Abstract
Objective To describe which variables were collected by rheumatologists to monitor patients with rheumatoid arthritis (RA) during teleconsultation and identify which ones have more impact on clinician intervention.
Methods Retrospective monocentric, routine care cross-sectional study including patients with RA seen in teleconsultation between March and September 2020. Available variables assessing disease status were collected in teleconsultation files. Clinician intervention was defined by treatment escalation and/or the need for a rapid face-to-face consultation or day hospitalization.
Results One hundred forty-three patients with RA were included (116 females, mean age of 58 [SD 16] yrs, mean disease duration of 14 [SD 11] yrs). The presence or absence of patient self-reported RA flares was mentioned in all medical files, followed by the presence and/or the number of tender joints (76%), the duration of morning stiffness (66%), the number of pain-related nocturnal awakenings (66%) and the C-reactive protein (CRP) value (54%). Teleconsultation led to a clinician intervention in 22/143 patients (15%), representing 51% of patients with self-reported flares (22/43 patients). Therapeutic escalation was necessary in 13 patients and/or face-to-face consultation or day hospitalization were organized for 10 patients. Multivariate analysis identified RA flares (odds ratio [OR] 15.6, 95% CI 3.37-68.28) and CRP values > 10 mg/L (OR 3.32, 95% CI % 1.12-13.27) as the variables independently associated with clinician intervention.
Conclusion Our study identified patient-reported RA flares and increased CRP values as 2 red flags in teleconsultation, independently associated with therapeutic modification and/or the need for a rapid face-to-face consultation. These indicators may help clinicians’ decision making in teleconsultation.
The sudden emergence of SARS-CoV-2 onto the world stage, along with the high morbidity and mortality associated with coronavirus disease 2019 (COVID-19) symptoms in a proportion of those infected, has accelerated a major change in the management of patients with chronic rheumatic diseases and has catalyzed the rapid emergence of telemedicine. Indeed, remote appointments largely replaced face-to-face consultations during the first wave of the pandemic.1 Although important for reducing viral spread, teleconsultations have been implemented rapidly, despite clinicians and patients having limited remote consultation experience.2 Thus, this rush to telemedicine prevented the possibility to reach a consensus in the use of existing tools to ensure standardized teleconsultation, leading to huge heterogeneities in telemedicine practices worldwide.
While facing the COVID-19 pandemic, rheumatologists stratified the risk of infection in people with rheumatoid arthritis (RA) according to their treatment regimen, age, and comorbidities3,4; many were advised not to come to hospitals and outpatient clinics because of an increased risk of infection. On the other hand, the importance of maintaining tight control of inflammation has been emphasized, both in order to prevent the risks of inadequately treated RA for the disease itself and because poorly controlled disease has been reported to be an independent risk factor for serious infections.3 The main challenge in this setting is the inability of clinicians to perform an accurate assessment of disease activity, such as tender and swollen joints; therefore, it is difficult to apply a treat-to-target strategy. Dealing with the absence of clinical examination can be challenging, as these objective variables are pillars for the calculation of composite indices, such as the Disease Activity Score in 28 joints (DAS28). Incorrect disease assessment may have potential harmful consequences, including ignoring disease flares or, conversely, misguidedly suspecting disease progression in patients in remission or with low disease activity but in whom certain subjective symptoms persist, such as pain, fatigue, or loss of physical function.
In total, 2 observational studies and 2 randomized controlled trials have previously assessed the effectiveness of telemedicine on disease activity outcome among patients with established RA.5-8 Telemedicine was found as noninferior to face-to-face visits in terms of disease activity and function. No study reported worse outcomes among patients who received telemedicine. It is important to note that these studies were all conducted in the pre–COVID-19 era by physicians trained in telemedicine, in select patient populations.9 These conditions markedly differ from the situation during the pandemic, in which physicians were not experienced or trained in telemedicine because of the severity of the health crisis. Thus, it is important to gather data reflecting this situation, with data obtained in clinical practice and from real-life conditions of teleconsultations.
Thus, our objective was to describe which variables were most frequently used by rheumatologists to monitor RA in teleconsultations during the first wave of the pandemic and identify the ones that more frequently led to a change in disease management.
METHODS
Study design. A retrospective, monocentric, routine care cross-sectional study was conducted between March and September 2020 in the Department of Rheumatology of Hôpital Cochin.
Study population. We included all patients aged ≥ 18 years with rheumatologist-diagnosed RA, and who were seen in teleconsultation by phone or video. These patients had a scheduled face-to-face appointment that turned into a teleconsultation because of restrictions placed during the first wave of the pandemic.
All included patients agreed to participate in the study after written informed consent, which was recorded in the medical source file. The protocol and the informed consent document received Institutional Review Board/Independent Ethics Committee approval before initiation of the study (“Comité de Protection des Personnes” Ouest VI, n°202-A02933-36).
Intervention. Teleconsultations were performed by telephone or video by 8 different physicians (Supplementary Table S1, available from authors upon request). These 2 modalities were fully available from the onset of the pandemic. The choice of modality was left at the discretion of the physician.
Data collection. Data were obtained after the review of the electronic medical record (EMR). In total, 8 different physicians were involved in this study and entered teleconsultation data into the EMR. For all teleconsultations, we collected demographic data, disease characteristics (ie, disease duration, antibody status, presence of erosive disease), ongoing RA therapy, and all available variables assessing disease status. No consensus among clinicians was decided on in advance regarding the nature of data to capture during the switch to telemedicine. We retrospectively collected information on the following items of interest: patient self-reported RA flares, defined as worsening of RA accompanied by at least 1 swollen and tender joint, as perceived by the patient10; tender joints (ie, presence or number of patient-reported tender joints); swollen joints (ie, presence or number of patient-reported tender joints); visual analog scale (VAS) for pain; VAS for fatigue; patient global VAS; pain-related night awakenings; morning stiffness; erythrocyte sedimentation rate and C-reactive protein (CRP) values; the DAS28 calculated by replacing provider-reported swollen and tender joint counts with patient-reported swollen and tender joint counts11,12; and treatment tolerability.
Outcomes. The first endpoint of the study was to describe which variables assessing RA were recorded during the teleconsultation. The second endpoint was to identify the variables that more frequently led to a change in disease management. To this end, we assessed which variables were associated with clinician intervention. We defined clinician intervention by treatment escalation (ie, introduction or increase in corticosteroids and introduction or increase in conventional synthetic or targeted disease-modifying antirheumatic drugs [DMARDs]) and/or the need for a rapid face-to-face consultation or day hospitalization to assess disease activity.
Statistical analysis. All data were expressed as mean (SD) values, unless stated otherwise. Statistical analysis was performed using MedCalc (version 18.9.1; MedCalc Software Ltd). The chi-square test was used to identify differences in frequency. Multivariate analyses by logistic regression were also performed to determine the factors independently associated with clinician intervention. This analysis included the clinician intervention in teleconsultation as the dependent variable. All relevant identified covariates with P < 0.10 in the single-variable analysis were then entered into each model in a single step. Odds ratios (ORs) and 95% CIs were then calculated. In this model, P < 0.05 was considered statistically significant.
RESULTS
Study population. A total of 143 patients (females: n = 116, 81%) with established RA were included; the patients had a mean age of 58 (SD 16) years and a mean disease duration of 14 (SD 11) years. Out of 139 patients, positive rheumatoid factor was reported in 100 (72%) patients and anticyclic citrullinated peptide (anti-CCP) antibodies were reported in 104 (75%) patients. Erosions were present in 75 out of 140 (54%) patients. Detailed characteristics of our study sample are provided in Table 1. Out of 143 patients, teleconsultation was performed by telephone for 106 (74%) patients and by video for 37 (26%) patients. Disease characteristics were similar between patients evaluated by telephone or video consultation (Table 1), except for a trend of higher teleconsultations performed by telephone vs video in patients with a low socioeconomic status (SES; 25% vs 13%, P = 0.11).
Patient characteristics.
Individual data collected during the teleconsultation. The different data collected during the teleconsultation are presented in Table 2. The most frequently reported items (> 50%) were the presence or absence of patient self-reported RA flares since the last visit (n = 143, 100%), the presence and/or the number of tender joints (n = 109, 76%), the duration of morning stiffness (n = 95, 66%), the number of pain-related nocturnal awakenings (n = 95, 66%), and the CRP value (n = 77, 54%). Treatment tolerability was reported for 63 patients (44%). Several differences were observed regarding age, sex, and SES. The presence and/or number of pain-related nocturnal awakenings were less frequently reported in patients aged > 65 years (55%) vs those ≤ 65 years (73%; P = 0.03) and in males (48%) vs females (71%; P = 0.03). The CRP value (39% vs 59%, P = 0.06), patient global assessment (29% vs 56%, P = 0.01), and DAS28 (11% vs 31%, P = 0.04) were less frequently reported in patients with low SES compared to those with higher SES (Table 2).
Frequency of individual data assessing RA status collected during the teleconsultation.
Value of patient-reported RA flares in teleconsultation. Patient self-reported RA flares concerned 43 out of 143 patients (30%). The presence of self-reported RA flares was associated with a more detailed evaluation of patients in teleconsultation. The presence and/or number of tender joints was reported more frequently, to a significant extent, in patients who reported a flare (39/43, 91%) compared to those who did not (70/100, 70%; P = 0.008). The presence and/or number of swollen joints was reported more frequently, to a significant extent, in patients who reported a flare (25/43, 58%) compared to those who did not (23/100, 23%; P < 0.001). A trend for higher reporting of the number of nocturnal awakenings in patients with a flare (33/43, 77%) compared to those without (62/100, 62%) was also observed (P = 0.08). In addition, presence (18/22, 82%) vs absence (25/121, 21%) of a flare was associated with a clinician intervention during the teleconsultation (P < 0.001). Of note, only a single patient experienced COVID-19 and temporarily interrupted their methotrexate (MTX) treatment, without the occurrence of a disease flare. No patient permanently interrupted their treatment because of side effects, COVID-19, or persistent remission.
Individual data associated with clinician intervention. Among all 143 teleconsultations, a clinician intervention was necessary for 22 patients (15%; Table 3). Therapeutic escalation was proposed to 13 patients—introduction or dose increase of corticosteroids for 8 patients, introduction or dose increase of MTX for 4 patients, and introduction of hydroxychloroquine for 1 patient—and face-to-face consultation or day hospitalization for early disease assessment was proposed to 10 patients.
Characteristics of patients with rheumatoid arthritis according to their need or lack of need for a clinician intervention.
Patients who requested clinical intervention, as compared to those who did not, had a shorter disease duration (mean 10 [SD 10] yrs vs mean 15 [SD 11] yrs; P = 0.049), a lower frequency of erosions (19% vs 60%; P < 0.001), more active disease (higher frequency or flares and CRP > 10 mg/L), and a higher likelihood of corticosteroid therapy (73% vs 42%; P = 0.007; Table 3).
The following variables were associated with clinician intervention during the teleconsultation as determined by univariate analysis (Table 4): patient self-reported RA flares since the last visit (P < 0.001), CRP > 10 mg/mL (P = 0.003), and morning stiffness lasting more than 30 minutes (P < 0.001). After multivariate analysis by logistic regression, RA flares (OR 15.6, 95% CI 3.37-68.28) and CRP values > 10 mg/L (OR 3.32, 95% CI 1.12-13.27) were the only variables independently associated with clinician intervention (Table 4).
Logistic regression analysis including clinician intervention as the dependent variable.
Outcome of patients requesting a face-to-face consultation. A face-to-face consultation or day hospitalization for early disease assessment was proposed to 10 patients who experienced self-reported RA flares; 4 of these patients also had CRP levels > 10 mg/L. Active disease was confirmed during this next face-to-face visit in 9 patients, with the DAS28 ranging from 3.35 to 5.62, leading to therapeutic modification (Table 5).
Disease activity and therapeutic modification of the 10 patients seen during the face-to-face visit or the day hospitalization.
The other 133 patients were seen in face-to-face consultations an average of 6 (SD 2) months after the teleconsultation. No DMARD modification was recorded during this next face-to-face consultation, particularly with the 22 patients who reported a flare that did not lead to a clinician intervention during the previous teleconsultation.
DISCUSSION
The COVID-19 pandemic had important consequences on decisions for the management of people with inflammatory chronic rheumatic disorders. A recent survey among European Alliance of Associations for Rheumatology (formerly European League Against Rheumatism) countries showed that measures related to containment of the COVID-19 pandemic led to a perceived delay between symptom onset and a first rheumatological visit, postponement of treatment decisions, and shortage of hydroxychloroquine and tocilizumab, thereby negatively impacting early treatment and treat-to-target strategies.13 Another study performed in Latin America revealed that patients with inflammatory rheumatic diseases were negatively affected by the COVID-19 pandemic, characterized by an increase in self-rated disease activity, a reduction in medication adherence, and hurdles to medical follow-up.14 To ensure continuous care of these patients, teleconsultation may represent a valid alternative to in-person visits during the pandemic. During the early days of the pandemic, DMARD interruptions were associated with an absence of telemedicine availability.15 The issue is that efficient teleconsultations must be conducted in order to continue to achieve optimum disease control. However, the abrupt transition from in-person visits to telemedicine during the first wave of the pandemic did not permit clinicians to adequately work on relevant variables to measure disease activity in teleconsultation.
A critical issue of teleconsultation is the reliability of clinician decisions and clinician intervention in identifying patients who require early face-to-face reviews and/or therapeutic adjustments because of insufficient disease control. Importantly, the decision made during the teleconsultation was mostly confirmed during the next face-to-face visit for 9 of the 10 patients with early disease assessment and for all 133 patients seen an average of 6 (SD 2) months later; this highlighted the accuracy of the teleconsultation-driven clinical intervention. The phase of the disease may be an important consideration in determining the appropriateness of telemedicine. Indeed, it was reported that clinicians prefer teleconsultations for patients with established RA, and our study population had long-standing disease, which may have increased the accuracy of the teleconsultation-driven clinical intervention. In addition, previous studies from the pre–COVID-19 era have shown that teleconsultation was as effective as face-to-face visits for patients with established RA in terms of disease activity.6,7
This review of teleconsultation files from 143 patients with RA allowed the identification of 2 red flags that mainly and independently drove clinical intervention: patient-reported RA flares and increased CRP levels. Patient self-reported flares, defined by worsening of RA along with at least 1 swollen and tender joint, were a major driver for pursuing an intervention on the part of the clinician in teleconsultation. This was consistent with a previous study showing that self-reported flares were substantiated by higher disease activity measures, independently associated with pain and swollen joints, and related to treatment escalation.10
CRP levels were identified as a second red flag; this, in combination with patient self-reported flares, resulted in an objective measurement of systemic inflammation of RA. However, CRP information was lacking in 46% of the medical files, emphasizing the need to remind physicians and patients of the importance of blood tests, particularly CRP, for monitoring RA. Patients may also have missed the opportunity to have laboratory tests done given the restricted access to labs during the pandemic, in agreement with a previous study.16
Teleconsultation does not make the clinical evaluation of swollen joints possible, which may explain why information about this variable was barely collected. One option would be to use patient-reported swollen joints, as suggested by a recent study that reported a good concordance between the patient and the clinician for the number of tender and swollen joints, especially in the case of low disease activity.17 The American College of Rheumatology (ACR) recommended that RA disease activity measures be adapted for use in telehealth settings to support high-quality clinical care. In particular, measures requiring formal joint counts can be calculated using patient-reported swollen and tender joint counts.12 This procedure was applied in this study to calculate the DAS28. Nurse-led programs of patient self-assessment of joint counts and disease activity have previously shown short-term benefits, and it may be helpful to include tender and swollen joint counts in teleconsultation.18,19
Since the evaluation of tender and swollen joints is challenging in teleconsultation, the use of patient-reported outcomes (PROs) could be one method of deciding which patients would be suited to a teleconsultation or a face-to-face consultation, while ensuring that disease activity is controlled and symptoms important to the patient are not missed. The Rheumatoid Arthritis Impact of Disease (RAID) score, which comprises 7 domains encompassing pain, fatigue, physical function, sleep, physical well-being, emotional well-being, and coping, may be an interesting candidate given its strong correlation with the DAS28.20,21
As this was a retrospective study, we were able to collect only variables used by physicians in their daily practice, explaining why many instruments were not administered, including most of the PRO measures.
Therapeutic modification was proposed in teleconsultation to 13 out of 143 (9%) patients with RA; this is close to what was previously reported in a cohort of 112 patients with chronic rheumatic disorders seen in teleconsultation, with 17% of patients experiencing treatment modification.22 Corticosteroids and MTX were the preferentially adapted drugs used to control disease activity in teleconsultation, highlighting their flexibility and clinicians’ confidence in their use, even at a distance from the patient. Interestingly, clinicians were reluctant to introduce or change targeted therapies during teleconsultation. These modifications were preferentially done during the face-to-face consultation or day hospitalization after a careful evaluation of disease activity by clinical examination, laboratory tests, and power Doppler ultrasounds. These modifications might also be influenced by the preferences of patients, who are more comfortable adapting the dose of an ongoing therapy than introducing a new one, particularly at the time of a health crisis.
Beyond technical issues, a critical challenge of telemedicine in terms of enduring after the pandemic and becoming a routine consultation modality is patient acceptance and satisfaction. Several reports suggest the successful use of telemedicine services in the evaluation and management of rheumatic diseases in the current pandemic situation. Time and cost savings were observed as beneficial factors for patients, with more than three-quarters of all patients with RA ready to use teleconsultation in the near future.23 Another study showed that it was possible to transfer rheumatological care activities to teleconsultation with a considerable degree of satisfaction from both the patient and the clinician.24 This is significant given the potential greater importance of teleconsulting in rheumatology in the future because of the lower number of physicians available and patients’ unwillingness to travel for a consultation.
This study has some limitations mainly inherent to its retrospective design. Despite the suggestion that PRO measures entered into the EMR may be highly relevant in teleconsultations,12 PROs are still barely used in daily practice in our department during teleconsultation. Given their importance, we aim to implement their use in the near future. To that end, we are now assessing their value for teleconsultation in a dedicated ongoing prospective study, with a specific focus on the RAID. Patient and clinician satisfaction have also not been evaluated. We did not use any validated questionnaire to collect the number of tender and swollen joints, but we modified existing measures, as recommended by the ACR.12 Although teleconsultation has the potential to expand the reach of rheumatology practice, some patients still lack the most basic resources required for this type of visit. This point was not analyzed in our study and will need to be taken into consideration in future work.
In summary, our study showed the reliability of clinician intervention in teleconsultation regarding identification of patients requiring face-to-face reviews and therapeutic adjustments. In addition, our study identified patient-reported RA flares and increased CRP values as 2 red flags, which were independently associated with therapeutic modification and/or the need for a rapid face-to-face consultation. These indicators may help clinicians’ decision making in teleconsultation and need to be confirmed in independent cohorts.
ACKNOWLEDGMENT
We acknowledge Ms. Carole Desbas for her expert secretarial assistance.
Footnotes
This work has received a research grant from Fresenius Kabi.
JA reports honoraria from Galapagos, Lilly, Pfizer, AbbVie, BMS, Sanofi, Roche-Chugai, Nordic Pharma, Medac, Novartis, Biogen, Fresenius Kabi, Janssen, and MSD, and research grants from BMS, Pfizer (Passerelle), Novartis (Dreamer), and Fresenius Kabi. AM reports honoraria from AbbVie, Biogen, BMS, Gilead, Janssen, MSD, Novartis, Lilly, Pfizer, and UCB, and research grants from Pfizer and UCB.
- Accepted for publication June 1, 2022.
- Copyright © 2022 by the Journal of Rheumatology
This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.
