errors.

Physical discomfort was assessed by Smith’s Pregnancy Discomfort Intensity Index (SPDII), developed by Smith in her preliminary work (Smith, 2001). The SPDII is a 6-item summated Likert scale that measures the intensity of physical discomforts experienced during pregnancy. Higher scores reflect a greater intensity of reported physical discomfort. Reliability of the index was assessed at .89 in a sample of pregnant women (n= 15) participating in an aquatic exercise program (Smith). Additionally, this scale was administered to nurses at obstetric clinics to determine content validity.

Institutional Review Board (IRB) approval was obtained prior to the initiation of the study. Participants were recruited from local health care providers (e.g., physicians and health departments). All participants completed the IRB-approved informed consent. The aquatic exercise program was a 6-week program, with three sessions per week. Each class was conducted at the recreation center pool and lasted 60 minutes. The exercise program was taught by an experienced obstetric nurse and two graduates of the women’s health nurse practitioner program. The investigator, a certified prenatal aquatic instructor, developed the exercises and trained the instructors. The university’s School of Allied Health Professions, Department of Physical Therapy, approved these exercises.

The first 10 minutes of the class involved warm-up exercises and stretches. Next, the continual movement phase lasted 25 to 30 minutes. During this portion of the class, exercises targeted large muscles of the body, such as the legs and buttocks. Participants were instructed to self-monitor their level of exercise intensity using the Perceived Level of Exertion Scale (PLES) (Borg, 1982). The PLES is a visual analog scale ranging from 6 to 20, with the safe intensity range for pregnant women between 12 and 16. The last phase of class (10-15 minutes) was designed to strengthen the abdominal muscles, stretch the muscles of the lower portion of the back, and promote flexibility. After this, a warm-down, stretch-out, and relaxation session concluded the class.

Pretesting of both control and intervention groups were scheduled at the School of Nursing 2 days prior to the start of the exercise program. Participants’ height and weight were measured on a digital calibrated balance scale to compute BMI. Participants completed the pretest research instruments, which included a demographic form, SPDII, PBSQ, HPLP, and Timed Get Up and Go Test. The test administrator was not aware of participant group assignment. Prior to leaving the testing area, participants were given instructions regarding posttesting. After pretests were completed, each participant received a $10 grocery store gift certificate.

Prior to the 1st exercise session, the prenatal aquatic instructor gave a demonstration of the prenatal aquatic exercises. Participants were familiarized with the safety guidelines and the sign-in procedure for the aquatic exercise attendance record. Transportation bus tickets or parking passes were given to assist participants’ attendance.

Statistical analyses were performed using SPSS version 12 for Windows; demographic and clinical characteristics were evaluated to detect group differences at the initiation of the study and during posttesting. The distributional characteristics of the health-related outcome variables were examined. No extreme values, outliers, or skewedness were detected. Therefore, the distributions are assumed to be approximately normal and meet the assumptions of analysis of covariance (ANCOVA). Internal consistency estimates (Cronbach’s alpha) were evaluated for each instrument. Independent t tests were analyzed to detect pretest group differences across all instruments and subscales of the HPLP. Posttest scale and subscale scores were evaluated by ANCOVA to detect group differences. Maternal discomfort level was thought to be a biasing factor in self-selection. The significance level was set at alpha ≤ .05.

The individual characteristics of the sample represented a deviation from most studies on exercise during pregnancy. Two perceived barriers to action, participating in an exercise program, for this sample were sedentary lifestyle history and obesity. The mean prepregnant BMI was in the obese range. The sample represented a deviation from most exercise during pregnancy studies. Table 1 describes sample characteristics. There were neither statistically significant differences among demographic and clinical characteristics between the groups, nor significant mean group differences across all instrument scales.

Comparisons of posttest clinical characteristics revealed that a significantly higher proportion of the exercising group participated in the health-promoting behavior of prenatal education classes as compared to the control group, χ²(1, N= 40) = 6.5, p= .01. No other significant group differences were identified.

Participants in the aquatic exercise program reported significantly less physical discomfort, improved mobility, improved body image, and more health-promoting behaviors.

Mean posttest comparisons of health-related outcomes by group indicated that women who had participated in the aquatic exercise program reported significantly less physical discomfort, increased participation in health-promoting behaviors, improved mobility, and positive body image as compared to control subjects (Table 2). Total HPLP posttest scores indicated that the participants were likely to participate in health-promoting behaviors (p= .05). Health responsibility (p < .04), physical activity (p= .006), and stress management (p < .001) scores indicated significantly more participation in health-promoting activities among women who had participated in the exercise group as compared to nonexercisers. Group differences regarding health-promoting practices related to nutrition, self-actualization, and interpersonal relations were not significantly different.

The aquatic intervention used in this study consisted of exercises designed to strengthen abdominal muscles that provided greater support for the expanding girth during pregnancy. Stretching exercises to reduce leg cramps and increase flexibility were included in each session. Study findings regarding overall physical discomforts were consistent with earlier evidence that aquatic exercise positively influences physical discomforts associated with pregnancy.

More than half of the participants were African American representing a departure from earlier studies related to pregnancy and exercise.

Podsiadlo and Richardson (1991) had classified individuals who were capable of completing the Timed Get Up and Go measure in less than 10 seconds as functionally mobile. Reflected by the study pretest scores, which ranged from 12 to 26 seconds, it is evident that pregnancy does impede level of mobility. While the groups did not differ on level of mobility prior to the intervention, the aquatic exercising group experienced significant improvement in mobility. Increased mobility among exercisers has provided important evidence, which has not been previously investigated.

At pretesting, while 75% of the sample could be categorized as obese, lower levels of negative body image were reported. This finding suggested that body image was not a great concern to these women. The racial diversity of the sample may explain why pregnant body image concerns did not reflect what has been previously reported in the literature (Hofmeyer et al., 1990; Walker, 1998). Recently, Morin, Brogan, and Flavin (2002) conducted a study among pregnant African American women to evaluate feelings about body image changes. When compared to White pregnant women, African American women tended to have a more positive body image irrespective of their body size. In conjunction with the differences in body image perception based on ethnicity, pregnant women who have participated in an aerobic exercise program have reported a more positive self-concept (Koniak-Griffin, 1994). The effect of the intervention did improve body image scores among exercisers (i.e., pretest M= 65, posttest M= 56) and may have been influenced by ethnicity.

Higher scores on the HPLP are associated with participation in health-promoting activities. Participants in the exercise group had statistically significant higher posttest scores than those in the nonexercise group (p= 05).

Evidence indicates that individuals who exercise take more responsibility for their health and are more likely to engage in proactive healthy behaviors (Bottorff, Johnson, Ratner, & Hayduk, 1996; Lannon, 1997). Study findings also support this evidence: Subscale scores were significantly better for health responsibility, physical activity, and stress management among exercisers. However, it appears that the exercise intervention had little impact on the health-promoting lifestyle dimensions of nutrition, interpersonal relationships, and self-actualization.

There were no significant group differences at pretesting, but postintervention scores reflected that a significant number of exercisers became engaged in prenatal classes as compared to nonexercisers. This is an example of assuming increased health care responsibility and most likely resulted from participation in the exercise intervention. Exercisers may have encouraged each other to participate in prenatal education classes, or the exercise intervention itself may have stirred a desire among the participants to learn more about how to take care of themselves and their developing babies.