Go through academic routine with a smile

Clinical Decision Support Systems

Clinical Decision Support Systems

Clinical Decision Support Systems (CDSSs) play an influential role in safeguarding medical data and therefore reducing medical errors. Also, CDSSs are important in improving the quality and efficiency of healthcare service delivery. Furthermore, widely promoted evidence-based medicine is able to ensure improved clinical outcomes. Evidence-based medicine means the practice of medicine which entails the most authentic and available scientific evidence. It also involves the management of particular patients via specific individual clinical experts and is relevant to judicious as well as conscientious current clinical care research. The approach is susceptible to mistakes such as incomplete or missing data, and low-quality evidence and will require the application of clinical judgment. Thus, the use of CDSSs can provide tangible assistance in improving and facilitating evidence-based medicine. The role of CDSSs is significant when regarding the improvement of the quality of health care. Hence, the policy challenges in capturing practice-based evidence in machine-interruptible repositories require the appropriate recommendations on the basis of their extensive analysis for improving the development and acquisition of clinical decision support for systems application in the field of evidence-based medicine.

Research Literature in Evidence-Based Medicine

Clinical decision support systems can only be as important and reliable as their respective evidence base. It means that the relevance and effectiveness of CDSSs depend upon the strength or deficiencies entailed in research evidence. Hence, when developing clinical decision support systems, it is quite important to acquire more clinical research evidence as well as to find high-quality, actionable, and useful evidence that is appropriate to the current case. Such evidence should also be accessible and machine-interpretable. However, the impossibility of the interpretation of such information shows its ineffectiveness in boosting practice-based evidence in machine-interpretable repositories.

Research shows that therapeutic interventions used in both outpatient and inpatient care concerning family and internal medicine have insufficient support with evidence of efficacy within the research literature. There are the areas that have only equivocal supportive evidence while the other areas remain unstudied at all. There are many problems encountered when examining the research literature for evidence-based medicine. One of the problems is that the efficacy studies of the clinical practice forming the foundation for evidence-based medicine constitute only a small part of the entire research literature (Kendall, Ryu, & Walsh, 2017). Additionally, the clinical research literature has deviated for many years with the study design and reporting problems. The problems still exist in the current randomized trials, guidelines literature, and systematic reviews. It is startling to note that the research literature continues to explode while the problems remain. It explains why many clinicians consider the research literature unmanageable. Some of them claim that the research literature shows limited applicability to their clinical practices.

The full expectation of CDSSs to facilitate evidence-based medicine is taking place; however, some challenges, such as in policy-making, still exist and derail the promise. If the CDSSs can maintain the systematic flow of literature, it is possible to improve the healthcare system greatly. It means that they can closely monitor new relevant studies, be selective in terms of high-quality literature, as well as incorporate the best evidence within the specific patient assessment (Mustafa & Abdullah, 2017). Automation of those particular approaches remains an open area for research. Some of the best evidence-based medicine systems are BMJ Best Practice, Covidence, Cochrane Interactive Learning, DynaMed, Up-to-date, (On-Campus Access Only) Clinical Evidence, PubMed, Best Evidence, and Cochrane Library, among others (Yu, 2018). They provide up-to-date research literature to ensure the solidity of the foundation of evidence-based medicine. The information available on such resources should always be machine-interpretable; otherwise, they would be inapplicable to the current CDSSs.

The practice-based evidence is always necessary when optimizing health outcomes; however, the research literature serves as the basis for evidence-based medicine. For instance, randomized trials have revealed that patients who have symptomatic carotid artery stenosis may have fewer strokes in case of a surgery known as carotid endarterectomy. If complication issues resulting from the particular surgery are above 6 percent, the benefits of the surgery will definitely come to a nullity. Also, the research shows that only about 30 percent of physicians know the CEA complication rates for the hospitals where they work or refer their patients to. Various clinical problems may entail variable aspects; thus, the creators of CDSSs should focus on obtaining local practice-based evidence in order to complement various literature-based evidence.

Concerning the development of practice guidelines, practice-based evidence becomes very useful. Despite the fact that evidence support for individual decisions usually arises from literature-based evidence, according to the information mentioned above, a guideline development process relies on experts’ opinions. If reliable practice-based evidence is available, guideline developers will definitely be in a better position to improve the flow of the various design processes. Thereby, the evidence entails real experience in operation areas. It becomes easy for the developers to design something linked to the practice. Hence, for the development of the practice guidelines, practice-based evidence is paramount for all practitioners of healthcare delivery.

Due to the importance of practice-based evidence, it may be uneasy to obtain it. The informatics communities may advance the sector by developing comprehensive technologies that are appropriate for practice-based evidence. Such research is much essential and should always encourage informatics to further development via constructive technology. The technological networks will make it possible and easy to capture the clinical process and events automatically. Then, deductions can be available from the recorded process and whatever transpired within the process (Cairney & Oliver, 2017). It would also be applicable in outpatient settings, thus making it possible to harmonize different findings to define their interrelationship. As usual, patient data privacy, standardization of data, and data ownership remain critical and an area of debate due to policy-making. Hence, it introduces a challenge in making policies when capturing practice-based evidence.

Apart from patient-based evidence and evidence-based medicine, patient-directed medicine is another interesting area that needs analyzing with respect to the prior two. While Internet sources and many others have provided platforms for patients and thus allow obtaining medical information, they have also created an arena for misinterpretation of data. It has made patients less dependent on clinicians for information; however, they still trust clinicians in choosing, appraising, applying, and understanding data for healthcare decisions. Clinical decision support systems can assist in fostering the relationship between patients and clinicians by making sure that the patients are accessing the most relevant information. The system should be able to provide both clinicians and patients with applicable, reliable, useful, and result-oriented healthcare information; it will result in comprehensive care decisions which are coherent with the current recommendations.

CDSSs have assured of the improvement of healthcare; however, the evaluations show that it still remains undone. They have been modest in improving intermediate measures such as drug dosing accuracy and guidelines adherence. Although the effects of CDSSs on clinical outcomes are clear, the expectations are far greater. It requires continuous assessing CDSSs in order to determine their impacts on the clinical process. The incorporation of both quantitative and qualitative methods in evaluating the organizational and clinical impact of CDSSs allows for providing complementary insight into the effects of CDSSs (Pal, Tomar, & Singh, 2013). All types of evaluation approaches require more attention and funding. Looking into the current mistakes of medicine, a special class of evaluation approach should appear to prevent them in the future. The applications of iterative evaluations in the ongoing studies, as well as the use of redesigns of CDSSs, are necessary. CDSSs should be able to capture and notify clinicians about the mistakes and dangers identified by the system.

There are several recommendations made to evaluators in order to determine the full use of a CDSS. The evaluators should assess the CDSSs with the use of an iterative approach that would identify both the pros and the unanticipated cons related. The iterative approaches are also valuable in identifying problems associated with the implementation of the CDSSs. Clinical decision support systems can significantly benefit from multiple stages, testing types, and other points of CDSSs’ lifecycle (Ibid). More CDSS evaluations are achievable in practice settings, such as ambulatory setups. The qualitative approach is able to encompass organizational behaviorists, ethnographers, and sociologists among other dimensions of researchers entailed within and without the medical informatics community. If the primary evaluations show that the CDSS can positively influence health outcomes, the CDSS requires further examination with the aim to define its extensive benefits. Furthermore, the process of evaluating necessitates partnerships between academic groups and community practices.

There are several challenges associated with the process of capturing practice-based evidence in machine-interpretable repositories. Such challenges range from the ethical point of view to the privacy of patients’ medical data. For example, a clinician has obtained the information that he or she believes comprises the best practices for the patient; it may require time to implement such practices. Thus, it means that the clinician should also have taken into account the patient’s considerations on the issue (Farley et al., 2009). Assuming that the clinician found that female juvenile sexual offenders usually benefit from trauma-founded models of intervention, it would be clear that the agency’s policies refuse to support the utilization of trauma-based models. It has already hindered the implementation of the model because the established policies avoid supporting such practices. It also becomes a challenge when convincing policymakers to amend the respective policies and incorporate such models. The clinician will have found themselves in an ethical dilemma to either go against the established policies or other practice guidelines. It would be unethical for them to provide an evidence-based intervention and may lead to ignoring the findings.

Although almost all agencies report that they support evidence-based practice, when it comes to establishing the best evidence-supported amendments in procedures and policies, substantial resistance appears. Some of the items that determine such amendments are leaders’ opinions, clients’ expectations, advocacy, the standard of the practice, the practice environment, the sense of competence, financial training, the compulsion to act, the perception of liability, and the organizational constraints, among others (Cairney & Oliver, 2017). When it comes to the practitioner’s implementation of evidence-based practice, organizational policy issues will definitely appear in the determinants mentioned above. Some of the other challenges are the lack of time and resources to perform practice-based medicine in the workplace. In a real sense, clinicians face an overwhelming amount of direct service guidelines, as well as caseloads.

The policy application of evidence-based practice normally takes various forms. Firstly, there is the usual use of evidence-practice programs that refer to outcomes and define the best essential practices in the attempt to secure funding organizations as well as to assist in managing liability. The extensive demonstration of evidence-based practice and policies arises from the evidence utilization in the construction and promotion of clinical policies, as well as the underwriting of the organizations which implement them (Head, 2017). Another form of policy application in evidence-based practice depends on the organization. It concerns professional and organizational policies which restrict practitioners to use standard practices as a part of their professional work. Thus it would warn clinicians of departing from the ethical guidelines, even though their findings entail the best practice evidence.

Learning healthcare systems (LHS) have proven their significance in improving practice-based medicine. As defined by the Institute of Medicine (IOM) in 2015, LHS is a system where sciences, cultures, informatics, and incentives are optimal for consistent improvement and innovation and incorporate the best practices embedded within the delivery process. It shows that learning healthcare systems has to address the issue concerning bolstering practice-based medicine via the research literature based on the best clinician practices. The lifecycle of learning healthcare systems involves five distinct stages: assembly of evidence, analysis of findings, interpretation of data, evaluation of feedback, and change of practices (Budrionis & Bellika, 2016). The first step requires gathering information from different fields and practitioners; it ensures that sufficient information is available for analysis. In the stage of analysis, the information becomes a subject of detailed analysis according to various parameters of practice. The next step assumes the interpretation of the results of the analysis into their real meaning and thus making them applicable for practice. After the implementation of the practices, the respective developers receive feedback on the effect caused by such practices. The suggestion and implementation of the appropriate changes fall within the feedback stage, as well as in the case of the new evidence discovered.

These five stages make learning healthcare practice very stable and reliable. The policy-making process also becomes quite easy due to the carrying out of the proper analysis. Sometimes in evidence-based practice, a tendency to bypass the analysis, interpretation, and feedback stages appears in cases when a clinician discovers the best practice evidence and attempts to utilize such practices without incorporating comprehensive and extensive analysis (Mullins, Wingate, Edwards, Tofade, & Wutoh, 2018). Thus, the adoption of the learning healthcare system into practice-based medicine will make it possible to advance the sector. However, learning healthcare systems avoid leaning entirely on machine interpretability. So, the findings discovered through such means should be proven to be machine-interpretable in order to use clinical decision support systems. Accordingly, it would become possible to merge the literature evidence with technological developments.

Predictive analytics could also be important for improving practice-based medicine. The predictive analysis comes with more accurate diagnoses, patient engagement, and improved treatment plans. Predictive analytics deal with huge amounts of data and information such as medical research (Linda & Miner, 2014). It means that the outcomes achieved are more accurate than searching through a limited number of sources. Notably, evidence-based practice focuses a lot on ensuring the relevant, useful, and high-quality evidence offered by predictive analytics. Again, predictive analytics allows clinicians to go behold evidence-based medicine by assisting them in constructing highly individualized treatment programs and plans. Predictive analysis can also play a significant role in getting patients involved in their healthcare. Activities, such as recording patients’ periods of sleep, can assist in getting important predictive information about their health. Hence, predictive analysis can play a major role in practice-based medicine; however, the information obtained should be machine-interpretable.


The difference between the current CDSSs and the role that the CDSSs should play in fostering both patience-based and evidence-based medicine requires more research and development agenda. If clinical research continues to improve clinical care, it means its relevance, accessibility, and high quality. The research should clearly display the aspects of efficacy, cost-effectiveness, and the effective applicability of typical inpatient and outpatient practice setups. As CDSSs are useful in translating research into real clinical practice, they must incorporate direct machine-interpretable access to the literature researched (Djulbegovic, 2018). It will allow incorporating and utilizing automatic methods which would assist in the update of information and the other available data. Hence, the creation of machine-interpretable content, shareable information, and relevant knowledge bases is quite crucial for both practice-based and research-based medicine.

Clinical and informatics researchers also need to improve their approaches. They should conduct efficient and effective research based on clinical interventions, specifically of primary care setup. Informatics should continue to develop better techniques for synthesized results with a wider view of study designs, randomized trials, and observational studies (Wells, 2012). Both studies should be in the stage of developing machine-interpretable repositories, up-to-date evidence, shareable, and of multiple types. They should also focus on the development of guidelines linked to up-to-date evidence repositories. Also, they should go ahead to establish standard interfaces among the respective repositories that would enable the evidence to display automatically. This would assist in decision modeling, maintenance, and guideline development into a better systematic flow of information.

It is also crucial to develop a comprehensive and expressive clinical language that supports scalability from administrative to the clinical decision support system’s needs. The developers should continue to construct shareable computer-based representations of the practice guidelines, as well as clinical logic. The created tools should make it possible for editors to incorporate new literature-based evidence into the CDSSs’ knowledge bases. The researchers should also specify the context, where the knowledge is applicable, as well as customize the literature-based evidence for local situations and other conditions deemed local (Baynouna & Ketbi, 2018). For example, it is crucial to factor out local surgical complications. The developers should explore and construct automatic approaches for updating CDSSs’ knowledge bases. It would allow the CDSSs to reflect the current state of the research and improve its quality. Additionally, there is a need for the development of more flexible models for decision-making in order to ease the accommodation of the clinical evidence for varying methodologies. Also, the variability can be in terms of strength and quality; the system should be in a position of determining the more reliable and high-quality data.

Clinical decision-making systems should also be able to simultaneously accommodate the values, beliefs, and perspectives of multiple decision-makers. Some of such decision-makers include physicians and patients as well. It is crucial to explain explicitly the care delivery setup as well as the clinical scenarios for which the CDSSs are applicable. For example, diabetes treatment CDSSs necessitate specification to define whether they are appropriate for the stable management of outpatients only. CDSSs should integrate with electronic healthcare records and other reliable systems that use the best interoperability standards. Currently, an insufficient number of CDSSs for outpatients is available; society requires the development of more CDSSs for outpatients and inpatients as well. There is a demand for treatment of an increasingly aging population that requires complex diagnostic procedures, treatment, as well as supportive services; the issue is highly essential when developing CDSSs.


Therefore, it is clear that practice-based medicine is important when it comes to improving healthcare delivery. The information collected via the literature should be reliable, high-quality, and machine-interpretable. Some policy challenges to practice-based medicine include the lack of support from organizations, insufficient funding, and ethical dilemmas. Clinicians, torn between implementing their best practice evidence and sticking by the organization’s standard practice, necessitate salvation of the problem. The recommendation states that developers should attempt to design easily updated models. Such models should be able to accommodate new information and be applicable diversely. Evaluators should carry out a continuous analysis of CDSSs to determine their effects. If CDSSs function inefficiently, the system requires the appropriate changes. Learning healthcare systems easily integrates with practice-based medicine because they involve the acquisition of the best practices. The predictive analysis also provides the CDSSs with accurate data, better treatment plans, and better patient engagement. Hence, practice-based medicine can face significant improvement through CDSSs that entail the best current information.