A dramatic increase in the number of cancer survivors in older adults is expected to occur over the coming decades (DeSantis et al., 2014). This growing group of survivors will suffer a reduction in their quality of life because of Cancer-Related Symptoms (CRS) and comorbidities and many potentially will die due to the treatment-associated side effects (Cho et al., 2013; Lötzke et al., 2015). Research directed towards clarifying the chronic effects of treatments and identifying risks of late effects during the survivorship phase of cancer is designed to prevent and reduce these late effects and reduce their impact on the continuing downward spiral of disability and decreased quality of life (Alfano et al., 2012). A new integrative evidence-based practice model for providing supportive care in this growing group of cancer survivors is urgently needed.

Complementary and Alternative Medicine (CAM) therapies are being used by approximately one-third of cancer patients (Wyatt, Sikorskii, & You, 2013). The development of an integrative model (Hu et al., 2015) using an evidence-based approach longitudinally and historically could aid scientists and healthcare providers to not only provide evidence-based care to this emerging group of cancer survivors but could also help guide studies that could be built on the model’s existing scientific evidence for the future.

The varying quality and quantity of the conducted research overtime can hamper a detailed analysis of any scientific field. While an expert in the field could develop a detailed analysis, it would nevertheless be a daunting and tedious task to undertake (Perez-Cabezas et al., 2018). For this reason, the use of scientific support tools, such as bibliometrics, can be used to facilitate the analysis of a research field by automatically classifying studies overtime into different themes and topics (Cobo et al., 2011b; Gutiérrez-Salcedo et al., 2017; Van Raan, 2005). Bibliometrics contributes to the progress of science in many different ways (Martinez et al., 2014): allowing assessing the progress made, identifying the most reliable sources of scientific publication, laying the academic basis for the evaluation of new developments, identifying major scientific actors, developing bibliometric indexes to assess academic output, etc. Thus, different actors could benefit from this kind of metrics, such as policy-makers, decision-makers, and the general public (Chinchilla-Rodrίguez, Miguel, & Moya-Anegón, 2015; Chinchilla-Rodríguez, Ocaña-Rosa, & Vargas-Quesada, 2016).

Taking into account the scientific literature, some papers analyzing the bibliometric impact of the CAM research field have been published. Danell and Danell (2009) analyzed the evolution of scientific production in academic journals from 1966 to 2007. Fu et al. (2011) analyzed the document types and the geographical and institutional distribution of the authorship from 1980 to 2009. Tam, Wong, Wong, and Cheung (2012) identified the most frequently cited articles published in the journals indexed in the Integrative and Complementary Medicine Web of Science (WoS) subject category. Moral-Munoz et al. (2014) described the thematic evolution of CAM in the Integrative and Complementary Medicine WoS subject category. Zyoud, Al-Jabi, and Sweileh (2015) analyzed the scientific production of Arab world in CAM. Huang et al. (2015) performed an analysis of the literature of traditional Chinese medicine in PubMed. Furthermore, several papers related to scientific production in cancer have been published. Ugolini et al. (2012) analyzed the scientific production in cancer rehabilitation. Hack et al. (2014) conducted a citation analysis of Canadian psycho-oncology and supportive care. Thonon et al. (2015) analyzed the trends and evolution of French cancer research. Powell et al. (2016) evaluated the 100 most influential manuscripts in gastric cancer. Singh et al. (2016) mapped the breast cancer research in India. Foley et al. (2016) performed an analysis of the 100 most cited articles in esophageal and junctional cancer. Brás et al. (2017) conducted a scientometric study of the oncology research produced by Portugal in the late twentieth century. However, a detailed scientific analysis of CRS and CAM therapies using science mapping has yet to be undertaken.

Therefore, the main aim of the present paper is to analyze the scientific evolution of CRS and CAM research area through the articles retrieved form the WoS database from 1980 to 2013. It is described how a longitudinal Science Mapping Analysis (SMA) (Cobo et al., 2011) was performed to identify the structure, conceptual evolution and future research directions of the analyzed field.

In order to perform the proposed study, a bibliometric analysis is carried out. The method and software SciMAT were used to perform the SMA under a longitudinal framework.

SMA (Cobo et al., 2011) provides a spatial representation of how disciplines, scientific fields, specialties, individual documents and/or authors are related to one other (Small, 1999). It can monitor the evolution of a scientific field and delimit research areas to identify and detail its structure and development over time (Börner, Chen, & Boyack, 2003; Morris S, Van Der Veer Martens, 2008; Noyons, Moed, & Luwel, 1999).

On the other hand, co-word analysis is used in SMA to study the conceptual structure of scientific fields, by identifying the most frequent keywords used in the documents (Callon et al., 1983). The open-source software SciMAT (Cobo et al., 2012) was used to carry out the SMA analysis presented in this paper. The followed workflow is shown in Figure 1.

SciMAT is based on a bibliometric approach that establishes four phases with which to analyze the themes and thematic evolution of a research field (Cobo et al., 2015; Moral-Munoz et al., 2018):

1. Research Themes Detection. A normalized bibliometric co-word network of keyword co-occurrence is first built (Callon et al., 1983), using the equivalence index (Callon, Courtial, & Laville, 1991) as similarity measure. Then, clusters of keywords into topics/themes using a simple centers algorithm (Coulter, Monarch, & Konda, 1998) are detected. These keyword clustering correspond to centers of interest and/or research problems that are the object of significant interest among researchers.

2. Low Dimensional Space Layout of Research Themes. A spatial layout of research themes is achieved by plotting each research theme using a two-dimensional strategic diagram based on their centrality (degree of interaction of the research theme with other research themes) and density (internal strength of a research theme) rank values (Callon et al., 1991). Once the research themes are mapped into a two-dimensional space, they can be classified into four groups: motor, basic and transversal, emerging or declining, and highly developed and isolated (Cobo et al., 2011).

Themes appearing in the upper-right quadrant are classified as motor themes and are considered to be well developed and important for structuring of a research field. Themes appearing in the lower-right quadrant are classified as basic and transversal and are considered to be important for a research field but are not yet developed. Themes that appear in the lower-left quadrant are classified as emerging or declining and are considered to be weakly or marginally developed. Finally, themes that appear in the upper-left quadrant are categorized as highly developed and isolated and are considered to be well developed but with marginal importance for the field.

3. Discovery of Thematic Areas. The evolution of the previously detected research themes over a set of different time periods is then analyzed to detect the evolution areas in the research field, their origins, and their interrelationships. The evolution of the research themes over the entire study period is determined by the overlapping of clusters between two consecutive periods. To this end, the inclusion index (Sternitzke C. & Bergmann, 2009) is used to detect the conceptual nexus between research themes from different time periods. It is worth noting that the research themes and their interrelationships might indicate that a particular research theme belongs to a unique thematic area or to more than one thematic area. It could also indicate that a particular research theme cannot be associated with any of the thematic areas identified, and thus could therefore be interpreted as the origin of a new thematic area in the research field.

4. Performance Analysis. The relative contribution of the research themes and thematic areas to the whole research field is measured quantitatively and qualitatively. In this way, the most prominent, productive and highest-impact subfields may be identified. To do this, the following bibliometric indicators are applied to the different detected themes and thematic areas: number of published documents, number of received citations, and h-index (Alonso et al., 2009; Hirsch, 2005).

To carry out the analysis, the data were retrieved from WoS by using three groups of terms related to Cancer, CRS and CAM. The first group of terms referred to the main concepts of cancer and was selected from the Medical Subject Heading (MeSH) thesaurus. These terms included: neoplasm, tumor, neoplasia and cancer. The second group of terms are related to the most common CAM treatments associated with cancer according to Ernst and Casileth (1998) These included terms such as, mind-body, acupuncture, acupressure, etc. The third group of terms included typical symptoms associated with cancer according to Cleeland (2000) (e.g. insomnia, pain, fatigue, etc.). To retrieve the overall number of published documents, the following search strategy was employed:

“TS=(neoplasm\$ OR tumor\$ OR neoplasia OR cancer\$) AND TS=(“complementary therap*” OR “complementary medicine” OR “alternative medicine” OR “alternative therap*” OR “integrat* medicine” OR “holistic medicine” OR mind-body OR mindbody OR acupuncture OR acupressure OR music OR massage OR laetrile OR macrobiotic* OR relaxation OR yoga OR “tai chi” OR taichi OR herb* OR beetroot OR homeopath* OR anthroposoph* OR meditation OR hypno* OR reflexology OR iscador) AND TS=(“relat* symptom$” OR insomnia OR pain OR fatigue OR depression﹩ OR anxiet﹩ OR hypervigilance OR nervousness OR dyspnea OR sleep OR appetite OR cachexia OR shoulder OR nausea OR vomit* OR lymphedema OR “arm morbidity”)”

This query retrieved an amount of 1989 documents (articles and reviews) under the period 1980 to 2013. Citation counts up to May 2014 are also used in the analysis.

The raw data was downloaded from WoS as plain text and entered into SciMAT (Cobo et al., 2012). To improve the data quality, a de-duplicating process was applied (the authors keywords and the WoS keywords plus were used as unit of analysis). Words representing the same concept were grouped. Furthermore, some meaningless keywords in this context, such as stop words or words with a very broad and general meaning were removed (e.g. disease, outcomes, system, etc.).

Next, using the SciMAT period manager, the entire study time period (1980-2013) was divided into two, three, and four subperiods, in order to test and establish the adequate number of periods to analyze. In that way, although it is common to use periods during the same time span, in the first years, there were low numbers of researchers and publications. Finally, a first subperiod of twenty-eight years (1980-2008), and a second subperiod of five years were established (2009-2013), since there was not enough document count to obtain an appropriate analysis. It allows us to have enough keywords to perform the analysis previously described.

In order to introduce the analysis and better understand the subperiod division, the distribution of publications by year between 1980 and 2013 is shown in Figure 2. As we can observe, in the last two years, the number of documents published presents a great growth, i.e. more than 250 documents per year. Therefore, CRS and CAM therapies is a topic which is attracting an increasing interest in the scientific community.

In this paper, a bibliometric analysis of CRS and CAM has been developed using SMA. As stated above, bibliometrics allows to analyze a research are in several ways (Santisteban-Espejo et al., 2018). Thus, in relation to the results obtained, some findings can be highlighted. The first remarkable finding is that the research field has attracted the interest of the scientific community throughout years. It is observed in the great growth of publications and citations received.

According to the analysis, it identifies for the first period twelve different main themes in CRS and CAM. All these findings are supported by the current scientific literature, which highlight the importance of these research lines. Anxiety is an important symptom as are its treatments. In the second period research is more focused on the physiological changes involved in anxiety. It is in relation to recent findings, where the repercussion of the anxiety in the emotional competences is highlighted (Balasooriya-Smeekens, Walter, & Scott, 2015; Baudry et al., 2018). Survivor and Palliative Care emphasizes healthcare that cancer patients receive following treatment and is highly focused on fatigue suffered by patients (Kassianos et al., 2018). Meditation is focused on the benefit of using this form of therapy in stressed patients and it is important to the concept of quality of life related to this thematic area (Haller et al., 2017). Treatment, Symptoms, and Cancer Types is a thematic area covering different kinds of topics: treatments (e.g. acupressure), symptoms (e.g. vomiting) and cancer types (e.g. prostate cancer). Postmenopause is mainly related to symptoms and hormonal changes that occur during postmenopause. It is also related to the term breast cancer in both periods (Rosner et al., 2015). Cancer Pain is mainly composed of themes related to treatments; in the first period it is centered on drugs and opioids, and in the second period it is focused on alternative therapies (Chiu, Hsieh, & Tsai, 2017; Kane et al., 2017). Low Back Pain is mainly focused on physiotherapy and therapies to reduce the pain in both periods (Ross & Bayer, 2005). Herbal Medicine is related to the use of herbal extracts to treat some CRS, and the presence of documents where herbal medicine is tested in animals (Lau et al., 2016). Children in the first period is related to different alternative therapies (e.g. hypnosis, imagery) and some symptoms (e.g. stress, behavior change) (Italia et al., 2015). In the second period it is less important and is related to more general terms. Depression and Insomnia is focused on those two CRS and their alternative treatments such as hypnotherapy and cognitive behavioral therapy (Cha et al., 2017). Inflammation Mediators is related to cytokines involved in immune processes, and is mainly focused on tumor necrosis factor (Fernandes et al., 2015). Lymphedema is primarily related to breast cancer and the physical treatment of this CRS (Shaitelman et al., 2015).

Regarding the thematic evolution, the thematic area Anxiety obtains the best impact index with the highest values in terms of the number of citations and h-index. This means that Anxiety is a growing research area with a good rate of published documents that captures the attention of the CRS and CAM research field. The thematic area Survivors and Palliative Care is another highly developed thematic area with a great number of documents published. The amount of documents published and the different issues related to it suggest that preventing the suffering of patients with alternative therapies is a major focus in this field.

Nevertheless, some study limitations can be highlighted. First, the most used CAM therapies, according to the literature, are analyzed, although some of them may not have been included. Second, the use of WoS as bibliographic source implies that articles not indexed in the database have not been analyzed. If they would be taken into account, the detected thematic areas could change.

Based on these study results, the following conclusions can be made. The scientific community can now identify thematic areas on which attention has been focused over time and others that are growing in terms of current research interest.
The citations and scientific production in CRS and CAM are primarily focused on two large topics, anxiety and the prevention of suffering in patients following cancer treatment. There are other thematic areas that are important in terms of research interest. These include meditation, symptoms as vomiting or nausea and postmenopause. These results suggest that this research field is growing rapidly and will continue attracting more research in the future. The study findings can be used to provide an historical perspective on research conducted in this emerging field and a scientific evidence-based practice model that could be used to develop future research.