Development and psychometric evaluation of a decision quality instrument for lung cancer screening decisions (DQI-LCS).
[BACKGROUND] Lung cancer screening (LCS) reduces lung cancer mortality among high-risk individuals.
APA
Tan NQP, Lopez-Olivo MA, et al. (2026). Development and psychometric evaluation of a decision quality instrument for lung cancer screening decisions (DQI-LCS).. BMC medical informatics and decision making, 26(1). https://doi.org/10.1186/s12911-026-03346-y
MLA
Tan NQP, et al.. "Development and psychometric evaluation of a decision quality instrument for lung cancer screening decisions (DQI-LCS).." BMC medical informatics and decision making, vol. 26, no. 1, 2026.
PMID
41723422
Abstract
[BACKGROUND] Lung cancer screening (LCS) reduces lung cancer mortality among high-risk individuals. As the screening procedure has both potential benefits and harms, the Centers for Medicare & Medicaid Services mandates a shared decision-making (SDM) discussion prior to the initial LCS. We evaluated the psychometric properties of a decision quality instrument for LCS (DQI-LCS) that assesses whether patients (1) were informed about their options (knowledge), (2) made a screening decision in line with their preferences (goal concordance), and (3) underwent an optimal SDM process (measured using the SDM Process measure [SDMP_4]).
[METHODS] Secondary analysis of survey data with patients who underwent LCS at tertiary care medical centers in the US within 12 months of the survey. Psychometric evaluation examined the DQI knowledge measure for item retention and deletion, feasibility, test-retest reliability, and validity of the three components of the DQI-LCS.
[RESULTS] The final sample ( = 264) was 48.0% female and had a mean age of 64.8. One knowledge item (KQ13) was dropped as it had a poor index of discrimination, resulting in a 15-item LCS knowledge measure (LCS-15). We also developed a brief knowledge measure with 5 items (LCS-5). Overall, the DQI-LCS was feasible to complete (under 5% missing data for all items). However, only the overall LCS-15 knowledge measure had acceptable test-retest reliability. The SURE score was positively associated with knowledge, demonstrating construct validity of the knowledge component. When there was goal concordance, there was a greater probability of the participant reporting that they would make the same decision again and be screened every year, demonstrating predictive validity of the goal concordance measure. Finally, higher SDMP_4 scores were associated with higher SURE scores, demonstrating predictive validity of SDMP_4. However, we did not find concurrent validity of SDMP_4 with CollaboRATE.
[CONCLUSIONS] Overall, there is evidence that the DQI-LCS was feasible to complete, had moderate test-retest reliability, and acceptable construct and predictive validity. A limitation is that our sample only included people who had undergone LCS. With further evaluation, the validated DQI-LCS can be a quality metric for assessing whether patients are receiving guideline-concordant care for LCS.
[METHODS] Secondary analysis of survey data with patients who underwent LCS at tertiary care medical centers in the US within 12 months of the survey. Psychometric evaluation examined the DQI knowledge measure for item retention and deletion, feasibility, test-retest reliability, and validity of the three components of the DQI-LCS.
[RESULTS] The final sample ( = 264) was 48.0% female and had a mean age of 64.8. One knowledge item (KQ13) was dropped as it had a poor index of discrimination, resulting in a 15-item LCS knowledge measure (LCS-15). We also developed a brief knowledge measure with 5 items (LCS-5). Overall, the DQI-LCS was feasible to complete (under 5% missing data for all items). However, only the overall LCS-15 knowledge measure had acceptable test-retest reliability. The SURE score was positively associated with knowledge, demonstrating construct validity of the knowledge component. When there was goal concordance, there was a greater probability of the participant reporting that they would make the same decision again and be screened every year, demonstrating predictive validity of the goal concordance measure. Finally, higher SDMP_4 scores were associated with higher SURE scores, demonstrating predictive validity of SDMP_4. However, we did not find concurrent validity of SDMP_4 with CollaboRATE.
[CONCLUSIONS] Overall, there is evidence that the DQI-LCS was feasible to complete, had moderate test-retest reliability, and acceptable construct and predictive validity. A limitation is that our sample only included people who had undergone LCS. With further evaluation, the validated DQI-LCS can be a quality metric for assessing whether patients are receiving guideline-concordant care for LCS.