Version 1.0 - March 2024

Hypothesis-Generating Research

Summary

Describe:

  • The specific research question
  • The precise and delimited purpose for identifying relevant data in relation to the research interest
  • The method by which data is collected and analyzed in accordance with scientific standards
  • The utility in terms of health benefits for society
  • The respect for participants' fundamental rights
  • The handling of any secondary findings following the feedback regulation
  • That health information and other confidential information are subject to confidentiality and legislation on personal data protection

Hypothesis-generating research projects aim to uncover potential correlations or patterns in extensive datasets with the goal of formulating new research questions or hypotheses that can be tested in future studies. In principle, the same legal rules and ethical principles of science apply to hypothesis-generating research projects as to hypothesis-testing research projects, but the former nonetheless requires special attention around a number of key points, which you can read more about in the following.

Going fishing?

Hypothesis-generating projects have traditionally been criticized for resembling a fishing expedition with a high risk of false findings. This can also be a problem with such exploratory or investigative studies if, to continue the analogy, one trawls a large section of the seabed without a closer plan for how to analyze and assess potential findings. However, exploratory studies can also contain significant scientific potentials if, so to speak, one selects a relatively precise (and delimited) area to investigate, uses relatively specific tools, and sorts thoroughly through the findings made along the way. Here, the analogy is rather a sport fisher instead of a trawler, who proceeds methodically in a delimited area with selected tools and a clear plan for how potential findings are sorted and handled. For hypothesis-generating research projects, it is thus crucial to specify where it is more specifically relevant to search for data, which methods are used to find relevant patterns or connections, and how results are analyzed and evaluated in the end.

Criteria for Approval

In addition to common ethical principles (as expressed through, for example, the Helsinki Declaration), a number of criteria for the approval of hypothesis-generating research projects have been established. There must thus be a concrete research question that differs from confirmatory (or hypothesis-testing) research projects, which intend to investigate a specific relationship. This could be, for example, the effect and safety of a drug when used as a treatment for a particular disease. Hypothesis-generating research projects have a completely different exploratory character. An example of this kind of research question could be: "How have genetic variations in human populations evolved over time, and how does this affect our health and adaptability" or "What is the cause of increased risk of allergic disorders in children with a family history of allergy".

A sufficiently good hypothesis-generating project is overall characterized by having 1) a precisely defined purpose, so it is clear what is relevant data and why this data is relevant 2) a delimited goal, so that no more data is collected than necessary with respect to the research interest, 3) an acceptable method, to ensure that data is collected and analyzed in accordance with scientific standards, 4) sufficient utility, regarding the overall potential gain for public health, 5) a fair nature, when it comes to participants' fundamental rights, 6) a considerate approach, regarding the handling of any secondary findings, and 7) a respectful approach, when it comes to dealing with participants' personal information.

Specified Purpose

Although hypothesis-generating projects by nature do not investigate a traditional hypothesis, such projects still have a purpose or a basic research question that can be more or less precisely defined. This purpose/question will usually be supported by theoretical assumptions based on relevant literature or clinical experiences, which can specify clear relevance criteria for the data one wishes to examine, as well as substantiate the health science potential and aim of the project. Thus, the good research question for a hypothesis-generating project is an expression of knowledge, and not solely an expression of ignorance, although hypothesis-generating research is primarily applied in areas where available knowledge is low.

Research Interest

As mentioned, hypothesis-generating projects also need to be delimited in relation to the research interest underlying the project. This means that the data one wishes to examine must reasonably be necessary to say something about whether there are interesting patterns or connections in the material – in relation to the project's overarching aim and purpose. This is so both to ensure that unnecessary data is not collected and to minimize the risk of false findings. If one searches for a sufficiently large number of different variables, it will likely be the case that some of these variables will, by chance, form patterns that appear significant. For example, if investigating whether there is a connection between the distance of residence to a water plant and a range of different diseases, it would almost be strange if there wasn’t at least one disease where the incidence randomly clusters around several water plants.

Method

Hypothesis-generating research projects must also have a clear method that can potentially deliver valid and valuable results. This means that the results (patterns or connections) generated through the project must be reliable (for example, because it is possible to replicate the results), and the project can reasonably be expected to bring new insights that justify initiating the project in the long term. It must thus be clear, 1) how relevant data is selected (or generated), 2) how (and when) data is collected, 3) how data is sorted and analyzed, and 4) how any scientifically interesting findings are handled. Furthermore, a description of the project's method should address potential risks for noise or bias in the interpretation of data – this could be, for example, sampling bias, meaning that the selected data does not represent the group one wishes to generalize to, or confirmation bias, where one ends up confirming one's initial expectations at the expense of alternative patterns or connections in one's data that are more plausible or interesting to investigate. The latter can be particularly challenging in hypothesis-generating projects.

Health Benefits vs. Potential Burden on Participants

Hypothesis-generating projects, like hypothesis-testing research projects, must overall create sufficient benefit to justify any disadvantages, nuisances, or burdens that might be associated with the project for both participants and the surrounding society. Hypothesis-generating projects should thus potentially create value for the broader society that matches any risks or negative consequences associated with the project. This includes ensuring that there hasn't already been a significant amount of research in the area that could reasonably form the basis for testable hypotheses, if the project is to contribute something useful. It should also be clear that the project will illuminate a fundamental issue that is actually important (or rather: useful) to shed light on. Naturally, it can be challenging to describe and assess the potential utility of exploratory studies, as the scope of possible outcomes for the project isn't as well known as with confirmatory research projects. Nonetheless, it is still possible to describe what potential new or more qualified hypotheses could contribute within a relatively delimited research area.

Participants' Rights Take Precedence Over Science

Hypothesis-generating projects must not only contain the potential to create value. Any risks, disadvantages, nuisances, and burdens associated with the project must also be acceptable in relation to the individual participant, whose fundamental rights take precedence over the pursuit of scientific progress. The principle is that hypothesis-generating projects should not be a burden to the individual participant. Generally, any potential negative consequences must be considered in light of whether the individual participant (or the group they are part of) can expect to benefit from the execution of the project – and in cases involving informed consent, potential participants should fundamentally be involved in accordance with the potential negative impact that the project could be expected to have on the individual.

In the field of registry research, the ethical review committee must take into account individual considerations and weigh any potential burden.

Possibility of Secondary Significant Health Findings?

Hypothesis-generating projects with extensive datasets often have the potential for the emergence of secondary significant health-related information during the project.

Significant secondary health findings are information that emerges in a project without being covered by the project's purpose, about the participant unexpectedly suffering from or with certainty or high probability is predisposed to get a life-threatening or clearly serious illness, which can be treated, prevented, or alleviated.

It is worth noting in this context that feedback can be greatly beneficial for participants, just as the lack of feedback can risk damaging trust in researchers if it later turns out there was something to report back. Conversely, feedback can also create unnecessary worry or pose a burden on the healthcare system, which has to follow up on the information that has been reported back. It must therefore be ensured that any feedback is handled in accordance with the rules in the feedback regulation[1] and established ethical principles. If it is assessed that there is a predominant possibility for the emergence of significant secondary health findings in the hypothesis-generating project, the researcher must describe the composition of an expert committee, which is to be established upon the emergence of such findings.

The starting point of the feedback regulation is that only serious secondary health findings are reported back, where

  1. the disease or disease predisposition can be significantly prevented, treated, or alleviated,
  2. the disease or disease predisposition has significant importance for the participant or research participant,
  3. the clinical validity of the finding, and
  4. the method for detecting the finding is secure.

Therefore, the expert committee must consist of a licensed health professional within the disease area being researched, and must additionally consist of members who possess the necessary expertise to assess the aforementioned 4 conditions.

It is the researcher's responsibility to ensure that no feedback on secondary findings is given to the participant if they have requested not to receive feedback.

Protection of Health Information and Other Confidential Information

Finally, in connection with hypothesis-generating projects, it must be ensured that participants' personal information is treated confidentially and with respect for the individual's right to privacy regarding their personal health information.

This involves compliance with legislation, including the EU regulation, on personal data protection as well as the health act. This includes, among other things, not collecting more data than absolutely necessary and having a clear and legitimate purpose for the personal information that one wishes to use in connection with the project.

Key Points that Elaborate on the Checklist on the Website

  • Describe the health science aim of the project
    • Which diseases/disease groups does the project target?
    • What is the nature of the potential connections that might be found?
    • What health benefits could potentially come from insights generated through the project?
  • Describe the project's theoretical assumptions about what counts as relevant data
  • Describe how data is collected
  • Describe how data is analyzed
  • Describe how results will be measured or assessed
  • Describe how potentially relevant results can be identified through the project
  • In consent projects: Describe for participants what a hypothesis-generating project is, and what is hoped to be achieved with the project
     

Footnotes

1. Executive Order No. 736 of May 24, 2022, on feedback on significant health findings from notifiable health science and health data science research projects, clinical trials of medical devices, performance evaluation of medical devices for in vitro diagnostics, and certain registry research projects

Last updated 20-03-2024
Icon of a scroll with text lines and a pencil.

Content on this page