Effects of Clinical Trial or Research Program Participation Status on In‐Hospital Mortality After Transcatheter Aortic Valve Implantation

Tadao Aikawa,Toshiki Kuno,Jef Van den Eynde,Alexandros Briasoulis,Aaqib H. Malik

DOI: https://doi.org/10.1161/JAHA.121.025920

2022-06-16

Abstract:The development of novel devices and the favorable results of several randomized clinical trials have allowed for the rapid expansion of transcatheter aortic valve implantation (TAVI) to elderly patients with aortic stenosis across all risk categories 1 ; however, the highly selected populations that are typically enrolled in randomized clinical trials may limit generalizability of the results to the real‐world population with aortic stenosis. Furthermore, clinical trial or research program participation itself can facilitate behavior change in patients and health care providers and may contribute to improved patient outcomes, which is known as the "Hawthorne effect." 2 Previous studies reported that research participation was associated with better survival in patients with acute coronary syndrome. 3 , 4 Given the lack of data exploring the effect of research participation on outcomes after TAVI, we compared the short‐term survival after TAVI between clinical research participants and nonparticipants using the Nationwide Inpatient Sample. The data that support the findings of this study are available from the corresponding author upon reasonable request. The Nationwide Inpatient Sample is the largest publicly available all‐payer inpatient health care database in the United States and did not require ethical approval. All patients who underwent TAVI between 2013 and 2019 (n=56 648) were identified from the Nationwide Inpatient Sample using the following International Classification of Diseases, Tenth Revision, Clinical Modification (ICD‐10‐CM) codes: 02RF37H, 02RF37Z, 02RF38H, 02RF38Z, 02RF3JH, 02RF3JZ, 02RF3KH, and 02RF3KZ. Patients with age ≤18 years (n=22), cirrhosis (n=760), end‐stage renal disease (n=2136), do‐not‐resuscitate status or palliative care involvement (n=383), and cancer (n=1952) were excluded with reference to previous trials. Patients with missing data (n=12) were also excluded. Research participation status was identified using ICD‐10‐CM code Z00.6, 4 which was restricted to code as the primary diagnosis or first secondary diagnosis to avoid overcapturing. The primary outcome in this study was in‐hospital mortality. A multilevel logistic regression analysis accounting for strata and hospital clustering was performed to examine the association between clinical research participation status and in‐hospital mortality. A sensitivity analysis was performed in the propensity score‐matched cohort, which was generated by 1:1 nearest‐neighbor matching using a caliper width of 0.01 adjusting for baseline characteristics. Variables for these analyses were chosen on the basis of clinical relevance and the previous work 4 and were included if they were significantly different in the baseline characteristics Table. Difference between groups were tested by 1‐way ANOVA and the chi‐square test, as appropriate. Statistical analysis was performed using Stata 16.1 (StataCorp) and R (R Foundation). Table . Baseline Characteristics and In‐Hospital Outcomes in Clinical Research Participants and Nonparticipants Undergoing Transcatheter Aortic Valve Implantation Characteristics Before matching After propensity score matching Research participants (n=14 311) Nonparticipants (n=37 072) Overall (n=51 383) P value Research participants (n=14 301) Nonparticipants (n=14 301) Overall (n=28 602) P value Age, y 80.0±7.9 80.0±8.4 80.0±8.3 0.736 80.0±7.9 80.1±8.5 80.1±8.2 0.344 Female sex 46.0% 46.9% 46.6% 0.075 49.7% 50.3% 50.1% 0.281 Race or ethnicity <0.001 <0.001 White 88.7% 87.7% 88.0% 89.5% 88.1% 88.8% Black 3.1% 3.9% 3.7% 2.9% 3.6% 3.2% Hispanic 4.2% 4.5% 4.4% 3.9% 4.2% 4.1% Asian 1.3% 1.2% 1.2% 1.2% 1.2% 1.2% Native American 0.3% 0.2% 0.3% 0.3% 0.2% 0.3% None of the above 2.4% 2.5% 2.4% 2.2% 2.8% 2.5% Insurance <0.001 <0.001 Medicare 90.9% 89.2% 89.6% 90.9% 89.3% 90.1% Medicaid -Abstract Truncated-