Document Type : Review Article
Authors
1 a) Assistant Professor, Nursing and Midwifery Care Research Center, Mashhad University of Medical Sciences, Mashhad, Iran. b) Department of Operating Room and Anesthesiology, Faculty of Nursing and Midwifery, Mashhad University Medical of Medical Sciences, Mashhad, Iran
2 a) Assistant Professor, Nursing and Midwifery Care Research Center, Mashhad University of Medical Sciences, Mashhad, Iran b) Department of Midwifery, School of Nursing and Midwifery, Mashhad University Medical of Medical Sciences, Mashhad, Iran
Abstract
Keywords
Introduction
Providing evidence-based health services requires sufficient knowledge to improve medical science and health; randomized trial studies provide valuable information (1).
Numerous Randomized Clinical trials (RCTs) are considered to be the 'gold standard' of evidence-based clinical practice. RCTs have validity for determining the efficacy of new interventions (2). RCTs are often the preferred trial design to evaluate the effectiveness and complications of therapeutic interventions (3). However, various biases may emerge during the designing, implementation, and reporting the stages of these studies (4, 5). In addition, evidence offers that the quality of the published RCTs in medical journals is suboptimal (6, 7). On the other hand, the acceptable reporting quality of RCTs is essential to decision-making of journal reviewers, expert advice and clinical application, and unbiased meta-analysis as an interpretation of evidence. Therefore, to progress the clarity and transparency of RCT reports and recognize the importance of such clarity in reporting of these studies, an international group of epidemiologists, statisticians, and editors developed the Integrated Standards Checklist for Consolidated Standards of Reporting Trials (CONSORT) for RCT reporting in 1996 (8), which was revised in 2010 and 2017. The purpose of CONSORT is to provide a guide for authors to improve the reporting of their trials. CONSORT checklist is a summary of all the essentials that must be reported in an RCT (9, 10).
Evidence shows that despite compiling a consort checklist, adherence to trials guidelines was not sufficient, and even the surveys reported poor adherence to a consort designed for non-pharmacological treatments (11). Chen et al. (2010) which evaluated the quality of abstracts in randomized controlled clinical trial in Chinese medical journals concluded that the quality of abstracts in randomized controlled trial trials needed to be improved (12). The results of the study by Tabatabai et al. (2018) showed that although the reporting quality of clinical trials related to herbal remedies has improved over time in the Middle East, it remains less than desirable (13). The study by Schulz et al. (1994) also indicated that clinical trials in the field of gynecology and obstetrics do not have a good reporting quality (14).
Breastfeeding is a key priority in public health all over the world. Breastfeeding is undoubtedly the excellent route of supplying favorable nutrition for infants. Breast milk is a convened biological fluid and the optimal nutrition for neonates, which provide the most beneficial nutritional balance in terms of quality and quantity (15-17).
According to the literature, the decreased rate and early discontinuation of breastfeeding could adversely affect maternal, neonatal, social health, thereby impose additional costs on healthcare systems. According to the World Health Organization (WHO), United Nations Children's Fund (UNICEF), and the American Academy of Pediatrics, exclusive breastfeeding is recommended as the most favorable manner of feeding for newborns during the first six months of life, and it should be continued until two years old (18-20). Despite that various health organizations have emphasized on the exclusive breastfeeding until six months for neonates, the rate of exclusive breastfeeding is unfavorable in many countries; for instance, the rate of exclusive breastfeeding in Iran has been estimated as 25%. The declining rate of breastfeeding is currently considered to be a main public health concern (21-25).
Given the importance of breastfeeding and its impact on the growth and development of infants, the promotion of breastfeeding has recently gained special attention worldwide with increasing human knowledge and growing body of research. Today, several approaches are available for the promotion of breastfeeding; one of the most effective approaches is to provide counseling-supportive services and breastfeeding training (26, 27). It has been suggested that family-centered training and support (especially prenatal support) could promote exclusive breastfeeding (28). In the study by Gholami Tabar et al. (2011), they reported that providing prenatal training, support system, and counseling before and after childbirth could enhance breast milk consumption in the infants (29). Similarly, other studies have indicated that properly training of pregnant women encourages breastfeeding and prolongs exclusive breastfeeding (30, 31).
Numerous RCTs have demonstrated the effectiveness of breastfeeding counseling and training in promotion and improvement of exclusive breastfeeding. Since the use of evidence-based care is critical in clinical interventions and breastfeeding care, and given the importance of RCT papers in evidence-based care, it is essential to appraise the quality of published studies in order to apply the results of RCTs in breastfeeding counseling and training. The present study was performed aimed to critically appraise the reporting randomized controlled trials which examined the effect of counseling-educational interventions on exclusive breastfeeding.
Materials and Methods
This critical review study was performed in 2021 to evaluate the quality of the RCT reports in Iran, and evaluate the effect of counseling-educational interventions on exclusive breastfeeding, which have been published in reputable databases.
Related articles were identified via searching in the English databases including Scopus, PubMed, Cochrane and Google Scholar as well as Persian databases of SID, Magiran, IranDoc, using the keywords in title and abstract such as (Breastfeeding OR Breast Feeding OR Lactation OR Human Milk) AND (Counseling OR education OR Training Programs) AND (Exclusive Breast Feeding OR Exclusive Breastfeeding OR breastfeeding exclusivity OR Breastfeeding Status) AND (Randomized Controlled Trials OR Randomized Clinical Trials OR clinical trial OR RCT) and their Persian equivalents with Boolean OR and AND operators. All the published articles were retrieved until the outset of the search phase. The search and selection processes of the trials were shown using the PRISMA flowchart (Figure 1).
The data extracted from the studies included in this study were presented in Table 1.
The inclusion criteria were all the RCTs published in Iran which examined the effect of counseling-educational interventions on exclusive breastfeeding. The exclusion criteria were letters to editor, conferences papers, and dissertations. All the articles were assessed by two independent reviewers, and the extracted data were recorded in the relevant forms. It is notable that in case of difference between the reviewers, the subject was re-examined with consensus in the presence of one observer.
Data collection forms included demographic information (journal name, year of publication, language, number of authors, specialty, and first author's degree and academic rank) and CONSORT checklist 2010.
The CONSORT checklist is a commonly used important assessment tool in clinical trial articles, which consists of 25 items and six main sections of clinical trial studies (Title and Abstract, Introduction, Methods, Results, Discussion and other information), each one consists of various subsets (9). In the scoring of checklist, each item is assigned one point if it is mentioned in the checklist and zero point if it is not.
The total scores assigned to the reported items are considered as the total quality score of the article, the minimum and maximum scores for each article is determined as 0 and 37, respectively.
Data were analyzed by SPSS (version 16) using descriptive statistics (Mean ± SD and Number (Percent)).
Results
In the initial search, 6095 articles were obtained. After removal of duplicate and irrelevant articles, 31 full-text articles were retrieved to assess eligibility and ultimately, 17 articles were criticized. The mean total score of quality was 18.35±3.33, with the minimum and maximum scores of 12 and 25, respectively. None of the published randomized clinical trials which examined the effect of counseling-educational interventions on exclusive breastfeeding reported all the sub-items listed in the CONSORT checklist.
According to the information presented in Table 2, among 17 published clinical trials, only one study (5.9%) was listed under the heading of 'randomized trial', while in 16 articles (94.1%), the item was not observed relied on the CONSORT checklist. However, 16 articles mentioned the use of randomization method. In the Method section, the sample size calculation was not reported in 12 articles (70.6%), the method of random allocation sequences was not reported in 11 articles (64.7%), and the type of randomization was not reported in 14 articles (82.4%). In the Results section, 9 studies (52.9%) provided no data on the number of lost samples and those excluded after randomization in each group. Table 3 showed the scores of each article divided by the six sections of the CONSORT tool. Table 4 showed the frequency distribution of the reported cases of the CONSORT items in the published clinical trials about the impact of counseling-training interventions on exclusive breastfeeding, as well as the mean, minimum, and maximum scores of the six sections CONSORT checklist (Title and abstract, Introduction, Methods, Results, Discussion, Other information).
Discussion
This study was performed aimed to critical appraise the reporting randomized controlled trials which examined the effect of counseling-educational interventions on exclusive breastfeeding based on the CONSORT checklist 2010. The overall compliance rate of the articles with the items in the CONSORT checklist was 49.2%. According to the findings, no studies were found based on international checklists to review the RCT reports on breastfeeding. Therefore, the researchers compared the results of the present study with the past findings regarding the critique of the clinical trials in midwifery. The appraised articles in the study by Sharifi et al (2021) had an average quality and the overall quality compliance rate with the CONSORT checklist which was estimated about 50% (49). Considering that Sharifi's study used the 2010 Consort, which has more details than the previous versions, the report presented on the desired quality of the articles can be promising. In the study conducted by Irani et al. (2017), the RCTs which examined the effects of massage on the severity of labor pain were assessed, and the compliance rate of the reviewed studies was estimated as 54% (50). In the studies performed by Sarayloo et al. (2018) and Bahri et al (2016), which reviewed the articles published on menopause, poor quality of the RCT reports has been reported (51, 52). Moreover, Talachi et al. (2012) confirmed the poor quality of RCT reports (53), Goenka et al. (2019) reported the overall compliance of the RCT articles in Indian medical journals is 54%, which was extremely poor (54).
In the assessment of the RCTs published in 2013 in JAMA and BMJ journals, Susvirkar et al (2018) reported a wide range for the percentage of the total score (PTS) of the articles (59.4-97.1%), showing the poor compliance rate of the articles (55). Based on the aforementioned studies, it could be concluded that the quality of the clinical trial reports based on CONSORT checklist is not favorable, which could be due to the inadequate knowledge of researchers regarding the methodology of clinical trials and their negligence to comply with CONSORT checklist in article writing. However, since the authors had access to the latest version of the Consort tool at the time of their study, they had to use it to report their findings. In addition, lack of emphasis of journal guidelines on the need to use the CONSORT items in writing reports may be the other cause of poor quality of RCT reports.
According to the present study, only one article (5.9%) mentioned the phrase of 'randomized trial' in the title. Not using the word 'RCT' could cause the study not to be classified as a clinical trial by the index (56). In the study by Joukar et al (2015), only 17.1% of the studies mentioned the word of 'RCT' in the title of the article (57). Moreover, in the studies by Irani et al (2017) and Sarayloo et al (2018), this rate was reported to be 0% and 14.9%, respectively (50, 52). The negligence in this regard could be diminished by the emphasis of journal editors and reviewers on the accurate and thorough implementation of the CONSORT checklist.
According to the current research, the lowest compliance rate in the published RCT reports regarding the effect of counseling-training interventions on exclusive breastfeeding belonged to the sections of Methodology (39.4%) and Results (48.82%). In the Methodology section, the compliance rate was zero in sub-items 3b, 6b, 7b, and 9. In fact, none of the assessed articles provided reports on the possible changes in the methodology and subsequent trial outcomes after the intervention. In the study by Wandalkar et al. (2019), they claimed that the difference between registered clinical trials and the final published articles in journals has a high effects coefficient, which necessitates the implementation of the CONSORT guidelines and presenting the reports on all the changes in the implementation or outcomes for all the authors and editors after starting trial (58). Furthermore, Susvirkar et al (2019) reported the compliance rate of the mentioned sub-items to be 2.1% (55). Zero compliance could be justified by two reasons; first, there are no differences between the reviewed articles in terms of methodology and trial outcomes after the initiation of the trial, second, lack of compliance could be due to the non-commitment of the authors to reporting changes.
In the Method and Randomization sections, the reporting rate of sub-items 8a and 8b was 35.3% and 17.6%, respectively, which was considered to be low. In addition, none of the assessed articles described sub-item nine, and their compliance rate was determined to be zero. Failure to apply the randomization principle or inaccurate description of the randomization report could definitely lead to bias and method distortion in RCT studies. In the study by Irani et al (2017), none of the reviewed articles referred to the subjects of randomization method and random allocation method (50). In the research performed by Samaan et al (2013), the reporting rate of randomization methods was estimated as 5-9% (59), while it was reported to be 23% in the study by Goenka et al (2019) (54). In addition, in the study by Sarayloo et al (2019), the randomization rate with its details was reported to be approximately 46.55% (52). Gupta et al (2022) in a critical review of trial studies that published in Indian journals demonstrated that some items in method statement were underreported including allocation concealment method and analysis of the data based on the Intention-to-treat (ITT) (11). According to the aforementioned studies, it could be inferred that many RCT authors are overlooking the importance of randomization, implementation, and accurate reporting of the research process, and even over time the accuracy of randomization reporting remains dissatisfactory due to the availability of tools such as CONSORT.
Randomization is a highly precise technique in RCT studies, and it is not sufficient to only refer to the word 'randomization' in the Method section; therefore, authors must also refer to the details of the randomization process (51). The other sub-items of the Method section (7a, 10, 11, and 12b) also had poor compliance. Among the 17 sub-items of the Method section in the current research, only six cases (35.3%) had higher compliance rates than 60%.
In the Results section, the report on sub-items 14b and 17b was determined to be zero, and none of the authors referred to these sub-items in their articles. Therefore, it could be inferred that none of the researchers faced with the potential problem of the untimely completion or termination of the study. In RCT studies, it is essential to provide an abstract of the statistical results in each study group and describe the difference between the groups, which is familiar as the effect size. As mentioned earlier, none of the studies provided data on the relative and absolute effect sizes. Susvirkar et al. (2019) has also stated that the effect size report is only 19% (55). Failure to report the relative and absolute effect sizes may be due to the inability to use the statisticians' views. Therefore, it is recommended that clinical science experts apply the views of statisticians and methodology of experts to design and conduct RCT studies.
In the present study, none of the assessed articles referred to the 'all important harms' or 'unintended effects' in each of the study groups in the Conclusion section, and the reporting rate was determined to be zero. This may be due to the nature of interventions in the areas of counseling and training (as opposed to the interventions such as procedures or drugs' prescription), as well as the fact that the researchers of the evaluated studies have not considered the potential risks of counseling and training interventions. The second reason is that the participants were exposed to any risks after the intervention.
Susvirkar et al., (2019) reported the harms of approximately 70% in their study (55). In the research by Joukar et al., (2015), adverse and important events of the intervention was only 6.8% (57). In addition to the recognition of the benefits of interventions, the articles' readers need to be aware of its possible risks. As such, the authors must accurately refer to the potential risks during the study, as well as the number of the participants that have been excluded due to complications (56). Moreover, describing the risks of interventions prevents the misinterpretation and misleading of the readers of the study's findings (57).
In the discussion section of the present study, only the 'limitations' sub-item had a low reporting rate (23.5%), which is considered to be poor compared to the study by Susvirkar et al., (2019) which reporting rate was 97.9% (55). Some journals designate specific structures for authors in the discussion section, and 'limitations of the study' is a component of such structures; this component must be mentioned by authors, along with methods to be used for the reduction of risks or their compensation.
The reporting rate of the 'trial registry' sub-item in the current research was estimated as 47.1%. In the study by Bahri et al. (51), the reporting rate of the 'trial registry' sub-item has been determined to be zero, while it has been reported to be 46.8% in the research by Sarayloo et al. (52) and 93.8% in the study by Susvirkar et al. (55). The International Committee of Medical Journal Editors in 2004 announced that the registration of trials before recruitment is essential for their publication. Nevertheless, the reporting rate of RCT studies remains low (56).
Critical evaluation is influenced by the individual's point of view who in this regard tried to overcome it by using a checklist as well as reviewing the scoring of articles and using a third researcher in scoring.
Conclusion
The quality of reporting RCTs regarding counseling-training interventions for breastfeeding is not favorable in Iran. Considering that clinical trials could be the optimal source for evidence-based practice, there is an urgent need for more commitment of researchers, authors, journal editors, and reviewers of health sciences journals to RCT studies. In addition, it is recommended that researchers comply with the use of various tools and CONSORT for training, implementation, and evaluation in RCT studies. Therefore, it is suggested that valid tools be designed and developed for research assessments by researchers, new statements and guidelines be accurately and thoroughly applied by authors, and instrument-based articles be reviewed properly and accurately by journal editors and reviewers.
Acknowledgements
This study was approved by the Local Ethics Committee, Mashhad University of Medical Sciences, Mashhad, Iran (code: 980162). The present study was adapted from the project confirmed by Mashhad University of Medical Sciences, Mashhad, Iran. Hereby, the research assistant of the university is sincerely appreciated.
Conflicts of interest
Authors declared no conflicts of interest.