This research is an applied study in terms of purpose, which employs mixed method (quantitative and qualitative) design for data collection. The research was conducted in two stages. In the first stage of the research, the literature was first reviewed, followed by semi-organized interviews with pregnant women, with the purpose of obtaining the criteria and sub-criteria for increasing physical activity in pregnant women. The method employed at this stage of research was that of qualitative. In the second stage of the research, based on the information obtained from the previous stage, the researcher finally implemented an E-learning program intervention to increase physical activity in pregnant women visiting health centers in Tehran based on the PEN-3 cultural model, and thus, the second stage of the research was a quasi-experimental study which was performed in a quantitative manner. This study was in fact a randomized control trial (RCT). Given the nature of the research, a quasi-experimental research design (pre-test/post-test) with a control group was employed for the purposes of the study.
The study population included all pregnant women aged 18 to 40 years with gestational age of 12–38 weeks visiting health-care centers in Tehran's District 5 out of all 22 districts of the city that had eliminated their physical activity during pregnancy.
In the qualitative stage, the sample size was determined based on the necessity of data saturation, and as such, interviewing continued until it was achieved. Sampling was performed using simple random sampling among those invited to participate in the study in the health center. Thus, after sending invitations to pregnant women who were in their first pregnancy and conducting telephone follow-ups by an expert and the head of the health center, 23 people were selected from the mothers who visited the health center.
Data Collection tools
A) Semi-structured interviews with semi-structured questions for small group discussion among pregnant women and their families, along with individual interviews with health care providers and health professionals
B) Questionnaires
1. Questionnaire for examining demographic variables including personal characteristics and disease information
2. Questionnaire designed based on the results of needs assessment and dimensions of PEN-3 model
This questionnaire was a researcher-made questionnaire, with three categories and 42 items. The categories include perceptual factors with 20 items, nurturing factors with 10 items, and enabling factors with 12 items. The items are scored on a 5-point Likert scale ranging from strongly agree to strongly disagree. The validity of the questionnaire was confirmed using content validity, while its internal reliability was confirmed following a confirmatory study among 250 pregnant mothers with similar traits to that of the study group. The reliability of the questionnaire was confirmed with Cronbach's alpha coefficient (0.85). The results of validity and reliability of the tools were presented.
3. Pregnancy Physical Activity Questionnaire (PPAQ)
This questionnaire is consisted of two parts, the first part pertains to the personal characteristics and the second part is consisted of 32 items on physical activity, which itself is divided into four groups of items, namely household/caregiving activities (16 items), transportation (3 items), occupational activities (5 items) and entertainment and sports (8 items). Activity intensity was calculated based on Metabolic equivalent of task (MET), which is a unit for estimating metabolic expenditure in physical activity per unit of time. Its reliability was determined by conducting a preliminary study on 3 eligible pregnant women, attaining a Cronbach's alpha of 0.85.
Validity and Reliability of the questionnaire
Face validity: The questionnaires were first administered among at least 15 pregnant women visiting health centers, then the face validity was determined based on the importance of the items and by calculating the impact score. Items with an impact score higher than 1.5 were deemed appropriate for subsequent analysis.
Content validity: The opinions of the panel of experts were extracted and used accordingly to determine the content validity. To quantitatively evaluate the content validity, two coefficients of content validity ratio and content validity index were used.
Content Validity Ratio (CVR): At least 15 specialists and experts responded to each item through 3 choices for each (i.e., item is necessary, item is useful but not necessary, item is not necessary). Items whose resulting scores were larger than the relevant value of the Lawshe table were considered as necessary items.
Content Validity Index (CVI): To ensure that items were optimally designed to measure structures, three criteria of relevance, clarity, and simplicity were examined using 8-point Likert scale, as each item was examined by at least 15 specialists. CVI scores higher than 0.79 was deemed appropriate.
Reliability test: To determine the reliability of data collection tool, a modified questionnaire was distributed among 30 people in the study population and Cronbach's alpha value was calculated. An alpha value higher than 0.7 was considered proper for the purposes of this study.
Data analysis method
In the first stage, qualitative data were coded immediately after each group discussion and individual interview, and these sessions were continued until data saturation. The main themes and topics were extracted using MAX-QDA software and based on deductive qualitative content analysis. Descriptive statistical methods such as mean and inferential statistics were employed to analyze the data.
Integration through inclusive themes
In the process of thematic analysis, the themes formed in the interviews were integrated under comprehensive themes, and a network of themes, namely “Factors Affecting Physical Activity Based on the PEN-3 Cultural Model” was formed, which is shown in Fig. 1.
Quantitative data analysis
In this section, structural measurement models were employed to examine the accuracy of measuring structures by the corresponding indicators.
Table 1 shows that the confirmatory factor analysis of the constructs of the questionnaires on the impact of E-learning program with the aim of increasing physical activity of pregnant women visiting healthcare centers in Tehran based on the PEN-3 cultural model has a suitable fit and thus the constructs of the questionnaire measure the relevant variables properly.
Table 1
Results of factor analysis of questionnaire structures
|
Item
|
Factor load
|
T-value
|
Awareness-related perceptual factors
|
During pregnancy, it is better to do less intense physical activity than before pregnancy.
|
0.25
|
3.41
|
To prevent pregnancy-related overweight, physical activity is crucial.
|
0.37
|
5.41
|
Exercise during pregnancy reduces the oxygen supply to the baby.
|
0.45
|
6.69
|
During pregnancy, exercises should be performed on smooth surfaces without slopes.
|
0.37
|
5.42
|
Pregnant women should avoid bending and hunching during exercise.
|
0.26
|
3.74
|
Those who do not exercise give birth to diabetic babies more often.
|
0.38
|
5.56
|
Exercise during pregnancy facilitates natural childbirth.
|
0.34
|
4.9`
|
Doing proper daily exercise will not harm the mother and fetus.
|
0.09
|
23/1
|
Physical exercise alleviates pregnancy-related complications such as low back pain and pelvic pain, constipation and extreme fatigue.
|
0.36
|
5.19
|
Heavy exercises should be avoided during pregnancy.
|
0.29
|
4.13
|
Performing light exercises to warm up is essential before starting exercise during pregnancy.
|
0.45
|
6.58
|
Stretching and strength exercises during pregnancy should be performed only in consultation with an expert.
|
0.34
|
4.89
|
Exercising during pregnancy avoids the risks of high blood pressure.
|
0.19
|
2.63
|
Exercise during pregnancy leads to a faster return to the mother's initial weight after delivery.
|
0.35
|
5.03
|
Pregnant women should try to avoid lifting any weights during pregnancy.
|
0.59
|
8.99
|
Exercising during pregnancy leads to fitness and weight control.
|
0.17
|
2.42
|
To prepare before exercise, performing 15 minutes of light exercises is necessary.
|
0.49
|
7.37
|
It is necessary to consult with a professional before doing any exercise during this period.
|
0.21
|
2.93
|
Exercise reduces the risk of musculoskeletal disorders.
|
0.30
|
4.28
|
In the third trimester of pregnancy, the intensity of exercise should decrease.
|
0.39
|
5.69
|
Attitude-related perceptual factors
|
I believe that by exercising I can easily cope with complications like gestational diabetes.
|
0.56
|
8.90
|
I believe that exercising during pregnancy reduces the fatigue caused by the pregnancy.
|
0.65
|
10.65
|
I believe that by exercising during this period, I can cope with the stress and anxiety of childbirth.
|
0.56
|
8.89
|
I believe that exercising during pregnancy helps me to do my daily activities better.
|
0.62
|
10.13
|
I believe I can easily maintain my fitness by exercising during pregnancy.
|
0.54
|
8.57
|
I believe that exercising can reduce my postpartum depression.
|
0.48
|
7.50
|
I believe that I can have an easier delivery if I exercise during pregnancy
|
0.43
|
6.58
|
I believe that having an active role in pregnancy care is of paramount importance in determining my health and performance.
|
0.48
|
7.51
|
I believe that despite economic problems, I can maintain my health by changing my lifestyle (such as exercise, diet, etc.).
|
0.37
|
5.68
|
I believe that cooperation with health care providers in doing pregnancy exercises can lead to improved health levels.
|
0.53
|
8.26
|
Given that all significance variables of model parameters except the item “doing daily basic exercise will not harm the mother and fetus” is greater than 1.96, the validity of the measurement structures for the relevant variables at a significant level is confirmed.
In the research model, the root mean square error of approximation (RMSEA) is 0.038 which is less than 0.1. Moreover, other fitness indicators are also listed in the table below. As can be seen, they confirm the significance and fit of the model.
Descriptive statistics of variables
The results of descriptive statistics on the pre- and post-test of the variables of “awareness” and “attitude” in the experimental and control groups are given in Table 2.
Table 2
Pre- and post-test results of variables in experimental and control groups
Variable (pre-test)
|
Experimental group
|
Control group
|
Mean
|
SD
|
Mean
|
SD
|
Awareness
|
24.30
|
12.51
|
22.45
|
13.11
|
Attitude
|
22.31
|
6.70
|
24.32
|
6.42
|
Perceptual factors
|
46.61
|
12.74
|
46.77
|
14.13
|
Increased physical activity
|
82.26
|
12.75
|
82.43
|
14.51
|
Post-test
|
|
|
Awareness
|
30.42
|
10.44
|
25.82
|
11.00
|
Attitude
|
43.56
|
5.31
|
34.79
|
4.57
|
Increased physical activity
|
127.11
|
13.17
|
100.00
|
12.88
|
Inferential statistics
Default values for analysis of covariance
According to the values of Table 3 and considering that the level of significance for all the variables of “awareness”, “attitude”, “perceptual factors” and “increase in physical activity” is higher than 0.05, it can be concluded that the relevant variable has a normal distribution.
Table 3
Default values of covariance analysis for research variables
Variables
|
Kolmogorov-Smirnov
|
Significance level
|
Awareness
|
0.208
|
0.000
|
Attitude
|
0.089
|
0.000
|
Increased physical activity
|
0.117
|
0.000
|
Homogeneity of variances
|
Levene value
|
df1
|
df2
|
Significance level
|
Awareness
|
0.001
|
1
|
198
|
0.971
|
Attitude
|
0.037
|
1
|
198
|
0.847
|
Increased physical activity
|
0.135
|
1
|
198
|
0.714
|
Homogeneity of regression slope
|
Df
|
F
|
Significance level
|
Awareness
|
1
|
1.723
|
1.191
|
Attitude
|
1
|
3.595
|
0.059
|
Increased physical activity
|
1
|
0.004
|
0.951
|
Furthermore, according to the same table and considering that the significance level of Levene test for all the variables of “awareness”, “attitude”, “perceptual factors” and “increase in physical activity” is higher than 0.05, it can be argued that the experimental and control groups have homogeneous variances.
Considering that none of the variables of “awareness”, “attitude”, “perceptual factors” and “increase in physical activity” are significant in examining the homogeneity of regression slope (P > .05), we conclude that the assumption of homogeneity of regression slope holds.
Examining the hypotheses
1. Intervention based on E-learning has a significant effect on perceptual factors (awareness, attitude) with the purpose of increasing physical activity in pregnant women in the experimental group.
To test Hypothesis 1, multivariate analysis of covariance was performed while controlling the possible effect of pre-tests, the results of which can be seen in Table 4. Covariance analysis is a comprehensive type of analysis of variance in which the effect of one or more control or interfering variables and covariates is excluded from the equation while comparing the means of one or more groups and estimating one or more independent variables. In fact, this analysis is a statistical method that allows the effect of one independent variable on the dependent variable to be examined while eliminating the effect of another variable.
Table 4
Multivariate statistical indices in analysis of variance of dependent variables
Trace
|
Value
|
F
|
Hypothesis df
|
Error df
|
Sig.
|
Size
|
Pillai's Trace
|
0.202
|
24.859
|
2
|
197
|
0.000
|
0.202
|
Wilks Lambda
|
0.789
|
24.859
|
2
|
197
|
0.025
|
0.202
|
Hotelling's Trace
|
0.752
|
24.859
|
2
|
197
|
0.025
|
0.202
|
Roy's Largest Root
|
0.252
|
24.859
|
2
|
197
|
0.025
|
0.202
|
The results from Wilkes lambda test indicated that the effect of interventions based on E-learning on the combination of variables of perceptual factors (i.e., awareness and attitude) was significant (F = 24.859, p < 0.05). The above test allowed the use of multivariate analysis of variance (MANOVA). The size of the difference effect is proper considering the eta-value (0.202). The results of multivariate analysis of covariance in Table 5 showed that the mean scores of perceptual factors, i.e., awareness (2080.125) and attitude (2284.880) were significantly higher in pregnant women in the experimental group than those in the control group, and as such it can be concluded that Intervention based on E-learning has a significant effect on perceptual factors (awareness, attitude) that aim to increase physical activity of pregnant women in the experimental group.
Table 5
Results of multivariate analysis of covariance on dependent variables
Dependent variables
|
Sum of squares
|
Df
|
average of squares
|
F
|
Sig.
|
Size
|
Awareness
|
2080.125
|
1
|
2080.125
|
4.592
|
0.033
|
0.023
|
Attitude
|
2284.880
|
1
|
2284.880
|
45.029
|
0.000
|
0.185
|
2. Intervention based on E-learning has a significant effect on increasing physical activity in pregnant women of the experimental group compared to those of the control group.
According to Table 6, pregnant women (F = 320.395, P < 0.05) received a significant impact from the intervention based on E-learning in the physical activity variable. As a result, the null hypothesis is rejected and thus the research hypothesis is accepted. In other words, E-learning-based intervention has a significant effect on the level of physical activity in pregnant women.
Table 6
Results of one-way analysis of covariance for physical activity
Dependent variables
|
Sum of squares
|
Df
|
average of squares
|
F
|
Sig.
|
Size
|
Physical activity
|
36996.391
|
1
|
36996.391
|
320.395
|
0.000
|
0.619
|