National and international guidelines outlining ethical conduct in research involving humans and animals have evolved into large and complex documents making the process of gaining ethics approval a complicated task for researchers in the area. Researchers, in particular those who are relatively new to the ethics approval process, can struggle to understand the parts of an ethics guideline that apply to their research and the nature of their ethical obligations to trial participants. With the scope of medical research likely to continue to expand in the future, it is clear that ethics guidelines will only increase in complexity and number. This paper describes one possible solution to the problem: the use of an internet-based expert system to intelligently and interactively distribute the information stored in ethics guidelines to individual researchers. This paper also details how one such system was designed and tested with respect to Australian medical research ethics guidelines.
Statistics from Altmetric.com
“… an experiment is ethical or not at its inception; it does not become ethical post hoc—ends do not justify the means” – Henry Beecher1
Guidelines on ethical conduct in medical research serve an important role in the development of medical devices, therapies and interventions. New products can be subject to a range of non-clinical (animal) and clinical (human) trials and the main purpose of the guidelines on ethical conduct is to protect participants in such research, human or animal. However, due to the continuing growth of the medical field, guidelines on ethical conduct are expanding in their scope and complexity with each new revision. Furthermore, different regions of the world have different guidelines and there is little hope of an internationally ratified guideline in the near future. Therefore, researchers must operate in an increasingly complex system of international guidelines if their research involves work with humans or animals.
Information distribution issues
Ethical oversight frameworks worldwide are immensely complicated in both documentation and organisational terms. Due to the large variety of research areas covered by ethical guidelines, applications must be reviewed on a case-by-case basis. Often this involves large amounts of work preparing and reviewing applications until the research is deemed ethically justifiable. As Howard states, “Gaining ethical approval can be an intimidating process, both for new and experienced researchers alike”.2
The current system cannot be simplified to any great extent. In fact, the system is likely to develop further in the future due to the ever growing number of research areas involved. In addition, each region (such as the USA, the EU or Australia) has its own system for ethical oversight. The end result of this level of complexity is that the researcher may struggle to understand their ethical obligations and the processes required to gain ethics approval for their study.2–4
The current system relies on knowledge diffusion to work effectively. This means that new or inexperienced researchers learn how to navigate the ethical oversight system from more experienced researchers. However, there is always the risk that new researchers may miss some critical piece of information. Henry Beecher’s observation suggests that a proposed trial is deemed ethical or not at its inception1 and it can be concluded that, for a trial to have the best possible chance of being performed in an ethical manner, the researcher must have a thorough understanding of their specific ethical obligations to trial participants when the research is in the planning phase. Thus, easy access to relevant information stored in ethics guidelines is critical during this preliminary phase.
Ensuring that researchers are highly trained in the theory and practice behind ethical oversight systems is an important factor in the success of the overall process and the protection of trial participants. The quality of information provided to participants has been identified as an issue of concern in some research5 and this may relate to issues with the quality of information and the way it is provided to the researchers who design the studies. Rather than research ethics committees (RECs) requiring researchers to re-evaluate their methodology until it is acceptable, the problem could be pre-empted by providing adequate knowledge to researchers during the study design phase. Of course, such an information distribution system would not be a replacement for the REC and would only be used as a tool to facilitate the process of ethics approval.
Such a system should be able to translate large standards and guidelines for the user and thus increase their understanding of the ethics approval process while reducing the time and effort expended in gaining the necessary knowledge. In essence, the system should be able to produce an acceptable and relevant solution with respect to the nature of the research being undertaken.
Another problem with the current system is the fact that ethics guidelines are constantly adapting. As new areas of medical research develop, guidelines in all regions go through several revisions. All codes are not updated at the same time and thus a researcher must keep abreast of any changes in the guidelines themselves. The information distribution system should also take this into account by providing a single location where the most recent changes to ethics guidelines are reflected.
Internet-based expert systems
An “expert system” is a program that can solve a real-world problem that would normally require a human expert.6 The concept of an expert system was developed in the 1960s and has since been refined and applied to many fields.
Traditional expert systems had some limitations. These included accessibility, software distribution and communication issues.7 However, the advent of the internet-based expert system has allowed these problems to be overcome and thus opened many new areas where expert systems could prove useful. One of the main benefits of internet-based expert systems is that they can be made extremely portable and platform independent.7
Expert systems also have the benefit of being able to provide information using hierarchical categorisation developed using the knowledge and analytical skills of an expert in the area.8 Therefore, in an area such as medical research ethics, where the ethical requirements of research trials must be evaluated on a case-by-case basis, an expert system can be used to provide useful advice even in cases where there is no obvious simple solution.
The internet-based expert system solution
Ethics codes can be integrated into an internet-based expert system to provide an intelligent, interactive portal for knowledge distribution. By asking the user a range of questions and analysing the responses, the system can deliver relevant information from ethics guidelines with respect to the user’s proposed research.
Furthermore, an internet-based expert system can be made to be easily updatable, easy to access and user friendly and go towards solving the issues with the current systems described above.
When a guideline or standard is created, it is often made with the intent of covering all possibilities in its area of application. However, when individuals or organisations use such guidelines or standards, they require a specific solution to their particular problem. Therefore, under such circumstances, guidelines and standards implemented as expert systems could prove to be very powerful tools.
An example of such a system, the Computer Aided System for Ethical Decision making (CASED), has been created for use with Australian research ethics guidelines.
Creation of CASED
In order to be useful to a wide range of users with different levels of expertise, CASED was designed as a hierarchical system comprising a high-level system and a low-level system. The high-level system contains general information about the core values behind medical research ethics guidelines worldwide. This includes explanations of the basic principles such as respect for persons, beneficence and justice, for human trial participants, and replacement, reduction and refinement, for animal trial subjects.9 10 The high-level system is intended to be used by those who are new to the ethics approval process.
The low-level system contains information which is specific to the Australian ethical guidelines for medical research. Due to the fact that there are separate guidelines for human and animal research, two separate low-level systems were created. They are based on the National statement on ethical conduct in human research and the Australian code of practice for the care and use of animals for scientific purposes, which are the research ethics guidelines produced by the Australian National Health and Medical Research Committee (NHMRC).11 12 The low-level system is designed to be used as part of the research design and ethics approval process.
Structure of high-level and low-level systems
The difference between the type of information provided by the high-level and low-level systems created a difference in the structure of the systems themselves. Figure 1 shows the structure of the two systems.
Both systems comprise two main components, an inference engine and a knowledge base as defined in table 1.
The inference engine
To create an internet-based expert system which was platform independent, a scripting programming language, PHP, was used. This means that all the processing or “logic” of the system happens on the web server where it is hosted and the system will operate regardless of the user’s operating system or web browser.
The knowledge base
As shown in fig 1, a wiki was chosen to be the knowledge base for the high-level system. A wiki is a website that allows (and encourages) users to add and modify content. Thus, the high-level system’s database is dynamic. The reason behind this choice is the nature of the information to be stored in the high-level knowledge base. High-level ethical concepts are open to wide interpretation and so constraining the system to use only the authors’ interpretations could detract from the system’s aim to be universally relevant. Rather than having users constantly contact the system developer asking for changes to the knowledge base, the wiki approach allows users to directly add their knowledge to the database.
A possible disadvantage to the wiki-based approach is that users may inadvertently or maliciously change content in a detrimental fashion. In order to prevent this from being a problem, users are required to open an account with the CASED wiki before they are allowed to edit content. Also, when any change is made to the wiki, the system developer is notified and can roll back the change if it is deemed inappropriate.
The inference engine for the low-level system requires a more robust database system. Also, the information stored in this database is not user modifiable. Therefore, MySQL (MySQL Inc, Cupertino, California, USA) was chosen as the database system for this component.
To make the expert system internet-based, a supporting framework was created along with the system. Firstly, a website to host the system was created. Secondly, the wiki which plays a dual role as part of the system and as part of the framework was created. Thirdly, a feedback system was established for the system and the supporting framework.
Feedback is critical to the success of such an expert system. As the system will need to remain relevant in the event of changes to the ethics codes, users need to be able to provide feedback to the system developer. Also, user feedback can provide information about bugs in the system and any changes that users may want to see in future versions of the system. These feedback requirements were addressed in two ways. Firstly, forums were added to the supporting framework. Not only do these forums provide a feedback path, but they also encourage discussion among users of the system. Secondly, the wiki plays a role in the feedback process. Users can add, remove or modify content as a form of feedback. They can also post comments on the wiki for the system developer or other users to read.
Using the system
When users enter CASED, they are presented with the choice of using the high-level system, the low-level human system or the low-level animal system. To improve the user friendliness of CASED, all three systems have a similar interface, an example of which is shown in fig 3.
The user is presented with a set of “critical questions” to determine the user’s progression through the chosen system. The answers provided to these questions allow CASED to gain an understanding of the research being conducted. The critical questions were mostly derived from the topic areas within the ethics guideline.
Using the answers provided to the critical questions, the system then provides the user with a set of “specific questions”. These questions were mostly derived from the specific clauses within the ethics guidelines and allow CASED to gain an understanding of what ethical issues the researcher has taken/will be taking into consideration.
Using the answers provided to both the critical and specific questions, CASED derives a unique solution of the ethical obligations which are relevant to the research being conducted. An average run through the CASED system takes 15–20 minutes.
A sample of the question and answer process of the CASED low-level system for human research in Australia is shown in box 1.
Box 1 Sample questions
Is there a requirement for Limited Disclosure during the informed consent process? (ie, Is it necessary to withhold information about your research from participants for any reason?)
If the user answers “YES” then the following Specific Questions are presented.
Specific Question 1:
Does your requirement for limited disclosure involve active concealment or planned deception?
Specific Question 2:
Is the reason behind your requirement for limited disclosure to expose illegal activity?
Specific Question 3:
Do you require a waiver for the requirement of consent?
An answer “YES” to Question 1 gives:
Information regarding the specific responsibilities researchers have for participants when performing active concealment or planned deception.
An answer “YES” to Question 2 gives:
A notice that users should consult the National Statement for further information as the particular portion of Statement relevant to this question is not covered by the CASED expert system.
An answer to “YES” to Question 3 gives:
Informs the user that only an REC can give a waiver for consent and points them to relevant portions of the National Statement to read.
An answer of “YES” to the Critical question but “NO” to all three Specific questions gives:
Information regarding limited disclosure that does not involve active concealment or planned deception.
RESULTS AND DISCUSSION
CASED was subjected to a limited evaluation by asking members of RECs to provide feedback on the expert system. In all, 52 requests for participation were sent to experts from five Australian universities. Of these, 12 responses from three universities were received.
The feedback indicated that the CASED concept was generally well received. The respondents suggested a number of areas where CASED could be improved including:
Allowing more effective access to the guidelines on the NHMRC website
Providing direct links to relevant sections within the guideline being referred to when CASED formulates a solution
Incorporating CASED into the ethics approval process, and
Expanding CASED to cover more qualitative research areas.
In response to these suggestions, some changes were made to the CASED system. A link to the relevant codes on the NHMRC website is now included when CASED provides its solution to a user. Also, in the next scheduled update to CASED, the system will be expanded to incorporate more qualitative research areas.
While including direct links to the relevant section of the code being referred to would be useful to researchers, the NHMRC provided its guidelines in pdf format only when the system was developed and so this was not possible. Subsequently, the National Statement on Ethical Conduct in Research involving Humans12 has been published in html format and future versions of CASED will include direct links to relevant sections of the human guideline and later, to the animal guideline.
Some respondents suggested that CASED should be incorporated into an online ethics approval process. However, it is important that CASED maintains the users’ autonomy and does not make decisions about the ethics approval application on their behalf. For this reason, CASED will remain strictly an information distribution system which can only advise users on their ethical obligations.
An interactive internet-based expert system (CASED) was created to provide information to researchers intending to conduct studies requiring ethics approval. Positive feedback from ethics experts who reviewed CASED suggests that this preliminary implementation of internet-based expert systems for this application has been successful. The model used in the development of the Australian system could be used to accommodate other ethics guidelines and standards from other international jurisdictions. In fact, the system could be scaled to suit many other types of standards or guidelines other than those related to medical research ethics.
The authors wish to acknowledge H Outhred for his contribution to the design of the CASED expert system.
Competing interests: None.