To our knowledge, this is the first study offering in-depth understanding of HNC patients’ preferences for disclosure of prognostic information, and the use of a prognostic model during treatment decision consultations.
1) Life-expectancy
Understanding the concept & using a tailor-made approach
While all patients considered it somewhat to very important to receive information about their life-expectancy, only some of them wanted to receive this in a specific quantitative manner, like 5-year survival rates. This is in line with previous research among patients with advanced or incurable cancer (28, 29). The majority of patients wanted to receive prognostic information from their doctor in general terms, like ‘your cancer can be well treated’. This kind of qualitative information would give these patients enough reassurance for the first moment. Even though doctors generally use the concept 5-year survival rate, participants often did not understand this concept or confused it with other terms, for example chances of cure, and thought it had a negative connotation. Overall, little is known about patients’ awareness, and understanding of prognosis(6). Previous research stressed that in some cases cancer patients misunderstand or fail to absorb the information given, cannot recall the status of their disease and often overestimate their survival chances (6, 30-32).
The need for receiving prognostic information was dependent on different circumstances. This means that sharing prognostic information requires a tailor-made approach. Patients had a stronger preference for quantitative information like months or survival rates, in the hypothetical case of cancer recurrence and/or a poor prognosis. This kind of information would enable them to weigh whether undergoing a second treatment to prolong survival would be worth the ‘costs’. Prognostic information is not a standalone concept according to patients and caregivers. Patients also expressed the need for information about their expected QoL, since this would be of significant importance in the decision-making process. Fried et al. asked 226 patients with a limited life expectancy whether they would choose a treatment with survival, but with severe functional or cognitive impairment. 74.4% of patients answered they would not accept severe functional impairment and 88.8% would not accept cognitive impairment, and thus rather face death (33). However, more recent research by Blanchard et al. among HNC patients showed that they overall prioritize survival over functional endpoints (34). Although we did not explicitly ask patients to prioritize survival and QoL, they did however mention that at a certain point the survival benefits would not weigh against the deterioration in QoL. On the other hand they mentioned that patients are prone to keep pushing their boundaries, and increasingly accept functional limitations in order to stay alive.
In case patients want to receive quantitative information, what would be the right timing to share this? Our focus group results suggest that the right timing and phasing are of key importance. It seems that life-expectancy should be best discussed after the conversation in which the cancer diagnosis is given. According to most patients and their caregivers, it would be too stressful to discuss this all at once.
Several patients addressed that it depends on personal preferences whether a patient wants to receive prognostic information. While on the one hand some patients gain an increased sense of control by receiving more information about their disease and prognosis, others want to receive very little information. The latter group often wants the doctor to take control and is not interested in the details on treatment or prognosis. Receiving unwanted prognostic information could destroy hope and therefore patients’ needs should be explored beforehand (35), instead of bluntly confronting them with unwanted information.
Who should take the initiative in exploring prognostic information needs? While some patients will take the lead, others aren’t capable or don’t want to. Therefore, according to the participants, the healthcare provider should be the one to introduce the topic, while the patient is given the opportunity to decide whether he or she wants to receive the information. This is in agreement with a qualitative research among advanced cancer patients: most patients and caregivers in this study said a physician should offer to discuss the prognosis, if the option to decline the information was also provided (36).
Communication skills professional
According to our participants, doctors should be honest while discussing the prognosis without taking away hope, and tailor prognostic information after exploring patients’ needs. The importance of being realistic and honest while maintaining hope is also identified in previous literature on patients with advanced or incurable cancer (37-40). For example, Kutner et al. found that while 100% of patients in their survey wanted honesty from clinicians, 91% also wanted them to be optimistic (37). Balancing between honesty while disclosing prognosis and maintaining hope can be a challenging task for healthcare providers (39, 41).
2) The prognostic model OncologIQ
After fully exploring patients thoughts and believes on the topic life-expectancy, the prognostic model OncologIQ was introduced. Some patients would appreciate counseling with OncologIQ as they thought it was clear and more personalized, while others were in doubt. Some patients didn’t want counseling with OncologIQ at all because of the need to maintain some ambiguity about the future. This need to maintain ambiguity about outcomes, is also identified in previous research among advanced or incurable cancer patients (29, 35, 38). Ambiguity could help to maintain hope and avoids a blunt confrontation with the facts. Participants shared several recommendations to improve the model. In three focus groups caregivers were concerned that the monthly health insurance premium would rise, if the insurance companies would also have access to an individuals’ prognostic estimate. Questions on this topic should be considered when using a prognostic model for counseling.
Visual formats of communication
Prognosis can be presented in various formats. While previous research showed that most persons find numbers and 100-person diagrams easiest to understand (42, 43), the HNC patients in this study preferred the pie chart. The pie chart was a favorite because they thought it was clear at a glance (see table 4) and less confronting than some of the other formats. The 100-person diagram was considered too confronting by both patients and caregivers. This is in line with previous research that explored this by using a 100-faces diagram (43). In addition, Davey et al. stated that the survival graph was considered negative, since it showed the constantly increasing mortality. In the current study, patients’ thoughts on the survival graph were also mostly negative. They found it too mathematical, since on must first must interpret the X- and Y-axis. Davey et al. also tested cancer patients’ understanding of the survival graph: only six out of 26 patients correctly interpreted the graph (43). Furthermore, we assessed that the included patients’ preferred to combine verbal explanation with visual prognostic information over a verbal explanation solely. This is also reported in previous research on this topic (44). Furthermore, it remains unclear as to what extent patients understand the uncertainty around prognostic models’ estimates (45). Presenting data uncertainty is difficult and there is no consensus in literature about the optimal way to communicate different types of uncertainty (45, 46).
Practice implications: a guideline for individualized prognostic counseling.
OncologIQ could take away physicians reluctance to discuss the prognosis and reduce ambiguity in case of conflicting opinions among healthcare professionals by providing individual estimates. Previous research showed physicians’ willingness to use prognostic models in end-of-life care, aiming to improve prognostic confidence (15). It also enabled physicians’ to take a more directive role in specific cases where the expected prognosis significantly differs from patients’ expectations, and it reduced ambiguity in case of conflicting opinions about prognosis among colleagues (15).
Based on the results of this focus groups study, especially the recommendations discussed in table 3, a clinical practice guideline was developed that includes basic steps for sharing individualized prognostic information (see figure 4). While our earlier published guideline for professional communication focuses on general aspects of sharing prognostic information with HNC patients (11), this guideline specifically focuses on how to share the information provided by the prognostic model OncologIQ. It could also be used for other similar prognostic models in HNC. Since the term ‘’5-year survival rate’’ seemed to confuse patients and caregivers, we recommend not to use it literally. We asked patients which survival period would be most appropriate if a patient wants quantitative prognostic information. Most patients preferred five years, as they deemed two years ‘’too short’’ and 10 years ‘’too far ahead’’.
Strengths and limitations
One must first listen to patients’ preferences and needs, to be able to provide patient-centered care. The use of a qualitative methodology provided us with rich data on HNC patients’ preferences on these vital but unexplored topics. However, it is difficult to make assumptions on its generalizability. This study focused on patients with HNC in the curative setting. Since each setting has its own concerns, the generalizability of these results to the incurable setting is not desirable. Also, our results may be different in other, non-Western, cultures or countries. A certain selection bias may have occurred as the included patients are willing to participate in a focus group with other patients and caregivers. In addition, while almost one third of the patient population in our center is single(27), only 10% of patients in the focus group were. The presence of family members or other caregivers adds complexity to prognostic discussions since they may have different information needs (47). We purposely chose to include caregivers in the focus groups, as they are also present during the treatment decision consultation.
Future perspectives
The results of the current study have been used to improve OncologIQ. Recently, the prognostic model has been updated.(27) In the first place because the original model was based on outdated data as the survival of HNC patients has improved in the past years (48). The second aim of the update was to test whether adding new prognostic factors would improve model performance, as recommended during the focus groups. Also, a visual format for patients has been developed, including a pie chart of the 5-year survival rate. The next step will be to evaluate the clinical impact of OncologIQ in a prospective clinical trial. The primary outcome of this trial is decisional conflict among HNC patients who are counselled with and without OncologIQ during the treatment decision consultation. The effect of the use of OncologIQ in our multidisciplinary tumor board meetings is also recently assessed in a pilot study.
A future aim would be to develop a prognostic model that includes both survival and QoL for HNC patients. Despite not addressing this future prospective during the focus groups, several patients stressed the importance of combining both survival and QoL, rather than focusing solely on survival. Due to the implementation of our Healthcare Monitor we will be able to meet this need soon(49). With this monitor we are collecting electronically patient reported outcomes (ePRO) on physical and psychosocial functioning since 2013, from intake until the last follow-up visit. In the first place this is done to improve patient care and counseling, although these data could also be used for research purposes.