**4. Synthesizing an Explanatory Theory**

Psychologist Kurt Lewin proclaimed, "There is nothing so practical as a good theory" [55]. This perspective article's method of knowledge synthesis borrows heavily from theoretical research methods such as grounded theory. Glaser and Strauss developed the grounded theory method to bring more rigor to qualitative research [56], although the methodological principles of grounded theory are also applicable to quantitative research [57]. The researcher's overarching aim in grounded theory is to begin an investigation with a clean slate and inductively construct a new theory through an iterative process of comparative data analysis. With a new theory in hand, the researcher can encourage the development of hypotheses to experimentally test concepts deducted from the theory. Eventually, systematic reviews and meta-analyses can critically assess results of experiments and clinical trials testing the concepts. Interestingly, Ioannidis [58] suggested that the number of systematic reviews of clinical trials may be currently higher than the number of clinical trials, implying a greater need for theory synthesis at the source in the flow of new scientific information. Of relevance, the number of physician-scientists has also been on the decline over the past several decades, further highlighting the need to increase physician medical education in scientific thinking and biomedical research [59], including a need for education in knowledge synthesis and theory development.

Researchers most often use grounded theory as a method to analyze, compare, and combine concepts from original data, but Wolfswinkel et al. proposed that grounded theory may also be used to conduct a rigorous literature review in which concepts from published research findings themselves are the data [60]. Reviewing literature in this manner is particularly useful for synthesizing new knowledge from current knowledge—similar to thematic synthesis [36], a qualitative method that combines thematic analysis with meta-ethnography across multiple studies. An important di fference is that the method proposed by Wolfswinkel et al. is a mixed method that combines both qualitative and quantitative research.

The following is an example of how I used a grounded theory approach to conduct a web-based synthesis in the field of epidemiology. I was interested in exploring online research literature investigating the association of phosphate toxicity with cancer [61]. The literature was reviewed using keyword searches in Google, Google Scholar, and other scholarly online databases, and all relevant studies associating phosphorus, tumorigenesis, cancer, etc. were identified. In addition to searching with keywords, I searched references cited in studies, which is known as citation analysis; a very useful method for evaluating a study's impact in a research area [62]. For example, The Hallmarks of Cancer [63] is an influential study often cited by other studies in a literature review of cancer research.

Included in my synthesis were studies selected from the fields of basic research, clinical research, and epidemiological research; the designs of the selected studies included case studies, cohort studies, laboratory animal experiments, in vitro studies, systematic reviews and meta-analyses. My investigation followed the evidence without boundaries, and my final synthesis was written in the form of a narrative review. Although many speculations, hypotheses, and explanatory theories were proposed by authors of the studies selected for my synthesis, I analyzed only concepts from the objective findings of the selected studies in order to foster a new grounded theory. The rigorous evidence-based grounded theory method helped assure the high internal validity of the synthesis, i.e., the analyzed concepts were based on trustworthy peer-reviewed findings rather than speculation. As concepts from the reviewed literature were analyzed and compared, certain themes began to emerge from the data which were eventually linked together into a cohesive theory that explained how phosphate toxicity from dysregulated phosphorus metabolism stimulated cancer cell growth. Interestingly, this inferred knowledge challenged several of the hypotheses proposed in The Hallmarks of Cancer. For example, evidence from my synthesis supported the concept that cancer cell growth is dependent on exogenous growth-rate factors, challenging the concept that cancer cells independently stimulate themselves to grow autonomously.
