Report from the Hand Osteoarthritis Working Group at OMERACT 2018: Update on Core Instrument Set Development

Domain pain and PtGA: progress in instrument validation

From the previous SLR, 32 relevant manuscripts were selected, providing data on VAS/NRS pain and/or PtGA⁶. Since January 2014, 18 relevant manuscripts were published and could be added (S1–S50, see reference list in the Supplementary File, available with the online version of this article). Details of all included manuscripts can be found in Supplementary Table 1. Summary results of the search (Supplementary Figure 1) and psychometric features of both scales within these domains were discussed by the WG (Table 1). VAS range 0–100 mm was the most studied scale (in 26/46 studies for pain and 10/15 studies for PtGA). No study reported test-retest reliability data on the use of either scale in these domains. For pain, good construct validity of VAS was shown (S3, S24, S50), while only limited data were available for NRS (S41). Twenty-two (S1, S2, S4, S6–S13, S15–S18, S21, S22, S26, S37, S38, S42, S46) and 8 studies (S15, S25, S28, S33, S34, S41, S45, S47) showed evidence for responsiveness of VAS and NRS, respectively, and 13 (S7–S12, S17, S21, S22, S26, S37, S38, S46) and 6 studies (S14, S28, S33, S34, S41, S47) for clinical trial discrimination for VAS and NRS, respectively.

For PtGA, construct validity was not studied. Evidence to support responsiveness for VAS was available in 10 studies (S3, S6, S12, S13, S15, S18, S22, S29, S38, S40), and 3 studies for NRS (S14, S28, S45). The capacity to discriminate in clinical trials was shown for VAS PtGA in agreement with the primary outcome in 5 studies (S12, S22, S29, S38, S40), while only 1 study supported this for NRS (S28).

Strikingly, phrasing of the question accompanying VAS/NRS in both domains was very heterogeneous, and details were often not reported. For pain, substantial variety existed in which aspect(s) of pain were assessed (e.g., pain at rest or upon exertion, average or worst pain), location and joint(s) referred to (e.g., target joints, dominant hand, both hands), and time of recall (undefined or ranging from current to 2 weeks; Supplementary Table 2, available with the online version of this article). Similarly, for PtGA, time of recall was undefined in most studies (3/15 studies did specify; all 48 h; Supplementary Table 3). After presentation of these findings at OMERACT 2018, the WG proposed that clear standardized phrasing accompanying these instruments be defined for pain and PtGA. It was proposed that PtGA should assess the effect of the disease on the patient’s general well-being. Review of results of previously held focus groups was suggested to analyze what is most relevant to patients¹⁷.

Results of the validation study of the modified ICOAP were discussed at OMERACT 2018. Detailed results are presented elsewhere⁹. In short, in patients with hand OA, constant and intermittent pain largely overlapped and were not separate constructs, in contrast to the situation in knee and hip OA^7,8. The existence of separate constructs in hand OA seemed clinically plausible but might be influenced by hand OA location (finger vs thumb base) and involvement of multiple hand joints at different disease stages. It was suggested to seek more patient input, since the development of ICOAP was based on focus group discussions with patients with knee and hip OA, but not hand OA. However, previous focus groups of patients with hand OA have already identified a range of pain concepts, such as fluctuating pain and psychological consequences of pain, which are not represented in the commonly used instruments to assess hand OA¹⁷.

Based on the available evidence, it was concluded that VAS and NRS are most likely the best instruments to measure pain and PtGA. However, evidence about some essential psychometric properties is missing, in particular regarding reliability, construct validity for NRS pain/PtGA, and clinical trial discrimination for NRS PtGA.

Evaluation of other potential core instruments and research agenda

The results of comparison of MHQ with AUSCAN and FIHOA for measuring domains of pain and function were discussed in light of OMERACT Filter 2.1⁴ (Table 2^10,12,13)¹¹. While displaying similar measurement properties, important advantages of MHQ above other instruments were that it can overcome issues of copyright (AUSCAN) and outdated questions (FIHOA). The possibility to propose more than 1 instrument for a core domain, with the accompanying risk of jeopardizing standardization, was discussed.

Assessment of HAMIS performance in comparison to other mobility instruments was published previously¹⁴. Though HAMIS appeared the most useful to measure hand mobility compared to other instruments, the WG debated that responsiveness data are weak. Over a 2-year period, limited change over time was observed¹⁴, either indicating that the domain itself does not change, or that the instrument cannot detect this change.

Progress in instrument development for joint activity is published in conference abstracts^15,16. Lack of a well-accepted definition hampers instrument development for this domain. Potential instruments include ultrasound and magnetic resonance imaging to detect inflammation, palpation to reveal pain, self-reported painful joint count, soft tissue swelling, and pain while gripping. In the WG discussion, it was suggested that some instruments complement each other, and a combination may be useful. Prediction of radiological progression was proposed as an anchor to assess suitable instruments.

Following discussion of these results, an agenda was developed to guide future research (Table 3).