Recruitment
This prospective study included patients participating in two studies ongoing at the Department of Obstetrics and Gynecology, Clinics Hospital of the Medical School of the University of São Paulo (Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo - HCFMUSP) comparing methods for preventing prematurity in twin and singleton pregnancies with a short cervix from May 2019 to January 2022. Participants in the above-mentioned studies were informed regarding the risks of prematurity and on the methods of investigation for preventing it. The study on singleton pregnancies compared the pessary and progesterone ovule methods and the patients were told that both treatments were effective. The study on twin pregnancies compared the methods of isolated pessary, P + P, or isolated ovules that could contain progesterone or placebo. Moreover, these pregnant women were informed that such methods have not yet been proven to be effective in this type of pregnancy. The patients were instructed on the use of ovules in the vagina and the method of placing the pessary through vaginal examination, maintaining the randomized treatment for up to 37 weeks in singleton pregnancies or up to 36 weeks in twin pregnancies. Before the randomization of the described studies, the patients were invited to participate in the research on patient preferences for prematurity prevention methods using self-administered questionnaires in Portuguese. Figure 1 shows the study flowchart 1. All the study participants signed an informed consent form and received prenatal care at this service. The project was approved by the HCFMUSP Research Ethics Committee (CAAE: 15278819.0.0000.0068).
Eligibility criteria
The patients participating in the studies “Comparison between natural progesterone and cervical pessary in the prevention of spontaneous preterm delivery in singleton pregnancies with a cervix less than or equal to 25 mm” (ClinicalTrials.gov Identifier: NCT02511574) and “Natural progesterone combined with a cervical pessary in twin pregnancies with a short cervix: Prospective and randomized study for the prevention of preterm birth” (ClinicalTrials.gov Identifier: NCT03058536) were eligible to participate in this study. The inclusion criteria for the study with singleton pregnancies were singleton pregnancies, gestational age between 20 weeks and 23 weeks and 6 days with asymptomatic short cervical length (less than or equal to 25 mm). For twin pregnancies, the inclusion criteria were twin pregnancy (dichorionic or monochorionic diamniotic), gestational age between 16 weeks and 27 weeks and 6 days, asymptomatic short cervical length according to the gestational age (measurement less or equal 30 mm between 16 weeks + 0 days and 22 weeks + 0 days, less or equal 25 mm between 22 weeks + 1 day and 24 weeks + 0 days and less or equal 20 mm between 24 weeks + 1 day and 27 weeks + 6 days). The inclusion criteria common to both studies were alive fetuses without fetal malformations at ultrasound scan, intact fetal membranes, absence of bulging fetal membranes into the vagina, and an absence of placenta previa. The exclusion criteria were patients who do not understand the questionnaires, or patients who do not speak or read in Portuguese.
Questionnaire
The questionnaire to evaluate the preferences (S1) on preterm birth prevention in pregnant women at real risk of prematurity was based on practical issues that could relate to the choice of methods studied. Therefore, this questionnaire included questions on the following aspects: (1) knowing about prematurity, (2) knowing about the methods studied, (3) whether they would use the methods, (4) preferable methods, and (5) concerns about the side effects of the methods. The questions 1, 2, 6, 7, 8, and 9 presented Likert-type8 five-point responses (not at all [0], a little [1], moderately [2], a lot [3], and extremely [4]). The statements 3, 4, and 5 were multiple-choice questions with a single answer (S1). This instrument was developed by the researchers specifically for the present study after a search in the scientific literature for tools that met the research objectives. The questionnaires were reviewed by three professors from the Department of Obstetrics, who discussed the topic and the wording of the elaborated items. The final version of the instrument was administered to ten pregnant patients who were not necessarily at risk of prematurity. These patients were asked to evaluate each item in the questionnaire, in addition to filling out their perception of prematurity and method preferences. A five-point numerical scale (from 0–4) was used to assess the easiness to understand the instrument as a whole and each question separately (0 = I did not understand anything, to 4 = I understood perfectly and I have no questions). Values above three were considered indicators of sufficient understanding. Comprehension was considered sufficient when at least 80% of the participants demonstrated full understanding of the evaluated item. Cronbach’s α was used to evaluate the internal consistency, with the nine items of the questionnaire showing high internal consistency (α = 0.927). The medians of the responses were between 3–3.5, reaching 80% of the expected understanding.
The questionnaires WHOQOL-bref 9,10 and STAI11,12 were subsequently administered.
The quality of life (QoL) questionnaire was assessed using the World Health Organization Quality of Life - Bref (WHOQOL-bref) instrument, validated for Portuguese.9,10 This instrument consists of 26 questions, of which two address the QoL globally (Questions 1 and 2) and the others are divided into four domains: physical (Questions 3, 4, 10, 15, 16, 17, 18), psychological (Questions 5, 6, 7, 11, 19, 26), social relations (Questions 20, 21, 22), and environment (Questions 8, 9, 12, 13, 14, 23, 24, 25). The responses ranged from 1–5 on a Likert scale8, from the worst to the best scenario.
The STAI questionnaire has two parallel scales, one to measure the T-Anxiety and the other E-Anxiety, each with 20 items.11,12 It is self-administered and uses the Likert scale,8 with scores for each individual item ranging from 1 (not at all) to 4 (very much). The pregnant women were instructed to indicate how they felt at a given moment.
Outcomes
The main outcome measure is the preferable method to treat short cervix (pessary, progesterone ovules, P + P, or no method) chosen by the pregnant women with short cervical length. The secondary outcomes included the influence of QoL, state, and trait anxiety on method preferences.
Sample size calculation
The sample size was calculated based on the hypothetical study by Ha and McDonald7, which evaluated the degree of preference for methods to prevent premature births in patients with no real risk for prematurity. In this study, the sample size was calculated using the method as the method based on the cited article did not obtain satisfactory values.
In single pregnancies, the sample calculation considered the difference between the proportions of progesterone (85%) or pessary (6%) choices in the study (1). Considering an alpha of 5% in a two-tailed hypothesis for a fixed sample, and reaching a minimum power of 80%, a sample of 32 patients was calculated for the pessary treatment and a sample of 30 patients was calculated for the progesterone treatment.
In twin pregnancies, the sample calculation considered the difference between the proportions of pessary (6%), P + P (50%), and progesterone (85%) observed in the study (1). Considering an alpha of 5% in a two-tailed hypothesis for a fixed sample, and reaching a minimum power of 80%, the sample sizes were calculated for the pessary treatment (n = 17), P + P treatment (n = 16), and for the progesterone treatment (n = 28).
Analysis of questionnaires
The analysis of question 1 in the preference questionnaire considered Likert scores 0 and 1 as low concern, 2 as moderate, and 3 and 4 as high. In question 2, the answers were grouped into yes (scores 1, 2, 3, and 4) and no (score 0). The statements 3, 4, and 5 were interpreted according to the patient’s choice. In questions 6, 7, 8, and 9, Likert scores 1, 2, 3, and 4 were considered as the presence of side effects and score 0 was considered to denote an absence of side effects.
In the STAI analysis, the total score was calculated by the simple sum of the questions ranging from 20–80 for each scale (trait and state). The anxiety levels were classified as low (score ≤ 32), moderate (score between 33–41), and high (score ≥ 42). 11
With respect to the QoL analysis, questions 3, 4, and 26 had their values recoded to make the entire scale positive (5 = 1, 4 = 2, 3 = 3, 2 = 4, and 1 = 5). The questions were calculated by the mean of the sum of each domain and multiplied by four, ranging from 4–20 points. The formula Domain X = (Domain X-4)*(100/16) was applied to transform the results into a scale from 0–100. Thus, the scores between 81–100 represent very good QoL; scores between 61–80 represent good QoL; scores between 41–60 represent neither bad nor good QoL; scores between 21–40 represent poor QoL; and scores between 81–20 represent very poor QoL. 10
Statistical analysis
A descriptive analysis of the data was performed using absolute (n) and relative (%) frequencies, central tendency measures (mean and median), and dispersion (standard deviation, minimum and maximum values, and percentiles).
The treatment preference distributions were compared using the binomial test for singleton pregnancies and the one-sample Chi-square test for twin pregnancies. Independent qualitative variables and qualitative outcomes (singleton or twin pregnancy, types of treatment methods, preference for methods, anxiety and QoL, and adherence to treatment according to preference) were compared using the chi-square, Fisher’s Exact, and Linear-by-Linear association tests, according to the characteristics of the variables. The Mantel–Haenszel test was used to compare the groups according to the preference of method. The Student’s t-test and ANOVA were used to compare quantitative independent variables and qualitative outcomes when the data were parametric. The Mann–Whitney or Kruskal–Wallis tests were used for non-parametric data.
The level of significance was set at 5% (P < 0.05). The data were entered in an Excel spreadsheet and analyzed using the SPSS software version 20.