Factor analyses
The women in samples I and II had similar characteristics (Table 1). All items were normally distributed in sample I. Items 2 and 5 were eliminated based on face validity. The Kaiser-Meyer-Olkin index was greater than .60 (.80) and the Bartlett’s test of sphericity value was significant (p < .001). The scree plot suggested a two-factor solution with a total explained variance of 47.4%, with a ‘women’s perception’ factor and a ‘social environment’ factor (Table 2). The component correlations between the two factors were found to be smaller than .30 (.04) with direct oblimin rotation, therefore varimax rotation was used. Items 4 and 6 loaded on both factors with a difference smaller than .20 and were therefore eliminated. Reliability analyses showed a Cronbach’s alpha of .79 for the six-item women’s perception subscale, which increased to .84 after deletion of items 1 and 3. The three-item social environment subscale had a Cronbach’s alpha of .60, and the total seven-item LPRAQ-p had a Cronbach’s alpha of .77.
Subsequently, CFA was performed on the seven-item LPRAQ-p in the second sample, and showed a moderate model fit (CFI = .97, NFI = .95, TLI = .95, RMSEA = .08, lower bound = .05). However, item 7 showed poor standardized residual co-variances. After removing this item, a two-factor structure with six items showed an excellent model fit (CFI: .99, NFI: .98, TLI: .99, RMSEA: .02, lower bound: .01). EFA with varimax rotation was repeated in sample II on the six-item LPRAQ-p, again resulting in a two-factor structure explaining 65.9% of the variance (Table 3). The Cronbach’s alpha was .78 for the three-item women’s perception subscale, .67 for the three-item social environment subscale, and .75 for the total LPRAQ-p. The items were recoded from 1-5 to 0-4, with total scores ranging from 0 to 24. Higher scores indicated greater willingness for primary pain relief at the start of labor.
Concurrent and construct validity analyses
For concurrent and construct validity analyses, both samples were merged (N = 861). Skewness and kurtosis values showed a normal distribution of the 6-item scale. As shown in Table 4, total EDS scores were significantly associated with the total labor pain relief attitude scores (r = .133, p < .001) as well as with the scores on the women’s perception subscale (r = .134, p < .001) and social environment subscale (r = .076, p = .025), all small effect sizes. The TPDS-NA subscale scores and the worries about delivery subcomponent scores were significantly associated with the total pain relief attitude scores (TPDS-NA, r = .223, p < .001; TPDS-NA worries about delivery, r = .250, p < .001; both small to medium effect sizes). The TPDS-NA and worries about delivery scores were also significantly associated with the women’s perception subscale scores (TPDS-NA, r = .243, p < .001; TPDS-NA worries about delivery, r = .267, p < .001; both small to medium effect sizes) and social environment subscale scores (TPDS-NA, r = .104, p = .002; TPDS-NA worries about delivery, r = .125, p < .001; both small effect sizes).
Next, construct validity was assessed by hypothesis testing. In the third trimester of pregnancy, 125 (14.5%) of the 861 women scored above the EDS cut-off score (EDS ≥ 10), categorized as depression. These depressed women had significantly higher attitude towards labor pain relief scores, compared to women without depression (Mean (SD) = 7.4 (4.2) and Mean (SD) = 6.2 (3.6) respectively, t(156) = -3.00, p = .003, Cohen’s d = .31, small effect size).
The cut-off score of the TPDS-NA subscale was defined at the 86th percentile of third trimester pregnant women, which resulted in a cut-off score of 11. In total, 141 (16.4%) women scored above the TPDS-NA cut-off score, categorized as pregnancy-related distress. Women with pregnancy-related distress scored significantly higher on the LPRAQ-p, compared to women without pregnancy-related distress (Mean (SD) = 7.9 (3.9) and Mean (SD) = 6.1 (3.6) respectively, t(859) = -5.37, p < .001, Cohen’s d = .48, medium effect size).
Nulliparous women scored significantly higher on the LPRAQ-p compared to multiparous women (Mean (SD) = 6.7 (3.6) and Mean (SD) = 6.1 (3.7) respectively, t(845) = 2.29, p = .022, Cohen’s d = .16, small effect size). Of the sample of pregnant women, 402 were multiparous. These women were asked for possible complications during a previous delivery. A total of 245 women (62.5%) reported no complications: group 1. Sixty-nine women (17.6%) reported complications regarding poor progression of labor (such as delayed dilation phase, secondary Caesarean section, use of ventouse or forceps and fetal hypoxia): group 2. Seventy-eight women (19.9%) reported miscellaneous complications (such as prolonged second stage of labor, primary Caesarean section and fetus in breech position): group 3. When comparing attitude towards labor pain relief scores between group 1 and 2, women who reported no complications had a significantly lower score (Mean (SD) = 5.8 (3.8) and Mean (SD) = 7.1 (3.7) respectively, t(312) = -2.57, p = .01, Cohen’s d = .35, small effect size).