Identifying the Barriers to Universal Cervical Length Screening for Preterm Birth Prevention at a Tertiary Hospital in Thailand (Physician Perspectives): Implementation Research

Objective: To identify physicians’ views on the barriers to measuring cervical length for preventing preterm deliveries. Materials and methods: This prospective, descriptive implementation study had three phases. In Phase I, 20 physicians were interviewed. Phase II comprised questionnaire development and data validation. The questionnaire was distributed to 120 Phase III participants. Results and discussion: All 120 participants responded. In 44 cases, the physicians received support from their local Maternal and Child Health Boards for preterm-birth-prevention programs; the other 76 physicians did not. The doctors tended to believe that cervical length screening plays no role in preventing preterm births (4/44 (9.1%) and 24/76 (31.6%); OR, 4.615; 95% CI, 1.482–14.373; p = 0.005). They were unsure about the correct measurement procedures (13/44 (29.5%) and 37/76 (48.7%); OR, 2.262; 95% CI, 1.028–4.977; p = 0.040). A lack of cost-free drug support (progesterone) for women with short cervices was identified as a barrier to preventing preterm births (30/44 (68.2%) and 32/76 (42.1%); OR, 0.339; 95% CI, 0.155–0.741; p = 0.006). Conclusions: Many physicians are unconvinced that measuring cervical length prevents premature births, and are unsure about the correct measurement procedures. There is a lack of government funding for hormone-usage programs.


Introduction
Globally, the chief cause of death among children younger than 5 years of age is a result of complications associated with preterm birth. In 2016, such complications accounted for 35% of neonatal deaths and approximately 16% of deaths in children up to age 5 worldwide [1]. Preterm births represented 12.98% of deliveries at Siriraj Hospital, Thailand, in 2008, [2] and the estimated global preterm birth rate for 2014 was 10.6% [3].
Premature babies who survive are at high risk of many short-and long-term illnesses. Common complications include respiratory distress syndrome, bronchopulmonary dysplasia, necrotizing enterocolitis, sepsis, periventricular leukomalacia, seizures, intraventricular hemorrhage, cerebral palsy, infections, feeding difficulties, and hypoxic-ischemic encephalopathy, as well as visual and hearing problems [4,5].
The financial burden of premature deliveries is substantial for the healthcare system and for parents [6]. About two-thirds of premature deliveries result in financial and emotional difficulties for the parents [6]. Research suggests that only a small number (7-15%) of spontaneous preterm births occur in women who have had a previous preterm

Materials and Methods
This was a prospective, descriptive, exploratory cross-sectional study on physicians' opinions and perspectives. It drew upon structured interviews and questionnaires derived from deep interviewing. The study was conducted at tertiary hospitals throughout the 6 regions of Thailand (northern, northeastern, southern, eastern, western, and central) from September 2019 to August 2020. Before the commencement of the research, ethics approval was obtained from the Siriraj Ethics Committee of the Faculty of Medicine, Siriraj Hospital, and the work was registered at the Thai Clinical Trials Registry.
To ensure the dataset was adequately sized to reach sufficiency for the details of barriers, we used a proportion for the results of interest of 50% (p = 0.5), an estimation error of ≤5%, and a 95% confidence level (type I error = 0.05; 2-sided). After factoring in the proportion of 1 physician to 3 patients in the healthcare system, it was determined that 120 physicians had to be questioned. The research was divided into 3 phases.

In-Depth Interviews
This phase collected information in the following 4 areas: (1) General physician information; (2) Physician attitudes to the performance of cervical length measurements and the provision of care for preterm births; (3) The decision-making process for performing measurements; (4) Frustrations experienced when deciding whether to prevent preterm labor when a short cervix is detected.
The interviewers traveled to tertiary hospitals in various provinces throughout all 6 regions of Thailand. In each region, 4 to 6 hospitals were randomly selected, and one physician working in each hospital was interviewed. If the doctor was unwilling to participate, another hospital in the same region was selected. A total of 20 physicians were interviewed.
Physicians willing to participate in the research project were invited to a private counseling room. After the details of the proposed project were described, the physicians were invited to ask questions and given time to consider whether they wished to proceed with their formal enrollment in the trial. The physicians were informed that they could decline to participate in the research and, if they agreed to proceed, could withdraw at any stage. Twenty physicians subsequently volunteered as research subjects.
It was requested that the participants sign an informed consent form before being interviewed. Permission was obtained from each participant for the structured interview to be audio recorded. The subjects initially completed an attitude assessment questionnaire: this dealt with the methods used to measure cervical length and the assessment of the degree of care to be provided in the event of preterm births. Several other aspects were then investigated in the interview-one related to frustrations that might be felt before performing a cervical length measurement. The total time from the commencement of the questionnaire until the completion of the comprehensive interview was approximately 30 min. The data integrity of the research questions was later verified.

Development and Validation of the Questionnaire
The data obtained from the questionnaire and in-depth interviews were analyzed in order to determine the means and standard deviations. This enabled the questionnaire and interview questions to be refined. The revised questionnaire and interview questions were tested for validity and reliability before their use in the next phase. The questionnaire's validity was checked by a statistician specialized in questionnaire construction and identifying double-barreled, confusing, and leading questions. To assess the test-retest reliability of the questionnaire, the same respondents completed the questionnaire again 1 month after first completing it. The data obtained from the questionnaire are detailed in "Supplementary File S1: Questionnaire for Physician's Perspective".

Administration of the Questionnaire
During the last phase of the study, the validated questionnaires were distributed to 120 physicians in tertiary hospitals. We wanted to find the barriers to cervical length screening at the level of tertiary hospitals, where adequate numbers of ultrasound machines and obstetricians are available. The tendency for Thai primary and secondary hospitals to not have adequate numbers of obstetricians or ultrasound machines is a known barrier to screening in the country.
The questionnaires were sent to various hospitals throughout the 6 regions of Thailand. In all, 24 tertiary hospitals were randomly selected using block randomization. The questionnaires were sent via registered mail. Each hospital was contacted to ensure that the questionnaires were completed and returned.

Statistical Analysis
Demographic data are summarized using descriptive statistics. Categorical data are presented as numbers and percentages, and continuous data are reported as mean ± standard deviation, or median and range. The statistical analyses were performed with PASW Statistics for Windows (version 18; SPSS Inc., Chicago, IL, USA). Hierarchical cluster analysis was employed because the variables used to group cases were "Yes" and "No." Group comparisons were made with independent t-tests, Mann-Whitney U tests, and Chi 2 tests.

Results
The Phase I interviews of the 20 physicians in tertiary hospitals throughout Thailand revealed that 0% to 7% of pregnant women underwent cervical length screening. The preterm birth rate also ranged from 9% to 15% (Figure 1). All questionnaires sent to the 120 participants were returned (a 100% return rate). Table 1 presents the personal information of two clusters of physicians: those perceiving that preterm births present a low to moderate level of problems, and those considering that preterm births present a high level of problems.
Of the 120 respondents, 108 physicians reported having performed cervical length screening, while 12 doctors had never conducted the screening. We found that screening was performed in conjunction with other work at most hospitals due to supportive policies (odds ratio [95% CI], 1.742 (0.105-28.840); p < 0.01; Table 2). Encouragement was given by local Maternal and Child Health Boards for implementing cervical length screening programs at the hospitals (odds ratio [95% CI], 1.742 (0.105-28.840); p < 0.01; Table 2).
Of the 120 respondents, 63 opined that preterm births have severe consequences, whereas 57 stated that the births have low to moderate consequences. We found that most hospitals had enough obstetricians who could accurately perform cervical length measurements (odds ratio [95% CI], 4.261 (1.312-13.834); p < 0.011; Table 3).
Of the 120 respondents, 108 people indicated that their hospital had an action plan for preventing preterm births, while 18 doctors stated that their hospitals did not have such a plan. The factors significantly associated with an action plan are presented in Table 4.
Even though local Maternal and Child Health Boards supported the implementation of programs for cervical length screening for preterm birth prevention, the surveyed doctors did not think that cervical length screening plays a role in preventing preterm births (odds ratio [95% CI], 4.615 (1.482-14.373); p < 0.005; Table 5). The doctors were significantly unsure about the correct procedures for the measurements (odds ratio [95% CI], 2.262 (1.028-4.977); p < 0.040; Table 5).
The significant problems when screening is performed for high-risk pregnant women are the skills and knowledge of the physicians and the knowledge of the patients (Table 6). Providing knowledge and skills relating to cervical length measurements for doctors who perform routine work is essential so that they can become certificated and undertake examinations confidently (odds ratio [95% CI], 2.400 (1.130-5.098); p = 0.022; Table 6).       Table 4. Factors associated with an action plan for preventing preterm births in the hospital.    Hierarchical cluster analysis was performed by grouping cases of a similar nature. We assumed that the doctors who reported that a heavy workload was a major barrier were the same as those who mentioned a lack of government funding. The results of the cluster analysis were placed into two groups. A comparison of the respondents' answers is given in Table 7. The answers with statistical significance were "other urgent and necessary tasks", "excessive routine tasks", and "insufficient number of personnel to support the performance of the procedure". Table 7. Cluster analysis by grouping physicians who stated that both heavy workloads and a lack of government funding were major barriers. Cluster #1: physicians who stated that a heavy workload was a major barrier. Cluster #2: physicians who stated that a lack of government funding was a major barrier.

Discussion
Our research found that the rate of cervical length screening at 20 tertiary centers was very low. About 90% of obstetricians were allowed to perform screening even if they had not received formal certification in the procedure. However, they required formal training to develop the knowledge and skills for cervical length measurements. Doing so would enable them to become certified and undertake examinations confidently while performing their routine work.
The current effective preventative measure for preterm deliveries is the use of progesterone [17]. Much research has supported that obtaining cervical length measurements is an effective screening method for pregnant women with short cervices. The procedure has also proven highly cost-effective with few risks [18,19]. Only a small proportion of women with preterm births have risk factors, and many preterm deliveries occur in nulliparous women. Therefore, universal transvaginal cervical length screening has been recommended in order to identify women prone to preterm birth [20].
One of the core barriers to the full implementation of universal screening in Thailand is the excessive volume of routine, urgent, and necessary tasks performed by physicians and nurses. Other perceived major barriers are the following: (1) Some physicians do not believe that the provision of universal screening justifies the requisite labor and funding. (2) There is inadequate funding by government agencies for both screening and the provision of cost-free progesterone.
Therefore, careful reconsideration of the need to perform universal screening is warranted.
Cervical length measurements can be safely performed during fetal structural assessments at 20 to 24 weeks of gestation. A transabdominal cervical length measurement should be offered to pregnant women with strong reservations about undergoing a transvaginal measurement [21,22]. Unfortunately, transabdominal measurements can be used only for some pregnant women [23]. When the procedure is performed, the cervical length will be longer than that determined by a transvaginal measurement. This is because the pregnant woman must have a full bladder in order to enable the ultrasound operator to obtain a clear field of view [23].
Cervical measurements are currently the most effective method, and transabdominal measurements should be reserved for women reluctant to undergo a transvaginal assessment. Regarding the cost-effectiveness of screening programs, transabdominal ultrasound should be performed for low-risk women during a fetal anatomy survey at 19 or 20 weeks of gestation, while the more accurate but relatively costly transvaginal ultrasound may be worthwhile for high-risk populations [24]. This approach has two benefits: First, the additional costs associated with transvaginal screening can be avoided [25]. Second, using dual methodologies improves the possibility that screening can be affordably performed for all pregnant women.
The vaginal administration of progesterone to women with a cervical length of ≤25 mm significantly reduces the risk of preterm birth [26]. The free supply of progesterone should be considered a national policy to prevent preterm births. However, one of many barriers to universal screening is the limited knowledge of the physicians involved in counseling pregnant women. If physicians do not believe in prevention strategies, the need for universal screening, or the benefits of progesterone treatment, screening utilization will be impaired [27,28]. The Maternal and Child Health Board can facilitate the implementation of universal cervical length screening. On the one hand, it could support the funding for training medical personnel in measuring cervical length, as well as the organization of regular training courses on preterm birth prevention for physicians and patients. Furthermore, it could also be responsible for providing the related medicines and medical supplies to all hospitals. These actions would ensure that screening is fully implemented, thereby reducing the preterm birth rate.
A physician's expertise in taking measurements markedly affects the results of cervical examinations. Incorrect results may lead to unnecessary treatment or missed opportunities in preventing preterm births by administering vaginal progesterone. The performance quality and the learning curve associated with obtaining accurate measurements are critically important [29,30]. Providing cervical measurement training to physicians will likely increase the screening rate in many centers.
Our study aimed to identify barriers to cervical length screening at tertiary centers in Thailand, where adequate human, material, and drug resources are available. A limitation is that there was a relatively small number of participants (120 doctors), all of whom worked at tertiary-level hospitals. Therefore, their questionnaire responses may only partially reflect the views of physicians at the many primary-, secondary-, and tertiary-care hospitals throughout Thailand. However, we ameliorated this limitation by randomly assigning the questionnaire to hospitals throughout all six regions of Thailand. The recommendations of our study can be modified for implementation at primary and secondary centers.

Conclusions
There are two major obstacles to achieving universal cervical length measurements. One is the skepticism of physicians that such screenings can stave off preterm births. The other is government agencies' lack of monetary support for hormone usage. Physicians are also unsure about the correct procedures for obtaining cervical length measurements. In order to overcome these barriers: • Workloads should be reduced by extending the screening program to secondary centers. • Government funding should be provided for progestogen usage. • Physicians should be trained in transabdominal and transvaginal ultrasound.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/healthcare11071039/s1. The questionnaire used to obtain physicians' perspectives is available in the File S1: "Questionnaire for Physician's Perspective." Table S1: Context evaluation of tertiary hospitals. Table S2: Availability of resources. Table S3: Impact of preterm births on hospitals, pregnant women, and families. Table S4: Assessment of project inputs. Table S5: Process evaluation of universal cervical length screening program. Table S6: Possible barriers to universal cervical length screening at hospitals. Table S7. Possible approaches to surmounting obstacles. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary Materials.