Lung cancer is the most common cause of cancer-related death in the United States with an annual incidence of over 226,000 and a 5-year survival rate of only16%. Veterans face a significant burden of disease from this condition due in part to higher rates of former and current smoking compared to the general population. To date, lung cancer control has focused primarily on smoking cessation. However, this clinical paradigm is shifting. A recent randomized clinical trial found that screening with low dose computer tomography (LDCT) was associated with a 20% reduction in lung cancer mortality. Despite the strength of this evidence, questions remain regarding the translation of findings to the real world setting including Veteran populations. Further, it is unclear how individuals will weigh the benefit, potential harms, and uncertainties associated with lung cancer screening. The mortality benefits of lung cancer screening must be balanced against possible harms and costs including the risk of false positive tests, over-diagnosis, and the resources required to implement a lung cancer screening program. This project will lay the foundation for the development of a patient centered approach to the implementation of a lung cancer screening program.
The objectives of this pilot study are: 1) to identify how best to communicate the benefits, harms, and uncertainties associated with lung cancer screening to Veterans, 2) to develop and administer a conjoint analysis experiment to assess patient preferences regarding lung cancer screening, and 3) to develop a conceptual framework for a shared decision making tool for use among Veterans considering lung cancer screening.
These specific aims will be accomplished through a mixed methods approach and collaborative efforts of an interdisciplinary team of investigators. Qualitative interviews will be conducted to identify attributes of lung cancer screening salient to Veterans and develop strategies for the communication of benefits, harms, and uncertainties associated with lung cancer screening. Communication formats will be designed to communicate statistical, evidentiary and stochastic uncertainty associated with lung cancer screening program based upon qualitative analysis of structured interviews in the Veteran population. A conjoint analysis experiment will be developed to evaluate attributes of lung cancer screening most important to older Veterans with a significant smoking history. The relative importance weights that Veterans place on lung cancer screening attributes will indicate if large differences are found by race in this population. Insights from the structured interviews and conjoint analysis experiment will be used to develop a framework for shared decision making tool for Veterans considering lung cancer screening.
Subjects for each phase will be recruited from primary care patients enrolled in clinics at the PVAMC. Eligibility criteria will include1) age of 55 to 74 years, 2) a history of smoking equivalent to 20 pack years or more, and 3) enrollment in a Patient Aligned Care Team at the PVAMC. Exclusion criteria will include 1) a prior history of lung cancer, or 2) a clinical history of cognitive dysfunction. A sample size of 100 subjects with 50 black and 50 white subjects will be recruited for the conjoint analysis experiment.
There were 10 Veterans enrolled in the qualitative interviews; 100% (n=10) were black, 60% (n=6) had up to a high school or GED degree, 20% (n=2) had up to some college, and 20% (n=2)had 4 years or more of college. Participants expressed a general understanding of random or stochastic uncertainty when presented as a pictograph conveying a risk estimate. There was also a general acceptance of statistical uncertainty, or imprecision, when presented with a pictograph that conveyed a range around a given estimate. Responses to evidentiary uncertainty, or controversy in the interpretation of evidence and guidelines, was mixed with many looking to a trusted provider to provide guideance. Based on the qualitative interviews, the additional attributes of overdiagnosis and need to attend a nodule clinic were added to the conjoint analysis design. There were 120 participants enrolled in the conjoint analysis study with all completing the study. Forty-four percent (44%, n=54) were white and 61% (n=61) were black. Niney-six percent (96%, n=115) were male and 4% (n=5) were female. Seventeen percent (17%, n=20) had up to an 11'th grade education, 56%(n=66) had a high school or GED degree, 24%(n=29) had some college, and 4% (n=4) had at least 4 years of college. Eighty-three percent (83%) had an annual household income of < $40,000. All participants completed the conjoint analysis experiment and the wieght of the utilities were determined and attributes ranked in terms of importance with respect to the decision to intiate lung cancer screening. The attributes that were most important to lung cancer screening decision making in decreasing order were 1)need to be followed in a nodule clinic, 2) reduction in lung cancer mortality, 3) need for invasive testing, 4) false positive findings, 5) risk of complications from follow-up testing, 6) health risks due to radiation, 7) cost, and 8) risk of overdiagnosis. The correlation of weights ascertained by the conjoint analysis and a ranking exercise was high (r=0.86). The mean fit statistics for the utility models was 0.46, within the acceptable range. We evaluated the effect of race (black vs. not black) on utility assessments. The only attribute that varied by race was cost, with concern about cost higher among blacks than non-blacks. A comparison of ranks found that black participants ranked overdiagnosis lower (p=0.049), need for invasive tests lower (p=0.023),and need to attend a nodule clinic higher (p=0.024) than non-black participants. These differences in utilities for outcomes associated with lung cancer screening can inform shared decision making interventions. Further, our data suggests that ranking of these attributes has a high correlation with utility assessment and may be a feasible and valid approach to value elicitation in a shared decison making intervention for lung cancer screening.
The goals of this project are consistent with those of the Veterans Health Administration National Center for Health Promotion (NCP) and investigators will work with the NCP to further develop and disseminate the informed decision making tools that emerge from this work. This pilot project provides preliminary data regarding the feasibility and acceptability of communication strategies and the use of a theoretically based preference assessment tool in VA populations. The findings in this study will inform approaches to preference assessment and value elicitation to be used for decsion support pertaining to lung cancer screening. Further work is needed to explain the concept of overdiagnosis and the experience and risk of follow-up testing and surveillance as a patient participates in a lung cancer screening program. The research team is planning to further refine preference assessment approaches and the decision support materials available to Veterans for lung cancer screneing and to evaluate the use of decision support in a multi-site clinical trial at VA Medical Centers.
- Schapira MM, Aggarwal C, Akers S, Aysola J, Imbert D, Langer C, Simone CB, Strittmatter E, Vachani A, Fraenkel L. How Patients View Lung Cancer Screening. The Role of Uncertainty in Medical Decision Making. Annals of the American Thoracic Society. 2016 Nov 1; 13(11):1969-1976.
- Schapira MM. Communication of disease/cancer risk to clinicians and consumers/patients. Presented at: Agency for Healthcare Research and Quality Dissemination Task Force Meeting; 2015 Jul 1; Rockville, MD.
- Schapira MM. Tailoring communication strategies to health numeracy level in the cancer consultation. Presented at: National Cancer Institute Behavioral Research Decision-Making Interest Group Meeting; 2015 May 1; Rockville, MD.