Prevention and Control
Conditions
Keywords
decision support, patient preferences, risk adjustment
Brief summary
The overall goal of this project is to test two strategies for implementing a shared decision making tool to be used by providers while talking to patients about lung cancer screening. Eight participating sites will be randomized to compare standard implementation with intensive implementation. Additionally, the investigators will determine the factors that were most important for successful implementation of the shared decision making tool. Finally, the investigators will survey patients to evaluate the effects of Decision Precision on patient's knowledge of the risks and benefits of lung cancer screening, the quality of their decision making, and their satisfaction with care.
Detailed description
The overall goal of this quality improvement project is to test two strategies for implementing shared decision making, which incorporates the Decision Precision lung cancer screening tool. The investigators will use multi-site, cluster-based randomization to compare standard implementation with intensive implementation. Additionally, the investigators will determine the factors that were most important for successful implementation of the shared decision making tool. Finally, the investigators will have a human subjects research component to evaluate the effects of Decision Precision on patient's knowledge of the risks and benefits of lung cancer screening, the quality of their decision making, and their satisfaction with care.
Interventions
OTHERWebinar, Promotion, and Tool Access
All sites will receive a professionally developed, 15-minute webinar that describes the tool's development (e.g., how the algorithm was designed) and a tutorial on how to use the web-site. The investigators will promote the webinar and using the tool through key local leaders. A web-link to use the tool will be placed within the lung cancer screening clinical reminder.
OTHERLEAP
Sites randomized to intensive implementation will be offered participation in the LEAP program. LEAP (Learn. Engage. Act. Program) is a multi-week, online learning collaborative using systems redesign techniques to help sites identify and overcome barriers to implementation of shared decision making (use of the Decision Precision tool).
OTHERAudit and Feedback
Providers at all sites will have access to this system. It will provide feedback on the screening and shared decision making process (e.g., number of provider's eligible patients screened for lung cancer, use of the tool, patient knowledge and satisfaction from patient surveys).
OTHERAcademic Detailing
All sites will be offered an academic detailing approach to implementation of Lung Decision Precision. A trained detailer will travel to all sites who agree to a visit. The detailer will meet with individual primary care providers whenever possible, and groups of providers as necessary. He will present evidence for a risk-based approach to screening and how to use the tool with patients to make tailored screening recommendations.
Sponsors
VA Office of Research and Development
Study design
Observational model
CASE_CONTROL
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
55 Years to 80 Years
Healthy volunteers
Yes
Inclusion criteria
* Patients at participating sites without documented exclusions for lung cancer screening who have had initial lung cancer screening clinical reminders resolved during the recruitment period
Exclusion criteria
Exclusions for initial lung cancer screening clinical reminders: * history of lung, pancreatic, liver or esophageal cancer * a health factor on file indicating they have a life expectancy of less than 6 months or who already had a chest CT in the past 12 months
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Odds Ratio of the Interaction Between Lung Cancer Risk and Implementation Arm | post implementation, an average of 15 months | First, we estimated screening's net benefit for an individual based on their baseline lung cancer risk, as estimated using the Bach et. al. annual lung cancer incidence model. Patients are considered high benefit if their annual lung cancer risk is between 0.3%-1.3%. Patients outside this range are considered preference sensitive. We fit a multilevel logistic regression model where receipt of screening is the outcome. Precision decision making is reflected in the association between baseline lung cancer risk and screening utilization: an increase in screening utilization for those at higher lung cancer risk indicates some degree of precision decision making. The primary outcome for the cluster-randomized design assesses the difference in precision decision making in the standard vs. intensive implementation arms. This is estimated as the effect on screening receipt of the interaction between risk and implementation arm. |
| Patient Satisfaction With Decision and Process | Survey mailed to Veteran several weeks after identified as having an initial discussion about lung cancer screening using VA administrative data | Obtained from patient surveys (for the subset of the overall participants who returned surveys). The unit of measurement is one unit on the scale \[scale of 0 (very poor) to 10 (very good)\]. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Number of Tool Assessments Where Patient Decision Aid Was Printed | post implementation, up to 25 months | Number of times during study duration where patient decision aid was printed from the Lung Decision Precision web-site. |
| Number of Times Dynamic Pictograph Depicting Personalized Benefit and Harm Was Opened for Display | post implementation, up to 25 months | Number of times dynamic pictograph depicting personalized benefit and harm was opened for display in the Lung Decision Precision web-site during the study period, collected as para data from Decision Precision web-site |
| Formative Evaluation to Determine the Factors Most Important for Successful Implementation of Decision Precision Tool | At least one year post-implementation of Lung Decision Precision web-site | Qualitative analysts will conduct telephone interviews with providers at each site who identify themselves as participating in shared decision making process with patients regarding lung cancer screening. Reporting for this report is number of interviews completed of number interviews requested. |
Countries
United States
Participant flow
Recruitment details
Eight VA sites participated in this Quality Improvement Project. Four were randomized to each arm. The number of participants in each arm reflects how many Veterans at the sites randomized to each arm who were eligible for lung cancer screening according to CDW data between October 2016 and December 2019.
Participants by arm
| Arm | Count |
|---|---|
| Standard Implementation Promotion, Tool Access, academic detailing + Audit and Feedback
Promotion, and Tool Access: All sites will be given access to a You Tube video that describes the tool's development (e.g., how the algorithm was designed) and a tutorial on how to use the web-site. The investigators will promote the webinar and using the tool through key local leaders. A web-link to use the tool will be placed within the lung cancer screening clinical reminder.
Audit and Feedback: Sites will be provided with reports to provide feedback on the screening and shared decision making process. Site teams were encouraged to send these reports to providers.
Academic Detailing: All sites will be offered an academic detailing approach to implementation of Lung Decision Precision. A trained detailer will travel to all sites who agree to a visit. The detailer will meet with individual primary care providers whenever possible, and groups of providers as necessary. He will present evidence for a risk-based approach to screening and how to use the tool with patients to make tailored screening recommendations. | 5,275 |
| Intensive Implementation Promotion, and Tool Access, academic detailing + Audit and Feedback + LEAP
Promotion, and Tool Access: All sites will be given access to a You Tube video that describes the tool's development (e.g., how the algorithm was designed) and a tutorial on how to use the web-site. The investigators will promote the webinar and using the tool through key local leaders. A web-link to use the tool will be placed within the lung cancer screening clinical reminder.
LEAP: Sites randomized to intensive implementation will be offered participation in the LEAP program. LEAP (Learn. Engage. Act. Program) is a multi-week, online learning collaborative using systems redesign techniques to help sites identify and overcome barriers to implementation of shared decision making (use of the Decision Precision tool).
Audit and Feedback: Sites will be provided with reports to provide feedback on the screening and shared decision making process. Site teams were encouraged to send these reports to providers.
Academic Detailing: All sites will be offered an academic detailing approach to implementation of Lung Decision Precision. A trained detailer will travel to all sites who agree to a visit. The detailer will meet with individual primary care providers whenever possible, and groups of providers as necessary. He will present evidence for a risk-based approach to screening and how to use the tool with patients to make tailored screening recommendations. | 11,758 |
| Total | 17,033 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 |
|---|---|---|---|
| Overall Study | Death | 482 | 855 |
Baseline characteristics
| Characteristic | Standard Implementation | Intensive Implementation | Total |
|---|---|---|---|
| Age, Categorical <=18 years | 0 Participants | 0 Participants | 0 Participants |
| Age, Categorical >=65 years | 2902 Participants | 5455 Participants | 8357 Participants |
| Age, Categorical Between 18 and 65 years | 2373 Participants | 6303 Participants | 8676 Participants |
| Age, Continuous | 66 years | 64 years | 64 years |
| Ethnicity (NIH/OMB) Hispanic or Latino | 59 Participants | 165 Participants | 224 Participants |
| Ethnicity (NIH/OMB) Not Hispanic or Latino | 4904 Participants | 10951 Participants | 15855 Participants |
| Ethnicity (NIH/OMB) Unknown or Not Reported | 312 Participants | 642 Participants | 954 Participants |
| Race (NIH/OMB) American Indian or Alaska Native | 60 Participants | 112 Participants | 172 Participants |
| Race (NIH/OMB) Asian | 5 Participants | 27 Participants | 32 Participants |
| Race (NIH/OMB) Black or African American | 314 Participants | 1903 Participants | 2217 Participants |
| Race (NIH/OMB) More than one race | 28 Participants | 58 Participants | 86 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 27 Participants | 63 Participants | 90 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 472 Participants | 807 Participants | 1279 Participants |
| Race (NIH/OMB) White | 4369 Participants | 8788 Participants | 13157 Participants |
| Sex: Female, Male Female | 196 Participants | 848 Participants | 1044 Participants |
| Sex: Female, Male Male | 5079 Participants | 10910 Participants | 15989 Participants |
| Travel Distance | 37.8 miles | 45.9 miles | 43.8 miles |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk |
|---|---|---|
| deaths Total, all-cause mortality | 482 / 5,275 | 855 / 11,758 |
| other Total, other adverse events | 0 / 5,275 | 0 / 11,758 |
| serious Total, serious adverse events | 0 / 5,275 | 0 / 11,758 |
Outcome results
Primary
Odds Ratio of the Interaction Between Lung Cancer Risk and Implementation Arm
First, we estimated screening's net benefit for an individual based on their baseline lung cancer risk, as estimated using the Bach et. al. annual lung cancer incidence model. Patients are considered high benefit if their annual lung cancer risk is between 0.3%-1.3%. Patients outside this range are considered preference sensitive. We fit a multilevel logistic regression model where receipt of screening is the outcome. Precision decision making is reflected in the association between baseline lung cancer risk and screening utilization: an increase in screening utilization for those at higher lung cancer risk indicates some degree of precision decision making. The primary outcome for the cluster-randomized design assesses the difference in precision decision making in the standard vs. intensive implementation arms. This is estimated as the effect on screening receipt of the interaction between risk and implementation arm.
Time frame: post implementation, an average of 15 months
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| Standard Implementation | Odds Ratio of the Interaction Between Lung Cancer Risk and Implementation Arm | 1.03 odds ratio |
| Intensive Implementation | Odds Ratio of the Interaction Between Lung Cancer Risk and Implementation Arm | 1.53 odds ratio |
p-value: 0.08195% CI: [0.58, 0.78]Regression, Logistic
Primary
Patient Satisfaction With Decision and Process
Obtained from patient surveys (for the subset of the overall participants who returned surveys). The unit of measurement is one unit on the scale \[scale of 0 (very poor) to 10 (very good)\].
Time frame: Survey mailed to Veteran several weeks after identified as having an initial discussion about lung cancer screening using VA administrative data
Population: Surveys were sent to eligible Veterans, the overall number of participants analyzed is a reflection of the subset of the overall participants who returned surveys. The unit of measurement is one unit on the scale \[scale of 0 (very poor) to 10 (very good)\].
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Standard Implementation | Patient Satisfaction With Decision and Process | 7 units on a scale |
| Intensive Implementation | Patient Satisfaction With Decision and Process | 7 units on a scale |
Comparison: Simple t-test comparing mean satisfaction ratings between armsp-value: 0.3595% CI: [-0.16, 0.46]t-test, 2 sided
Secondary
Formative Evaluation to Determine the Factors Most Important for Successful Implementation of Decision Precision Tool
Qualitative analysts will conduct telephone interviews with providers at each site who identify themselves as participating in shared decision making process with patients regarding lung cancer screening. Reporting for this report is number of interviews completed of number interviews requested.
Time frame: At least one year post-implementation of Lung Decision Precision web-site
Population: Of sites randomized to each Arm, the number of providers involved in lung cancer screening who were completed an interview about DecisionPrecision and the risk-based approach of the number of providers who were asked to complete an interview.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Standard Implementation | Formative Evaluation to Determine the Factors Most Important for Successful Implementation of Decision Precision Tool | 25 Participants |
| Intensive Implementation | Formative Evaluation to Determine the Factors Most Important for Successful Implementation of Decision Precision Tool | 13 Participants |
Secondary
Number of Times Dynamic Pictograph Depicting Personalized Benefit and Harm Was Opened for Display
Number of times dynamic pictograph depicting personalized benefit and harm was opened for display in the Lung Decision Precision web-site during the study period, collected as para data from Decision Precision web-site
Time frame: post implementation, up to 25 months
Population: The para data from the website is not reliable - unable to assess this planned outcome.
Secondary
Number of Tool Assessments Where Patient Decision Aid Was Printed
Number of times during study duration where patient decision aid was printed from the Lung Decision Precision web-site.
Time frame: post implementation, up to 25 months
Population: Unable to complete - para data from website did not collect this information as planned.