Background/Purpose:

Patient reported outcomes (PRO) represent an important part of outcomes assessment in rheumatoid arthritis (RA) patients. The Patient Global Assessment (PGA) is an essential PRO; pain and fatigue can also be assessed by visual analogue scales. Since RA disease activity fluctuates, it is unclear how estimates of the smallest detectable difference (SDD) are influenced by the time elapsed between assessments, and if SDD differs in relation to disease activity. Thus we investigated if reliability estimates of PROs change when using different test-retest intervals, and if they can be better specified based on patients’ disease activity.

Methods:

Forty RA patients were asked to report their PGA, pain and fatigue daily over 28 days in this prospective observational study. We calculated variability of PROs comparing the initial measurement to ones performed 1-27 days thereafter. We calculated reliability using the intra-class correlation coefficient (ICC), which determines the amount of measurement variability attributable to true scores. The SDD was calculated as the standard deviation of the difference between two measurements multiplied by 1.95. We finally divided patients into tertiles by baseline simplified disease activity index (SDAI) and investigated the differences in SDD based on starting points.

Results:

Data of forty patients (85% female, 60% rheumatoid factor positive, mean SDAI: 12.8±8.4, mean disease duration 11.3±8.3yrs) was used. Mean ICC between day 1 and each consecutive day was 0.63±0.13 for PGA, 0.56±0.14 for pain and 0.48±0.15 for fatigue. The mean SDDs of PGA, pain and fatigue were 26.5±5.7, 27.2±6.4 and 42.1±5.7mm, respectively (Figure). As expected, higher reliability (ICC) coincides with a smaller SDD, but no temporal trend was found for any PRO.

Ranges for the tertiles by baseline SDAI were 3.2-8.1 (mean=5.3), 8.3-16.8 (mean=12.5), and 17.4-37.7 (mean=23.9). For PGA, higher SDAI coincided with higher SDD (p<0.001), resulting in (mean, 95% confidence intervals) 18.2 (14.9-21.4), 27.14 (23.4-30.9), and 30.1 (26.1-34.1) mm, respectively. The same trend could be found for SDDs of pain (p=0.002): 22.5 (19.7-25.4), 28.1 (24.3-32.1), 29.9 (26.1-33.7). For SDDs of fatigue the trend was inverse, so that patients with higher SDAI had lower SDDs (p<0.001): 45.6 (39.5-51.7), 39.7 (34.8-44.5), 18.7 (16.6-20.8).

Figure: smallest detectable differences (SDD) between day 1 and consecutive 27 days. Blue line: development of SDD for pain; the red line: for patient global assessment; the green line: for fatigue.

Conclusion:

We demonstrated here, that reliability and SDDs of neither PGA, pain nor fatigue are dependent on the time between assessments. PGA showed the best reliability of all tested PROs. Cut-offs to identify true changes seem to be dependent on disease activity. Much lower SDDs of pain and PGA can be applied in patients with low disease activity.