Remote care through telehealth for people with inflammatory bowel disease

Abstract Background People with inflammatory bowel disease (IBD) require intensive follow‐up with frequent consultations after diagnosis. IBD telehealth management includes consulting by phone, instant messenger, video, text message, or web‐based services. Telehealth can be beneficial for people with IBD, but may have its own set of challenges. It is important to systematically review the evidence on the types of remote or telehealth approaches that can be deployed in IBD. This is particularly relevant following the coronavirus disease 2019 (COVID‐19) pandemic, which led to increased self‐ and remote‐management. Objectives To identify the communication technologies used to achieve remote healthcare for people with inflammatory bowel disease and to assess their effectiveness. Search methods On 13 January 2022, we searched CENTRAL, Embase, MEDLINE, three other databases, and three trials registries with no limitations on language, date, document type, or publication status. Selection criteria All published, unpublished, and ongoing randomised controlled trials (RCTs) that evaluated telehealth interventions targeted at people with IBD versus any other type of intervention or no intervention. We did not include studies based on digital patient information resources or education resources, unless they formed part of a wider package including an element of telehealth. We excluded studies where remote monitoring of blood or faecal tests was the only form of monitoring. Data collection and analysis Two review authors independently extracted data from the included studies and assessed their risk of bias. We analysed studies on adult and paediatric populations separately. We expressed the effects of dichotomous outcomes as risk ratios (RRs) and the effects of continuous outcomes as mean differences (MDs) or standardised mean differences (SMDs), each with their 95% confidence intervals (CIs). We assessed the certainty of the evidence using GRADE methodology. Main results We included 19 RCTs with a total of 3489 randomised participants, aged eight to 95 years. Three studies examined only people with ulcerative colitis (UC), two studies examined only people with Crohn's disease (CD), and the remaining studies examined a mix of IBD patients. Studies considered a range of disease activity states. The length of the interventions ranged from six months to two years. The telehealth interventions were web‐based and telephone‐based. Web‐based monitoring versus usual care Twelve studies compared web‐based disease monitoring to usual care. Three studies, all in adults, provided data on disease activity. Web‐based disease monitoring (n = 254) is probably equivalent to usual care (n = 174) in reducing disease activity in people with IBD (SMD 0.09, 95% CI −0.11 to 0.29). The certainty of the evidence is moderate. Five studies on adults provided dichotomous data that we could use for a meta‐analysis on flare‐ups. Web‐based disease monitoring (n = 207/496) is probably equivalent to usual care (n = 150/372) for the occurrence of flare‐ups or relapses in adults with IBD (RR 1.09, 95% CI 0.93 to 1.27). The certainty of the evidence is moderate. One study provided continuous data. Web‐based disease monitoring (n = 465) is probably equivalent to usual care (n = 444) for the occurrence of flare‐ups or relapses in adults with CD (MD 0.00 events, 95% CI −0.06 to 0.06). The certainty of the evidence is moderate. One study provided dichotomous data on flare‐ups in a paediatric population. Web‐based disease monitoring (n = 28/84) may be equivalent to usual care (n = 29/86) for the occurrence of flare‐ups or relapses in children with IBD (RR 0.99, 95% CI 0.65 to 1.51). The certainty of the evidence is low. Four studies, all in adults, provided data on quality of life. Web‐based disease monitoring (n = 594) is probably equivalent to usual care (n = 505) for quality of life in adults with IBD (SMD 0.08, 95% CI −0.04 to 0.20). The certainty of the evidence is moderate. Based on continuous data from one study in adults, we found that web‐based disease monitoring probably leads to slightly higher medication adherence compared to usual care (MD 0.24 points, 95% CI 0.01 to 0.47). The results are of moderate certainty. Based on continuous data from one paediatric study, we found no difference between web‐based disease monitoring and usual care in terms of their effect on medication adherence (MD 0.00, 95% CI −0.63 to 0.63), although the evidence is very uncertain. When we meta‐analysed dichotomous data from two studies on adults, we found no difference between web‐based disease monitoring and usual care in terms of their effect on medication adherence (RR 0.87, 95% CI 0.62 to 1.21), although the evidence is very uncertain. We were unable to draw any conclusions on the effects of web‐based disease monitoring compared to usual care on healthcare access, participant engagement, attendance rate, interactions with healthcare professionals, and cost‐ or time‐effectiveness. The certainty of the evidence is very low. Authors' conclusions The evidence in this review suggests that web‐based disease monitoring is probably no different to standard care in adults when considering disease activity, occurrence of flare‐ups or relapse, and quality of life. There may be no difference in these outcomes in children, but the evidence is limited. Web‐based monitoring probably increases medication adherence slightly compared to usual care. We are uncertain about the effects of web‐based monitoring versus usual care on our other secondary outcomes, and about the effects of the other telehealth interventions included in our review, because the evidence is limited. Further studies comparing web‐based disease monitoring to standard care for the clinical outcomes reported in adults are unlikely to change our conclusions, unless they have longer follow‐up or investigate under‐reported outcomes or populations. Studies with a clearer definition of web‐based monitoring would enhance applicability, enable practical dissemination and replication, and enable alignment with areas identified as important by stakeholders and people affected by IBD.


Description of the condition
Inflammatory bowel disease (IBD) is an umbrella term that encompasses three main disease subtypes that a ect the gastrointestinal tract: ulcerative colitis (UC), Crohn's disease (CD), and IBD unclassified. It a ects approximately 1/1000 people in Western countries, and its incidence is rapidly rising in developing countries (Gasparetto 2013). It has no known cure but can be managed; therefore, it places a huge financial burden on healthcare systems (Ghosh 2015). Approximately 25% of cases are diagnosed before 18 years of age, and the main treatment modalities are pharmacological therapy, dietary therapy, and surgery. Guided management and care can improve disease activity, symptoms, clinical outcomes (e.g. need for surgery), and quality of life (Elkjaer 2012). A er diagnosis, intensive follow-up is required to optimise IBD care, necessitating the need for frequent consultations, at least for some stages of the disease course (Bernstein 2011).

Description of the intervention
IBD telehealth management is the delivery of healthcare management from the healthcare professional to the person with IBD, at a distance (McLean 2011). It includes consulting by phone, instant messenger, video, text message, or webbased services. It can take place live, such as a telephone conversation, or with delayed communication, such as email communication (McLean 2009). During a telehealth session, the person provides information about their condition and their health status. The information becomes electronically available to the clinician or other healthcare professional, and they use it to provide feedback to the person, based on their professional judgement (McLean 2011; Sood 2007). Telehealth can be beneficial for certain subgroups of people with IBD who might face problems with accessing traditional healthcare resources that require their physical presence, such as older people, people from socioeconomically disadvantaged backgrounds, and people with physical or learning disabilities. However, accessing telehealth resources for these subgroups might have its own separate set of challenges (Choi 2014;Forducey 2012; Rimmer 2013).

How the intervention might work
Telehealth consultations work similarly to face-to-face consultations. The only di erence is that any procedure that requires physical presence cannot occur (for example, blood tests or physical examination (Heida 2018)). Therefore, while they might be a useful substitute in cases when face-to-face consultations are not possible or recommended, it is not known how e ective they are compared to face-to-face consultations. The breadth of available telehealth options also means that each option has its own advantages and disadvantages.
Telehealth consultations provide the potential to reduce potential barriers to multi-disciplinary team communication across multiple team members and organisations, and achieve this in real time. This can lead to improved outcomes of consultation. It can facilitate more timely data monitoring, and more timely sharing of questions and concerned voiced by the person with IBD with the entire multidisciplinary team, including the primary care professionals (Cross 2012)

Why it is important to do this review
It is important to systematically review the evidence on the types of remote or telehealth approaches that can be deployed for IBD care, and their e ectiveness. This is particularly relevant given the COVID-19 pandemic, which has necessitated increases in selfand remote-management, which these means can facilitate (Al-Ani 2020). It is also key to ascertain the attributes of remote or telehealth packages, so they can be replicated and disseminated.

O B J E C T I V E S
To identify the communication technologies used for remote healthcare sessions, how they are used, their accessibility, and their potential benefits and drawbacks for people with inflammatory bowel disease.

Types of studies
All published, unpublished, and ongoing randomised controlled trials (RCTs) that compare the use of telecommunication technologies for the management of inflammatory bowel disease (IBD), with face-to-face interventions, or no interventions. Crossover studies and cluster RCTs will be included if identified.
We will not include studies on digital patient information resources (e.g. information on IBD organisation websites, such as Crohn's and Colitis UK), or education resources alone, unless they form part of a wider package that includes an element of telehealth as defined in this review. A concurrent review, focussing on education resources for people with IBD is being conducted separately (Gordon 2021).

Types of participants
People with a confirmed IBD diagnosis of all ages

Types of interventions
We will include studies on IBD management interventions that take place via phone, instant messaging, video, text message, or web-based services, or any other means of remote communication, whether they use live communication (e.g. telephone conversations) or delayed communication (e.g. email communication).
We will consider any control intervention, such as face-to-face interventions or no intervention, including studies in which the control intervention is another telehealth intervention.
We will separately analyse trials that compare the addition of telehealth to traditional consultations, and those that replace traditional consultations with telehealth.

Types of outcome measures
Both dichotomous and continuous outcome measures will be included in the review.

Library
Trusted evidence. Informed decisions. Better health.

Cochrane Database of Systematic Reviews
• Flare-ups or relapses, measured clinically, endoscopically, or histologically • Quality of life at study end, using validated scales or tools

Secondary outcomes
• Number of episodes of accessing health care (outpatient, remote, or inpatient) • Change in disease activity, using a recognised score • Change in quality of life, using a validated tool • Medication adherence • Attendance or engagement rate (number of planned appointments attended, number of planned interactions attended)

Qualitative outcomes
• Programme attributes (technology type, design, cost, user guidance, live contact, and management of delayed contact, contact with other members of the multidisciplinary team, time to response, data security) • Programme requirements (cost, so ware, infrastructure, training needs, user requirement to access -for the person with IBD and healthcare provider)

Search methods for identification of studies Electronic searches
We will search the following sources from the inception of each database to the date of search and will place no restrictions on the language of publication: • We will search the following trial registries by combining terms related to IBD and telemedicine: • Cochrane Gut group Group Specialized Register • ClinicalTrials.gov (www.clinicaltrials.gov); • World Health Organization International Clinical Trials Registry Platform (ICTRP; wwww.who.int/trialsearch).

Searching other resources
As complementary search methods, we will carefully check the references of relevant systematic reviews for studies that we may potentially include in our review. We will also scrutinise the references of studies we included in our review. We will seek unpublished trials by contacting experts in the field, and we will scan the internet and relevant conference abstracts that identified in the search (EMBASE and CENTRAL) , to capture any studies presented, but not yet published in full.
We will attempt to obtain translations of papers when necessary.

Data collection and analysis
We will carry out data collection and analysis according to the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020).

Selection of studies
Two review authors will independently screen the titles and abstracts identified from the literature search. We will discard studies that do not meet the inclusion criteria. We will then obtain the full report of studies that appear to meet our inclusion criteria, or for which there is insu icient information to make a final decision. When these articles are obtained, two review authors will independently assess them to establish whether the studies meet the inclusion criteria. We will resolve disagreements by discussion, and consult with a third review author if resolution is not possible. We will enter studies rejected at this or subsequent stages in the 'Characteristics of excluded studies' tables and record the main reason for exclusion. We will outline the selection process in a PRISMA flowchart (Page 2021).

Data extraction and management
Two authors will independently extract data, using piloted data extraction forms. We will extract relevant data from full-text articles that meet the inclusion criteria. If reported, we will collect information on: • Trial setting: country and number of trial centres • Trial registration details: registration number, date of registration, registered outcomes • Methods: study design, total study duration and date • Participant characteristics: age, sociodemographics, ethnicity, disease status, disease type, diagnostic criteria, and total number • Eligibility criteria: inclusion and exclusion criteria • Intervention and comparator: type of telehealth and control intervention, who is delivering the intervention, required resources for the delivery of the intervention, time to response, who has access to the intervention, data security • Outcomes: outcome definition, unit of measurement, and time of collection • Results: number of participants allocated to each group, missing participants, sample size • Funding source and conflicts of interest

Assessment of risk of bias in included studies
During data extraction, two review authors will independently assess all studies meeting the inclusion criteria for risk of bias, using criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions Version 5.1 (Higgins 2011). The domains that will be assessed are as follows:

Cochrane Database of Systematic Reviews
• Selective reporting (reporting bias) • Other bias We will judge the studies to be at low, high, or unclear risk of bias for each domain assessed, based on the guidance in Higgins 2011.
A er data extraction, the two review authors will compare the extracted data, to discuss and resolve discrepancies before the data are transferred into the 'Characteristics of included studies' table in Review Manager 2020.
We will judge risk of bias for cluster-RCTs as prescribed in Section 16.3.2 of the Cochrane Handbook for Systematic Reviews of Interventions Version 5.1 (Higgins 2011).

Measures of treatment e ect
For the dichotomous outcomes, we will express treatment e ect as risk ratios (RR) with corresponding 95% confidence interval (CI). For continuous outcomes, we will express the treatment e ect as mean di erence (MD) with 95% CI. However, if the studies assess the same continuous outcome di erently, we will estimate the treatment e ect using the standardised mean di erence (SMD). We will present SMDs as standard deviation units and interpret as follows: 0.2 represents a small e ect, 0.5 a moderate e ect, and 0.8 a large e ect.

Unit of analysis issues
The participant will be the unit of analysis. For studies comparing more than two intervention groups, we will make multiple pair-wise comparisons between all possible pairs of intervention groups. To avoid double counts, we will divide shared intervention groups evenly among the comparisons. For dichotomous outcomes, we will divide both the number of events and the total number of participants. For continuous outcomes, we will only divide the total number of participants, and leave the means and standard deviations unchanged.
We will include cross-over studies, but we will only pool their data if they are reported separately before and a er cross-over, and we will only use pre-cross-over data. In the case of cluster-RCTs, we will only use study data if the trial authors have used appropriate statistical methods in taking the clustering e ect into account.
We will exclude cluster-RCTs in a sensitivity analysis to assess their impact on the results.

Dealing with missing data
We will contact authors when there are missing data, or studies have not reported data in su icient detail. We will attempt to estimate missing standard deviations using relevant statistical tools and calculators if studies report standard errors. We will judge studies that fail to report measures of variance at high risk of selective reporting bias.

Assessment of heterogeneity
We will scrutinise studies to ensure that they are clinically homogenous in terms of participants, interventions, comparators, and outcomes. To test for statistical heterogeneity, we will use a Chi test. A P value of less than 0.1 will give an indication of the presence of heterogeneity. Inconsistency will be quantified and represented by the I statistic. We will interpret the thresholds as follows (Higgins 2020): • 0% to 40%: might not be important • 30% to 60%: may represent moderate heterogeneity • 50% to 90%; may represent substantial heterogeneity • 75% to 100%: considerable heterogeneity We will examine possible explanations for heterogeneity when su icient data are available, including factors such as participant characteristics (e.g. age, sex), condition severity, healthcare system, and country.
We will not pool data in a meta-analysis if a considerable degree of statistical heterogeneity is detected (I > 75%). In the case of considerable statistical heterogeneity, we will investigate whether this can be explained on clinical grounds or risk of bias, in which case, we will conduct sensitivity analyses. If we cannot find reasons for the considerable statistical heterogeneity, we will present the results narratively, in detail.

Assessment of reporting biases
Our use of an inclusive search strategy will minimise most reporting biases. We will investigate publication bias using a funnel plot if there are 10 or more studies. We will determine the magnitude of publication bias by visual inspection of the asymmetry of the funnel plot. We will also test funnel plot asymmetry by performing a linear regression of the intervention e ect estimate against its standard error, weighted by the inverse of the variance of the intervention e ect estimate (Egger 1997).

Data synthesis
We will summarise the study characteristics in a narrative synthesis of all the included studies. We will then carry out a metaanalysis if two or more studies have assessed similar populations, interventions, and outcomes. We will analyse studies on paediatric populations, adult populations, and di erent sub-intervention types separately, using Review Manager 5.4 (Review Manager 2020). We will synthesise data using the random-e ects model. We will combine e ect estimates of studies, which report data in a similar way, in the meta-analysis. We will pool risk ratios (RR) for dichotomous outcomes and mean di erences (MD), or standardised mean di erences (SMD) for continuous outcomes, alongside 95% confidence intervals (CI). When we are unable to carry out a meta-analysis (e.g. due to lack of uniformity in data reporting), we will present a narrative summary of the included studies.
We will group qualitative outcomes by the key attributes defined above, and present them in an additional table. We will also present summary descriptive statistics (number of specific remote telehealth solutions used, mean costs, resources, etc.), to help readers ascertain the core attributes across studies. We will present this data narratively.

Subgroup analysis and investigation of heterogeneity
If we detect heterogeneity, we will investigate possible causes, and address them using methods described in Higgins 2020. We will also undertake subgroup analyses of potential e ect modifiers if enough data are available, such as technology type used, age, gender, and disease type. We may also use key qualitative outcome data to inform specific subgroup analysis, or to investigate heterogeneity, if these attributes appear to be the source of such issues.
We recognise that the nature of the studies likely to be included in this review may be capricious and heterogeneous in a number of key clinical and methodological ways that cannot be fully predicted. If such factors are identified and become relevant to ensure the integrity of the analysis, we may need to modify this list. The review authors will fully report these changes.

Sensitivity analysis
Where possible, we plan to undertake sensitivity analyses on the primary outcomes to assess whether the findings of the review are robust to the decisions made during the review process. In particular, we intend to exclude studies at high or unclear risk of bias due to allocation bias and performance bias from analyses that include studies with di erent risk of bias judgements. Where data analyses include studies with reported and estimated standard deviations, we plan to exclude those with estimated standard deviations, to assess whether this a ects the findings of the review. We will investigate whether the choice of model (fixed-e ect versus random-e ects) impacts the results to explore heterogeneity in case of major inconsistencies between the results of the two models.

Summary of findings and assessment of the certainty of the evidence
We will present the main results in a summary of findings table. We will export data for each comparison and primary outcome to GRADEpro so ware so we can assess the evidence for certainty (GRADEpro GDT). We will include all three primary outcomes in the summary of findings table.
Based on risk of bias, inconsistency, imprecision, indirectness, and publication bias, we will rate the certainty of the evidence for each outcome as high, moderate, low, or very low. The GRADE Working Group has defined these ratings as follows: • High certainty: we are very confident that the true e ect lies close to that of the estimate of the e ect.Moderate certainty: we are moderately confident in the e ect estimate; the true e ect is likely to be close to the estimate of the e ect, but there is a possibility that it is substantially di erent. • Low certainty: our confidence in the e ect estimate is limited; the true e ect may be substantially di erent from the estimate of the e ect. • Very low certainty: we have very little confidence in the e ect estimate; the true e ect is likely to be substantially di erent from the estimate of e ect.
We will justify all decisions to downgrade the certainty of the evidence using footnotes, and we will make comments to aid the reader's understanding of the review where necessary.

A C K N O W L E D G E M E N T S
The review authors would also like to thank the following editors and peer referees who provided comments to improve the protocol: Dr. Paul Moayyedi (Contact Edtor), Dr. Yuhong Yuan (Managing Editor), Dr. Berkeley Limketka (Peer Reviewer), an anonymous doctor (Peer Reviewer), and Sara Blake (Consumer Reviewer), and Victoria Pennick, for copy editing the protocol.
Dr. Yuhong Yuan (Information Specialist at the Cochrane Gut Group) designed the search strategies.