Background

Several studies have estimated the potential economic and social impact of the mHealth development. mHealth is an abbreviation for mobile health, a term used for the practice of medicine and public health supported by mobile devices. According to WHO [1], nearly 90 % of the world population could benefit from the opportunities offered by mobile technologies and with a relatively low cost. Considering the latest study by Institute for Healthcare Informatics (IMS) [2], more than 165.000 apps of health and medicine are offered including all the stores from different platforms. Thus, the global mHealth market was an estimated $10.5 billion in 2014 and is expected to grow 33.5 percent annually between 2015 and 2020s [3].

Also, the IMS Institute indicates that 70 % of health apps is focused on general population, offering tools to reach and maintain wellness and to improve physical activity. The remaining 30 %, were designed to more concrete areas such as professionals or people affected by specific diseases.

Despite this situation, it is important to note that more than 50 % of the available apps received less than 500 downloads and only five of them comprise 15 % of all those in the health category. The IMS attributed this situation to different causes, which include: poor quality in many of them, the lack of guidance on the usefulness of the app and a low level of support from health professionals.

However, it is well-known that health apps, solving the problems detailed above, could represent a very useful tool for monitoring chronic diseases will account for 65 % of the global market for mHealth in 2017 [3].

This fact will represent revenue of 15.000 million dollars. The pathologies with a higher potential to increase business are in order: diabetes and cardiovascular disease. They will also play an important role related to diagnostic services (they will reach 15 % and will generate 3.400 million of dollars) and medical treatments (10 % of the market and revenues of 2.300 million). By the other hand, it is estimated that business will increase from 4.500 million in 2013, to 23.000 million in 2017. Continents with largest market share are, in descending order, Europe and Asia (30 %), United States of America and Canada (28 %) [3].

However, we do not know if the apps available to the population are based on scientific knowledge and therefore, it is difficult to assess the real impact of this spectacular development on the health of populations. On the other hand, we do not know how the great spread of the phenomenon of Health 2.0 (that is a term presented in the mid-2000s, as the subset of health technologies mirroring the wider Web 2.0 movement, offering possibilities for changing health care which started with the introduction of eHealth following the emergence of the World Wide Web [4, 5] ) that is reaching the scientific field (medical or computer), which should occur in parallel in order to offer products that positively affect the health of citizens.

Therefore, this study aims to identify, study and evaluate the role of interdisciplinary research teams in the development of articles and applications in the field of mHealth and the impact that the development of mHealth has had on the health and computer science field, through the study of publications and the composition of the research teams in specific databases for each area, which have been published until nowadays. According to Yadros et al. [6] interdisciplinary research seems to be a supplier of creative and innovative approaches. It is able to produce new lines of research and renew scientific field. In this sense, the justification of using interdisciplinary research is thus particularly strong and crucial in scientific programmes addressing grand societal issues or challenges that require an holistic approach including biological, physical and social factors.

Methods

This work is extended and based on the previous work [7]. A systematic review was conducted in two stages during November 2014. The first one was focused on locating papers available in databases from Health Sciences. After this step, we repeat the search but in the Computer Science field because we wanted to find the different penetration in each area. As recommended in the PRISMA Statement [8] for systematic reviews. The PRISMA Statement consists of a 27-item checklist and a four-phase flow diagram. The aim of the PRISMA Statement is to help authors improve the reporting of systematic reviews and meta-analyses. We describe the search strategy and the number of papers located, discarded and finally selected for review using a for-phase flow diagram. In the first stage, we have consulted the PubMed database. We used “mHealth” and “mHealth AND app” terms as search strategies. Finally 79 items were selected for reviewing (see Fig. 1). We also consulted “Science Direct” and “Scopus” using “mHealth” but reducing the search to “Computer Science” area. Having initially located 375 publications, only 27 were chosen for our study (see Fig. 2). Thus, a total of 106 items were reviewed. The impact factor of the journals that published the papers selected was consulted using the journal citation report from web of science (WoS). Since this impact factor usually varies for each year, we took the corresponding to the year when the article was published.

Fig. 1
figure 1

Review and selection criteria of papers (health science)

Full size image
Fig. 2
figure 2

Review and selection criteria of papers (computer science)

Full size image

As noted previously, the execution of this work has been carried out by means of the recommendations given by the PRISMA [8] statement. So, this work includes a study through the following information:

  • The summaries and results of all reviewed papers after performing his complete reading.

  • The departments that participate in the development of the works and their categorization for subsequent statistical analysis.

This last categorization has been performed including the departments in ten large groups, which are: ‘Research Center’, ‘Nursing and other health professionals’, ‘Engineering and Technology’, ‘Finance and Statistics’, ‘General Medicine and Specialties’, ‘Agencies and Institutions’, ‘Health’, ‘Health Care and Community’, ‘Physiotherapy Associates’, ‘Pharmacy and Associates’.

In order to analyze and evaluate the impact of health apps in health and computer science publications in a precise way, additional features which are not used to be considered into systematic reviews, have been included in this study. All the information taken into account has been clustered into two main categories:

  1. 1.

    Publication characteristics:

    • Journal name.

    • Type of journal (According to 22 categories taken from ISI Web of Knowledge).

    • Journal ranking (quartile).

    • Journal impact factor.

    • Article publication date.

    • Type of study.

    • Number of received citations.

  2. 2.

    Interdisciplinary nature:

    • Departments working on the contribution.

The evaluation of the last item (interdisciplinary nature) has been performed using the Rao–Stirling index as explained by Rafolds and Meyer [9], among others. Interdisciplinary research has been defined as a mode of research that integrates techniques, tools, and/or theories from two or more disciplines to advance fundamental understanding or to solve problems whose solutions are beyond the scope of a single discipline or area of research practice [10]. The advantage of the Rao–Stirling measure is that it takes into account the distribution of references across disciplinary categories on journals in the WoS for 220 WoS categories or subject categories (SC) and also considers how cognitively distant these categories are.

For statistical analysis of the data we have used own descriptive statistics (frequency tables, measures of central tendency and dispersion, Pearson correlation coefficient as well as graphic representation) and analytical techniques, using as evidence contrast hypothesis Chi square, t Student, ANOVA and Kruskal–Wallis as non-parametric version. Processing and analysis of data was performed using the SPSS version 22.0.0 [11].

Results and discussion

A brief summary of the main features, topics and contents in the reviewed papers is shown in Tables 1 and 2. We found papers focused on mHealth in 51 different journals, being most of them (68.6 %) indexed in the ISI web of Knowledge. The higher proportion of papers published in journal indexed in JCR, is located in four WoS categories: medical informatics (17.9 %), Healthcare Sciences & Services (12.8 %), computer science interdisciplinary applications (11.1 %) and mathematical and computational biology (8.5 %), showing a great concern about developing mHealth research in two fields of ISI that we think that must be strongly linked to this topic: clinical medicine and computer science. Taking into account not indexed journals, we found same development patterns, because most of the researches belong to different departments and institutions, including professionals from health and computer science. This fact could be explained by the concentration of articles in just three journals, journal of medical internet research mHealth and uHealth, (JMU) containing 21.7 % of the published papers, journal of medical internet research (JMIR), covering 7.5 % and international journal of medical informatics, reaching 4.7 % papers. Thus, for example, journal of medical internet research is classified by journal citation report (JCR) into several categories (“Health Care Sciences & Services” and “Medical Informatics”) and the international journal of medical informatics is listed by JCR into three categories, the two previously mentioned, as well as in “Computer Science, Information Systems” that clearly belongs to a non-health area. Another important result is the great impact factor of these publications. This way, in any category where they could be classified of these two journals belong to the first quartile, excepting JMU which is not indexed because it was created in 2013 like an spin-off from JMIR. This may also explain that the average impact factor of the found publications was 1.54 (±1898) including all the papers, and 3.0248 (±1.6014) if we select just the papers published in indexed journals. Moreover, 27.4 % of the articles are located in first quartile journals, 13.2 % in the second, 7.5 % in the third and 2.8 % in the fourth quartile. The 49.1 % of the papers appear in not indexed journals, specially in JMU. From our point of view, this fact highlights the high impact and newness that scientific work based on the use of mHealth technology represents for editors, allowing researchers to access to high impact publications and how new categories of are opening progressively.

Table 1 Main features of reviewed papers
Full size table
Table 2 Main results and conclusions of reviewed papers
Full size table

However, the number of citations received of these articles is relatively small and widely dispersed. The sample shows an average of 6.35 (±13,56) references, excluding 2015 and 4.71 (±8.77) counting this year. Two papers highlight with 74 [12] and 63 [13] citations, facing many that have never been referenced. Interestingly, these two articles were published in 2011 and 2012 (respectively), when the exponential growth in the number of publications on mHealth began. Thus, we can find just 2 papers in 2011 (1.9 % of total found) while we locate 45 in 2014 (42.5 %). Similarly, only four articles were found in 2012 (3.8 %) and 26 in 2013 (24.8 %). Until October of 2015 we located 29 (27.4 %) papers published in 2015. In 2010 there were no relevant papers included in the search. This could be a sure sign that the scientific community is increasingly aware of the potential and the need to focus some research into this field.

Similarly, the mean number of citations received for years (excluding 2015), shows a clear downward trend (F = 36,72, 3 df; p < 0.001). As the years pass, the average is down from the 62 items on average in 2011 till 1.53 in 2014. Temporal matters could explain this, i.e., it stands to reason that more recent articles have had less time to be referenced (and that is the main reason because we decided to exclude 2015 from this analysis), but it is a trend that is also seen in 2012 (with an average of 23 references) and 2013 (7.85). Post hoc tests found significant differences between every year (p < 0.01).

In addition, considering the types of papers, we found that original researches are more prevalent (63.2 %), Besides this, we evaluated the type of the research carried out. We found descriptive studies (23.6 %), researches focused on technological development (21.7 %) and reviews (16 %) highlighting over the rest (see Fig. 3). This fact happens because most of the articles are focused on the user data collection and evaluation from apps designed to assess adherence or modification of habits to healthier lifestyles (studies with larger sample sizes). Another percentage of the original researches is represented by small pilot studies or clinical trials with smaller sample sizes (5.7 %). As we mentioned above, reviews are also frequent, something justifiable given the exponential growth of publications. This situation force researchers to conduct periodically synthesis and evaluation of trends in the development of apps and main findings, as well as the errors most frequently committed (especially as the methodological design is concerned). Another significant point was to find nine papers from qualitative research which represent 9.4 % of the types of research. It is understandable if we comprehend that this type of studies is very helpful in trying to expose the different realities of the new or unknown phenomena which we have a very little information or experience. A good example could be the expected results of interactions between mHealth apps and populations who want to improve their health. Furthermore, we were surprised because two of these articles had been published in computer science journals, where papers highly focused on applied and/or quantitative research are traditionally accepted. No statistically significant relation between the type of research and the number of citations received (p = 0.297) were found. Similarly, the type of research and the impact factor of the journal where the article is published did not show to be related (p = 0.094).

Fig. 3
figure 3

Type of research

Full size image

We also wanted to know the link between the field of knowledge where the publishing journal is classified and the impact factor (see Fig. 4). We found a clear relation (p < 0.05), highlighting “Environment/Ecology”, “Chemistry” and “Multidisciplinary” with higher average of impact factor. In any case, it is very important to evidence how fields of knowledge, very far from health, a priori, accept papers focused on mHealth, which, from our point of view, demonstrate the deep impact that this new topic is generating in many fields of knowledge, related or not to health sciences.

Fig. 4
figure 4

Impact factor average related to ISI 22 Fields

Full size image

Research teams and the role of interdisciplinarity

The results show that more than half of the reviewed papers (67.9 %) had the participation of at least two departments/institutions focused on different topics, reaching the maximum amount of five different topics (see Fig. 5). Only 32.1 % [13–41] of the reviewed papers had participants focused on just one topic.

Fig. 5
figure 5

Number or departments/institutions focused on different areas involved in the research

Full size image

Besides this, we considered the interdisciplinarity, calculating the degree of diversity through Rao–Stirling (multidimensional index which includes the analysis of variety, balance and disparity). We found a very low and dispersed rates (0.07 ± 0.05). From our point of view, and related to the different topics where the departments involved in each paper are focused, these results mean that, although the diversity of the participants could seem high because they belong to different areas, finally they do not use the specific knowledge of each one, finding many references in each paper that belong to the same WoS categories (causing a low Rao–Stirling diversity).

About interdisciplinary research teams, interdisciplinary publications, impact factor and number of citations received

Although the visual examination of Fig. 6 shows a clear tendency for researches carried out by more than two departments/institutions from other fields of knowledge, are published in major journals, statistical analysis shows that these differences are not significant (p = 0.33), obtaining an average impact factor of 1.81 (±1.93) for researches carried out by more than two teams focusing on different fields, versus 1.42 (±1.88) for teams with two or fewer different fields of knowledge. However, we evidenced a light but positive and significant relation between the Rao–Stirling diversity and the impact factor (r = 0.182; p = 0.031) which means that a higher degree of interdisciplinarity increase of impact factor of the paper. Anyway, as we mentioned above, it is important to remember that the participation of multidisciplinary research teams does not necessarily means the presence of interdisciplinarity. This fact is supported by the analysis of the relation between the number of departments focused on different research areas for each paper and the value of Rao–Stirling diversity. In this case, we can not assume that the participation of research teams from different fields lead performing a multidisciplinary study. (p = 0.108) (see Fig. 7).

Fig. 6
figure 6

Relation between teams of more than 2 different departments/institutions focused on different areas and impact factor

Full size image
Fig. 7
figure 7

Relation between number of departments/institutions involved in the research and Rao–Stirling diversity

Full size image

Another method to evaluate the impact of a paper consists in evaluating the number of citations received. Neither the Rao–Stirling (p = 0.063), the field where the publishing journal is classified according to 22 ISI web of knowledge categories (p = 0.811), or the number of departments focused on different fields of knowledge (p = 0.869), seem to be related to the number of citations received. The average number of citations for teams with more than two departments focused on different fields of knowledge is 6.52 (±9.9), versus an average of 6.14 (±9.6) citations for teams composed of two or less. From our point of view, this could be explained by the novelty of the research on mHealth, which seems to take precedence over the quality or focus of each investigation. We think that this could be also confirmed by the absence of statistical relations between the impact factor of the journals in which the papers were published, the number of citations received and the research design. That is, it would be logical to think that clinical trials, technological developments, and other quasi-experimental studies with higher levels of evidence, should be published in major journals or be more referenced. However, the result described above, indicates that this type of research designs is not better valued by publishers or other research investigators than others with lower levels of evidence (descriptive studies, qualitative researches, etc.).

A priori, it would seem that the interdisciplinary nature of the research teams is directly linked with access to publication in journals indexed in higher quartiles. This fact is confirmed by finding significant relations (p = 0.027) between the quartile of the journal in which the articles were published and the number of departments focused on different areas involved in the investigation (see Fig. 8). By contrast, we did not find relation with Rao–Stirling diversity (p = 0.475); showing that the fact of accessing higher quartiles may be related to the presence of multidisciplinary teams in the investigation, which does not necessarily mean that researches show an interdisciplinary approach.

Fig. 8
figure 8

Relation between the quartile of publication and the number of departments/institutions focused in different areas involved in the research

Full size image

Another aspect to take into account is referent to the number of articles that are included in our study, which covers until October of 2015. Because of that, it is difficult to stablish the progression in the number of papers in this year. We considered that the number of papers in the last quarter of this year (2015) continue with a increasing tendency, but, it seems that in a lineal level (not exponential as in previous years). This fact could be only estimated in studies that will be carried out in 2016.

Limitations

The results should be taken with caution given various limitations. In this sense, Yadros et al. [6] stated some limitations as: “the inaccuracies in the WoS categories used to define subdisciplinary categories may create biases in the indicators of citation impact (since citation impact is highly affected by normalisation) and may have an important effect as well in diversity measures.”, or “the inclusion or not of some control variables such as number of co-authors, institutions or article length is open to debate and these may have an effect on results.”, among others.

On the other hand, from the revised bibliography it is possible to extract that for calculating certain bibliometric indexes (as Rao–Stirling diversity), it is usual to consider a higher quantity of the number of papers than the number of them that we have considered in this work. This fact is because the main objective in our paper was to realize a systematic review and no the calculation of bibliometric indexes. The number of revised papers is considerably minor, so, it would be advisable to take with caution the results of this bibliometric index.

Conclusions

There are mHealth papers of all kinds (trials, analytical, descriptive, reviews, etc.) and its number grows at an exponential rate. This could show how technologies related to mHealth are reaching the scientific field. These technologies are reaching the population even faster. Therefore, mHealth is becoming an interesting research topic in both fields, health and computer sciences (hasta 2014 siendo más moderado en 2015).

These articles are published in high impact journals. We have found them in specific journals (focused on eHealth) and in generalist health and computing journals. Generalist journals have just begun to accept research based on the application of mHealth technologies. This highlights the growing importance of this topic.

There are already studies that warn of methodological deficits in some studies in mHealth, low accuracy and no reproducibility. Studies of low precision and poor reproducibility, coupled with the low evidence, provide low degrees of recommendation of the interventions targeted and therefore low applicability. The market apps, mostly lack of scientific evidence or participation of health professionals. This should compel publishers and researchers to be more stringent on the design of experiments and in the publication of results.

Fundamentally, it is possible to extract two main conclusions from our study. First, it is evident the increased interest that publishers of scientific journals are showing to mHealth, given the steady increase in publications focusing on this issue, regardless of the subject area. Second, it seems clear that the participation of multidisciplinary teams (with a variety of professionals focused on different areas of knowledge) is not necessarily linked to the presence of interdisciplinary approaches. Finally, this interdisciplinarity plays a limited role in the impact of the papers (measured by the impact factor of the journal where they are published and the number of citations that they received), this fact could be more related to the novelty of the topic of research.