From the idea to the research question and hypothesis

The scientific study process unfolds through three distinct phases: the discovery phase (involving the development of the research idea, question, and hypothesis), the justification phase (encompassing study planning, methodology, statistical analysis, and results presentation), and the exploitation phase (centered around discussion and data interpretation) [1]. This article focuses on the pivotal steps within the discovery phase, namely specifying a research idea and formulating the research question, hypothesis, and research goal. A research idea may originate from various sources. On the one hand, daily clinical practice can provoke new inquiries, while on the other, conducting another clinical study or engaging in peer review processes for different study groups can reveal new, unanswered questions. Regardless of the catalyst, there must be an intrinsic interest in the chosen area. What fuels this interest in knowledge? To sustain enthusiasm for scientific work alongside routine clinical responsibilities, curiosity is paramount. According to Manuel R. Theisen, “scientific [and innovative] work is the systematic and comprehensible satisfaction of curiosity.” Beyond interest, a comprehensive understanding of the selected topic is crucial. This understanding is essential for the development and execution of an innovative and clinically relevant study. To achieve this, one must be aware of the questions already addressed within the chosen topic area. While initial insights can be gleaned through interviews and discussions with experts or colleagues, a thorough, systematic, and detailed literature search is indispensable. Ensuring the relevance and significance of the research idea is crucial.

In today’s digital age, the Internet stands as the primary source of information, with scientific literature accessible through various online repositories such as MEDLINE, Google Scholar, Embase, and Cochrane, among others. The corpus of scientific literature is broadly categorized into primary, secondary, and gray literature [2]. Primary literature encompasses original works addressing specific topics and providing scientific answers, while secondary literature, like review articles, summarizes findings from primary sources. Both primary and secondary literature undergo rigorous scientific review before publication, affirming their reliability and citability. In contrast, gray literature, including preprints and unpublished manuscripts, lacks a formal review process and is therefore not citable [1].

Prior to delving into a literature review, it is imperative to narrow down the research topic and formulate precise research questions, facilitating the extraction of key terms for investigation. The subsequent literature research can be conducted using two distinct methods:

The bibliography method relies on a deep understanding of the existing literature in the desired subject area, allowing for targeted searches for specific articles by author or title. On the other hand, the snowball system, requiring no prior detailed knowledge, involves exploring new articles and sources in a pyramid form from a foundational work, preferably a systematic review, using its bibliography. Both methods demand a precise and relevant list of terms to execute the most efficient and targeted analysis possible. It is imperative to critically examine all sources used. The quality of a scientific work is not determined by the quantity of studies cited but rather by the quality of individual publications. Thorough and critical literature research can give rise to innovative study ideas with explicit questions that may not have been adequately addressed in previous studies.

Distinguishing itself from the broader study idea, the question involves the precision of formulating a specific query addressed through an empirical scientific approach. For instance, while the idea could be framed as “What is the impact of anterior cruciate ligament reconstruction on the rate of osteoarthritis in patients with anterior cruciate ligament rupture?” a focused question might be: “Do patients with conservatively treated anterior cruciate ligament rupture have an increased risk of osteoarthritis compared to patients who underwent anterior cruciate ligament reconstruction?” This example illustrates that a single study idea can generate multiple questions, allowing exploration not only of the comparison between conservatively and surgically treated anterior cruciate ligament ruptures but also identification of general patient-specific risk factors for osteoarthritis development. It is imperative, however, to ensure that each formulated question can be effectively examined in a study [3].

The explicit study question holds paramount importance as it profoundly influences the population under investigation, the chosen methodology, the evaluation of results, and the interpretation of findings [4].

To craft a robust question, Hulley et al. introduced the FINER criteria, a set of guidelines aiding in study question formulation (Table 1; [5]). The acronym emphasizes that a study question should be feasible and answerable within the empirical scientific project’s framework. It should not only align with personal interests but also contribute to the broader scientific community. The question must tackle an open issue in scientific literature, aiming to fill knowledge gaps. It must adhere to basic ethical rules, subject to scrutiny by an ethics committee through an ethics application. Importantly, every study question must be relevant to the chosen topic area. Relevance is key for achieving increased visibility of study results, ensuring that the findings contribute to innovation in the field of medicine.

While the FINER criteria help outline a general design for formulating a research question, the PICOT criteria can be beneficial in developing a specific question for an empirical scientific study [6]. In this context, consideration must be given to both the population under investigation, the intervention conducted, the comparison group, the desired outcome, and the timeframe (Table 2; [7‐9]).

Guided by these criteria, the development of a study question is facilitated, ensuring the formulation of an adequate study methodology. The PICOT criteria in particular offer a structured approach, prompting considerations regarding the target population, the intervention and its alternative (comparison group), and the research goal, thereby aiding in the selection of an appropriate measurement instrument [10].

Undoubtedly, the significance of a well-designed research question cannot be overstated. A poorly crafted question can detrimentally impact study design, lead to fruitless efforts, and impede the identification of clinically meaningful results. Ultimately, this can adversely affect the likelihood of scientific publication. Neglecting the careful formulation of a robust research question can compromise the quality of the study and its outcomes. Thus, it is imperative to allocate sufficient resources during the initial phase of the study to develop a question that holds clinical relevance and can be effectively investigated using empirical scientific methods [6].

Once the research question has been clearly articulated, the subsequent step involves seeking an answer. To facilitate this process, it is essential to articulate the presumed answer to the study question at the study’s outset. This “hypothesis” (derived from Greek or Late Latin, meaning “assumption”.) represents a provisional statement serving specific purposes until validated or refuted. A hypothesis is not a proven explanation for an observed phenomenon; rather, it is a tentative assertion grounded in observations, experiences, or existing theories. Its role is to serve as a starting point for subsequent research and experiments aimed at testing and refining it. Importantly, both the question and the hypothesis should be formulated before the study is planned and should not be generated “retrospectively” based on data already collected [5, 6, 9]. While it is possible to identify a statistically significant difference through various statistical comparisons within a database to “retrospectively” formulate a question based on the already-found answer, this approach is counterintuitive. The research question was specifically posed to collect data “prospectively,” and adopting a retrospective approach may lead to erroneously considering an effect occurring purely by chance within a database as an answer. This, in turn, could have no impact or even a negative impact on the current state of science. Therefore, every robust hypothesis (and question) must be posed before the data collection process.

When formulating scientific hypotheses, it is crucial that they meet specific criteria. For instance, hypotheses should be generally valid (not just applicable to an individual case), falsifiable, and logically consistent. Additionally, hypotheses should be logically derived and operationalized, ensuring that observations can be transformed into measurable variables. Hypotheses are often framed as conditional statements, such as “If–then” or “The higher X is, the higher or lower Y is.”

In the realm of a scientific study involving statistical significance testing, the initial hypothesis takes the form of a null hypothesis [3]. This asserts the absence of any difference or connection between two groups or variables. For instance, a null hypothesis might posit that there is no difference in the incidence of osteoarthritis between patients who underwent anterior cruciate ligament reconstruction and those who did not following an anterior cruciate ligament rupture. Subsequently, an alternative hypothesis is crafted, presenting the opposing view to the null hypothesis. In our specific scenario, it suggests a divergence in the incidence of osteoarthritis between patients with and without anterior cruciate ligament reconstruction for treating an anterior cruciate ligament rupture. Both the null and alternative hypotheses are stated in the study protocol or manuscript, but the statistical evaluation pivots on the null hypothesis, subjected to statistical testing. If a statistical difference is detected, the null hypothesis is rejected, and the alternative hypothesis is accepted. Conversely, if no significant difference emerges from statistical testing, the null hypothesis stands. It is noteworthy to mention the concept of one- or two-sided hypothesis testing. A two-sided hypothesis posits a difference between two groups without specifying the direction of the difference, considering whether, for instance, the outcome in group 1 is better or worse than that in group 2. The basis for testing a null or alternative hypothesis should thus always be two-sided, given the unknown direction of the potential difference, and the choice between one- or two-sided hypothesis testing can significantly impact statistical significance [6]. In contrast, a research hypothesis can be formulated one-sidedly; for example, positing that patients who undergo cruciate ligament reconstruction following an anterior cruciate ligament rupture exhibit lower rates of osteoarthritis than those who do not receive such reconstruction.

A well-crafted research question, coupled with a sound research hypothesis, serves as the cornerstone of a study’s methodology, exerting a profound influence on the design of the scientific work. Once these foundational elements have been meticulously addressed, the subsequent step involves determining the research objective.

The research objective serves as a guiding beacon throughout the entire research process, defining the purpose of a study. It should be distinctly articulated in the introduction of a research protocol or manuscript [10]. Unlike the hypothesis, the research objective fundamentally outlines how the research question will be addressed, often incorporating the study design [6]. Establishing a clear and precise research objective lays the foundation for a methodical implementation of the study, aiding in focusing on essential aspects and avoiding unnecessary effort. Building upon the earlier hypothesis example, the research objective might state that the study intends to compare the osteoarthritis rate in patients with and without cruciate ligament reconstruction after anterior cruciate ligament rupture over a follow-up period of at least 10 years. Notably, the research objective defines the study’s outcome parameters, which can be categorized into a primary (“primary objective”) and a secondary (“secondary objective”) research objective. In this instance, the primary goal is to examine the rate of osteoarthritis, while the secondary goal could involve collecting clinical outcome scores to assess both the osteoarthritis rate and the clinical outcomes of these patients. By outlining the outcome parameters and study design, the research objective also contributes to the calculation of the study power [10].

During the study’s execution, it is imperative to keep the research objective at the forefront to ensure that all research activities align with the intended goal.

In conclusion, the value of a research objective lies in conferring clinical relevance to the study through the selection of an appropriate outcome parameter. This significantly influences the impact of the study’s determined results on medicine and future research.