Article Text
PDF
Abstract
Objectives To evaluate the methodological quality of randomised controlled trial (RCT) abstracts in leading critical care nursing journals against the Consolidated Standards of Reporting Trials-Abstracts (CONSORT-A) checklist and to identify variables related with abstract reporting quality.
Design Descriptive methodological quality review.
Data source We searched the PubMed database and the websites of each included journal.
Eligibility criteria We included RCT abstracts published between 2011 and 2021 in the first 11 Scopus-ranking (2021) critical care nursing journals that reported the results of RCTs in English that referred to the care of adult patients with acute/critical illness or their relatives and/or caregivers and conducted in ICUs.
Data extraction and synthesis Two independent investigators extracted the data using a prespecified 17-item checklist directly derived from CONSORT-A. For the primary outcome, each item was evaluated whether it was adequately reported or not and descriptive statistics were reported. An overall score was calculated by summing the results of all items and multivariate linear regression was conducted to detect potential predictors.
Results Seventy-eight RCT abstracts were included in this review. The items with the highest CONSORT-A adherence were authors, objective, conclusion, participants, interventions and outcome. The randomisation item had the lowest CONSORT-A adherence, followed by trial registration, funding source, harms or side effects, recruitment, blinding and outcome results. The average CONSORT-A score was 8.5±1.5 points (of the maximum 17.5 points). Multivariate analysis indicated that the categorised word score and publication date were positively associated with the overall CONSORT-A score, while the first author’s country being in Asia was negatively associated with the overall CONSORT-A score.
Conclusions The 78 RCT abstracts showed poor overall adherence to CONSORT-A. The results indicate that the methodological reporting quality of RCT abstracts in critical care nursing journals requires improvement to facilitate assessment of the applicability and relevance of the results reported.
- Adult intensive & critical care
- STATISTICS & RESEARCH METHODS
- EPIDEMIOLOGY
Data availability statement
The data that support the findings of this study are available from the corresponding author (MV), on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
This study is a systematic evaluation of the quality of randomised controlled trial (RCT) abstracts published in critical care nursing journals between 2011 and 2021.
The evaluation instrument is directly derived from the Consolidated Standards of Reporting Trials-Abstracts (CONSORT-A).
The search methodology for RCTs in PubMed is sensitive and reproducible.
Two reviewers independent screened all bibliographic records identified by the search string and assessed the compliance of each abstract with CONSORT-A using a standardised method.
Because we gave equal weighting to each of the 17 CONSORT-A items, we could not weight the different contributions of the most significant items in the overall quality score.
Introduction
The abstract is a key element of scientific reporting. It has to include core study information so the readers can quickly assess the originality of the topic, the method, the validity and the generalisability of the results for possible application of these results in their clinical context.1
The abstract is usually the first section read in a scientific publication and determines the reader’s decision to continue in reading. Often many practitioners read only the abstract due to the difficulty and/or cost of retrieving the full text or the limited time available. However, inappropriate and/or incomplete data reporting could lead to improper application of results with negative consequences for patients.2 3
In recent decades, the emergence of evidence-based practice has witnessed the increasing application of randomised controlled trials (RCTs) by the nursing discipline to estimate intervention effects, representing the gold standard for treatment questions and for evaluating the effectiveness of a nursing intervention.4
The reporting could also represent a problem for the nursing practice, which is often based on experience, tradition and intuition rather than scientific validation practice, even if in recent decades there has also been an increasing application of evidence with the conduction of nursing RCTs.5 6
RCTs’ lack of methodological rigour and adequate reporting leads to bias in estimates of intervention efficacy.7 8 In particular, non-relevant reporting prevents a reliable assessment of the study method and can also determine bias.9 Moreover, adherence to reporting guidelines can minimise bias, improve reproducibility, and facilitate clear, transparent and comprehensive reporting of research results.10
In 2008, the Consolidated Standards of Reporting Trials-Abstracts (CONSORT-A)11 was developed, with 17 items that authors have to consider during the reporting of RCT abstracts. Subsequently, this checklist was integrated into the 2010 CONSORT statement to promote clear, transparent and detailed RCT abstracts.11 12
The aim of the review is to understand the abstract reporting quality of RCTs in critical care nursing. This is because many RCT abstracts published in nursing and other scientific journals, often present insufficient information and do not include the key points for high-quality reporting.13–16 Several methodological reviews have noted various critical issues specifically related to the reporting of primary outcome results, randomisation methods and allocation concealment.17 18
Methods
Study design
The study was a descriptive methodological review of the quality of abstracts of RCTs published in critical care nursing journals.
Objective
The primary aim of the study was to estimate the reporting quality of RCT abstracts published in the last 10 years in the highest-ranking critical care nursing journals by examining the adherence to the CONSORT-A checklist items. Second, the study aimed to investigate the possible factors associated with abstract reporting quality.
Inclusion and exclusion criteria
To conduct the RCT abstract review, research was conducted in critical care nursing journals according to the 2021 Scopus ranking considering first, second and third quartile and as indexed in MEDLINE. In order of ranking, the journals included were as follows: (1) Nursing in Critical Care; (2) Intensive and Critical Care Nursing; (3) Australian Critical Care; (4) American Journal of Critical Care; (5) Critical Care Nurse; (6) Dimensions of Critical Care Nursing; (7) Critical Care Nursing Clinics of North America; (8) Journal of Trauma Nursing; (9) Enfermería Intensiva; (10) AACN Advanced Critical Care and (11) Critical Care Nursing Quarterly. Given that AACN Advanced Critical Care replaced the journal AACN Clinical Issues in 2013, the search also included the name of the previous journal.
Inclusion criteria
The inclusion criteria were as follows: abstracts reporting the results of RCTs of parallel and/or cross-over investigations; English written; conducted in the ICU (Intensive Care Unit) setting, referred to critically ill adult patients care, included their relatives and/or caregivers.
Exclusion criteria
The exclusion criteria were: manuscripts reporting the results of drug RCTs; pilot or feasibility studies; cluster trials; observational or cohort studies; interim analyses with intermediate results; economic analyses of RCTs; post-trial follow-up studies; subgroup and secondary analyses of previously published RCTs; editorials and RCTs without an abstract, such as RCTs published as letters to the editor and single-subject clinical trials.
Search strategy and study selection
In June 2022, an in-depth search was conducted in MEDLINE via PubMed for RCTs published between 1 January 2011 and 31 December 2021 in the journals included. The search was performed by two authors experienced in literature searching (MV and TC) and used the Cochrane Highly Sensitive Search Strategy for identifying randomised trials19 in MEDLINE adapted by removing terms referring to drug therapy and using a temporal filter (online supplemental file S1). To improve sensitivity, a second research was performed on the website of each journal via ‘Advanced Search Mode’ using the words ‘RCT’ in the title field.
Supplemental material
Two authors (MV and TC) independently examined the bibliographic records identified by the search string to include the RCT studies. When the title or abstract was insufficient to determine eligibility, the full text was evaluated. Disagreements were resolved through discussion with a third author (MB) until consensus was reached. The identified abstracts were archived in EndNote V.X20 (Clarivate Analytics) and subsequently imported into an Excel data set. The double abstracts were removed with the appropriate function of the reference management software.
Data extraction and assessment
The abstract reporting quality was assessed with the 17 items of CONSORT-A.11 Each point was assigned a score of 1 if it was adequately reported and 0 if it was valued as inadequate. For each abstract, the overall score (CONSORT-A score) was calculated by summing the results of all items.
Despite the fact that in the original 200811 version, it was only intended for conference abstracts, item 2, ‘authors’, was included to be evaluated; in 2010, the same item was included in the CONSORT-A checklist.20 From the experiences and reflections reported in the literature by various authors of methodological analyses in different health journals1 21 22 and according to the present research team, specific indications were established, to start the review and also to evaluate the items: (A) 4 ‘participants’; (B) 8 ‘randomisation’; (C) 9 ‘blinding’ and (D) 13 ‘outcome results’.
In item 4, ‘participants’, both the presence of the inclusion–exclusion criteria and the description of the setting were assessed. One point has been attributed where described in the abstract while, if only one of the criteria is present, 0.5 points has been assigned.
Item 8 was considered adequate with the attribution of 1 point when the randomisation details were specified in the text, that is, concealment and randomisation list generation method. If only one of the two aspects was present, 0.5 points were assigned.
Item 9, ‘blinding’, was rated as partially compliant (0.5 points) when the presence or absence of blinding was generically reported but without specifying the subject (patient, caregiver, outcome assessor). Full marks (1 point) were awarded when the authors specified which subjects were blinded.
In item 13, ‘Outcome results’, 0.5 points were assigned to each of the three key elements: (1) primary outcome of each group (ie, number of participants with or without the investigated outcome); (2) effect size in relation to the primary outcome (eg, relative risk, OR, risk difference, HR or median survival time, mean difference); (3) CI of the contrast between the two groups (95% CI limits). Because of these 13 items only, the maximum score was 1.5 points. Given these specifications, the CONSORT-A range score was 0–17.5 points (online supplemental file S2).
Two authors (MV and MLP) blindly and independently assessed the compliance of each abstract with CONSORT-A using a structured form (online supplemental file S3). Any critical issues were analysed and the other two expert reviewers (MB and TC) were consulted until consensus was reached.
Abstract characteristics
The following data and descriptive information on each RCT abstract were collected to analyse the secondary end points: publication year, field of investigation, type of intervention (device/substance, relational/educational, care intervention), first author’s country, word count, number of authors, multicentre study, structured abstract, study population (nurses, relatives, patients) and sample size.
The journal websites were also searched for any support from reporting guidelines (ie, CONSORT or EQUATOR network - Enhancing the QUAlity and Transparency Of health Research) because not all journals required compliance with a specific guideline.
Statistical analysis
Assuming a symmetrical distribution, the continuous variables are presented as the mean±SD or the median and IQR. The categorical variables are summarised as frequencies and percentages. Inter-rater agreement was studied using the Cohen’s kappa between the evaluators for the abstract inclusion and the evaluation of each checklist item. We assessed the adherence of abstracts to the items of CONSORT-A and calculated the overall proportion for each item (percentage and 95% CI).
To identify the association between abstract reporting quality and potential predictors we initially performed a univariate analysis with simple linear regression with the following variables: (A) scientific journal; (B) publication year; (C) intervention type under study (device/substance, relational/educational, care treatment); (D) first author’s country; (E) word count (< 200, 200–250, 250–300, >300); (F) multicentre study; (G) structured abstract; (H) subject number; (I) study subject and (J) number of authors.
Factors that were individually significant with p<0.1 were included in the multivariate modelling. Multicollinearity was assessed by calculating the variance inflation factor (VIF) for each variable in the model and only variables with VIF <10 were accepted. A p<0.05 was considered statistically significant. The statistical analyses were performed using SPSS Statistics V.22.0 (IBM).
Patient and public involvement
None.
Results
The study initially identified 953 records in the 11 critical care nursing journals (figure 1); of these, 272 duplicates were excluded, while 80 were excluded because they did not meet the inclusion criteria, 493 because the studies were not RCTs and 30 were not assessable according to CONSORT-A (pilot, feasibility or cluster studies). For this reason, the review was conducted in 78 RCTs with a kappa=0.93 of the researchers’ agreement on the abstract inclusion criteria.
PRISMA flow diagram depicting the study selection procedure. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses35; ICU, Intensive Care Unit.
Of the 78 abstracts included, 68 (88.5%) reported being single-centre studies, 53 (68.8%) of the included abstracts reported a favourable and statistically significant result for the intervention under study, and 34 (43.6%) of the studies were conducted in an Asian region (table 1, figure 2).
Characteristics of included abstracts
Characteristics of included abstracts: region, publication year and Journal title.
The author guideline consultation performed on 1 June 2022 on the 11 journal websites revealed that just under half of the journals (5/11, 45.5%) clearly indicated that they promote and/or adopt the CONSORT statement for RCTs or refer to the EQUATOR network for guidelines; only 2 journals (18.2%) require registration with international trial registries for publication. Additionally, 97.4% (76/78) of the abstracts were structured and 78.2% (61/78) contained from 200 to 300 words.
The percentage match between evaluators was high (>90%) for 12 out of 22 items assessed, with a range from −0.013 to 0.893 of Cohen’s kappa (see online supplemental file S4).
Overall, the average CONSORT-A adherence score was 8.5 ± 1.5 points out of a maximum 17.5 points. The lowest CONSORT-A adherence (<20%) was recorded for: (1) randomisation (item 8) (0/78, 0%; however, 2/78, 2.6% of the subitems complied with random sequence generation and 0/78, 0% of the subitems had allocation concealment); (2) trial registration (item 16) (3/78, 3.8%); (3) funding source (item 17) (0/78, 0%); (4) harms or side effects (item 14) (7/78, 9%); (5) recruitment (item 11) (12/78, 15.4%); (6) blinding (item 9) (9/78, 11.5%; however, 23/78, 29.5% blinding and 9/78, 11.5% subject blinding) and (7) outcome results (item 13) (7/78, 9%). Of the outcome results (item 7), the subitems were primary outcome (39/78, 50%), effect size (15/78, 19.2%) and precision (10/78, 12.8%).
The authors (item 2), objective (item 6) and conclusion (item 15) domains were reported in all abstracts. The other items with the highest adherence were participants (item 4) (56/78, 71.8%), interventions (item 5) (76/78, 97.4%) and outcome (item 7) (63/78, 80.8%). Moreover, 35.9% (28/78) of the articles did not identify the study design as an RCT (see table 2; online supplemental file S5).
Adherence to CONSORT-A in the critical care nursing journals
The univariate analysis showed that the only factors significantly associated with the overall CONSORT-A score are word count, publication date and first author’s country was in Asia. Including these variables, the multiple linear regression model statistically significantly predicted the CONSORT-A score: F(6, 71)=4.944, p<0.0001, adjusted R2=23.5% (see table 3).
Univariate and multivariate linear regression of factors associated with the CONSORT-A score
No evidence of multicollinearity was found in the final model, with tolerance values below 10. Multivariate analysis indicated that a higher CONSORT-A score was positively associated with a higher abstract word count (β=0.45; 95% CI 0.068 to 0.833) and the publication year (β=0.17; 95% CI 0.061 to 0.28), and was negatively associated with the first author’s country being in Asia compared with authors based in Europe (β=−1.08; 95% CI −1.846 to −0.319) (figure 3).
Factors associated with CONSORT-A. CONSORT-A, Consolidated Standards of Reporting Trials-Abstracts.
Discussion
The strength of evidence emerging from robust RCTs can strongly influence clinical practice. It is, therefore, crucial that abstracts strictly follow the CONSORT-A checklist without under-reporting. The principal aim of the present methodological review was to examine the CONSORT-A checklist adherence of RCT abstracts published between 2011 and 2021 in critical care nursing journals.
Overall, the abstracts included in the review are not so adherent to the CONSORT-A items, with more than seven of the 17 items insufficiently reported (<20%). However, these results are in line with the findings of other authors in both nursing and other health science journals, which highlight a modest accordance within CONSORT-A and the structure of the RCT abstract papers.17 18 23
This review highlights that in some cases the literature on nursing in critical care, particularly in the RCT papers, may be affected by un underestimation and/or a poor CONSORT-A guideline implementation by authors and scientific journals. This could determine an abnormal transfer of the theory into practice, caused by an incomplete methodological reporting of the evidence obtained from the studies.
Nevertheless, the results indicated a trend over time towards an improvement of the methodological rigour. Indeed, the multivariate analysis determined that the publication year was a variable independently associated with the abstract quality (β=0.17 per year; 95% CI 0.061 to 0.28). This trend was similar to that reported by Guo and Iribarren,18 who analysed 68 abstracts in oncology nursing research (multiple linear regression β=0.05, 95% CI 0.01 to 0.05, p<0.05), in other research in anaesthesiology17 and in other specialty journals.24 It should be noted that the number of published RCTs overall has also gradually increased over the years, contradicting the common belief that RCTs are not useful for nursing science but only for other healthcare disciplines.25
The word count is also positively associated with the reporting quality. Given the difficulty in determining an upper limit, the CONSORT-A statement considers 250–300 words sufficient to report all elements of the checklist.11 In nursing science, this need may be even more significant, as the description of nursing interventions and their evaluation may be more complex than other disciplines—for example, pharmacological treatments. Other similar studies have also identified this positive association between the word count and abstract quality.18 26 Knippschild and colleagues analysed 212 abstracts of dental implantology trials and reported that the word count had a significant impact on the number of CONSORT criteria reported in the abstracts (Poisson regression-based incidence rate ratios: 1.001, 95% CI 1.001 to 1.002).24 However, it should be noted that the choice of word count does not depend only on the author’s decision, but that different journals define different criteria, which have changed over time.
The third factor that influences the abstract quality, in addition to the publication year and word count, is the geographical area. This regional difference in study reporting was also already mentioned by Khan and colleagues of top-tier cardiovascular journals, where it emerged that the ‘mother culture’ could also influence the way of writing scientific papers.3 14
In the study analysis, the best-reported items, according to CONSORT-A, were ‘authors’, ‘intervention’, ‘objective’, ‘outcome’ and ‘conclusion’. This is consistent with the only available analysis in nursing science by Guo and Iribarren,18 who nevertheless also identified the items ‘participants’ and ‘funding’ among the best-reported criteria (>80%). In the abstracts included in the review, the item ‘participants’ is positively reported (71.8%) but is penalised by shortcomings in reporting the subitems ‘study setting’ (75.6%) and ‘eligibility criteria’ (84.6%), which combined thereby reduced the overall score.
One of the worst-reported items is ‘blinding’. Almost 30% of the abstracts generically indicated the presence of blinding (masking) but only 11.5% comprehensively fulfilled the required criteria. It should be specified that, in contrast to other disciplines, blinding study participants (ie, patients, nurses, assessors and/or caregivers) is not often possible or appropriate in nursing interventions. However, in these cases, the authors ought to indicate the form of blinding used, such as blinding of the outcome assessors only, to minimise bias and maximise the validity of the results.11 This methodological aspect should be emphasised in the nursing abstract reporting as a meta-epidemiological study highlight in which it emerged that RCTs without blinding overestimated the treatment effect sizes by an average of 0.56 SD.27 Guo and Iribarren also noted this lack of attention, reporting only a 2.6% adherence in cancer nursing trials.18
Another aspect that emerged is that only two abstracts (2.6%) reported the generation of the randomisation sequence and no abstracts included detailed the allocation concealment. Masking the allocation sequence prevents investigators from deliberately or inadvertently influencing the participant allocation to one intervention group rather than to another. For the authors, briefly mentioning these aspects in the abstract would allow the reader to quickly understand some information about the risk of selection bias. Inadequate allocation list concealment reporting was associated with a 7% result overestimation.28 This difficulty was estimated to a lesser extent by Guo and Iribarren in cancer nursing trials, with only 28% of adequate reporting in this item. This critical issue was also present in the RCT abstracts published in other high-impact factor health science journals.1 Khan and colleagues recently reported that only 3.6% of the abstracts published in cardiovascular journals adequately reported the randomisation specifications required by CONSORT-A.3
Even though in nursing care, experimental treatments can produce undesired and unintended effects in addition to the expected benefits,29 clinicians need information about the usefulness and effectiveness of the interventions and the possible undesired effects. Briefly describing the extent of possible adverse or unexpected events in the abstract or specifying their absence would be very helpful for the reader.11 30 The fact that only 9% of the studies reported adverse events in the abstract is particularly concerning. Reporting the risks clearly in the abstract is particularly useful both for the reader and for facilitating effective querying of biomedical databases.11 22 This gap overlaps in percentage terms with the report by Guo and Iribarren18 of 2.6%, but was significantly worse than the 20% reported by Janackovic and Puljak in the abstracts of seven of the highest-impact anaesthesiology journals.17
In the item ‘outcome results’ reporting, critical points emerged particularly for the statistical measures of effect size and precision. Often, authors only report conclusions such as ‘there was no difference’ or ‘there was no association’ because the p value was larger than a threshold, such as 0.05. Frequently, the estimated risks, ORs, HRs, mean differences and CIs are not reported. This could be misleading to the reader, to whom the correct information for evaluating the study results is not provided, based on the recent stance of the American Statistical Association regarding the misuse of statistical significance and p values.31 Adequate reporting of the results through the use of effect size indicators and relative CIs would enable assessment of whether the results refute or support the hypothesis under study, thereby avoiding an interpretation exclusively based on p values in a conventional and dichotomous manner.
In 2014, the AllTrials campaign exhorted all those involved in ‘medical’ research, and therefore, also nurses, to ensure that all clinical studies were registered and that all results were reported.29 Unfortunately, only three abstracts in this study referred to the trial registration, although the other RCTs could nevertheless have reported it in the main text. Given the importance of this point and the minimal effort that would be involved in its inclusion in the abstract, stakeholders involved in review and publication should address this quickly. However, as the item ‘funding’, which had a very low compliance rate, the trial registration might also not be reported within the abstract because journal editors typically require these details to be mentioned separately within the publication and not in the abstract.24
Another consideration arises from the fact that the titles of only one-third of the abstracts define the research as a randomised study and indicate that the participants were randomly assigned to the comparison groups.11 This deficiency not only hampers the clarity and immediacy of the abstract but also limits the ability to identify the report as relevant within an indexed electronic database. This difficulty has emerged to varying degrees in nursing literature14 15 18 but to a lesser extent in other health science journals.3 17 32
One expected criticism is that abstract quality analysis is of marginal relevance for nursing research, as abstracts are generally considered a secondary part of trial reporting. This belief is erroneous for two reasons: first, because abstracts are an essential part of articles and should help the reader quickly identify the purpose of the article and anticipate its essential points33; second, abstracts are crucial for the initial literature screening to identify which studies should or should not be included in meta-analyses or systematic reviews. Inadequate reporting would therefore prevent an RCT from being included in a review. Moreover, given that not all journals are open-access and many continue to be based on a paid subscription model,34 abstracts remain the only rapid open-access resource. Abstracts are also a frequently used resource for clinicians to implement decision-making, overcoming potential economic, technical, time and language barriers that may prevent or reduce the use of the article’s full text.2 22
This study has some limitations. First, some items were clearly more important than others—for example, the methodological items of randomisation, blinding or funding as compared with authors.1 The evaluation instrument used in this study is based on the presence of the 17 items reported and is therefore unable to account for other methodological differences. However, the reported analytical results and discussion present the possibility of evaluating each individual item; this method has been used extensively for analysing the abstracts in several health science journals.1 22
Second, we could not include in the multivariate analysis two variables recognised in the literature as being associated with better reporting—specifically, the impact factor, because some journals do not have one, and the journal support for CONSORT, because the exact date of membership of each journal is not available. For this reason, including the abstracts with these missing variables in the multivariate analysis determines their exclusion from the analysis, significantly reducing their power.
Another limitation is the fact that the present quality-reporting study is limited to abstracts published in critical care nursing journals. The journals included in this study are predominantly read by critical care nurses and therefore represent an important vehicle for transmitting knowledge, professional models and research in this nursing area.
Despite these limitations, the present analysis has several strengths: (1) it seems to be the first evaluation of the quality of RCT abstracts published in critical care nursing journals; (2) the methodology of searching RCTs in PubMed is very sensitive and reproducible; (3) the evaluation checklist is directly derived from CONSORT-A and (4) the inter-evaluator agreement after the training showed a high degree of concordance.
Conclusion
In recent years, an increasing number of studies have used an experimental design to measure nursing intervention effectiveness. The present work revealed the problem of the low methodological quality of RCT abstracts reported in critical care nursing journals, a problem that has been reported and described in other healthcare journals but that has not been described in this discipline.
Based on these findings, a greater awareness of the problem among all stakeholders involved is desirable for improving adherence to RCT reporting checklists, particularly regarding the methodological domains. This would be an additional element to ensure that nursing studies are rigorously conducted and in accordance with international standards for such research.
Data availability statement
The data that support the findings of this study are available from the corresponding author (MV), on reasonable request.
Ethics statements
Patient consent for publication
Acknowledgments
The authors are grateful to Paula Deelen and Stefano Bernasconi (Cardiocentro Ticino Institute – EOC) for their support in the preparation of this manuscript.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors Conception and design of the study: MV and MB. Screening abstracts searched through PubMed and journal websites: MV and TC. Data extraction for evaluation of the abstract quality and performed data entry: MV and MLP. Abstract assessment: MV and MLP. Statistical analysis: MV. Data interpretation: all authors. Drafting of the article: MV, MB, TC. Final approval of the article: all authors. MV is responsible for the overall content as guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.