AB0145 SMARTPHONE- AND SMARTWATCH-ACQUIRED DAILY STEPS, ACTIVITY, AND BAROMETRIC PRESSURES ASSOCIATED WITH SUBJECTIVE MEASURES OF RHEUMATOID ARTHRITIS: A PROSPECTIVE STUDY FOR RA DIGITAL PHENOTYPING

Background: Symptoms in patients with rheumatoid arthritis (RA) are potentially influenced by exercise load and meteorological change, and often vary from day to day, especially in unstable condition of RA. Patients with RA not infrequently experience worsening of joint symptoms when the load on the joint, such as walking and doing housework, exceeds a moderate range. However, the worsening of joint symptoms is often not observed in the midst of the loading of the joint, but often becomes apparent after a few hours or days. Objectives: To elucidate the relationship between smartphone- and smartwatch-acquired daily objective data (barometric pressures, steps, and activity) and daily subjective patient reported outcomes of RA. Methods: A smartphone (iPhone 8) and a wristband-type smartwatch (Fitbit Versa 2) were lent to each patient for free. A mobile app was developed and installed into the smartphones to collect patients’ daily subjective RA symptoms including Pt-P-VAS (patient-pain-visual analogue scale), Pt-G-VAS (pt-general-VAS), PtTJCount(68)(patient self-determined tender joint count among 68 joints), PtTJCount(28), PtSJCount(66)(patient self-determined swollen joint count among 66 joints), PtSJCount(28). Also, the smartwatch data and physicians’ assessment were collected from the same subject. Physicians’ and patients’ assessment of TJC, SJC, and G-VAS was independently performed without seeing each other’s assessment. We conducted a simple linear regression analysis with outcome variables of Pt-P-VAS, Pt-G-VAS, PtTJCount(68), PtTJCount(28), PtSJCount(66), and PtSJCount(28). The independent variables included smartphone-acquired daily steps and barometric pressure of the reported day and the previous day, and smartwatch-acquired minutes of “lightly active (1-3 METs equivalent)”, “fairly active(3-6 METs equivalent)”, and “very active(>6 METs equivalent)” of the reported day and previous day. We defined low barometric pressure as below 1000 hPa. The level of activity was measured by the smartwatch. Patients were blinded to daily barometric pressure data and their daily active time when the patients answered daily symptom questions on the smartphones. Results: A total of five patients were enrolled. At baseline, mean (± standard deviation (SD)) age was 50.8±14.8 years; all patients were females; mean disease duration was 6.6±4.9 years; mean SDAI was 18.6±25.5; mean DAS28-CRP was 3.23±1.85; mean morning stiffness was 134±116 min; mean HAQ-DI was 0.7±0.9. Mean observation period was 77.8 days. Because of the missing data, the sample size (N) for the regression analysis varies with the outcomes: Pt-P-VAS and Pt-G-VAS are 250 while PtTJCount and PtSJCount are 260. The table 1 showed that the patients’ assessment of TJC, SCJ, and G-VAS was correlated well with the physicians’ assessment. The figure 1 showed the change associated with one SD increment in each independent variable with 90% confidence intervals. Low barometric pressure was associated with bad health conditions (high Pt-G-VAS, Pt-P-VAS, and SJCount). Moreover, longer very active time in the previous day (“veryactive_1” in the Figure 1) was associated with bad health condition (high SJCount). Many steps were associated with good health conditions (low Pt-G-VAS, Pt-P-VAS, and SJCount). Conclusion: High barometric pressure was associated with good health conditions, and longer very active time in the previous day was associated with bad health condition. Barometric pressure data and physical activity data acquired by mobile digital devices may predict the change in RA symptoms. Further investigation in larger patient numbers is warranted. Acknowledgements: The authors would like to thank Harumi Kondo for her assistance. Disclosure of Interests: Keisuke Izumi Speakers bureau: Abbvie, Asahi Kasei Pharma, Bristol Myers Squibb, Chugai Pharmaceutical, Eli-Lily, Mochida Pharmaceutical, Ono Pharmaceutical, Grant/research support from: Abbvie, Asahi Kasei Pharma, Daisuke Moriwaki Employee of: CyberAgent, Inc., Takamichi Toda Employee of: AI Shift, Inc., Misako Higashida-Konishi: None declared, Manami Koyama: None declared, Hisaji Oshima: None declared, yutaka okano Speakers bureau: Asahi Kasei Pharma, Yuko Kaneko Speakers bureau: AbbVie, Astellas, Ayumi, Bristol–Myers Squibb, Chugai, Eisai, Eli Lilly, Hisamitsu, Jansen, Kissei, Kirin, Novartis, Pfizer, Sanofi, Takeda, Taisho, Tanabe-Mitsubishi, and UCB, Shigeru Ko: None declared, Tsutomu Takeuchi Speakers bureau: Abbott Japan Co, Ltd, Bristol–Myers KK, Chugai Pharmaceutical Co, Ltd, Eisai Co, Ltd, Janssen Pharmaceutical KK, Mitsubishi Tanabe Pharma Co, Pfizer Japan Inc, Takeda Pharmaceutical Co, Ltd, Astellas Pharma and Daiichi Sankyo Co, Ltd., Consultant of: Astra Zeneca KK, Eli Lilly Japan KK, Novartis Pharma KK, Mitsubishi Tanabe Pharma Co, Asahi Kasei Medical KK, Abbvie GK and Daiichi Sankyo Co, Ltd., Grant/research support from: Abbott Japan Co, Ltd, Astellas Pharma, Bristol-Myers KK, Chugai Pharmaceutical Co, Ltd, Daiichi Sankyo Co, Ltd, Eisai Co, Ltd, Janssen Pharmaceutical KK, Mitsubishi Tanabe Pharma Co, Pfizer Japan Inc, Sanofi–Aventis KK, Santen Pharmaceutical Co, Ltd, Takeda Pharmaceutical Co, Ltd, Teijin Pharma Ltd, Abbvie GK, Asahikasei Pharma Corp and Taisho Toyama Pharmaceutical Co, Ltd.