The webpage went online in 2015. This study consists of 4 substudies that are still ongoing and had separately been approved by the ethics committees. Eligible patients were recruited and screened in 6 University Medical Centers, and all patients provide signed informed consent. In Figure 1, the 4 clinical trials are described.

Due to the diseases studies being different, the primary outcomes and the inclusion and exclusion criteria vary between the substudies. In Table 1 the primary and secondary outcomes of the substudies are summarized. The HELP study (hepatic inflammation and physical performance in patients with nonalcoholic steatohepatitis [NASH]) and the COMMED study (cystic fibrosis online mentoring for microbiome, exercise, and diet) are prospective single-arm trials. The EXDEP study (exercise in depression) and the multicenter iPEP study (internet-based perioperative exercise program) are randomized controlled trials. Patients in the control groups had no access to the webpage (treatment as usual).

Data of patients in the iPEP study were collected at 3 time points. Baseline (t0) was at diagnosis (8-12 weeks before surgery, depending on the date of surgery). The first follow up (t1) was immediately before surgery to show the impact of the exercise program. The final examination (t2) was performed 12 weeks post surgery. In the EXDEP study and HELP study, the patients were tested at study start (t0) and 8 weeks later (t1). In the COMMED study, the study participants were compared with 3 points in time: to study start (t0), 12 weeks later (t1), and after 12 months (t2).

The inclusion criteria in the iPEP study were (1) histologically proven adenocarcinoma of the esophagus or adenocarcinoma of the esophagogastric junction type I according to Siewert’s classification, clinical stages IIB-IIIC (T3/T4 and/or N+; M0) according to Union Internationale Contre le Cancer, 7th Edition; (2) resectable stage according to discussion in the local multidisciplinary tumor board of the participating centers and patient medically fit for multimodality therapy (Eastern Cooperative Oncology Group performance status at least 1 or better, no severe impairment of cardiac, renal, hepatic, endocrine, bone marrow, and cerebral functions); (3) planned abdominal-thoracic esophagectomy with gastric pull-up and intrathoracic or cervical anastomosis; and (4) cognitive ability of the patient to understand the perioperative program and to participate actively.

The exclusion criteria were (1) the presence of a second malignant tumor (unless curatively treated >5 years ago); (2) chemotherapy or radiochemotherapy in patient's history; (3) orthopedic, rheumatologic, cardiovascular, or neurologic (epilepsy, stroke, Parkinson disease, muscle wasting diseases such as amyotrophic lateral sclerosis or multiple sclerosis) contraindications for the sports program; (4) inability to use the internet or no internet access; (5) inability to communicate in German; (6) each active disease that hinders completion of the study; and (7) active alcoholism or illegal drug consumption within the last 6 months before study entry.

The inclusion criteria for the HELP study were (1) histologically proven NASH or fatty liver disease.

The exclusion criteria were (1) bariatric surgery within the last 5 years, (2) body mass index (BMI) <18.5 kg/m² or >45 kg/m², (3) heart attack or stroke within the last 6 months, (4) higher grade coronary artery disease (CADIII-IV), (5) chronic obstructive pulmonary disease (asthma, COPD), (6) renal insufficiency, (7) uncontrolled hypertension or metabolic abnormalities, (8) alcohol consumption >30 g/day (male) and >20 g/day (female), (9) pregnancy, (10) concomitant medication able to cause a secondary NASH (eg, tamoxifen, corticosteroids), (11) concomitant medication able to affect inflammation (eg, tumor necrosis factor antagonists), (12) concomitant anticoagulant medication (eg, phenprocoumon; novel oral anticoagulants, NOAC), (13) other immunological or inflammatory diseases (eg, systemic lupus erythematosus), and (14) musculoskeletal disorders, preventing sport physiological investigations.

The inclusion criteria for the EXDEP study were (1) ability to understand the purpose and risks of the study and provide signed and dated informed consent and authorization to use confidential health information in accordance with national and local subject privacy regulations; (2) aged 18 to 65 years, inclusive, at the time of informed consent; (3) Montreal Cognitive Assessment >26 to exclude cognitive impairment; (4) apart from a clinical diagnosis of major depression or bipolar affective disorder, the subject must be in good health as determined by the Investigator, based on medical history and physical examination; (5) Quick Inventory of Depressive Symptomatology scores >5.

The exclusion criteria were (1) use of antidepressive medications or benzodiazepines at doses that have not been stable for at least 6 weeks before screening; (2) psychotherapy that started less than 8 weeks before screening; (3) any clinically significant psychiatric illness other than major depression or bipolar affective disorder; (4) transient ischemic attack or stroke or any unexplained loss of consciousness within 1 year before screening; (5) any uncontrolled medical or neurological/neurodegenerative condition that, in the opinion of the investigator, might impair treatment compliance and adherence; (6) history of unstable angina, myocardial infarction, chronic heart failure (New York Heart Association Class III or IV), or clinically significant conduction abnormalities (eg, unstable atrial fibrillation) within 1 year before screening; (7) clinically significant 12-lead ECG abnormalities, as determined by the investigator; (8) uncontrolled hypertension defined as: average of 3 systolic blood pressure/diastolic blood pressure readings >65/100 mmHg at screening; (9) history of malignancy or carcinoma, with the following exceptions: (i) subjects with cancers in remission more than 5 years before screening, (ii) subjects with a history of excised or treated basal cell or squamous carcinoma, (iii) subjects with prostate cancer in situ; (10) history of seizure within 2 years before screening; (11) recent history (within 1 year of screening) of alcohol or substance abuse as determined by the investigator, a positive urine drug (due to nonprescription drug) or alcohol test at screening; (12) clinically significant systemic illness or serious infection (eg, pneumonia, septicemia) within 30 days before or during screening; (13) history of HIV, hepatitis C virus, or hepatitis B virus; (14) any other medical conditions (eg, renal disease) that are not stable or controlled, or, which in the opinion of the investigator, could affect the subject's safety or interfere with the study assessments; (15) female subjects who are pregnant or currently breastfeeding; (16) participation in another study; (17) other unspecified reasons that, in the opinion of the investigator or Biogen, make the subject unsuitable for enrollment.

The inclusion criteria for the COMMED study were (1) age >12 years (2) forced expiratory volume in 1 second (FEV1) <90% and >28% of the set point or FEV1 >90% or/and lung clearance index LCI >9.

The exclusion criteria were (1) orthopedic, rheumatologic, cardiovascular, or neurologic contraindications for the sports program; (2) inability to use the internet or no internet access; (3) inability to communicate in German; (4) the absence of consent; (5) further cystic fibrosis–specific criteria: (i) severe pulmonary exacerbation, (ii) FEV1 <27% of the set point (standard value Global Lung Initiative), and (iii) acute infection.

Before the start of the study, all eligible study participants performed a stepwise cardiopulmonary exercise test until volitional exhaustion. Each stage of the modified walking protocol lasted for 3 min and intensity was increased by speed and elevation of the treadmill. During the test, heart rate (HR) and respiratory gas analysis were continuously monitored. Furthermore, blood samples from the earlobe were taken at the end of each stage to determine lactate concentration. Subjective degree of exhaustion was measured utilizing the Borg scale (6-20) 30 s before the end of each stage [43].

For study design and content structure of the website, experiences from other studies were considered [29,44-46]. Similar to Barak et al (2009), we took the key components of a Web-based intervention into account. We delivered program content, used multimedia aspects, promoted interactive Web-based activities, and provided tailored feedback [37]. Participants were registered on the home page by the administrator and a printed (also online available) instruction manual was provided. Once registered and logged in, users were able to change their profile by editing or deleting information such as username, profile picture, or password. Relevant aspects of data protection were taken into account. The various disease intervention groups obtained access to different parts of the website (Figure 2).

Besides the interaction with a counselor, peer support was considered as a basic principle of our concept. Therefore, each subgroup had its own discussion forum and chat room to improve social support and adherence [46]. Furthermore, nutritional experts or supervising physicians distributed quickly and easily the disease-specific questionnaires and informative documents.

To improve the current condition, sport scientists were responsible for training management and weekly recommendations. Individually tailored exercise plans were sent weekly by an internal email in the secure area of the website. The program consisted of walking or running recommendations, muscle strengthening and stretching exercises, as well as relaxation exercises. An HR monitor (Polar, FT1) was provided to the study participants to monitor the endurance training. The resistance training was carried out in a home-based environment with body-weight exercises and with elastic resistance bands (in different strengths; Pinofit; Pharmazeutische Präparate GmbH, Hamburg) and lasted for approximately 45 min. In addition to an illustrated tutorial, all exercises were additionally deposited as a video file on the home page and could be downloaded or viewed online (Figure 3).

Participants were encouraged to provide training-related information (eg, average HR, duration, subjective perceived exertion) to the supervisor at the end of each week while filling out a schedule. This information helped the trainers to adapt the training load for the upcoming weeks. Due to the individual feedback of study participants, tailored recommendations focused on the needs, problems, and limitations of each participant were possible. Strength and endurance training were examined separately and allowed specific increase or decrease of training content in terms of duration and/or intensity (Figure 4).

Due to a steady contact with the patients, a missing schedule was noticed quickly by the supervisor sending an email to identify possible problems with the program. Patients were able to contact the trainer at any time by an internal email. In case of questions and feedback, the supervisor answered within 24 hours.

The data on using behavior were assessed by evaluating the log-in rate and log-in duration during the intervention period of 8 weeks. The exercise implementation was assessed by evaluating the training time, exercise interruptions, self-chosen alternative exercise programs instead of the weekly recommendations, and adverse events. The user satisfaction with the exercise concept and the webpage was assessed based on a short questionnaire (8 Likert items). Furthermore, the participants graded the concept and were asked if they would continue to use the webpage.

Preliminary descriptive statistics were used to present data on the baseline characteristics of each single trial. Descriptive statistics were also used to show data on utilization and satisfaction with the Web-based exercise concept, as well as exercise adherence and training interruptions and log-in rate and log-in behavior. No data on exercise effects will be presented, due to the heterogeneity of primary outcomes, study collectives, and study designs. Statistical analysis was performed using SPSS (version 22.0, Chicago, IL, USA) and P values <.05 were considered significant. Data were not normally distributed and showed heterogeneous variances. Therefore, nonparametric Kruskal-Wallis H-tests were carried out to examine trends of physical activity over time among the substudies. The Dunn-Bonferroni test was additionally used as post hoc test to further determine, where exactly the differences between the groups or time points were located.

All study participants were recruited by physicians in the cooperating centers. Due to different recruiting processes of the single trials and different intervention periods, with regard to the present analysis, we present the first 5 recruited study participants of each study over an 8-week period. As summarized in Table 2, characteristics of the 20 study participants and primary outcome measures at baseline are presented. The patient flow of each trial is shown in Figure 5.

In Table 2, the main baseline characteristics are summarized. The majority of participants were males (n=12) and showed similar distributions with respect to cardiorespiratory fitness. Mean age was 42 years (SD 14.54 years), and mean BMI was 26 kg/m² (SD 3.91 kg/m²). Patients in the HELP study tended to be obese (BMI: 30 kg/m²; SD: 3.02 kg/m²), whereas patients in the COMMED study tended to be underweight (BMI: 22 kg/m²; SD: 1.42 kg/m²). There was a difference in age across all groups. Patients of the COMMED study were younger (32 years; SD: 7.91 years) due to the congenital disease, whereas Barrett cancer patients in the iPEP study were older (55 years; SD: 5.45 years; P<.001). Additionally, there was a significant difference in body weight (P<.001) and BMI (P<.001) between patients of the COMMED and HELP study. However, a low cardiorespiratory fitness level was common in all substudies compared with sex-specific normative data according to the Heywood classification [47] and the comprehensive investigation by Herdy et al [48].

This investigation evaluated the feasibility of the Web-based exercise concept for different diseases. The participants of all studies were able to take advantage of the information material and the individual recommendations, irrespective of types of diseases. A quick and easy access to the website enabled a continuous and regular support for the patients. However, a reduced log-in behavior (number and duration) was observed during the time course of the first 8 weeks.

In accordance to Eysenbach et al (2005), less-frequent log-in rates over time are a serious problem of Web-based interventions [42]. According to Couper et al (2007), technical challenges (eg, no access to email, problems accessing or submitting the survey) and problems with the survey (eg, lack of interest in or lack of effectiveness of intervention, no time or bad timing, survey was boring or too long) are the main reasons for noncompletion [49]. Nevertheless, the only common feature of Web-based interventions is the delivery channel [50]. Therefore, the particular role of the website must be evaluated carefully. Regardless of the underlying issue, several features could be frequently used, from computer-generated feedback and general information, such as pamphlets and regular newsletters through to fully tailored feedback and regular contact with peers and the study team (eg, chat or forum function) [51-60]. Besides regular updates, peers on the Web, and general information, the communication to a counselor and the regular monitoring of results by a professional sports therapist seem to be a key reason for compliance and adherence in Web-based settings [34,44,45,61,62]. The log-in rates or durations during the intervention period are influenced by the features offered and are of particular importance in investigations where programs for health education or behavioral change demand regular interaction. In our setting, patients obtained weekly individually tailored exercise recommendations and were encouraged to perform the training in their home environment. The exercise manual for strength training and detailed video files for each exercise could be downloaded and used offline. In accordance to Farvolden et al (2005), this opportunity could explain the high attrition rate over time [53]. Study participants of our investigations got used to the home page and probably preferred viewing the study materials offline. The patients were not obliged to log in several times a week or to be online for special appointments, to complete the study. Furthermore, the patients were able to communicate with the study team via other channels such as apps (WhatsApp), email, or phone contact. The average length of a visit was 14 min and 38 s (SD: 11 min, 29 s), and thus similar to a cognitive behavioral therapy investigation by Farvolden et al (2005) in a collective of patients with panic disorders [53]. A comparable log-in behavior could also be seen in a Web-based physical activity investigation in a cohort of diabetes patients over an 8-week period. The average session duration was 13 min, and the investigators also identified a steep decline in usage of the program toward the end of the study [63]. On the contrary, it is particularly interesting to note that the patients with depressive disorders of our EXDEP study checked the webpage 2.6 times a week and stayed online for more than 23 min. A review article by Vandelanotte et al (2007) summarized the effects of Web-based physical activity interventions in different collectives. Physical activity improvements could be detected in 8 of 15 studies reviewed from 2000 to 2007 [57]. However, essential differences exist between support strategies and the integration of the website content [58-60]. To examine the participants’ perceptions of a physical activity promotion website, Sciamanna et al (2002) provided support in terms of general information materials such as tips for overcoming barriers or instructions for measuring HR and setting activity goals [64]. Furthermore, a printable daily self-monitoring chart of physical activities was available [64]. Similar to our investigation, McKay et al (2001) provided tailored feedback on the basis of the baseline activity level [63]. Study participants received tailored messages to increase the daily activity minutes [63].

In our substudies, regular feedback and tailored recommendations were provided. However, not all study participants followed our instructions. Due to individual situations (eg, well-being vs cancer with neoadjuvant radio/chemotherapy), participants performed alternative exercise programs (eg, relaxation exercise, yoga, or hiking) instead of our recommendation. Therefore, the results in terms of exercise progress during the intervention period should be interpreted with caution. Despite the observed decline in log-in rates and durations toward the end of our substudies, most of the participants (80%, 16/20) were still satisfied with the structure of the concept, which is in accordance with the findings of McKay et al (2001) [63]. It remains unanswered by our present data whether a significant and ongoing decline in log-in behavior, as shown, necessarily means a progressive reduction in exercise adherence.

The main strength of this study is the easy access to the website, the opportunity to reach a large group of patients, and the possibility to provide individually tailored exercise support. Furthermore, an immediate reaction on questions, problems, or postings was feasible. Log-in data could be easily collected by tracking the account activity in an objective manner. Finally, the intervention was free of charge and due to minimum investment in human effort, cost effective.

The fact that study participants gave subjective response on their physical activity during a week without an objective evaluation, can be seen as the main disadvantage. The study team had to rely on the participants’ report. Another aspect that needs to be considered, is the website itself. Neither was the forum moderated, nor were regular newsletters distributed, and this possibly led to a reduced log-in activity. Finally, Web-based interventions do reach only selected participants who have access to the internet. Lack of control groups should also be seen as a major limitation. In addition, comparison of the single trials must be interpreted with caution due to the small sample size and the heterogeneous collectives. However, the common platform and same procedure for exercise support might show a general and uniform applicability of the concept used.

Regular updates, intensified monitoring, moderated discussion forums, and additional information materials from other areas of interest, associated with coping with the specific disease, such as nutrition and relaxation, could be added to sustain the website usage in the long run. Furthermore, the integration of app support and modification of the website toward a more intuitive structure as well as the integration of extended functions, such as Web-based forms or data entry, should be realized with regard to subsequent investigations. Finally, investigations with other lifestyle-related diseases, mainly with the metabolic syndrome and its consequences, such as diabetes, obesity, or heart disease, should be focused upon in the future due to its rising incidence.

The universal use of the Web-based concept appears to be applicable across different diseases and age groups. Although the development of physical activity shows only moderate improvements, flexible communication, timely response to patients’ needs, and tailored support could be easily integrated into patients’ daily routine. However, because of different application habits of the website among the substudies, designing a website that is suitable and sustainable for most users will be a challenging target for future studies. Despite the existing limitations, Web-based approaches can be a helpful supplemental method to bridge the gap between inpatient and outpatient rehabilitation and home treatment for chronically ill patients. An ongoing development in telemedicine makes this kind of intervention with its cost- and time-effectiveness especially interesting for the future.