A Comparative Study of Different Methods of Handling Missing Data in Patient Reported Outcomes, Burdens and Experiences (PROBE) Score Algorithm among People with Hemophilia

Powered by the Voice of Patients.

Ibrahim Q, Iorio A, Curtis R, Nichol M, Noone D, Stonebraker J, Skinner M, Germini F, and the PROBE investigators.
A Comparative Study of Different Methods of Handling Missing Data in Patient Reported Outcomes, Burdens and Experiences (PROBE) Score Algorithm among People with Hemophilia. THSNA 2022.

Background

The patient reported outcomes, burdens and experiences (PROBE) questionnaire measures quality of life in people with hemophilia and healthy controls. A score is calculated as the average of nine core questions. There is currently no validated method for calculating the PROBE score when data are missing.

Objective

Our objectives were to identify domains within the questionnaire and compare strategies for estimating the PROBE score when:
1. ≥50% item scores within a domian were available,
2. only one item score within a domain was available,
3. ≥50% item scores were available irrespective of any domain,
4. 8 out of 9 item scores were available.

Method

Scores with ICC ≥0.5 were grouped into a domain. We created one simulated data set for each combination of 1) types of missing data: MCAR, MAR, and MNAR, 2) 10, 15, and 20% missing values, and 3) the four strategies described above. A strategy with mean of absolute errors (MAE) + SD <0.05, and calibration intercepts and slopes not systematically different from 0 and 1, respectively, was considered acceptable.

Results

Among 3217 participants four item scores had ICC ≥0.5 and were grouped into a domain, pain or difficulty in daily activities (Cronbach’s alpha of 0.8). Strategies 1 and 4 performed better (MAE ± SD: 0.02 ± 0.02 for MCAR and MAR) followed by strategies 2 and 3 (Table 1). Strategy 1 estimated PROBE score accurately for MCAR and MAR data (Table 2), and slightly under-estimated the score in case of MNAR data.

Conclusion<

We recommend estimating PROBE scores as a simple average of the available item scores if at least one item within pain/difficulty in daily activities domain is available or 8 out of 9 items are available irrespective of the domain.

View Poster: HERE