Surveys in the health field in the digital age. Reality or fiction

Quality research demands effort, dedication, time and resources, both financial and in terms of personnel^1,2. Fortunately, tools are being developed that facilitate research and make fieldwork easier and more reliable. However, these new tools cannot justify poorer quality work. There has been a major shift in recent years in the management of research project data, from manual to largely digitalised handling, where protocolisation is much faster and more reliable. In addition, the widespread use of e-mail and video conferencing allows almost immediate correspondence between people in general, and between researchers in particular.

This situation, which is a great help, has cast a blight on psycho-social research conducted by means of questionnaires. The combination of large databases of contacts, immediate access by e-mail to any person, and online databases for filling in questionnaires, has resulted in a large number of psychosocial studies. This has been accentuated by the COVID-19 pandemic, where this type of strategy can be carried out without leaving the office.

For all these reasons, I would like to highlight the national psycho-social work on attitude published in this same issue of Cirugía Española³ and which I have had the good fortune to direct as president of the International Collaborative Donor Project (PCID)⁴, supported by the enthusiasm of Ms López-Gómez at the start of the project, the perseverance of Dr. Belmonte who managed to complete it and turn it into his international doctorate, and not forgetting the support of almost a hundred members of the PCID psycho-social research group⁴. All these “accolades”, although well-deserved, are not gratuitous; they are to show how a psycho-social study that can provide real and useful data for the design of health policies requires time, the dedication of several professionals with experience in the subject, and the support of specialists in field research work in this area^5–9. In other words, a great deal of effort, design, fieldwork and data analysis are involved. This study shows that more than 80% of the sample selected did not participate in the study, i.e., there was a positive selection bias. If we only analyse the questionnaires obtained, we can conclude that 42% of this population is in favour of donation, which would not explain the situation where organ donation from this social group is rare. However, a study such as this one, which helps us determine the causes of this bias, shows that only 7.8% are in favour of donating their organs after death, which does reasonably explain the clinical reality we are facing. This is not an isolated example; other publications have already shown this situation¹⁰. Therefore, for some years now, our psycho-social research group (PCID) has been highlighting the need to conduct controlled, stratified psycho-social studies, in which the degree of compliance is fundamental^4,10, if we do not want a distorted view of reality.

There is currently is a high volume of articles based on surveys with unrepresentative results, which has generated a great deal of confusion in scientific information. It is not difficult, therefore, to find studies that contradict each other, and if really analysed, these are studies with a high risk of non-comparable bias, where any result is possible. This being the case, why are they still being conducted, and moreover, why do editorial teams continue to accept these studies? It is clear that in any problem simple answers are usually false or at least half-truths. And therefore, assessment is not easy and the answer is complex. Based on this premise, it should be noted that there are three factors that have a definite influence on their conduct and publication.

The first is that they convey a sense of high quality. This is because they are conducted using large contact databases, where there are often thousands of people registered, which makes it possible to contact many potential respondents, and thus obtain a high sample even if the completion rate is low. There are hundreds of examples available in the literature, but I would like to highlight the study conducted on 1541 medical students in 104 countries around the world and supported by the International Federation of Medical Students’ Associations on attitudes towards donation¹¹. In other words, a study with a large, international sample, with the participation of more than 100 countries and endorsed by an international association, all of which are theoretically parameters of quality. However, if analysed in depth, we can see that an e-mail was sent to thousands of medical students through the student associations of each country, within the International Federation of Medical Students’ Associations, where 1541 respondents are a tiny percentage, less than 2%, of the respondents who were asked to complete the survey. This explains why this study shows a significantly better attitude towards organ donation than most studies of this type¹². In addition, there is a selection bias of the organising group, as 13.9% of the respondents are from the country of the researchers conducting the study, while in the rest, the vast majority of countries have a completion rate of less than 0.5% of respondents.

The second is that they are usually conducted in collaboration with the corresponding scientific societies, as theirs are the databases, and this endorsement-support favours their implementation and publication. Especially if we consider that most scientific societies have their corresponding scientific journals, and endorsement of these societies facilitates their evaluation in them.

The third is that these studies are easy and quick to conduct; all that is needed is to write an e-mail request, attach the questionnaire linked to an online database, and send it (generally via the secretariat of the scientific society) with that society’s logo. Once the deadline for data collection has passed, the database is downloaded, analysed and published.

I do not want to finish without mentioning the reasons why editorial teams continue to accept this group of papers. The first reason is that in most cases the articles have been generated from the databases of the scientific society on which the journal depends. This situation favours their evaluation, and sometimes their acceptance, as they are considered general data of interest to the scientific society, its affiliates and the journal in question. Moreover, it is striking that, because there is a positive selection bias, a particular issue is often hypertrophied and this often results in recommendations to the scientific society. The second reason is that they are potentially citable articles, since by hypertrophying a certain “issue”, due to the positive bias of the topic, it is given visibility and therefore rendered citable, even if this is only temporary.

All of the above explains why many authors have started publishing this type of study, as it is an easy and quick way to publish. These studies are easy to design, quick to execute, as they usually require mass e-mails linked to databases with a high volume of records, which are completed directly in an online database, and within a few months they can be fully completed, most within scientific societies. This does not correspond with the time, dedication and resources, both personal and financial, required for a quality psycho-social study. All this has been accentuated in the COVID pandemic, as these are studies that can be conducted in an office, with no more than a computer and permission to access the database of contacts.

As a knock-on effect, the current lack of prestige of psycho-social studies has rightly been accentuated and has led many Q1 journals to be increasingly strict in accepting these articles or even to systematically reject them, hindering the acceptance of articles that are important for health policies. As a side effect, this has made it more difficult for psycho-social research groups to publish quality articles.

Finally, I would like to encourage all psycho-social research groups to continue their great work, and those who dare to enter this exciting world to do so rigorously and meticulously^13–16. I urge editors of scientific journals, myself included and I assume my share of responsibility, to select these studies better and avoid those with a high risk of bias, even if they are linked to scientific societies.

With this editorial I am not encouraging researchers to work without internet or e-mail, in fact most psycho-social research groups use all these online tools that facilitate their work and data analysis, but this does not mean there should be a reduction in the quality of the projects or their design. Online is a support but not a substitute for quality criteria.