Telemedicine application in patient with Diabetes, Hypertension and Rheumatoid arthritis: A systematic review and meta-analysis

Background: Under the global epidemic condition (cid:0) the telemedicine was widely used in the world, especially in the period of long-term care and treatment. The aim of study is to verify the effectiveness of telemedicine in the management of chronic disease by using scientic methods such as hypertension, diabetes, rheumatoid arthritis by using Meta-analysis and systematic review methods. The purpose of this study was to systematic review of the effect of telemedicine in chronic disease, so as to provide inspiration for chronic disease management in the future. Methods: Article searching were performed using Web of Science, PubMed, MEDLINE, EMBASE and other library or database to retrieve articles which published from database and library establishment to October 31 st , 2020. Literature quality assessment, systematic review and Meta-analysis were then performed. Results: 12 articles were included in the literature quality assessment (7 diabetes,3 hypertensions,2 rheumatoid arthritis), the article included in this study is of high quality. There are 7 articles included in the Meta-analysis, the result shows there have effect in glycosylated hemoglobin after 12 months of intervention (95%CI=-1.53, -0.16; Z=2.42; P=0.02), compare with 6 months (95%CI=-1.32, -0.01; Z=1.99; P=0.05). It also showed that there was no signicant difference in fasting blood glucose after 6 months of intervention (95%CI=-1.19,0.21; Z=1.37; P=0.17). Both systolic blood pressure (95%CI=-12.79, -3.69; Z=3.55; P= 0.0004) and diastolic blood pressure (95%CI=-9.90, -0.43; Z=2.14; P= 0.03) showed statistically signicant. Moreover, we also found positive inuence of telemedicine about good behaviors and rehabilitation for rheumatoid arthritis patients. Conclusion: The results showed that telemedicine had a positive effect on the management of diabetes, hypertension and rheumatoid arthritis, especially on the management of glycosylated hemoglobin and blood pressure. technologies great promotion to access the services and improve the quality of care, especially for people with chronic diseases. effects model due to its high statistical heterogeneity (I 2 =78%). The results showed that there was no signicant difference in fasting blood glucose between the experimental group and the control group (MD=-0.49;95%CI=-1.19,0.21; Z=1.37;

2030, (14). Moreover, the incidence of rheumatoid arthritis in adults worldwide is 0.5%, it has become one of the top 10 chronic diseases in China, and the incidence of rheumatoid arthritis in adults ranged from 0.5% to 1.0% in the United States (4,10).They are common chronic diseases in the population.
In recent years, some studies on the intervention of chronic disease management through telemedicine have been carried out gradually, but there is no consensus about the effect of telemedicine which used in chronic disease management. the aim of this study focused on the effectiveness of the telemedicine intervention by systematic evaluation and meta-analysis.

Selection of studies
Inclusion and exclusion criteria of the literature are as follows. Inclusion criteria for participants:(1) used randomized-controlled trial or quasiexperimental designs (2) telemedicine was the main intervention in this investigation. (3) the disease type must be chronic. (4) the articles published in English and Chinese. Exclusion criteria for participants: participants cannot use related remote monitoring software tools.

Search strategy
The literature search was performed in the Cochrane, CINAHL, EBSCO, Medline, PubMed, EMBASE, Web of Science, JBI, NICE, Sinomed, CNKI (Chinese database), Weipu (Chinese database) and Wanfang (Chinese database) database. The articles searching time interval from database and library establishment to October 31 st , 2020.
English searching terms used in the literature retrieval include: ("Telemedicine" OR "Remote Consultation" OR " Telehealth" OR "Telemonitoring " OR " Web-based" OR " mobile monitoring" OR "mobile health" OR" internet-based" OR " promoting monitoring "OR" mHealth" OR "Telecare")AND("Chronic Disease" OR "Arthritis" OR "Rheumatic Diseases" OR "Hypertension" OR "Digestive System Disease" OR "Coronary Disease" OR "Pulmonary Disease, Chronic Obstructive" OR "Dyslipidemias" OR "Kidney Diseases" OR "Diabetes Mellitus" OR "Asthma" OR "Liver Diseases" OR "Stroke" OR "Cognition Disorders" OR "Mental Disorders" OR "Neoplasms")AND("Disease management" OR "Management"). Corresponding Chinese searching terms were used during the literature search in Chinese databases.

Data extraction and critical appraisal
Inclusion criteria and exclusion criteria were piloted by 2 researchers and disagreements were resolved by discussion with all authors. The two researchers screened the literature independently according to the inclusion criteria and exclusion criteria, and made a preliminary screening according to the title, keywords and abstract of the literature.

Literature quality assessment
The quality of each RCT was evaluated by the risk-of-bias assessment tool for RCT recommended by the Cochrane Handbook for Systematic Reviews of Interventions 5.0.2 (15). Two researchers made the evaluations independently. A third researcher was consulted if there was any discrepancy. The quality of each QRCT was evaluated by the QRCT-assessment tool from the Joanna Briggs Institute, and the assessment was independently assessed by two researchers and. If they have inconsistent opinions after evaluating the quality of the literature, a third researcher will be invited to evaluate the quality of the literature again.

Data retrieval
The literature data were extracted into a detailed table, which included the following items: (1) Author, country, year of publication.

Data synthesis and analysis
Meta-analysis was performed using Revman 5.3 software according to inclusion and exclusion criteria for each study. We conducted a metaanalysis of three or more articles describing the same disease, which had the same outcome indicators and similar intervention duration. The nal measured outcome measures are all continuous variables. We also calculated 95% con dence interval (CI). A xed-effect model was applied if the heterogeneity from multiple studies was small (consistency test had a p > 0.05 with I2 ≤ 50%). Otherwise, a random-effects model was adopted if there was high heterogeneity (p < 0.05, I2 ≥ 50%).

Study selection and characteristics
The gure below shows the chart of the article selection process. The search resulted in 3585 articles, 2756 articles did not address after the title and abstract were assessed. 501 articles full text were read. 489 articles were excluded in the nal stage assessment. 12 articles were included in the study. ( Figure.2 Literature screening ow chart)

Results of literature retrieval
The quality evaluation of 12 articles included in the literature(16-27) ( Table 1 Characteristics of included studies). It was published between 2006 and 2020. There were seven references included in the nal meta-analysis, four articles focused on diabetes (16, 19,20,23), and three articles studies hypertensions(25-27)

Risk of bias assessment
Baseline characteristics were reported in all the articles included in the literature quality assessment, there were eight literatures (17, 18, 20, 22-24, 26, 27) with low bias risk by random method, eight articles (17, 18, 20, 22-24, 26, 27) have achieved the generation of standard random sequence and proper allocation concealment, only one article of allocation concealment is classi ed as having a high risk of bias (25). Four of them used blinded methods for subjects, trainers, and evaluators (16, 17, 21, 24), the two articles clearly mentioned that the study subjects, interveners and outcome measurers were not blinded due to intervention methods (22,26). Most of the literature reports the missing data and the reasons except for one article(19) ( Figure.3 Chart of quality risk assessment of related literature) All of the literature included in this study is considered high quality after assessment.

Outcome measures
There are 4 articles(16, 19, 20, 23)used glycosylated hemoglobin and fasting blood glucose as the primary outcome. High fasting blood glucose concentration (FBG) has been considered as one of the potential risk factors for small arterial stiffness, it is tend to cause diabetic complications(28). glycosylated hemoglobin, an indirect measure of mean blood glucose index, it could re ect the level of blood glucose over the previous 2-3 months (29). Meta-analysis of glycosylated hemoglobin at 6 months(19, 20, 23) and 12 months(16, 20, 23) was performed in the four literatures according to the different intervention duration.
The type of disease studied in the other three literatures is hypertension (25)(26)(27), and telemedicine is the main intervention method. The uctuation of blood pressure become a major risk factor for stroke and heart disease in the population (30), all of these three articles used systolic blood pressure and diastolic blood pressure as primary outcome, so we conducted a meta-analysis to test the effect of intervention.

Meta-analysis of diabetes
Jun Yang Lee(20), Han Yun(23)and Feng ya kun et al (19)reported the experimental data of the intervention for 6 months in the literature, we found no improvement in glycosylated hemoglobin and fasting blood glucose of three RCTs. However, according to the other studies reported that the signi cant difference in glycosylated hemoglobin after intervention of 12 month(16, 20, 23). Meta-analysis was conducted at two time points of 6 months respectively according to the duration of intervention.
Meta-analysis of glycosylated hemoglobin after 6 months of intervention Because of the high statistical heterogeneity (I 2 =94%), the random effect model was used. The results showed that there was no signi cant difference in glycosylated hemoglobin between the intervention group and the control group after 6 months of intervention. Meta-analysis of glycosylated hemoglobin after 12 months of intervention A random-effect model was adopted since there was high statistical heterogeneity (I 2 = 99%). However, analysis of three other 12-month intervention duration studies(16, 20, 23) showed statistically signi cant differences in glycosylated hemoglobin between the intervention and control groups (MD=-0.84;95%CI=-1.53, -0.16; Z=2.42; P=0.02)( Figure. 6 Forest plots of glycosylated hemoglobin after 12 months of intervention between the intervention group and the control group)

Meta-analysis of hypertension
This meta-analysis included three articles studied on hypertension (25)(26)(27). The results showed that under the intervention of telemedicine measures, the difference between the experimental group and the control group in systolic blood pressure and diastolic blood pressure was statistically signi cant after 6 months of intervention, and the difference in systolic blood pressure was remarkable.
Meta-analysis of systolic blood pressure after 6 months of intervention The systolic blood pressure in the experimental group and the control group after 6 months of intervention in the included literatures were analyzed. A random-effects model was used because there was high statistical heterogeneity (SBP: I 2 = 91%). The results showed that there were statistically signi cant differences in systolic blood pressure between the group receiving telemedicine and the control group (SBP: MD=-8.24;95%CI=-12.79,-3.69;Z=3.55;P= 0.0004) Figure. 7 Forest plots of systolic blood pressure after 6 months of intervention between the intervention group and the control group Meta-analysis of diastolic blood pressure after 6 months of intervention Three articles (25)(26)(27) report the diastolic blood pressure at 6 months of intervention. The random effect model was used because of the high statistical heterogeneity (DBP: I2=98%). The results showed that there were statistically signi cant differences in diastolic blood pressure between the intervention group the control group (DBP: MD=-5.16;95%CI=-9.90, -0.43; Z=2.14; P= 0.03) Figure.  From the result of meta-analysis, there was no statistically signi cant difference of fasting blood glucose levels between the experimental group and the control group. Angeles 's study (31)also showed that compared with the usual group, the meta-analysis of fasting blood glucose after intervention of 3 months, 6 months and 12 months was not statistically signi cant. In a study of type 2 diabetes(20), 120 people were included in the intervention group, it was found that the difference between the intervention group and the control group was not statistically signi cant. However, Han Yun et al (23) gure out that signi cant differences in fasting blood glucose levels at the 3th ,6th and 12th months after intervention, and the fasting blood glucose levels in the intervention group were signi cantly lower than those in the control group. Perhaps, the reason for the difference may be related to the duration of diabetes and the severity of the disease in participants of the intervention group. (31).

Duration effect of telemedicine on glycosylated hemoglobin in patients with diabetes
Our result shows there was signi cant difference in glycosylated hemoglobin after intervention of 12 months, but there was no signi cant difference in glycosylated hemoglobin between the intervention group and the control group after 6 months. Therefore, we supposed that changes in glycosylated hemoglobin may have time effects.
Our results are similar to Viana et al(32) they performed a meta-analysis of glycosylated hemoglobin with telecare intervention, a total of 1,782 people were included in the study, the length of most studies was 6 months, and they also found no signi cant improvement in glycosylated hemoglobin. Moreover, we also nd the long-term duration of intervention has positive effects on the control of glycosylated hemoglobin, the consequence is similar to the meta-analysis of Lee and Timpel et al. Lee et al gure out that telemedicine system can be positive in type 1 diabetes patients after intervention (33), the telemedicine system used in most trials were relatively simple and involved data transmission of blood glucose data with feedback, it turned out that the studies lasted at least 6 months in duration or longer, and that the effect of telemedicine was more obvious. Furthermore, Timpel et al (34) pointed out the great bene t in glycosylated hemoglobin reduction after telemedicine intervention, especially the communication and interactive functions of telemedicine, they nd that digital health education was statistically signi cant in the long-term intervention duration of 12 months. Some of the original RCTs also re ects the time effects. Wake eld et al (35) conducted original research, the intervention group included home remote monitoring and nursing management over 6 months, the result shows that the difference between the intervention group and the control group was signi cant at 6 months, but if the intervention was withdrawn 6 months after the intervention, the glycosylated hemoglobin indicators of the intervention group and the control group will gradually tend to be consistent. Similar results also re ected in the study of Stelios Fountoulakis et al (17), telemonitoring which combined with a management and feedback system based on transmitted data has been proven effective in rapidly reducing glycosylated hemoglobin, the results showed signi cant reductions in glycosylated hemoglobin at both 3 months and 6 months compared with baseline, and the slightly attenuated at 6 months after its discontinuation. In the ve-year follow-up study by shea et al (21), a total of 1665 people were recruited, the telemedicine group regularly implemented home unit management, nurse case management, and remote consultation activities, follow-up examinations were conducted at 1-year intervals, nally, the intervention group had net improvement in glycosylated hemoglobin compared with usual group.
In conclusion, most studies have shown that telemedicine has a positive effect on the management of diabetes, but the best time to intervene is different, from our study we found that telemedicine has a positive effect at 12 months. Telemedicine is a process of long-term intervention and will be more effective over time, this improvement could be attributed to better compliance due to a more frequent contact by using telemedicine(17).

Blood pressure can be improved by telemedicine intervention
The results of our meta-analysis showed that there was signi cant difference in systolic blood pressure and diastolic blood pressure after intervention of 6 months. Many studies have found that blood pressure can be improved by telemedicine intervention, although the duration of the intervention in many studies was different.
Our results of the study are similar to Verberk et al(36), it was a meta-analysis included nine randomized clinical trials, their research showed that the systolic blood pressure and diastolic blood pressure of the intervention group are signi cantly decreased by using telecare, this measure not only controls blood pressure effectively, but also enables patients to adhere to treatment. Another system review gure out that home remote monitoring does a better job of controlling blood pressure than other approaches, they selected some articles that examined populations of patients with hypertension, it is worth noting that in most cases the studies found a signi cant drop in blood pressure in the rst 3 months of remote monitoring (37). However, an RCT study(38)which involved 387 participants, the interventions were instructed by trained nurses, such as blood pressure monitoring, answer questions, link device to the participant's home, both systolic blood pressure and diastolic blood pressure were found to have signi cance after 12 months of intervention.
Most articles in RCTs have con rmed the positive effect of telemedicine on blood pressure, it can help support doctors and nurse for closer, continuous follow-up of patients with hypertension, although the duration of the intervention was different. (39)(40)(41)

Integration of telemedicine studies in patients with rheumatoid arthritis
Rheumatoid arthritis is one of the most prevalent chronic in ammatory diseases. The main symptoms of rheumatoid arthritis include rheumatoid nodules, pulmonary involvement or vasculitis, and systemic comorbidities (42). Patients with rheumatoid arthritis have a longer disease duration, it can cause joint deterioration and functional disability, eventually leading to unfavorable disease outcomes and seriously affected the activity of daily life (43). Therefore, the long-term management of rheumatoid arthritis patients with telemedicine as the main intervention method is developed gradually.
The main intervention in Ferwerda 's research(18) is a cognitive behavioral methods to manage the patients with rheumatoid arthritis. The results showed the effectiveness of Internet-based tailored cognitive behavioral interventions for rheumatoid arthritis patients with psychological risks. Yuqing Song (22) conducted a RCT study to observe the impact of telemedicine education on drug compliance and disease activity of rheumatoid arthritis patients, the results showed that it has not improved symptoms of patients by using remote medical education, but it can enhance the medication compliance of patients. Both articles (18,22) showed that telemedicine could promote positive behaviors in patients with rheumatoid arthritis, the positive behaviors in uence of telemedicine is bene t for rheumatoid arthritis patients, Optimistically, By searching the relevant literature, telemedicine will have the opportunity to become an important tool in the management of patients with rheumatoid arthritis, at the same time, there are some uncertainties in the application of the Internet for rheumatoid arthritis patients. (44), so we still need more RCT study to prove the effect of telemedicine on rheumatoid arthritis, which will be a key point for the future research.

The advantages and disadvantages of telemedicine
Nowadays, the technology of telemedicine is frequency to be used. Telemedicine technology has also been applied to various elds of medicine, and some diseases can be treated and cared at home (1). telemedicine can provide long-term care and treatment for people with chronic diseases through the website or other telecommunication equipment, which is helpful to reduce the risk of virus spread in period of COVID-19 global pandemic. Many articles have con rmed telehealth and telemedicine technologies have the potential to increase access to healthcare services and make full use of medical resource , it could also help to change the treatment plan in time and improve the service quality. (45). Telemedicine has also been be considered as a cost-effective methods for the management of chronic diseases(46). However, there still has a big barrier for the patient who has di culty and anxiety to use computer or mobile especially for elderly people.(47).

Conclusions And Recommendation
This study conducted a meta-analysis and systematic review to verify the effect of telemedicine application on patients with diabetes, hypertension and rheumatoid arthritis. The results indicated that telemedicine had a positive effect on the management of diabetes, hypertension and rheumatoid arthritis, Telemedicine technologies have great promotion to access the medical services and improve the quality of care, especially for people with chronic diseases (48) Although telemedicine also has some vulnerabilities such as data leakage and other network security issues, it can break through the distance between time and space and reach the insu cient medical resource area, this is the bene t of telemedicine (5). 6. Declarations 6.1 Ethical approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and material
All data generated or analysed during this study are included in this published article [and its supplementary information les].

Competing interests
None.

Funding
The work described in this paper was supported in full by a grant from the Fu Xing Nursing Scienti c Research Fund of Fudan University, Shanghai, China [grant number FNF202063]. The sponsors had no role in the design and conduct of the study, in the collection, management, analysis or interpretation of the data, in the preparation, review, or approval of the manuscript, or in the decision to submit the manuscript for publication.

Author contributions
Yue Ma assess the quality evaluation of literature and Writing -Original draft preparation; Yan Zhao Assess the quality evaluation of literature ; Chongbo Zhao, Jiahong Lu and Hong Jiang review and revise the manuscript ; Yanpei Cao and Yafang Xu designed study and revise the manuscript.

Acknowledgments
The authors thank all the participants in the study.  Forest plots of glycosylated hemoglobin after 6 months of intervention between the intervention group and the control group Forest plots of fasting blood glucose after 6 months of intervention between the intervention group and the control group Forest plots of glycosylated hemoglobin after 12 months of intervention between the intervention group and the control group Forest plots of systolic blood pressure after 6 months of intervention between the intervention group and the control group Forest plots of diastolic blood pressure after 6 months of intervention between the intervention group and the control group