Design
The study protocol is structured following Spirit 2013 [24]. PROACTIVE-19 is a pragmatic, randomized controlled clinical trial with adaptive “sample size re-estimation” design. This design allows interim analyses and necessary modifications of the sample size of the ongoing trial to ensure adequate power.[25]
Legislative amendment and ethical approval
In Hungary, Act CLIV of 1997 on Health and Decree No. 23/2002 (of 9th May 2002) of the Minister of Health on Biomedical research on human individuals (as amended) stipulates the procedure for non-interventional investigation, according to which: 1) the leader of investigation or the investigator shall inform the subject both orally and in writing, prior to obtaining the consent of the subject to participate in the clinical research; 2) the subject’s informed consent shall be written. This Act and Decree would not have allowed to commence the clinical trial as it would have amounted to a criminal offense. Based on our request sent to the Prime Minister of Hungary to amend the Decree, the Government of Hungary, within three days, issued Government Decree No. 63/2020 of 24 March 2020, according to which the new decree amends: 1) in addition to Section 159 of Act CLIV of 1997 on Health, subjects with full disposing capacities can be informed about the non-interventional investigation qualified as a clinical research on coronavirus via means of telecommunications; 2) subjects may consent to participate in the clinical research by means of telecommunications; 3) subjects may withdraw their consent via means of telecommunications.
Ethical approval: Scientific and Research Ethics Committee of the Hungarian Medical Research Council (IV/2428- 2 /2020/EKU).
The trial organization, committees and boards
The corresponding centre and designer of the PROACTIVE-19 trial is the Centre for Translational Medicine at the Medical School, University of Pécs (coordinating institution and sponsor, www.tm-centre.org).
The Steering Committee (SC) will be led by PH (principal investigator, gastroenterologist, specialist in internal medicine and clinical pharmacology). SC members will be BE (gastroenterologist, specialist in internal medicine and primary care), ASz (interdisciplinary unit), ZM (intensive care specialist), and ZH (pharmacologist, specialist in clinical pharmacology). There will be independent members as well, and the SC will include a patient representative.
The SC will supervise the trial primarily and will make decisions regarding all critical questions (e.g., premature termination of the study, dropouts, etc.).
Adjudication Committee (AC): The committee will include a specialist in infectious diseases (ZP), pharmacologist (AV), and a paediatrician (BZ).
The study was designed by the SC and AC and was supported by the Medical School, University of Pécs. The sponsor had no role in the design of the trial and will have no access to the randomization codes or the data.
The study will have independent members, including physicians and a safety manager (LC), to comply with current ethical regulations.
Patient and public involvement
We will perform the assessment and minor modifications in the structure and wording of data collection and the interventions based on the operators’ and participants’ feedback after testing the protocol on 100 potentially eligible subjects. Data of these subjects will not be recorded; only anonymous feedback will be given.
Patients were not included in recruitment to and conduct of the study. Immediately after publications, study results will be disseminated to the population above 60 years of age via the electronic media when, depending on which study arm will better, either general or personal lifestyle intervention will be delivered.
Our interventions do not impose considerable financial burden on patients; therefore, such compensation will not be required. The volunteered test patients claimed that the time and efforst required to be invested to complete the recommendations of the interventions are fully acceptable.
Study population
Inclusion and exclusion criteria
The inclusion criteria of our selective primary prevention programme are: (1) age over 60 years (that is, high-risk individuals), (2) informed consent to participate. The exclusion criteria are: (1) confirmed COVID-19 (active or recovered); (2) hospitalization at screening for eligibility; (3) someone was already enrolled in the study from the same community/household (to avoid potential crosstalk between the study arms).
Recruitment
The population will be informed about the study and the contact details via social media platforms, newspaper, radio, and television advertisements.
Flow and timing
A toll-free phone number will be available for all who are interested in participation. By dialling this number, the participant will be informed about the trial through a pre-recorded voice message, including the study rationale, conditions of participation, the process of the study, and the information on data protection. Willing participants will be redirected to an available operator, who will ascertain eligibility. Following verbal consent and randomization, the operator will obtain key personal information of the participants and all study-related information (Fig.1.).
Interventions
Participants will be randomized into two groups: (A) general health education; (B) personalized health education. They will go through questioning and recommendations in 5 domains: (1) mental health, (2) smoking habits, (3) physical activity, (4) dietary habits, and (5) alcohol consumption. Both groups will receive the same line of questioning to assess habits concerning these domains (Suppl. files 1,2).
Group A: questioning will be done in the order as mentioned above, followed by a general health education aiming towards improvement of these factors with general recommendations (the expected mean duration is approximately 10 min).
Group B: questioning will be done in the same structured order, but an assessment of each domain will be followed by personalized recommendations (the expected mean duration is approximately 20 min).
After the first contact, there will be follow-up calls in both groups, with a matching schedule: every week in the first month, every second week in the second month, then monthly. During these encounters, all change in all five domains since the last call will be assessed. The structure, script and algorithm of the initial and follow-up lifestyle interventions are detailed in Suppl. files 3,4, respectively.
The operators will be personnel who received (or are currently undergoing) any healthcare education. Before enrolling participants, the operators have to complete a standard training program consisting of seminars on the interventions held by medical professionals, followed by practice of scenarios. The operators will be trained not to give additional healthcare advice, and we will not secure other information sources, including electronic and printed material.
Since standard delivery of the interventions and data collection are essential, the first three and every 50th call of each operator will be assessed. Besides, random calls with various scripts will be made by the study staff to test the operators’ reactions (who are unaware of the test session), followed by detailed assessment and feedback to ensure quality control.
Outcomes
Based on literature data [5, 26][20], the primary endpoint will be defined as the composite of any of the followings in COVID-19 cases (an accredited laboratory should verify positivity) the rate of:
- ICU admissions
- hospital admissions (longer than 48 hours) for the following reasons
- arrhythmia (causing hemodynamic instability and requiring continuous monitoring and/or cardiac support, as indicated by mean arterial pressure <65 mm Hg, and/or serum lactate >2 mmol/L) and/or
- ARDS (severe hypoxaemic respiratory failure indicated by a PaO2/FiO2 <300 mm Hg according to the Berlin definition) [27] and/or
- circulatory shock (the requirement of continuous vasopressor support to maintain mean arterial pressure ³65 mmHg and/or serum lactate £2 mmol/L) and/or
- deaths
Secondary endpoints are the followings:
- the number of general practitioner visits,
- the number of emergency, hospital, and intensive care admissions;
- the length of hospitalization and ICU stay,
- the number of organ dysfunctions and failures (central nervous system, cardiovascular, respiratory, renal, liver, haematological),
- the measurable lifestyle changes (including physical and mental health),
- the costs of care.
Randomization and blinding
Computer-generated random sequence randomization (central) will be performed, after giving informed consent. Due to the expected large sample size, we will use simple randomization. The allocation ratio will be 1:1.
In the study, participants will be blinded to the knowledge of the details of differences between the interventions. Everyone else (outcome assessors, caregivers, and data analysts) will be blinded regarding the allocation.
Sample size calculation, interim and final analyses
The primary outcome is estimated to occur in 20% of COVID-19-infected cases (≥60 years of age) receiving the standard of care based on Chinese reports [5]. Due to the lack of data, we hypothesized that our intervention would result in a 50% risk reduction. Considering one interim analysis on efficacy (with the Pocock correction), 90% power, 5% alpha (superiority design, two-sided), a drop-out rate of 20% [28, 29] and assuming 10% incidence of COVID-19 in the target population, the estimated sample size is 3788 (rounded up to 3800) subject per study arm. The calculation was performed by Stata (version 15, Philadelphia, the USA).
We plan to hold three interim analyses: the first for sample size re-estimation at 5% of the target sample size due to the drop-out rate, the second for safety assessment at 10% of the target sample size and a third for efficacy assessment and sample size-reestimation at 50% of the target sample size. Early stopping will be executed if (1) safety concerns arise during the interim analysis, (2) the statistical power reaches at least 90% and p<0.05 at the efficacy interim analysis (stopping for benefit), (3) the statistical power does not reach 10%, p>0.05, and the event number does not reach the assumed 10% for the whole population at the efficacy interim analysis (that is, 380 events for the primary outcome - otherwise, the interim analysis is postponed and repeated when the event number reaches 380 events) (stopping for futility), (4) the consequences of the pandemic make further recruitment or follow-up impossible (stopping for unfeasibility).
In the final analysis, the intention-to-treat analysis will be favoured over per-protocol (or "as-treated”) analysis. We expect a full dataset for the primary endpoint (since the Hungarian Ministry of Interior will provide these data). If, for any reason, data will be missing for the primary outcome, we will use available case analysis. The “last observation carried forward” strategy will be followed to impute missing data for other outcomes measured during the study. Missing more than one consecutive interventions after the initial assessment or withdrawal of consent during follow-up result in the drop-out of the patients unless hospitalization is required in the meantime.
In descriptive statistics, the count and percentage will be provided for each treatment arm for binary outcomes. For continuous outcomes, n, mean, median, interquartile (Q3–Q1), standard deviation, minimum, and maximum values will be provided for each treatment arm. In a univariate comparative analysis, we will calculate relative risk with 95% confidence interval (CI) when comparing the primary endpoint between two groups (alpha=5%) with a reference arm using non-repeated intervention complemented with chi-square or Fisher’s exact test (the same strategy will be followed for binary secondary outcomes). For continuous variables, we will use t-test assuming unequal variances or the Mann-Whitney test. We will perform univariate (Kaplan-Meier and Cox-regression) and multivariate (Cox-regression) survival analysis for binary outcomes. An adjustment will be carried out at least for age, sex, and education. Mixed effect logistic regression will be conducted to estimate the effect of the multicomponent intervention on the outcomes, where the subject IDs will be used as a random subject. The model will be adjusted for changes in smoking habits, alcohol consumption, physical activity, and dietary habits (or body mass index).
All analyses will be carried out with SPSS version 26 and Stata version 15.
Study duration
The planned starting date of the study is 1st April 2020, and the anticipated finishing date is the end of the pandemic or development of the vaccine, but no more than one year from the enrolment of the last participant.
Data management
Data handling
Confidential and anonymous data handling will be performed by the Data Monitoring Committee (DMC). To be able to trace data to an individual subject, a subject identification code list will be used. A Personal Identification Number (PIN) will be generated to identify the data of the participant. This PIN will be present on all forms and documents of each individual. Electronic case report forms (eCRFs) will be used. The Investigator will ensure that the data in the eCRFs are accurate, complete, and legible. Detailed data flow will be described in a Data Management Plan (DMP). Data from completed eCRFs will be validated under the direction of the Data Manager on the DMC according to a Data Cleaning Plan (DCP). Any missing, implausible, or inconsistent recordings in the eCRFs will be referred back to the Investigator using a data query form (DQF) and will be documented for each subject before clean file status is declared. All changes to eCRFs will be recorded.
The DMC will perform an independent assessment of trial-related documents and activities to ensure respect for subjects’ right, safety and well-being and to guarantee the plausibility of clinical data. The similarity of groups at baseline will also be checked.
After verbal consent of the subjects, the data will be recorded by the investigator. Clinical research data are processed separately from participants’ personal data under pseudonyms. Data may only be accessed by persons acting under the authority of the controller and in accordance with the authorization system established within the controller’s organizational structure, only to the extent and in the manner necessary for the performance of tasks. Personal data are not accessible to third parties.
Safety
Due to the nature of the multi-component moderate-intensity lifestyle intervention, we do not expect serious adverse events. However, minor or moderate adverse events may develop, such as alcohol and nicotine withdrawal, weight change exceeding the optimum, and the need for change in regular medications (antihypertensive or antidiabetic drugs). Participants will be advised to consult their primary care physician if any non-lifestyle-related health issue arises except for COVID-19-related concerns when the call will be transferred to the COVID-19-specific national helpline immediately. If a participant develops a potentially serious health problem, the chairman of the Safety Monitoring Board (LC) will be notified. After the first interim analysis for safety at 10% of the target number, the board will revise the charts of all visits to health facilities and assess if any event is related to the interventions (see, early stopping for safety).