Disclosing neuroimaging incidental findings: a qualitative thematic analysis of health literacy challenges
All stakeholders involved in disclosure of incidental findings from MRI research have significant concerns regarding the incomprehensible medical jargon in radiology reports; however, the challenge for investigators and radiologists is that simplifying the radiology report information alone would reduce its clinical utility for follow up care. Incidental finding reports written by radiologists in complex language may diminish the autonomy of participants with low health literacy skills. Autonomy is traditionally considered a principle underlying informed consent. In order to optimize participant autonomy, there are ethical principles that support the provision of unexpected information to participants, including IF [29]. In the context of returning radiology reports, autonomy requires that investigators grant individuals sufficient information to make educated decisions regarding their follow up care. However meeting this autonomy principal is challenging for investigators when their research participants have low health literacy skills. Investigators could consider reconfiguring their approach to reporting incidental findings by focusing on those participants with the most limited health literacy [30]. Doing this would require ongoing measurement of health literacy levels. Absent this measure of health literacy abilities, the most ethical approach is to offer all participants their research findings in comprehensible language, or at an 8th grade reading level [30]. However the difficulty is how to provide this information while still retaining the critical clinical specificity required for any necessary follow-up.
There is consensus among all stakeholders in our study that the only approach is the ethical approach: IF need to be reported to participants. This finding is comparable to the existing literature describing participant’s expectations for participating in research studies [10–13]. In work by Kirschen and colleagues, over 90 % of their research participants wanted IF information communicated to them [10]. This demand for information was recently echoed in the Presidential Comission for the Study of Bioethical Issues’ report for investigators to anticipate and communicate their approach in managing IF [14].
Our own empirical work reveals that the complexity of the information in research radiology reports contributes to the challenges of making appropriate follow up decisions. Despite very clear recommendations of whether or not to seek medical care, 10 % of participants in our preliminary research sought follow up care without referral and 67 % did not seek medical care when directed to do so. Similar to the retrospective survey participants, the parents of child participants who took part in the focus groups shared that if they do not understand information pertinent to their child’s health, they will seek medical evaluation – regardless of the report’s recommendation that follow-up medical care is unnecessary. These findings suggest participants may be making inappropriate healthcare decisions based on their inability to understand their research radiology report. Measuring broad outcomes in literature concerning health literacy suggests poor health literacy skills are detrimental to patient health in many ways, including inappropriate use of medical services [21–23]. It is important to note that our survey did not query participants about whether they already knew about the finding listed in their radiology report, so it is possible some participants did not seek follow up care for findings about which they and their physician were already aware. Still the potential consequences of misunderstanding the information in the research radiology report puts a considerable burden on investigators to communicate that information clearly and accurately.
In this manner, participant’s health literacy skills present a challenge for investigators to determine the optimal disclosure of IF information. Even as participants reported they did not understand the radiologist summary, they also reported high health literacy skills. Within the retrospective survey, those participants who reported having higher health literacy were likely to be older and less anxious about their health. These skills do not necessarily reflect an ability to successfully interpret complex radiology review information. Having self-confidence (equated with high health literacy) in their health navigation skills or in their ability to complete medical forms does not indicate an ability to dissect neuroimaging jargon (nor should that be expected of them).
Furthermore the results of our study highlight one of the challenges in measuring participant health literacy using brief instruments – such as the Chew instrument we used in our surveys [28]. Researchers have several choices: administer a face-to-face instrument or rely on participants’ abilities to independently complete a self-administered measure. Further, within each type, researchers have to choose between a briefer health literacy measure that may provide more limited and potentially misleading information, or a longer instrument that provides more information but may increase participant burden. A shorter measure, such as the Chew instrument, asks participants to rate their self-confidence to perform a variety of health care tasks like “completing medical forms” or “reading the instructions on a prescription bottle.” While the Chew questions address participant self-efficacy beliefs regarding their health management skills, this instrument falls short of the more in-depth, face-to-face instruments such as the Test of Functional Health Literacy in Adults (TOFHLA), that measure tangible health literacy skills such as interpreting the instructions on the label of a bottle of Tylenol [31]. In the context of neuroimaging and returning IF information, inaccurate measurement of health literacy leads to the conclusion that investigators must design radiology reports to meet the needs of individuals, whose health literacy skills can vary widely.
A consistent theme from participants who participated in the focus groups was their desire to have the ability to follow up with someone, to ask questions or to receive more insight to help explain their findings. This finding is similar to other measures of research participant expectations from receiving IF – participants desire someone they see as reputable to help explain the reports [10–13]. In a preliminary effort to address this issue, our research institution includes a cover letter accompanying all radiology reports that explained the general nature of IF in lay language, listed the contact information for our Medical Director and invited participants to contact the him in the event that they had any questions about their radiology report or would like to discuss their findings. Unfortunately the feedback we received from participants in the retrospective survey and the focus groups leads us to believe that the cover letter was not maximally effective in communicating how and from whom participants could seek additional information about their IF.
We would advocate for materials designed for participants with the lowest level of health literacy skills and the use of other delivery mechanisms to ensure the information is accessible to all individuals. The health literacy literature advocates that all information should be written at the 8th grade level or lower, despite whether the population is college educated or not.
The discussion in the focus groups demonstrates the concern physicians have about the burden health literacy challenges can have on the healthcare system. Physicians expressed they did not have the time to address patient anxieties raised by research reports about benign IF. Physicians feel stretched to meet their current patient loads; the imposition of additional medical appointments for this purpose could further burden an already overtaxed system. To date, our research locally suggests the potential workload for the healthcare system caused by reporting IF is minimal [15], but in general, participants with low health literacy skills represent a constant challenge for the healthcare system. From the participant perspective, it is well documented that health literacy affects health outcomes [18–23, 27–30]. The health system should address the broader challenge of managing effective patient education and targeting health communications to the needs of those with limited health literacy.
Similar to the challenges expressed by physicians in the healthcare system, investigators must manage the potential research burdens created by participants’ inability to understand their radiology reports. First, there is a cost to investigators when time must be allocated to explain the findings to individual participants. This reflects some concern in the existing literature regarding the existing logistical, and financial, cost to managing participant’s IF reports [5]. However, unique to our imaging center, where the investigators in our study perform MRI research, is the cost-effective and logistically stream-lined process of radiology review and report (for full details see, [15]). Within this system, there is already a process where investigators can refer participants to talk to the Medical Director, a certified neurologist, to help explain their findings. Also, investigators fear that one participant’s negative experience with a radiology report might deter other participants from signing up for future studies. This concern is also noted in the literature, as investigators predict receiving ambiguous IF information would cause participants unnecessary fear and anxiety [3–8]. If research participants know each other, or studies are recruiting from a specialized sample, this anxiety may be shared with other prospective participants. However experience shows this rarely occurs; other research finds that 90 % of participants believe receiving IF is a benefit to study participation [13]. Still, investigators fear these types of challenges when providing reports to participants. For this reason, many research facilities do not return findings or only return findings that require immediate clinical follow-up, even though they have been encouraged to do so by recent federal policy initiatives [17].
Until recently, the topic of health literacy has not been openly discussed as a concern of radiologists or investigators who return IF. Our institution’s experience in disclosing IF has revealed the importance of considering a participant’s ability to understand IF information and make appropriate decisions. In order to fully address the burdens created by disclosing complex health information to those with limited health literacy skills, investigators should listen to and learn from participants’ experiences [32]. Investigators face a complex conundrum: they don’t expect participants to understand the radiology jargon, yet they believe that participants ethically deserve to receive the information. Our research shows that participants report a strong desire to receive all medically relevant IF information. Returning radiology reports then, on one hand, creates a burden that investigators and the healthcare system must manage. On the other hand, investigators choose to provide participants with the information to advance patient autonomy. Our research leads us to conclude, as did Volandes Paasche-Orlow, that disclosure methods should be reoriented to a level that accommodates the needs of participants with the most limited health literacy abilities [29].
Based on participant feedback, our institution modified the recommendation rating scale to a clear binary “Please see your doctor” or “You do not need to see your doctor” in order to meet the needs of those with limited health literacy. Additional investigation is still required to determine whether this relatively new system will result in more appropriate healthcare resource utilization. These findings also indicate a need to establish clear, and descriptive consent language for those research studies that conduct MRI scans and return IF reports. We recommend that investigators use the initial consent conversation to help outline the possible outcomes from an IF report, and direct research participants interested in learning additional information to educational resources. We intend to develop and measure the effectiveness of additional resources (such as an easily navigable website that provides information about the most common brain imaging IF) to enable participants to learn and make appropriate decisions about their specific IF. These informational resources, written in lay language, may address the gap between participant health literacy skills and the complex clinical information provided in the radiology reports, while limiting the amount of burden placed on either investigators or the healthcare system.
This study has several strengths: the diversity of data collection methods, the relatively large sample size, and a participant population that has the experience of receiving IF reports. The study was conducted within one of the few institutions that returns IF reports to every participant. However there are also several limitations to this study. First, the retrospective survey data collected may not accurately reflect current institutional practice. The desire to grant participants in our imaging studies autonomy and improve the accessibility of the reports has resulted in numerous updates to the “radiology review and disclosure” process in a short period of time. Further empirical investigation is still required to fully characterize the burden on the healthcare system and investigators involved in the disclosure process. Second, our desire for a brief health literacy assessment led us to use the Chew measure. However, Chew suggested that the instrument was best used to identify individuals with low to inadequate health literacy and would not be as accurate at estimating distinctions among individuals with higher health literacy [28]. Therefore our ability to accurately assess the range of health literacy among our research subjects was limited. Third, as with all research data emerging from purposive samples, the individuals who participated may not be representative of others who were not included.