Over the last 90 years, Canadians have gained an average of 24.6 years of life expectancy according to the Conference Board of Canada [1]. Over the same period, the most common place of death has shifted from home to hospital [2]. Therefore, a growing number of older and sicker patients are being admitted to an intensive care unit (ICU) where either they or their family members will have to make crucial decisions about their goals of care [3]. One of these decisions is to decide whether or not to conduct cardiopulmonary resuscitation (CPR) in the event of a cardiac arrest.

Data from the Get with the Guidelines-Resuscitation registry, the world’s largest database on in-hospital cardiopulmonary arrest from over 400 participating hospitals in the United States, estimate that the incidence of treated cardiac arrest is 0.92/1000 patient days. Almost half of these events (48%) occur in the ICU [4]. Over the years, CPR has become a standard intervention used in almost all cases of cardiac or respiratory arrest unless a do-not-resuscitate order (DNR) is recorded in the patient’s chart [5]. However, estimated survival to discharge after treatment for in-hospital cardiac arrest is 18%, with about half of these patients being neurologically intact or having only mild neurologic deficits at discharge [4].

Because cardiac arrest can have potentially devastating outcomes even if CPR is performed, making the decision about whether or not to initiate CPR should involve the patient or his/her surrogate decision maker(s) in cases where the patient is unable to participate in decision making [4]. Discussions about advance directives for CPR in the ICU are a major source of stress and grief for patients, their family members, and their health care professionals. Shared decision making (SDM), a process in which health care professionals and patients work together to make health care choices, offers opportunities to facilitate discussions about advance directives regarding CPR in ICU settings [6].

A common way to support SDM is the use of patient decision aids (DAs). DAs are generally used in situations where a range of treatments are possible, with each treatment having its own advantages and disadvantages. They are intended to be used by patients to complement health care professionals’ counselling. DAs aim to clearly present the best available evidence and to support patients’ reflections about their values and preferences [7].

There exists a series of multiplatform DAs about CPR developed and owned by ACP Decision, an American nonprofit foundation, and adapted to various care settings, illness, and decision makers (patients or surrogate decision makers). Although these DAs are available in multiple languages, there currently lacks a French version that could be used by the population in the Province of Québec where the predominant language is French. Moreover, ACP Decision’s DAs are currently only available to organizations within the United States. We were granted permission to adapt another DA about CPR that was developed in Canada, the “Cardiopulmonary resuscitation: decision aid for patients and their families” [5]. However, in addition to being only available in English, it was initially designed for patients hospitalized in the general hospital ward. Patients admitted to the ICU are critically ill and have specific contextual factors (eg, urgency to make a decision, difficulty reading, unstable condition) that need to be considered in the creation of a DA designed for their use. Moreover, this DA only presents population-level statistics about the outcomes of CPR classified by age category and by the reason of cardiac arrest. It does not present a precise and tailored survival estimate for each patient. The Good Outcome Following Attempted Resuscitation (GO FAR) clinical prediction rule [4] has recently been developed to help health care providers produce better estimates of patients’ likelihood of survival after an episode of in-hospital cardiac arrest, with these patients being neurologically intact or having only minimal deficits at discharge (Cerebral Performance Category 1 [CPC 1]). The GO FAR clinical prediction rule assigns scores to 11 risk factors such as major trauma, renal insufficiency, and septicemia among others and 1 protection factor (ie, being neurologically intact before cardiac arrest). Using these variables, a final estimate of the survival rate with minimal neurological deficits after in-hospital CPR is generated. The final survival-to-discharge rates with CPC 1 range from a maximum of 27.7% to a minimum of 0.9% [4].

Adapting knowledge tools such as DAs to local context is a crucial step in knowledge translation because it (1) reduces duplication of effort and optimizes use of existing resources; (2) encourages consideration of implementation and “fit” with the local context; (3) enhances applicability; and (4) engages local knowledge users, thus increasing the likelihood of uptake [8]. This step therefore requires engaging local knowledge users to ensure that the knowledge is relevant and applicable to their needs, from its generation to its implementation [9]. Although this is a key step in knowledge translation, little is known about how to accomplish this step effectively [10-12]. Experts have referred to this problem as an “evidence-based crisis” [10,13] and have challenged researchers to find innovative solutions that support the involvement of local knowledge users in adapting knowledge tools to their contexts [14-17].

We submit that using a wiki, a website that can be consulted and edited by anyone who is granted access, could be an effective strategy for adapting DAs to local contexts, as wikis were precisely designed to involve users interactively in the generation and application of knowledge [18]. Wikis—highly accessible, interactive communication vehicles—have also been shown to increase professionals’ self-efficacy with regard to their use of various types of knowledge tools [18-20]. The use of a wiki to perform rapid prototyping has successfully helped in the creation of knowledge tools adapted to the needs of patients and health care professionals [18,21-26]; however, to the best of our knowledge, our team is the first to use a wiki to support the creation of a patient DA.

The primary objective is to use an existing DA about CPR to create a wiki-based DA that is adapted to the context of a single ICU and tailorable to individual patient risk factors while employing a user-centered design. The secondary objective is to document the use of a wiki platform for the adaptation of knowledge tools such as patient DAs.

The main researcher (AP) will conduct 3 weeks of ethnographic observation of patients, families, intensivists, and other allied health professionals’ daily interactions. Ethnography is a qualitative method involving the immersion of the researcher into the setting to be studied. Traditionally used in cultural anthropology, ethnography is increasingly being used in health care research. Ethnography allows discovering rich information that could not have been discovered through surveys or other quantitative methods [30]. Because the presence of the main researcher in the ICU will not interfere with the usual care process and that the main goal of these observations is to learn how to improve the decision-making process in the ICU, only verbal consent will be required by the Ethics Review Committee to complete these observations.

During the observation sessions, the main researcher will specifically observe 5 dyads of attending intensivists and alert and oriented ICU-admitted patients, and their family members if available, who need to clarify whether or not to initiate CPR in case of a cardiac arrest. Patients who are unstable, aged less than 18 years, have cognitive impairments, and who do not speak French will be excluded. Potential participants will be identified by attending intensivists. Aptitude of the patients to participate will be confirmed at the beginning of the observation by their attending intensivist using a short evaluation of their orientation in time, space, and person. Using an observation grid developed by a human factors engineer (HW), the main researcher will observe the discussion and note users’ needs, goals, strengths, and limitations of the actual process of decision making without the use of a DA. The first 5 eligible patients will be enrolled in the study. The following data will be collected from the patient’s chart: date of birth, sex, reason of ICU admission, and level of medical intervention requested on admission to the ICU (full code, limited intervention, DNR, patient undecided or information not available).

In addition to ethnography, the main researcher will conduct one 60-minute semistructured interview with each of the 5 intensivists involved in the project. The development of the interview grid will be informed by the data collected during the ethnography. We will also collect the following demographic information about the 5 intensivists: age, sex, clinical specialty, and number of years working in the ICU.

Potential participants will be identified by the main researcher who will first obtain consent from the attending intensivist to approach the patient. Although we will purposively attempt to target a wide range of different types of patients based on their age, sex, and disease process, these participants will form a convenience sample based on their availability at the time of recruitment. Once the intensivist judges that the eligible patient is capable and fit to participate, the main researcher will obtain written consent from the patient to participate and will then present the DA to the patient and/or a family member. A minimum of 3 hours will be given to the participant to read the document. A follow-up meeting will be arranged 3 hours later with the attending intensivist so that the patient can provide his/her feedback on our DA and engage in a discussion about their goals of care.

The patient/family member and the intensivist will be requested to review together the clinical content and visual aspect of the wiki-based DA. Using an observation grid developed by a human factors engineer (HW), the main researcher will observe users’ needs regarding our DA and will assess its strengths and limitations. Then, the main researcher will conduct interviews with the patients and/or their family members. The interview will start with 3 sociodemographic questions about the participants’ highest level of education completed, their profession, and their religion. Then, 7 open-ended questions will follow about the DA concerning the following:

Finally, each participant will be asked if they had previously discussed their goals of care and their preferences about CPR. We will also verify if our DA changed their previous advance directives (eg, changing from wanting to receive CPR to now refusing CPR). After each use of our DA with any patient, intensivists will also be requested to provide feedback about their experience using the DA with their patient. All interviews will be recorded and then transcribed verbatim.

We will also collect the following data from the patient’s chart: date of birth, sex, reason for ICU admission, and the code status preferences upon admission to the ICU.

After each rapid prototyping cycle, 2 researchers (AP and PA) will perform qualitative content analysis of the verbatim transcripts, audio recordings, and researcher interview notes and observations to identify the usability problems (eg, visual design, format, layout of information), and need for content clarification. Any problem identified will be addressed prior to the next round of user testing. All changes to the prototype will be done online using the wiki platform to keep track of the changes and the different prototype versions produced.

We will perform descriptive statistical analyses of the patient and intensivist sociodemographic data. We will also measure the proportion of patients who state that they changed their goals of care after reading our DA and will note if this change increased the level of care or decreased it. We will also compare patient’s final decision about CPR after reading our DA with the level of care documented in the medical chart.

This study was approved by the Research Ethics Board of the Centre Intégré de Santé et de Services Sociaux de Chaudière-Appalaches on January 28, 2015 (CER-1415-019). Informed written consent will be obtained from patients and family members participating in Phase II. Participants will have the opportunity to withdraw from the study at any time. Quantitative data gathered about the participants and qualitative information gathered during the observation sessions and interviews will be kept strictly confidential and the study results will not allow to identify participants. Five years after the publication of the study results, all information about participants will be destroyed. If the participant or a family member of the participant demonstrates by his or her words, gestures, or behavior any kind of discomfort with respect to information contained in the decision tool, the situation will be reported to the attending intensivist. Patients will not be paid for their participation. Because family members participating in the prototyping phase could be asked to travel to the ICU to meet our research team, they will be offered a small stipend to pay for parking or a meal at the hospital’s cafeteria.

Our study protocol has some limitations. First, although studies on usability testing suggest that 15 patients are sufficient to address most usability issues, we cannot ensure that our final DA will address all the issues about end-of-life decisions that are very complex, emotionally charged, and culturally sensitive. Our wiki platform will however offer an interesting solution to facilitate making improvements to our local DA if new usability or content issues arise in the future. Second, the vast majority of patients admitted to our ICU setting will be white, Catholic, and French-speaking while the issues of end-of-life decisions are deeply linked to culture and religion [35]. This will limit the broader use of our context-adapted DA to settings that are more multicultural. Nevertheless, the use of our wiki platform and methodology will still help other researchers and health professionals to adapt our open-source and free tool to these other cultural settings. Third, although we aim to recruit as many family members as possible to make our DA fully adapted to their needs as well, their participation will be optional for feasibility and ethical reasons. Future studies will have to explore the development of DAs adapted to the needs of surrogate decision makers in situations where patients are unable to make decisions for themselves. Finally, this study will not measure the impact of creating a context-adapted DA. A prospective study assessing the impact of our DA on patient satisfaction, quality of care, and patients’ decisional conflict will be needed.

This study will enrich the emerging literature on the use of wikis and other collaborative writing applications to adapt knowledge tools to the local context, as well as applying user-centered design to develop DAs that better support SDM. We expect that using a wiki will help other centers and researchers adapt our DA to their own context and culture, thus reducing duplication and accelerating the dissemination of such DAs.