IntroductionTheoretically, doing what has been proven to be effective or stopping was is not working has the potential to provide health benefits or to reduce patient risks and waste of health care resources. As a consequence, it is essential that we find ways of implementing this evidence into practice. One possible way to implement evidence into practice is the use of instruments such as clinical practice guidelines (CPGs).Giving the broad spectrum of chronic illnesses treated in ambulatory care, the potential benefits of well-developed and effectively implemented CPGs and the various opportunities of electronic approaches, the objective of this thesis was to study the barriers and facilitators of electronic guideline implementation in ambulatory settings (part I) and to evaluate an innovative electronic development method (part II) for clinical practice guidelines.MethodsFor part I, we first performed a systematic review to synthesize current evidence on the effectiveness of computer-based guideline implementation systems in ambulatory care settings. A survey study was carried out to explore future opportunities of evidence-based practice guidelines implementation in occupational health and insurance medicine. We then evaluated family physicians acceptance of an electronic clinical decision support system (the EBMeDS system) in routine practice.For part II, we developed and evaluated a new consensus method based on human computation techniques. The human computation method is comparable with an online Delphi approach packaged as a game. The method has been proven to be effective in building large knowledge bases. Differences between the human computation method and an informal consensus method were measured by means of a randomized controlled trial in a population of students following a master of nursing and obstetrics and a population of trainees in family medicine. Primary study outcomes were differences in the changes in evidence scores and in agreement across multiple rounds for four clinical scenarios of low back pain.ResultsTwenty-seven studies were selected for analysis in our systematic review. With success defined as at least 50% of the outcome variables being significant, none of the studies were successful in improving patient outcomes. Only seven of seventeen studies that investigated process outcomes showed improvements in process of care variables compared with the usual care group. No incremental effect of the electronic implementation over the distribution of paper versions of the guideline was found, neither for the patient outcomes nor for the process outcomes.Our qualitative research showed that the majority of occupational health and insurance medicine physicians had a positive attitude towards evidence-based medicine and clinical practice guidelines. However, there were major barriers in the legislative framework and the information infrastructure that did not allow successful implementation within the given time-frame. Therefore, a shift to family medicine was made, where a new evidence-based decision support system, EBMeDS, was in its implementation phase.The respondents of our EBMeDS-study demonstrated a relatively high degree of acceptance. Mean intention to keep using the system was high. Their perception of the ease of use of the system, usefulness and the facilitating conditions was in general positive. Part of the statements gave mixed results and could be identified as important points of interest for future implementation initiatives and system improvements. In our RCT of guideline development, we demonstrated that our innovative method based on human computation may be a time efficient and acceptable method for guideline development specifically for the type of scenarios where the evidence shows no resonance with participants beliefs. When clear evidence was available and when some controversy about the evidence existed, changes in answers across rounds were more evidence-based in the HC groups compared to the IC groups. The human computation method was able to create a consensus to the same degree as an informal face-to-face method. Because it often takes years to achieve general acceptance of evidence by the medical profession, the HC method could also be very useful when new evidence emerges and existing guidelines need to be updated. When no evidence was available or when the evidence supported participants beliefs or usual practice, a higher degree of agreement was reached in the IC groups compared to the HC groups. Arguments used in the decision making process of the informal consensus method were mainly based on personal conviction or preferences. Very few evidence arguments were quoted during informal group discussions (7% in the population of students following a master of nursing and obstetrics versus 1% in the population of trainees in family medicine).DiscussionA number of studies have shown positive findings for electronic decision support systems in areas such as drug-dosing systems and computer-based reminder systems for e.g. preventive services. However, -although increasingly used and commercialised- little evidence exists at the moment for more complex electronic guideline implementation. This thesis made a significant contribution to the understanding of electronic guideline implementation barriers from a real-life perspective. Guideline recommendations with a high degree of certainty and agreement will more easily be implemented. More complex guidelines will not pass in a spontaneous way and calls for adaptive behaviors as well for the development as for the implementation process. We were able to demonstrate that the choice of the consensus method in guideline development might have an important influence on the guideline content. Our new method of human computation proved very useful in the introduction of clinical evidence arguments, while neutralising for psychosocial influences by authoritarian opinions.