Is the Ratio of Retracted Iranian Papers Compatible With Their Research Growth Rate? An International Database Survey

1Gastroenterology and Liver Diseases Research Center, Research Institute for Gastroenterology and Liver Diseases, Shahid Beheshti University of Medical Sciences, Tehran, Iran 2Department of Anesthesiology and Critical Care, School of Medicine, Hamadan University of Medical Sciences, Hamadan, Iran 3Department of Anesthesiology and Critical Care, Khatamolanbia Hospital, Zahedan University of Medical Sciences, Zahedan, Iran 4Trauma Research Center, Nursing Faculty, Baqiyatallah University of Medical Sciences, Tehran, Iran


Background
If it is determined that the published studies have significant defects, these studies should be retracted in order to avoid misleading potential readers. 1 Retraction is a mechanism for the correction of literature and a warning to readers in relation to publications containing such serious defects or incorrect data. Unreliable data can be the result of an honest error or research misconduct. 2 Evidence shows that there are several reasons for retracting publications, the most important of which are research misconduct, data fabrication, ethical misconduct, plagiarism, data falsification, multiple submissions and duplicate publications, copyright infringement, and ethical research ethics. [3][4][5][6] The number and frequency of retraction publications are important indicators of the health of the scientific enterprise. Therefore, the study of retracted publications can provide valuable information about the status of adherence to ethical principles in scientific work. 7 A considerable amount of literature has been published on retracted articles; some of these studies focus on general medicine, 8,9 or specific subject areas, including dentistry, 10 cancer, 6 pharmacology, 11 radiology, 12 mental disorders 13 or on scholarly literature and life sciences topics. [14][15][16][17][18] Previous studies show an increasing trend in the number of retraction publications in journals indexed on the Science Citation Index Expanded, as well as in the Medline database between 1990 and 2009. 4 The main concern about increasing number of retraction publications in biomedical research is that unreliable information can put patients at risk. 19 Since the field of medical sciences deals with human health, this issue may have significant dimensions and consequences. 11 In addition, evidence suggests that many of the retraction medical articles have been cited. 19 This indicates that not only are patients included in retraction studies at risk but also studies that cite these articles may present risks to patients.

Objectives
In the last decade, Iran's scientific articles have grown significantly and according to the statistics published from the Scopus database in 2012, it was ranked 17th in the world. 20,21 However, scientific misconduct has also grown. 22 As mentioned, the retraction of articles is mostly due to intentional errors that means scientific misconduct, thus reviewing the retracted articles can help both researchers and scientific research policy makers in the country to find appropriate solutions to solve this problem. The increasing rate of retracted publications in the scientific literature is an important emerging phenomenon of which clinicians and researchers should be aware. Considering the importance of reviewing retracted publications, we conducted this descriptive observational study to identify the frequency of retracted Iranian publications, compared to those of other countries and adjust them with influential parameters such as research growth and growth of retracted publications worldwide.

Database Sources
For the present study, the statistics of publication history were extracted from Scopus® database. Since there are a lot of database indexes (including the overlaps of publication), we selected Scopus as one of the largest index abstract and citation databases of peer-reviewed literature to extract the number of published papers annually for each country and region under study. Additionally, we used the SCImago Journal and Country Rank (http://www.scimagojr.com) as a portal that includes the journals and countries' scientific indicators developed from the information contained in the Scopus® database. On the other hand, we used Retraction Watch database (http://retractionwatch.com) to extract the absolute number of retractions for the corresponding countries and regions under study. Retraction Watch database is a blog that reports on the retractions of scientific papers and related topics.

Period Under Study
All statistics belonged to the years 2008 -2017, however, we broke up a decade to two departed five-year periods were defined to make the growth for both publication and retraction, first the period of 2008-2012 and second the period of 2013-2017. The first date is assumed as the baseline to calculate the scientific growth which happened in the next date (including a five-year periodical time for each).

Science-Metric Indexes
Retraction growth: To calculate the growth of retraction for each proposed country (or region), first the rate of an absolute number of retracted papers over the years 2013-2017 was divided into similar number for the years 2008-2012, then this rate was adjusted to the retracted rate which similarly computed for the total retracted statistics of the world. Retraction growth per 10,000 publication: This index was calculated from the absolute number of retracted papers for both periodical times (2008-2012 and 2013-2017), adjusted to 10,000 published papers for the same date (from Scopus ® database).
Publication growth and adjusted publication growth (world): Similar to the retraction growth, first we calculated the rate of publication increasing (according to the statistics extracted from Scopus ® database) for these mentioned period time (2008-2012 and 2013-2017), then based on the same method, we adjusted the same rate of publication for the total world, which means the rate of increase in the number of articles (or abstract) indexed in Scopus® database to previous five-year (2013-2017), compared to the other periodical five-year (2008-2012), adjusted to the world's publication rate (publication growth adjusted to world's publication growth).
Retraction growth was adjusted to publication growth: The index was calculated by dividing the retraction growth into corresponding publication growth, for each proposed country (or region).
Regional Comparison of Retraction: To compare the number of Iranian retraction papers with that of other regions in the world, a Poisson Regression Model was used adjusted to the number of published documents, growth domestic production (GDP), income (high, low and middle, low, lower-middle, middle, and upper middle income according to the world bank categorization) and H index (according to the http://www.scimagojr.com) during the period under study. The statistics of retraction growth, adjusted retraction growth (for publication growth), and adjusted publication growth (for the world's publication growth) from 2013 to 2017 compared to 2008-2012 are presented in Table 1. Also, Figures 1 and 2 show the indicators shown in Table 1. The results show the status of retracted Iranian publications among other countries and the region. According to the results, the retraction growth of Iran (for the second period under study, compared to the first one) was higher than some western and developed countries, but less than some developing countries, including Turkey, Egypt, Mexico, Saudi Arabia, etc. The retraction growth adjusted to publication index showed less than 1% (0.82) of all scientific articles published by Iranian researcher have been retracted. Although in countries such as Turkey (1.04), Ukraine (1.09), Russia (1.11), South Africa (1.18), Poland (1.73), Nigeria (1.73), and Colombia (2.69), the retracted rate of articles was higher than 1% compared to a scientific publication produced in those countries, but this index (retraction growth adjusted for publication growth) for most countries, such as Iran, has been less than 1%. In addition, the rate of retraction of scientific articles in Iran The Poisson regression analysis showed that the number of retracted documents for Iranian scientists was significantly higher than those of other regions (adjusted for the number of published documents, GDP, income, and H index during the period under study), except for Western Europe whose result indicated the same pattern. The results of Poisson regression model are presented in Table 2. The results indicated the number of retracted scientific publications of Africa (OR: 0.09, 95% CI: 0.07-0.11, P < 0.001), Asiatic Region (OR: 0.87, 95% CI: 0.77-0.99, P = 0.03), Eastern Europe (OR: 0.11, 95% CI: 0.09-0.13, P < 0.001), Latin America (OR: 0.14, 95% CI: 0.11-0.16, P < 0.001), Middle East (OR: 0.21, 95% CI: 0.18-0.24, P < 0.001), Northern America (OR: 0.13, 95% CI: 0.09-0.17, P < 0.001) and Pacific Region (OR: 0.45, 95% CI: 0.37-0.54, P < 0.001) were significantly lower than retracted Iranian publications.

Discussion
The annual number of Iranian publications in the scientific literature is increasing rapidly, and consequently, the number of retracted publications has increased. To date, four studies have been conducted in the field of retraction Iranian publications. Two studies on retracted Iranian publications have been published in the Web of Science database, 23,24 and two other studies on retracted Iranian publications in the PubMed database. 25,26 The most important aspects of the current study, compared to the      four studies studying the retracted publications of Iran, is that in addition to identify the frequency of retracted Iranian publications and compare them with other countries, we adjusted the rate of retracted publications with influential parameters such as research growth and growth of retracted publications worldwide. According to the results of the current study, less than 1% (0.82) of all scientific articles published by Iranian researchers were retracted. However, the rate of retracted scientific articles in Iran compared to the rate in the world and many other countries and regions was significantly higher. Although there is a substantial increase in the rate of retracted Iranian published papers during the study period in almost one decade, the adjusted rate of this growth was at a moderate rate, compared to other regions and countries. This growth was less than the same developing countries in Middle East and a bit higher than developed countries in the West. Meanwhile, according to adjusted Poisson regression model, the pattern of this retraction is almost similar to that of Western Europe.
A descriptive cross-sectional study conducted by Masoomi and Amanollahi 25 identified 145 retracted Iranian articles on PubMed database and over 80% of articles were retracted due to scientific misconduct and duplicate publication (19.31%). A the result of a study by Moradi and Janavi 23 showed that the retraction rate has increased from the year 2008, but there was no significant correlation between this rate and publication rate the highest type of scientific misconduct is introduced as plagiarism. A study by Mansourzadeh et al 26 to retrieve all retracted Iranian publications in PubMed database revealed 164 retracted Iranian publications. "Islamic Azad University" and "Tehran University of Medical Sciences (TUMS)" were two affiliations that have received the highest number of retracted publications. The most issuer of retraction publications was editor-in-chief and the most mentioned reasons for retractions were authorship issues, plagiarism, and redundant publication. In addition, a study by Moradi et al 24 suggests that the retracted articles have been on rise in recent years. Overall, Iran, India, and Pakistan had the highest rates of scientific misconduct.
Although Iranian retracted publications represent a small fraction of all Iranian publications, even one retracted publication is a lot and is considered to be unethical behavior. Therefore authors should be more aware of the consequences of misbehavior in scientific publications. 27 The editor-in-chief for Journals should follow the instructions set out in the Committee on Publication Ethics (COPE) guidelines, 28 and new technologies for plagiarism detection should be used. 29,30 Moreover, implementing monitoring of pre-publication processes can certainly be useful that consequently reduces the scientific misconducts.

Conclusion
Although retracted Iranian publications make up a small percentage of all Iranian publications, the number of retracted publications is increasing. It is important to recognize the reasons for scientific misconducts and to provide researchers with proper education. By encouraging researchers to engage in writing and study ethics training while penalizing those who misconduct, these actions can be avoided.

What Is Already Known?
• Retraction of articles occurs as a result of scientific misconducts or honest errors. • Retracted Iranian publications are on the rise.

What Does This Study Add?
• Overall, 545 Iranian retracted publications from 2008 to 2017 were identified from Scopus® database and Retraction Watch database. • The results showed that less than 1% (0.82) of all scientific articles published by Iranian researchers were retracted. • The rate of retracted scientific articles in Iran compared to the rate in the world was significantly higher in many other countries and regions. • We suggest; authors should be more aware of the consequences of misbehavior in scientific publications. The editor-in-chief for Journals should follow the instructions set out in the COPE guidelines, and new technologies for plagiarism detection should be used. Moreover, implementing monitoring of pre-publication processes can certainly be useful.