This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on http://www.researchprotocols.org, as well as this copyright and license information must be included.
The National Lung Screening Trial demonstrated that low-dose computed tomography (LDCT) screening could be an effective way to reduce lung cancer mortality. Informed decision-making in the context of lung cancer screening requires that potential screening subjects accurately recognize their own lung cancer risk, as well as the harms and benefits associated with screening, while taking into account their personal values and preferences.
Our objective is to develop a Web-based decision aid in accordance with the qualifying and certification criteria in the International Patient Decision Aid Standards instrument version 4.0 that will assist patients in making informed decisions with regard to lung cancer screening.
In “alpha” testing, a prototype of the decision aid was tested for usability with 10 potential screening participants in focus groups. Feedback was also sought from public health and health risk communication experts external to the study. Following that, improvements to the prototype were made accordingly, and “beta” testing was done in the form of a quasi-experimental design—a before-after study—with a group of 60 participants. Outcomes tested were knowledge, risk perception of lung cancer and lung cancer screening, decisional conflict, and acceptability of the decision aid as determined by means of a self-administered electronic survey. Focus groups of a subsample of survey participants will be conducted to gain further insight into usability issues.
Alpha testing is completed. Beta testing is currently being carried out. As of 2014 December 7, 60 participants had completed the before-after study. We expect to have results by 2015 January 31. Qualitative data collection and analysis are expected to be completed by 2015 May 31.
We hypothesize that this Web-based, interactive decision aid containing personalized, graphical, and contextual information on the benefits and harms of LDCT screening will increase knowledge, reduce decisional conflict, and improve concordance between patient preferences and the current US Preventive Services Task Force’s screening guidelines.
Lung cancer remains the leading cause of cancer death in the United States [
Implementing lung cancer screening in an environment where patients do not have the tools or information to understand disease risks and the harm-benefit balance of screening will most likely be counterproductive. In addition to providing information for individuals regarding lung cancer screening that allows them to weigh the potential harms of LDCT in accordance with the benefits, we also recognize that the decision to be screened is preference-sensitive. In light of this, there is a need to assist individuals with making informed decisions regarding lung cancer screening in which their personal values are also taken into account.
The USPSTF defines informed decision-making as “an individual’s overall process of gathering relevant health information from both his or her clinician and from other clinical and nonclinical sources, with or without independent clarification of values” ( [
Evidence shows that decision aids can improve decision quality as a result of better knowledge of options and their associated harms and benefits; decrease decisional conflict; and reduce the overuse and increase the underuse of screening options [
Whereas numerous decision aids exist for prostate, colon, and breast cancers, there are only a handful of tools that fulfill the functions of a patient decision aid, either partially or fully, with regard to LDCT screening [
A prototype was developed based on the most recent clinical guidelines provided by the USPSTF [
The distinguishing factor of the Web-based version is that it allows individuals to compute their individualized lung cancer risk using an established risk model. Although all cancer screening is based on some risk factor identification, for the most common evidence-based cancer screening (e.g., colorectal, breast, and cervical cancer screening), risk factor identification involves little more than noting the relevant age and sex data. Lung cancer screening differs in very important ways in that proper application of current recommendations involves measuring risk by also identifying the individual’s cumulative tobacco exposure measured in pack-years and the timing of tobacco use. Whereas smoking history accurately identifies risk on a population basis, it is less useful for individuals because there is such great variability of lung cancer risk, even among smokers with similar exposure. Current models that more accurately quantify individual lung cancer risk incorporate up to 10 clinical and demographic variables that provide a more accurate, though complex, determination of risk. This requires sophisticated lung cancer risk prediction models [
In addition to accurately characterizing individual lung cancer risk, there is evidence that individuals may prefer to have such tailored information, which, in turn, may affect health-care-seeking decisions [
We believe that putting LDCT screening in context with other common screening recommendations in terms of reduction of disease-specific mortality (i.e., reduction of breast cancer death by screening with mammography) will allow the individual to gauge how LDCT screening compares to other widely accepted screening practices. This has not been done with the available decision aids for LDCT lung cancer screening [
Our target population comprised of potential users of LDCT screening. The specific inclusion criteria are given in
These criteria applied to all 3 phases of the study detailed below: Phase I alpha testing focus groups, Phase IIa beta testing before-after survey, and Phase IIb beta testing postsurvey focus groups.
Current and former smokers
Aged 45-80 years
No history of lung cancer
No previous chest computed tomographic scan in the past year
A combination of passive and active recruitment was done to form 2 convenience samples of participants: 10 for alpha testing and 60 for beta testing. An advertisement was placed on the University of Michigan’s (UM’s) online portal for volunteers of clinical studies [
We solicited feedback from public health and health risk communication experts with regard to the content and wording and how risk is expressed in the prototype. After incorporating their suggestions, we conducted focus groups with potential users of LDCT screening to test the usability of the prototype as part of the decision tool’s iterative development process. Specifically, we pretested our tool for comprehension of the content, as well as the acceptability of the design, layout, and messages conveyed. A total of 10 people participated in 1 of 2 focus groups. The eligible and willing participants were sent a Web link to work through the Web-based decision aid prototype via e-mail. All participants were asked to participate in the focus group within the week they reviewed the decision aid. Version 1 was the product of alpha testing.
The focus group was conducted by a trained external facilitator and a study team member using an interview guide developed by study team members with expertise in qualitative research and lung cancer screening decision-making (see
Audio recordings were transcribed verbatim. Data analysis took place simultaneously with data collection, which, in turn, assisted in the iterative development of both the interview guide and decision aid. Using the transcriptions and field notes, a brief report was given by the study team member present at the focus groups to the rest of the study team. The study team then decided what changes to incorporate into the decision aid, forming version 1.
Following alpha testing, we conducted a pilot study of version 1 of our decision aid with 60 individuals using a quasi-experimental design: a before-after study. This study design is consistent with the development of a decision aid and is an accepted method to test the effectiveness of decision aid tools under “real-life” conditions [
Following a successful screen for eligibility, a participant was invited to come to UM to complete a series of surveys administered on a computer by Qualtrics, an online survey tool. In particular, a participant began with the knowledge survey (see
The outcomes measured were adapted from the Ottawa Decision Support Framework: knowledge of the benefits and risks of lung cancer screening, acceptability [
Following completion of the survey, participants were asked if they would like to participate in a focus group to allow them to give the study team more feedback about the decision aid they just viewed. If the answer was affirmative, the participant was told that he or she could be contacted within the month to make an appointment. They were also asked whether the decision aid had indicated that they were eligible for screening (based on USPSTF guidelines) and what the risk of dying due to lung cancer was as computed by the calculator.
Focus groups will be stratified by self-reported eligibility for LDCT screening. Given that one of the main goals of the tool is to increase concordance with USPSTF eligibility guidelines, it will be useful to have specific focus groups consisting of screen-(in)eligible individuals entirely. The aim is to have 4-8 individuals per focus group. Where possible, focus groups will also be stratified by sex and age, given eligibility for LDCT screening.
The same steps will be followed for beta testing as those used for the focus groups conducted for alpha testing. The current version of the interview guide can be seen in
Thematic analysis will be done on the data yielded from the focus group sessions. Two study team members with qualitative analysis expertise will code the data separately and compare codes to establish the themes to be explored. A report will be given to the rest of the team, and all members will decide what changes to include for the next iteration of the decision aid.
All study participants will be asked to complete consent forms. Specific consent for audiotaping of focus groups will also be sought. This study was approved by the University of Michigan Medical School Institutional Review Board on 2014 June 18 (Study ID: HUM00088232).
Alpha testing is completed. Beta testing is currently being carried out. As of 2014 December 7, 60 participants had completed the before-after study. We expect to have results by 2015 January 31. Qualitative data collection and analysis are expected to be completed by 2015 May 31. The current iteration of the results page with personalized risk generated by the decision aid can be seen at
Preliminary results from the before-after study indicate that the decision aid improves knowledge about lung cancer screening, decreases decisional conflict, and increases concordance between USPSTF recommendations and the screening option preferred by the user. Therefore, we anticipate that the decision aid will be helpful to individuals in making informed decisions about lung cancer screening.
An example of the current iteration of the results page.
Focus group interview guide - alpha testing.
Focus group participant survey.
Focus group exit survey.
Before-after survey: knowledge.
Before-after survey: demographics and risk factors.
Before-after survey: risk perception.
Before-after survey: decisional conflict.
Before-after survey: values clarification.
Before-after survey: acceptability.
Focus group interview guide – revised for beta testing.
International Patient Decision Aid Standards instrument
low-dose computed tomography
National Lung Screening Trial
University of Michigan
University of Michigan Health System
US Preventive Services Task Force
YKL, TJC, DA, and RM conceived of the study and sought funding and ethical approval. YKL and RM are responsible for the management of the alpha and beta testing phases and planned the statistical analysis. STC and PC assisted in the development of the prototype and subsequent iterations of the decision tool. PC and MW have assisted with data collection in both phases of testing. All authors have been involved in drafting and revising the manuscript and approved the final version.
This work has been supported by the Elizabeth A Crary Fund of the University of Michigan Comprehensive Cancer Center.
None declared.