Test-Retest Reliability Analysis of the Patient Reported Outcomes Burdens and Experiences (PROBE) Study

Powered by the Voice of Patients.

Curtis R, Wu J, Iorio A, Nichol M, Germini F, Kucher A, Skinner M. Test-Retest Reliability Analysis of the Patient Reported Outcomes Burdens and Experiences (PROBE) Study. (2023), POSTER ABSTRACTS (PO077). Haemophilia, 29: 24-202. https://doi.org/10.1111/hae.14715

Objective

To investigate the test-retest reliability of the PROBE questionnaire mobile application (MyPROBE).

Methods

People with hemophilia (PWH) including carriers and individuals without a bleeding disorder who attended hemophilia related workshops were invited to participate in this study. Additional participants were recruited through social media. Participants completed PROBE anonymously 3 times: twice on the MyPROBE app (Time 1 and Time 2, ~24 hours apart) and once on the web portal (Time 3, ~14 days after T1). Test-retest reliability was measured calculating the Cohen’s Kappa coefficient for categorical variables, and the correlation coefficient for continuous variables. 

Results

Forty-eight participants were enrolled with a median age (range) of 56 (27-78) years. Of these, 13 (27.1%) were PWH, 12 (25.0%) were carriers of hemophilia A or B and 23 (47.9%) were individuals without a bleeding disorder. On general health domain questions, Kappa coefficients ranged from 0.72 to 1.00, indicating substantial to almost perfect agreement using Cohen’s Kappa for all items (T1 vs T2). T2 vs T3 values ranged from 0.64 to 0.97 with only the acute pain-related questions scoring less than almost perfect agreement. For hemophilia-related domain questions (T1 vs T2), Kappa coefficients ranged from 0.49 to 1.00. Of these, 5 of 8 items were in almost perfect agreement. Values for T2 vs T3 ranged from 0.34 to 1.00, with the time-based bleeding question showing the only coefficient that scored below substantial agreement. For overall health-related quality of life, EQ-5D-5L index scores had a paired mean difference of -0.01 for T1 to T2 and 0.01 for T2 to T3 indicating a near perfect correlation. The correlation coefficients for these two time points were 0.89 and 0.83 respectively. Reliability of the MyPROBE app showed substantial to almost perfect agreement with the web version (T2 to T3). Correlation coefficient of the EQ-visual analog scale (EQ-VAS) for T1 to T2 was 0.90 (0.83 – 0.94) and 0.71 (0.53 – 0.83) for T2 to T3.

Conclusions

The test-retest exercise showed substantial to almost perfect agreement in a majority of questions for the app vs app, and a high correlation for the web vs app. The results suggest the MyPROBE app is a reliable tool to assess patient reported outcomes for PWH and control populations independently of the platform used for its completion.

View Poster: HERE