TOP RATED ABSTRACTS
Method(s): UK family physicians (N=151) read a description of a patient seeking advice regarding screening for a hypothetical cancer X. In the description, we manipulated the numeracy of the patient (low vs. high vs. unspecified), the effectiveness of the screening for reducing mortality (effective vs. ineffective), and the presence of a clinical guideline recommending screening (present vs. absent). We measured physicians’ risk communication, recommendation to the patient, understanding of screening statistics, and numeracy.
Result(s): Consistent with best practices, family physicians generally preferred to use visual aids rather than numbers when communicating information to a patient with low (vs. high) numeracy. However, 20% of physicians recommended a screening that was not effective and 44% offered incomplete risk information. Nevertheless, physicians with high (vs. low) numeracy offered more meaningful risk communication: they were more likely to mention mortality rates, OR=8.55 [95% CI 1.77, 41.41], and harms from overdiagnosis, OR=8.82 [1.34, 60.25]. Physicians with high numeracy were also more likely to understand that increased survival rates do not imply screening effectiveness, OR=6.05 [1.27, 28.72].
Conclusion(s): Screening patients for numeracy may help physicians tailor risk communication to patient needs and abilities. However, many well-intentioned physicians have low numeracy and are prone to communicating incomplete information to their patients. Although less numerate physicians know how to make risks easier to understand for patients, they themselves are likely to misunderstand risks and can unintentionally mislead patients. High-quality risk communication and shared decision making can depend critically on factors that can improve the risk literacy of physicians (e.g., numeracy, visual aids).
Our inherent drive to formulate coherent judgements can lead to biased information processing: incoming evidence may be distorted to favour an emerging judgement, before a final decision is reached. This “predecisional information distortion” (PID) has also been found in medical diagnosis: physicians may interpret patient information in a way that favours their leading diagnostic hypothesis. The role of PID in misdiagnosis has not, however, been investigated.
To assess the role of PID in misdiagnosis.
We constructed two patient cases, each with two competing diagnoses. One diagnosis was common and non-serious, the other rare and serious. Each case consisted of a brief patient description (demographics and health complaint) and several cues (symptoms, signs, and investigation results). Based on the available cues, the serious diagnosis could not be ruled out and warranted specialist referral. We presented 148 family physicians with one of the two patient cases, at random. After reading the patient description, physicians chose one of the two competing diagnoses. They then elicited further information: cues were arranged as labelled buttons on an information board that participants could click to reveal the answer. Each time a cue was revealed, participants evaluated it in relation to each competing diagnosis (0=“no support” to 10=”strong support”) and updated their diagnostic choice. When they felt ready, they made their final choice of diagnosis. We measured PID against the cue evaluations of a control group, and assessed its contribution to the final diagnosis via mediation analysis.
Initial choice of diagnosis (non-serious vs. serious) predicted final choice (OR=4.78, P<0.001). Magnitude and direction of PID fully mediated this relationship: an initial non-serious diagnosis was associated with PID to support it, which in turn increased the odds of a non-serious final diagnosis. Final diagnosis predicted management: most physicians who provided a non-serious final diagnosis did not refer the patient (70%), whilst only 3% of those who provided a serious final diagnosis failed to refer. We identified no differences in the number of cues elicited by physicians who selected a non-serious vs. a serious final diagnosis.
Our findings shed light on some of the cognitive causes of diagnostic error that can impact patients. Initial diagnostic hypotheses are important, but the interpretation of subsequent information may be more so.
Conventional tornado diagram has several challenges: it is based on deterministic one-way sensitivity analysis (1-way SA) which produces biased results when parameters are correlated. Moreover, it uses the incremental cost effectiveness ratio (ICER) which has a number of problems such as non-uniqueness. Our purpose is to generate a modified tornado diagram that addresses the above problems to improve decision-making.
We develop a modified tornado diagram that allows the result of probabilistic 1-way SA to be presented clearly to decision-makers through these processes: (a) for each of three parameters (probability, utility and cost), we fixed several (outer loop) values. For each outer loop value, we allowed the other parameters to undergo simulation so that when simulation is complete we observe the costs and effects for two strategies (21 gene assay versus ROR used in genomic tests for early-stage breast cancer) under comparison. (b)We proceeded to calculate the incremental net monetary benefit (INMB) of gene assay against ROR, corresponding to each outer loop value. (c) Next, we plotted the probability density function (PDF) for each parameter with INMB on the x-axis and probability on the y-axis; a vertical line is then drawn at the point where the INMB is zero (decision-switching point). (d) Finally, we calculated the Expected Value of Perfect Parameter Information (EVPPI) for each parameter and arranged the PDF graphs with highest-EVPPI parameter on top and lowest-EVPPI parameter at the bottom.
The modified tornado diagram we have developed is shown below:
The graph shows PDF for the two willingness-to-pay (WTP) values we used in calculating INMB. The EVPPI for each curve and the proportion of each curve that lies above 0 on the x-axis (thus, the probability that gene assay is cost-effective) are reported on the graph. For WTP of £20,000, only the probability parameter contributes towards decision uncertainty. For WTP of £30,000 the order of importance of parameters to decision-makers is probability, utility, and cost.
We have executed a probabilistic 1-way SA which, while examining the sensitivity of model conclusions to changes in only one parameter's value, simultaneously takes into consideration its correlation with other parameters. We have also used INMB as model output instead of the ICER. Our approach yields a more reliable decision-making tool.
TITLE: Probabilistic sensitivity analysis of cost-effectiveness analysis of the Breast cancer screening programme in the Basque country: A multi-cohort discrete-event simulation model.
Purpose: The aim of this study was the evaluation of the breast cancer early detection programme in the Basque Country from 1996 to 2011 in terms of probabilistic cost-effectiveness analysis and the probabilistic sensitivity analysis.
Method(s): A discrete event simulation model was built to reproduce the natural history of breast cancer (BC). We estimated for lifetime follow-up the total cost of BC (screening, diagnosis and treatment), as well as quality-adjusted life years (QALY), for women invited to participate in the evaluated programme during the 15-year period in the actual screening scenario and in a hypothetical unscreened scenario. The probabilistic feature of the model was based on varying the main variables randomly at the same time. Uniform distributions were adopted to vary time between successive invitations and the mean duration of the pre-clinical phase, Beta distributions for sensitivity and specificity of the programme and Dirichlet was the distribution selected for the detection stages classification in screen-detected cancers were the variables we varied in the probabilistic sensitivity analysis. Therefore, we were able to examine the effect of joint uncertainty in these variables through cost-effectiveness plane and calculate the expected value of perfect information.
Result(s): The actual screening programme involved a mean cost of 1,123 million euros and provided 6.7 million QALYs over the lifetime of the target population, resulting in a gain of 10,110 QALYs for an additional cost of 22.3 million euros, compared with the unscreened scenario. Thus, the incremental cost-effectiveness ratio was 2,209€/QALY. All the model runs in the probabilistic sensitivity analysis resulted in an incremental cost-effectiveness ratio lower than 10,000€/QALY. The expected value of perfect information associated to a 5,000€/QALY threshold was a population opportunity loss of 163,620€. Cancer stage distribution in screen-detected cancers was the variable with greater impact on the final incremental cost-effectiveness ratio.
Conclusion(s): The BC screening programme in the Basque Country proved to be cost-effective during the evaluated period. No addition research on the main parameters was necessary. These results confirm the epidemiological benefits related to the centralised screening system and support the continuation of the programme.
Figure 1: Cost-effectiveness plane for the period from 1996 through 2011