Ethical Considerations during the Informed Consent Process for Acute Ischemic Stroke in International Clinical Trials

Literature Search

Literature searches with the aid of a University of Michigan research librarian were performed using the following platforms: PubMed (https://pubmed.ncbi.nlm.nih.gov/; Bethesda MD US), clinicaltrials.gov (https://clinicaltrials.gov/; US National Library of Medicine), National Institute of Public Health Clinical Trials Search (https://rctportal.niph.go.jp/en/; Japan), Europe PMC (https://europepmc.org/; Europe), EU Clinical Trials Register (https://www.clinicaltrialsregister.eu/ctr-search/search; The Netherlands), the Australian New Zealand Clinical Trials Registry (https://www.anzctr.org.au/; Australia), and International Clinical Trials Registry Platform (apps.who.int/trialsearch/; World Heath Organization). Peer- reviewed scientific articles published in English language were identified using the following search terms: “alteplase” OR “tPA” OR “Activase” and “ischemic stroke” OR “acute ischemic stroke”. Publication dates were restricted to January 1991 through March 2020. The PubMed search was further filtered by selecting Clinical Trials and Humans. The Clinicaltrials.gov search was further filtered by excluding recruitment status listed as suspended, terminated, withdrawn, or unknown. The Australian New Zealand Clinical Trials Registry search was further filtered by selecting “interventional” trial type. Inclusion criteria for trials were studies of patients with acute ischemic stroke who would be eligible for IV tPA administration within 4.5 hours of symptom onset and comparison of IV tPA versus an alternative treatment. Exclusion criteria were trials where IV tPA was combined with an alternative treatment, or where an alternative treatment was compared to a placebo. Three authors assessed all trials independently for eligibility and subsequently collaborated to agree upon the final trials included. Of the 401 articles uploaded to Rayyan (https://rayyan.qcri.org/; Doha, Qatar) for researcher review, 14 trials met inclusion and exclusion criteria (Figure 1).(18–31)

• Download figure
• Open in new tab

Figure 1.

Flowchart of results from the literature search conducted.

Studies Contacted

The corresponding author and trial coordinators listed on clinicaltrials.gov from each of the 14 trials were emailed by the study team with a request for a copy of their patient consent form, other documents related to informed consent, and to complete a survey developed as described below. The email also included a reminder that participation is voluntary and that emails would be collected at their own discretion. Individuals receiving the survey were asked to share the survey with other members of their trial team who were knowledgeable about informed consent for the trial. The survey contained 24 questions pertaining to the informed consent process, including multiple choice, and free response questions (Supp 1). 43 representatives from the 14 trials were contacted. 6 studies completed the survey in its entirety, 1 study (AcT) partially completed the survey and therefore results were excluded, and the remaining 7 studies did not respond after a third contact.(18–31)

Survey Development

An anonymous online survey developed using Qualtrics (SAP software company; Utah, USA) assessed how informed consent was obtained in a population that may have challenges with standard consent due to absence of capacity in the setting of acute stroke. Survey questions included demographic questions (year of result publication, location, participants, and median NIHSS of trial arms), multiple choice (yes/no/unknown) questions regarding specific aspects of the informed consent process, and free response questions concerning challenges trials may have faced. A draft of the survey was presented to an acknowledged collaborator, Dr. Lesli Skolarus, where specific feedback was provided on content and format. The survey was then revised and distributed to the 43 contacts mentioned above.

Statistical Analysis

Informed consent documents from each trial were analyzed for presence or absence of informed consent elements (Table 1). If data was published on median DTT times and the survey respondents used estimates of DTT times, the published DTT times of the trials were used in the statistical analysis. Descriptive statistics were displayed for continuous variables as either mean ± standard deviation or median (interquartile range) depending upon data distributions, and as frequency (percent) for categorical variables. Nonparametric methods (including Wilcoxon Rank Sums and the Kruskal-Wallis test) were used to evaluate factors potentially impacting DTT time. Statistical analyses were performed using R version 3.62 (r-project.org) and Excel (Microsoft Corp; Redmond, Washington, USA).

Patient Capacity

The median NIHSS score for each completed trial was at least 8, meaning most participants experienced a moderate stroke (Table 1). 5 trials allowed for consent to be obtained from a designated patient proxy if the patient was deemed to lack capacity (Figure 2). Of those 5 trials, 2 required assent from the patient if the patient was consented by proxy and the other two trials did not require patient assent. Trials reporting direct consent (defined as not consent by proxy) of 75% to 100% of their participants were associated non-significantly with shorter DTT times than trials reporting direct consent of 25% to 50% of patients for both the experimental and control arms with a trend towards significance (group = median[IQR]; 75% to 100% experimental = 32[16-48], control = 34[17.5-50.5]; 25% to 50% experimental = 117.5[65-170], control 117.5[65-170]). However, 2 of the 6 trials did impose a time limit for when informed consent needed to take place (Table 1). When trials were asked to comment on challenges that arose during study protocol development, all comments referenced problems delineating specific exclusion or inclusion criteria (Table 2). Additionally, when asked about challenges related to obtaining consent, 3 out of the 5 comments mentioned difficulty with obtaining proxy consent.

• Download figure
• Open in new tab

Figure 2.

Patients were either directly consented to the trial, consented by proxy with required assent from the patient, or consented by proxy without assent. A) Number of trials requiring assent in addition to consent by proxy. DTT of the B) control arm and C) experimental arm was plotted in relation to percentage of patients directly consented divided into quartiles: 0% to 25% (n=1), 25% to 50% (n=2), 50% to 75% (n=0), 75% to 100% (n=3).

Logistics of Obtaining Informed Consent

All trials allowed research attendings to obtain informed consent in the ED, but no trial allowed consent to be obtained by nurses, EMTs, undergraduate students, or residents other than a neurology resident (Figure 3). Although some trials allowed consent to take place in the hospital and in 1 case over the phone, no trial allowed for consent to be obtained in an ambulance or while at an outside hospital (OSH) prior to transfer. When asked to describe the required formal training received specific to the informed consent process, only 2 trials mentioned specific training with regards to informed consent skills and 1 trial did not require formal training.

• Download figure
• Open in new tab

Figure 3.

A) The types of professionals obtaining consent and B) the location consent was obtained. The Emergency Department (ED) was considered a separate location from the hospital itself. Outside hospital (OSH) was defined as a hospital other than where the study was being conducted.

Foreign Language

The 6 trials took place in various continents and 4 out of the 6 trials made translators and translated consent documents available. Interestingly, trials that offered translated consent documents had longer DTT than trials that did not offer translated documents in both the control and the experimental arms (Figure 4). Similarly, trials that offered translation services had longer DTT than trials that did not in both the control and experimental arms.

• Download figure
• Open in new tab

Figure 4.

The effect of translator services on the DTT for trials. The DTT for the A+C) control arm and B+D) experimental arm were plotted in relation to having A+B) translated consent documents available (n = 2) or having C+D) a translator available (n=2).

Demographics of Trials

The 6 trials included were conducted at major stroke institutions internationally, with representation from North America, Asia, Europe, Australia, and South America. Because of this variation in location, different institutional and cultural values may have influenced the methods and parameters used for obtaining informed consent, as evidenced in previous studies.(38) One such unique challenge was that a centralized IRB was not yet established before some trials took place (Table 2). Therefore, each participating trial center created an individual consent protocol to be approved by the home institution IRB, which decreased reproducibility of the informed consent process across trial sites. Another significant variation observed was the number of participants: trials ranged from 7 to 3206. It is likely that trials with larger cohorts faced more challenges, but these multi-center trials may have had the motivation and resources to implement more standardized approaches at multiple locations. Although the differences between the trials examined may be related to trial design, legal, institutional, and cultural effects, these trials overall provide evidence of the diverse landscape of informed consent.

Reporting the Consent Process

Information regarding how consent was obtained, consent forms, timing, and personnel involved across these trials was highly variable, and often could not be located publicly through protocol papers or within the published manuscripts. Interestingly, only half of the trials mentioned the involvement of a research ethics committee in their published results. Ethical oversight in the design of future studies, as is the case in the ongoing AcT QuICR trial, may be a useful approach to mitigate ethical dilemmas and oversee abbreviated methods of informed consent.(26,39) 1 of the 6 trials enrolled patients via a waiver of informed consent. None of the 14 trials made a copy of the informed consent document publicly available. Even after contacting trials for a copy of the informed consent document, only 2 provided forms for further examination. The absence of information highlights the need for research ethics involvement in trial design and greater transparency concerning informed consent in published manuscripts.

Patient Capacity and Consent by Proxy

Given that patients experiencing a stroke are a vulnerable population on the basis of the neurologic nature of the disease, understanding patient capacity is critical. In this study, 3 of 6 trials reported consenting patients directly for 75 to 100 % of participants, while only one trial reported consenting patients directly for less than 25% of participants (Figure 2). It remains important, however, to include cognitively impaired subjects to determine the generalizability of the proposed interventions.(40) NIHSS scores reported indicate most participants experienced a moderate stroke (Table 1). In the moderate NIHSS group, the ability to cognitively understand circumstances and make an informed decision varies widely.(41,42) This potentially discordant finding between reports that the majority of patients directly consented and had a “moderate” stroke could be due to alternate definitions of direct consent outside of the USA.(43) 4 of the 6 trials in this study were conducted outside of the USA, where deferred consent is commonly practiced.(44,45) Deferred consent involves randomization of the patient into the protocol based on the discretion of the investigator, followed by informed consent during a later phase of care. In that case, investigators may have received assent from the stroke patient in the acute phase and formal written consent from the legally authorized representative. Given this information, there may have been a difference in interpretation of the survey question, where trial representatives responded with a focus on who was actually signing the consent form rather than who initially provided consent before treatment was administered (Supp 1). This may explain why although most patients experienced a moderate stroke, trials responded that over 75% of stroke patients were directly consented. Although deferred consent is one possible solution to ensure patients receive treatment in a timely manner and are appropriately aware of therapies, patient autonomy may become an issue: the investigator cannot take away the medications or treatment already provided, even if the patient were to later decline participation in the trial.

While it is unclear how trials in this study determined capacity for consent, other trials have used aspects of the NIHSS score to assess capacity, including the sections regarding “level of consciousness”, “best gaze”, and “best language.”(15) More research needs to be conducted about ways to best assess capacity in moderate NIHSS patients to determine whether examining aspects of the NIHSS is sufficient to determine capacity.(46) When capacity is lacking, many stroke trials use consent by proxy rather than direct consent to enroll patients.(43) From an ethical perspective, there is concern as to whether surrogates are aligned with the wishes of patients in research trials.(47) To address this issue, patient assent can be required in addition to proxy consent, which was performed in some of the trials in this study.(35) Although surrogate decision makers must make difficult choices, the research team should still involve the patient to their fullest extent by attempting to gain assent when patient capacity is determined to be lacking.

Logistics of Obtaining Informed Consent

Although communication techniques employed during the consent process between research staff and patients are vital for ethical and efficient obtainment of consent, only a subset of the trials indicated a dedicated training process specific to the trial (Table 2). All trials in this study sample relied on highly educated and experienced providers as well as dedicated research coordinators to obtain consent (Figure 3a). A dedicated training protocol for delivering informed consent can ensure a high standard of communication and create more time to discuss the trial with patients and families.(46,48) A third party involved in obtaining consent may also place less pressure to enroll.(48)

In the trials studied, consent was obtained in a variety of settings; however, only one trial allowed consent to be obtained over the phone and no trials obtained consent upon patient transfer or in an ambulance. Obtaining consent over the phone in a prehospital setting has been performed in the FAST-MAG trial and provided more time for patients to consider consent as well as increased efficiency.(49) We suggest that consent in the pre-hospital setting should be considered for future trials. In addition, 3 out of 5 trials reported challenges obtaining consent by proxy, some of which may be mitigated by phone consent (Table 2). Phone consent is important when no proxy is available within the time window due to geographic constraints.(36,50) However, it can be challenging to reach patient families with distressing information and ask for a decision when they are most vulnerable. Another solution to consider in the case of acute stroke trials may be a waiver of consent, known as an exception from informed consent (EFIC) in the USA, as delays in treatment due to the time it takes to achieve informed consent can lead to worse outcomes for patients.(51,52)

Recent revisions to the Common Rule in the Federal Rules for the Protection of Human Subjects emphasize the responsibility of researchers to deliver information with concise and simple language during the consent process.(53–55) Consent forms in the USA now should emphasize key points for participation up front.(56) This is important because enrollment decisions in stroke trials can occur in a highly stressful environment that may increase cognitive load and diminish comprehension. A third of the trials reviewed implemented a time limit in which consent had to be obtained (Table 1). This approach may be appropriate in the acute setting to indicate the time sensitive nature of the process to research staff and to promote efficiency in the consent process. One trial made note of the decision to allow patients to provide verbal consent to initiate treatment followed by deferred written consent (Table 3). Interpersonal communication may be a more efficient means to obtain consent and produce better understanding in comparison to the medium of paper forms.(48,57) As such, verbal consent has previously been proposed as a function-based approach to informed consent.(58–60) Shortening the consent form is another consideration to reduce complexity, however, a longer consent form may be appropriate for proxy decision making.(35,61,62) Overall, a more targeted consent model may be preferable on an ethical basis due to the time sensitive nature of benefit to the patient and the potential to enhance patient understanding.(17)

Foreign Language

Translators and translated consent documents may provide an avenue to greater patient diversity in stroke trials. Trials which included patients with potential language barriers requiring translators and translated consent documents demonstrated a trend towards increased door-to- treatment time, suggesting that efficiency and efficacy may be impacted where such barriers exist (Figure 4). Efforts to increase the availability of rapid bedside interpretation services, such as with interpreter phones and professional translators, may help correct well-documented, ethnicity-based disparities in thrombolysis.(63,64) Additional barriers in the consent process may also exist for patients who speak foreign languages. Patients with limited English proficiency may have lower health literacy, availability of surrogate decision makers, or knowledge of concepts related to study design, such as randomization and clinical equipoise.(65,66) These additional barriers increase the need for rapid translation services and education at the bedside to ensure informed consent is achieved in a timely manner and to promote representation of a diverse patient population.

Limitations

The findings from this study are limited by the small sample of trials that met inclusion criteria. We also did not have responses from 8 of the 14 trials that met inclusion criteria and could not locate trial protocols or procedure manuals that outlined the informed consent procedures. We welcome any trial investigators or staff to complete our survey through this link http://bit.ly/strokeconsent or by contacting the corresponding author.

Some of the trials were ongoing, and the respondents could not complete all data fields. In cases where the trial was conducted across multiple clinical sites, responses may not reflect site- specific variation in how informed consent was obtained. Trials were additionally limited to studies listed in trial registries in English, thus excluding the pool of international studies which otherwise satisfied inclusion criteria. Besides informed consent, unreported clinical factors, such as blood pressure management, may also influence the DTT times in the included trials.(67)

Finally, the informed consent process for stroke trials may change in the future due to altered interpretation of EFIC.(68) EFIC was drafted in 1996 and finalized in 2013, but has not been frequently implemented in stroke trials.(52,69) Standard of care trials, such as those examined in this study, are typically not candidates for the waiver because of the proven moderate efficacy of tPA.(17,69) Other countries have different regulations and models of consent not captured by this study. A new landscape for informed consent in the acute setting may be necessary to ensure medical care does not stagnate and that research can allow for the development of the next standard of care.