JMIR Publications

ABSTRACT

Background:

According to the 2016 Global Burden of Disease (GBD)(1), despite of the increase in both quality of life and access to health care observed since 1990, there are still many countries in which health inequalities still remain, particularly in relation to cancer and noncommunicable diseases (NCDs), such as asthma, chronic obstructive pulmonary disease and diabetes. These diseases have a significant impact on quality of life(2) and this is particularly true in the case of diabetes, because it has a range of serious complications in situations where it is poorly controlled by the individual, such as cardiovascular disease, chronic kidney disease, blindness and lower limb amputation(3), many of which could be avoided by providing better access to health care. In 2016, Brazil was in the 96th position among 195 countries, regarding access to health and quality of life, according to the GBD(1). Despite a slight improvement in the health index (from 46.5 in 1990 to 63.8 in 2016), access to health is still one of the largest indicators of social inequality(4) . While the inhabitants of São Paulo, Brazil´s largest city (Figure 1), for example, have access to 2.81 doctors per thousand inhabitants, in the northeast region, this proportion is about 1.41 doctors, and even in some states can reach less than one doctor per thousand inhabitants (4). When considering physicians of all medical specialties, almost 70% of these professionals are concentrated in the south and southeast regions of the country (5). To improve medical access, there is a need to develop strategies that can not only enhance local primary care but can also improve the regulatory processes and the organization of specialized healthcare in Brazil(6)(7)(8). Although there is an increasing evidence for the use of remote consultation worldwide, in Brazil, under Resolution number 1643/2002(9) of the Federal Council of Medicine, medical consultations by telephone or internet can´t be conducted directly between health professionals and patients. Thus, currently, teleconsulting is only allowed when there is a health professional at both ends of the communication channel. If teleconsulting directly between the physician and the patient can be shown to be safe and effective, this would provide strong evidence to change the current regulations, freeing up one professional from the consultation and driving down costs. There are no national studies in Brazil that address the effectivity and the costs involved in teleconsultation service in the public health system.

Objective:

The main objective of this study is, therefore, to test the hypothesis that remote consultation can be as effective as standard face-to-face consultation in type-2 diabetes in the Brazilian public health system and to assess the associated costs related teleconsultation in public heath scenario.

Methods:

STUDY DESIGN TELEConsult Diabetes is a pragmatic, phase 2, single-center, open-label, non-inferiority trial with central randomization that will evaluate the efficacy and safety of specialized remote consultation compared to a face-to-face consultation in patients with type 2 diabetes mellitus referred by primary health care units to specialist care in the Unified Health System (SUS)(10). The trial has been approved by all the research ethics committees of the institutions. It´s registered on ReBEC (http://www.ensaiosclinicos.gov.br/) by number RBR-8gpgyd. STUDY POPULATION The inclusion criteria will be: adults of both sexes; over 18 years old, have an established diagnosis of type 2 diabetes; either, not insulin dependent (with any level of glycemic control measured by HbA1c accordingly to local protocol (11)(12)(13) (14) or insulin dependent (with HbA1c greater than 8). Exclusion criteria will be: patients under 18 years old or with type 1 diabetes, women with gestational diabetes or diagnosed during pregnancy, and patients with chronic renal failure with estimated or measured creatinine clearance less than 30 ml/min/m2(15) will be excluded (Figure 2). SETTING Brazil has a universal public healthcare system called SUS – the Unified Health System, which is structured around the principle of primary care, and operates integrating other services and levels of care, such as specialized care, into the system for whole population. The main goal of primary care in Brazil is to provide basic care, and to coordinate other levels of care throughout the health system network, regulating referrals to specialist care(8). The Telehealth Center of Santa Catarina state uses a system that does not represent the reality in most Brazilian states. First, the primary care physicians request a teleconsultation with a specialist. This consultation involves both, the primary care physicians, and the specialist. Following this first contact, the primary care physician decides whether the patient needs to be referred to a face-to-face consultation with the specialist(16)(17). This system of compulsory flow established and operating in Santa Catarina state is the ideal setting to investigate the potential of remote consultation in primary health care, compared to face-to-face consultations (Figure 3). RECRUITMENT OF PARTICIPANTS The participant will be identified by the primary care physician and the Telehealth Center of the state of Santa Catarina in conjunction with the regulatory department of the Municipal Health Office of Joinville will identify his/her eligibility criteria. Patients with type 2 diabetes mellitus may be recruited from any of the city's 70 primary care centers. A researcher will contact the patient by telephone to tell them about the trial and invite them to participate in the study. If the patient agrees, they will be told that they need to provide their written consent, and an informed consent form will be posted to them prior reading and assessment of their participation in the study. Two to five days after sending the informed consent form, the researcher will again telephone the patient to confirm their interest in participating in the trial. If the response is still positive, they will schedule a home visit at a convenient time with the patient, in the presence of a legal representative or relative if they so wish. During the visit, any questions related to the study will be answered. After signing and agreeing to participate in this study, the researcher will tell the participant about the next steps: their data will be randomized in a system created especially for this trial and they will be scheduled for a face-to-face consultation or for a remote consultation with the specialist. By the end of this home visit, the participant will have been advised about the study, their participation, the date, and place of the research consultation, and that there are no costs involved (Figure 3). COSTS OF TELECONSULTATION To calculate the real cost of a teleconsultation service to compare with the cost of the face-to-face service, and to generate a cost parameter for the primary health care units for this type of service, the time-driven activity-based costing (TDABC) method will be used (18)(19) In addition to the costs of the resources used, this method also considers the amount of time spent on each stage of the process, enabling the identification of stages where the teleconsultation speeds up or slows down the process, reducing or increasing costs in relation to face-to-face consultation. This study will also include other indicators, such as the transportation costs for the patients and physicians in a “real-word” scenario, to carry out cost-effectiveness estimates. Finally, we intend to discuss issues related to the reduced costs that an increased access to medical care, via teleconsultation, can generate for the system. RANDOMIZATION The randomization list will be generated electronically using appropriate software. Randomization will be performed in blocks of 4 at a 1:1 ratio. The confidentiality of the randomization list will be ensured by setting-up a central database, and the use of an electronic case report form (e-CRF). Access to the system will be granted through specific usernames and passwords given to each investigator or study team. The patient will be allocated to one of the treatments (remote consultation vs face-to-face consultation) only after being registered in the system. Given the nature of the intervention, blinding is not feasible. However, data analysis will be performed by a statistician blind to patient allocations. PATIENT RECRUITMENT Participants in both groups will be assisted by one of a team of four endocrinologists from the public outpatient clinic. They will take turns to provide the assistance in the two modalities. The remote consultation with the intervention group will involve only the specialist and the participant. These consultations will take place in one of the six primary care units chosen to cover the main regions of city. In each of these units the research team set up a room with a computer, a microphone, and a camera to deliver the remote consultation, assisted by the same team of endocrinologists. The same physicians will undertake both the remote consultation and the face-to-face consultation, with the same duration (30 minutes - 60 minutes). All the physicians were all specialists medically trained in endocrinology and experience in caring for patients with diabetes in different settings and as part of the city´s specialist care network. A protocol based on national and international guidelines specifically designed for the study (13)(20)(21) (22) will be used to ensure that all the consultations are as similar as possible. The physicians were instructed to follow the protocol to avoid any bias related to the consultation. The protocol includes instruction on the questions to be asked about the participant's health and the medicines they are using, the scale to be used for hypoglycemia evaluation(23) and the medications to be prescribed (following current guidelines. At the end of the consultation, participants from both groups receive, in addition to medical advice and drug prescriptions, will receive guidance on any laboratory tests that need to be carried out before the next consultation as shown in Figure 2. STUDY OUTCOMES The primary outcome will be a change in glycated hemoglobin (HbA1c) levels at 6 months after randomization. However, we will perform a pre-specified analysis at three months after randomization. Secondary outcomes will be fasting blood glucose, complete blood count, urea, creatinine, total cholesterol, triglycerides, HDL cholesterol, LDL cholesterol, systolic and diastolic blood pressure, body weight, body mass index, frequency of hypoglycemia and incidence of adverse events. In addition, in the intervention group we intend to evaluate the satisfaction of the endocrinologists and the patients with the video conference system using a structured questionnaire(24). The specific quality of life in diabetes questionnaire will be applied to patients with diabetes mellitus in the first and last research consultation of both groups(25)(26). In addition, we will estimate the actual cost of the remote consultation service, using the time-driven activity-based costing (TDABC) method (27)(28) to obtain the unit cost of remote consultation in the primary health care units. The trial is registered with the Ethics and Research Commission with the number: 03434218.1.2001.5362. STATISTICAL ANALYSIS The primary efficacy endpoint in this study is the change from baseline in HbA1c (%) at 6-month .The assessment of non-inferiority of remote consultation in relation to face-to-face will be given, according to the CONSORT (Consolidated Standards of Reporting Trials)(29)(30), by the confidence interval approach to the difference of the mean of the primary variable between the two groups. If the upper limit of the 90% bilateral confidence interval is lower than the established non-inferiority margin (NIM)(0.5%), the non-inferiority of the remote consultation group relative to the face-to-face group is declared at the 5% significance level(31) (32) An analysis of covariance (ANCOVA) model will be constructed using main effects of treatment and baseline HbA1c as covariate. Adjusted means (least square mean) by treatment will be presented as well as an estimate of the difference between adjusted means. A 90% two-sided confidence interval, based on the ANCOVA model, will be computed for the difference of two-groups: • If the upper bound of the CI is no higher than NIM, then the null hypothesis (Remote is inferior to face-to-face) will be rejected. • Else, if the upper bound is greater than NIM, the non-inferiority is not demonstrated. Least square means with 90% confidence intervals (CI) will be reported for ANCOVA. A general mixed analysis of variance (ANOVA) model for repeated measurements (linear mixed effects model) will be considered for secondary efficacy endpoints. The model will be applied to absolute values from baseline to posterior time points. The model includes fixed effects for treatment, time, and the interaction between treatment and time, and subject as a random effect. The primary efficacy endpoint will be analyzed on the ITT and PP population. All other endpoints will be analyzed on the ITT population. Multiple imputation will also be used as a sensitivity analysis for primary efficacy endpoint. If proportion of missing greater than 5%, sensitivity analyses for missing data imputation will be performed using a standard multiple imputation approach under the Missing At Random (MAR) assumption (33) (34) (35) (36). All demographic variables and baseline characteristics will be summarized treatment group for the ITT population. Categorical data will be summarized by numbers and percentages. Continuous data will be summarized by mean, SD and range if data are normal and median, IQR and range if data are skewed. Tests of statistical significance will not be undertaken for baseline characteristics; rather the clinical importance of any imbalance will be noted. Normality will be assessed by visual inspection of histograms and with Shapiro-Wilk normality test or D'Agostino-Pearson normality test (37). The safety data will be summarized using descriptive statistics by treatment group in the Safety population. Proportion of adverse events will be compared between the two groups via Fisher's exact test. Statistical analyses will be performed using SAS version 9.4 (SAS Institute Inc, Cary, NC). Multiple imputation will be performed with SAS or IVEware (last version) (38).

Results:

Brazil is a country of continental proportions, with great heterogeneities and gaps in access to health, types of health services and specialized medical professionals. Lack of access to health services is one of the main indicators of social inequality in Brazil. Thus, to improve access, there is a need to build strategies that impact primary health care, the processes that regulates access, and the organization of specialist care. In this context, studies have shown telemedicine to be equivalent to face-to-face care and it can be an effective solution to increasing patient access to services, especially to specialist doctors. Providing evidence of the efficacy and safety of remote treatment for different condition in Brazil will contribute to improving patients access to the public health system, including the specialist doctors. This evidence can also help to remove restrictions placed on remote consultation by the Brazilian Federal Medicine Council, which currently restricts direct specialist-to-patient consultation. This model of improved access can help to meet the health needs of the population, breaking the geographical barriers that a country like Brazil imposes on the provision of health services. In addition to greater access to health care, the use of telemedicine has potential economic benefits for health systems and can be used safely to deliver a quality service.

Conclusions:

Our expectations are that providing remote access to healthcare will result in improvements in the health and quality of life of type 2 diabetes patients, reduce costs, and that both patients and clinicians will benefit and be satisfied with this technology. Clinical Trial: The trial has been approved by all the research ethics committees of the institutions. It´s registered on ReBEC (http://www.ensaiosclinicos.gov.br/) by number RBR-8gpgyd.