4J
ORAL ABSTRACTS: AFFECTIVE EXPERIENCES AND DECISION MAKING
Method: Adult smokers (N=293; mean age=33.7) who did not plan to quit were recruited from two US cities. Participants were stratified on the basis of age, gender, and amount smoked, and then randomly assigned to receive their own brand of cigarettes in modified packages for four weeks. Packaging was modified to feature either the nine basic-text warning statements mandated by the Family Smoking Prevention and Tobacco Control act (text only), the basic-text warnings plus the nine images selected by FDA in their 2011 Final Rule to be required on American cigarette packaging (graphic images), or the nine images selected by the FDA plus elaborated text which provided additional details about smoking risks (elaborated text). Participants returned to the lab each week to receive additional cigarettes and complete questionnaires. Affective reactions and risk scrutiny were reported after one week of exposure to the warnings. Perceived warning credibility, risk perceptions, quit intentions, label memory, and risk knowledge were assessed after four weeks of exposure to the warnings. Risk knowledge was assessed again approximately one month after the experiments conclusion.
Result: In structural equation models, the presence of graphic images (compared to text only) indirectly influenced risk perceptions and quit intentions by means of an affect heuristic (image->negative affect->risk perception->quit intention). Negative affect from graphic images also influenced risk perceptions and quit intentions by motivating greater processing of risk information (e.g., image->negative affect->risk scrutiny->label credibility->risk perception->quit intention). We further predicted and found that warnings with graphic images were more memorable than text-only warnings, and this increased memory for label content mediated increased smoking risk knowledge at the conclusion of the study and one month later. Finally, increased smoking risk knowledge was associated with greater quit intentions, but only among participants who perceived warning labels as credible.
Conclusion: Graphic warning labels are more effective than text-only warnings in encouraging smokers to quit and in educating them about smoking’s risks. Negative affective reactions, thinking about risks, and perceptions of label credibility are important mediators of their impact.
Methods: Women (N=148) aged 40-74 years of age were recruited from a large participant registry to complete an online experiment. Participants rated 10 side effects--randomly selected from a pool of 20--on each of 15 characteristics (e.g., frightening, gross). In addition, for each side effect they read a hypothetical scenario in which an effective and necessary medical treatment conferred a 1 in 100 risk of experiencing a side effect (e.g., nausea). Aversiveness of each side effect was measured in four ways: participants indicated their willingness to take the medication (i.e., choice), willingness to pay to avoid the side effect (WTP), the amount of negative affect associated with the side effect, and they ranked the side effects in terms of their desirability. Each of the 20 side effects was rated by at least 45 participants.
Results: A principle component analysis of the ratings of the side effects’ characteristics yielded a four-factor solution, which together accounted for 82% of the total variance: dread (27%), shame (24%), disabling (19%), and coping (12%). Regression analyses further indicated that dread was the strongest predictor of each measure of aversiveness. High dread was associated with statistically significantly (p<.05) lower choice (b=-.52) and desirability rankings (b=-.64), and higher WTP (b=.59) and negative affect (b=.73). Side effects that were perceived as disabling were associated with significantly higher WTP (b=.47) and negative affect (b=.51), and lower desirability (b=-.61), but not choice (b=-.16). Side effects perceived as shameful or associated with coping difficulties were not statistically significantly associated with any of the aversiveness measures (ps>.05).
Conclusions: These findings reveal that affect is a key factor in patients’ perceptions of side effects and this aspect has a strong impact on their decisions regarding treatments that involve side effects. Decision support tool developers should consider adding components to address the affective nature of medical decision making.
Methods: We recruited 147 adults from the U.S. on a web-based portal (MTurk). Participants were told to imagine they were just diagnosed with psoriasis, and randomized to vignettes in a 2 (High DA vs. Moderate DA) X 2 (High MA vs. Moderate MA) factorial design. Participants were either given a picture of severe psoriasis (High DA) vs. no picture (Moderate DA). Participants were told the medication was either an inflammation reducing pill (Moderate MA) or an immune suppressing injection accompanied by an injection picture (High MA). Participants completed three questions: Treatment Choice (treat now or wait), Willingness to Trade-Off (WTO) if Disease Risk Changed (yes, no), and WTO if Medication Risk Changed (yes, no). A series of logistic regressions examined the influence of DA, MA, and their interaction on Treatment Choice, and DA, MA, Treatment Choice, and their interactions on WTO if Disease or Medication Risk Changed.
Results: The majority of participants chose to treat (61.2%). High MA was associated with choosing to wait (e.g., defer treatment) (OR = 2.89, p = .002). Treatment Choice (OR = .14, p = .005), MA (OR = .24, p = .046), and the interaction between these two variables (OR = 8.46, p = .011) were associated with WTO if the risk of disease progression was changed (see Figure). People randomized to the high MA condition were less willing to trade off if medication risk changed (OR = .43, p = .019).
Conclusions: These results have significant implications for patient decisions regarding treatment escalation. People who approach a treatment decision highly activated by a treatment are likely to refuse treatment escalation. Further, people who are highly activated by a treatment, and therefore more likely to refuse treatment escalation, are also unlikely to reconsider their decision in light of new information about the risks associated with either the disease or the treatment.
Purpose: People typically overestimate the unpleasantness of medical experiences and may avoid important screenings (Dillard et al., 2010) or medical procedures with long-term health benefits (Angott et al., 2013). We sought to determine whether targeted narratives could reduce these mispredictions or 'affective forecasting errors.'
Method: In Study 1 (N=196), college students were surveyed about 10 common medical events (e.g. Pap test, donating blood) and provided ratings of predicted discomfort (if they had never experienced it) or actual discomfort. Participants making predictions were randomly assigned to either the control condition (no narratives) or the targeted narrative condition (4 narratives describing experiences with the medical event provided by participants in a pilot test; narratives were chosen to target the direction of bias observed in prior work) before making predictions.
In Study 2, college students (N=150) made predictions (Time 1) about the discomfort associated with the cold pressor task (keeping your hand in ice water 0-1°C for up to 2 minutes). Before making predictions, participants were randomly assigned to one of three conditions: 1) control (no narratives), 2) positive narratives (2 stories describing the task as not painful), or 3) negative narratives (2 stories describing the task as painful). Narratives were selected from an earlier cold pressor study. All participants completed the cold pressor task and then immediately provided ratings of the discomfort experienced (Time 2). Participants also reported their memory for the experienced discomfort one month later (Time 3).
Results: In Study 1, affective forecasting errors were observed for 8 of the 10 medical events. Specifically, predicted discomfort was significantly greater than reported discomfort, p<.05. Targeted narratives successfully reduced affective forecasting errors in 5 of the 8 events where bias was observed.
In the Study 2 cold pressor task, predicted discomfort was significantly less than reported discomfort, and negative narratives (which targeted the direction of the bias) again eliminated affective forecasting errors (Figure 1). However, participants in the positive narrative condition reported significantly less discomfort from the experience despite exhibiting forecasting errors, p<.05.
Conclusions: Affective forecasting errors can be improved with the use of narratives that target the direction of bias in prediction errors. However, stories that paint an overly positive impression, while still resulting in prediction errors, resulted in less experienced discomfort during an unpleasant task.
Method: We recruited 1230 participants (mean age=34, SD=11; range=18-71, 83% white) to complete an online survey. Participants were randomly assigned to either (1) read the Gardasil Vaccine Information Statement from the Centers for Disease Control, (2) additionally learn how many deaths and disabilities were reported for Gardasil in 2013 (31 events total), or (3) view all of the above information and additionally read the 31 VAERS event reports. Participants who read these reports rated whether they thought that the vaccine caused these events. Afterward, participants reported their willingness to vaccinate a child (1-6 Likert scale) and their belief that the CDC is faithfully reporting vaccine risks (0-100 slider scale).
Result: Expansive disclosure of the VAERS reports caused participants to be significantly less willing to vaccinate a child (M=3.93, SD=1.84) compared to both the summary data (M=4.67, SD=1.64) and the CDC information statement (M=4.41, SD=1.76), p<.001. Participants in the expansive disclosure condition were also less likely to believe that the CDC is faithfully reporting the risks (M= 4.53, SD=1.51) compared to the other conditions (Ms=4.71, 4.85, SDs=1.36, 1.33), p=.004. Participants with preexisting negative vaccine attitudes were more likely to believe that the vaccine caused the health events reported in VAERS, r=-.39, p<.001.
Conclusion: Contrary to predictions, expansive disclosure of the VAERS reports decreased evaluations of HPV vaccine safety and reduced trust. By contrast, the VAERS summary data did not negatively affect judgments. One possible reason for these findings is that when participants read the VAERS reports, their preexisting vaccine attitudes shaped their interpretation of whether the vaccine caused each event.
Method: A Markov decision process (MDP) model is formulated to identify optimal treatment decisions when breast cancer is diagnosed at an annual mammogram exam. The underlying Markov chain represents a patient’s health status at each screening. We assume follow-up tests (e.g., biopsy) are used to confirm the presence of breast cancer after a positive mammogram. Type II error of mammography and the possibility of spontaneous regression of early stage cancers are considered in the model. While the patient can choose to wait on treatment, there are various types of treatments including different combinations of surgeries and adjuvant treatments considered in this study: breast conserving surgery (BCS), mastectomy, radiation therapy, chemotherapy, and hormone therapy The goal of the decision model is to maximize a patient’s life score, which is a metric that seeks to take into account the patient’s expected remaining life years as well as her personal preferences. Life scores depend on various factors: age, cancer stage, hormone receptor status, type of treatment, and the patient’s personal opinion (as measured by emotional weights) on treatment side effects including cancer recurrence, change in appearance, figure and general pain.
Result: The results show that the optimal treatment policies vary with the patient’s personal preferences. No treatment decision is found to be optimum for elderly patients, when side effects outweigh the gains in life years. The threshold of no treatment decisions depends on individual patient preferences. The average life score of the invasive cancer patients is around one-fourth of the average life score of in situ patients. When emotional weights of side effects are high, the treatment combination of a BCS plus radiotherapy and chemotherapy is often recommended as optimal.
Conclusion: This study provides a dynamic decision framework to evaluate and identify optimal treatment decisions based on individual patient preferences. It shows how patient preferences can be incorporated as an outcome measure in the MDP model.