Digital interventions have the potential to improve health care delivery by improving effectiveness, accessibility, the ability to reach patients, cost-effectiveness, and personalization [1-3]. The prevalent use of digital interventions and devices is constantly creating new possibilities for connecting people and providing services [4-6]. There are a variety of digital interventions used in a health care setting. In this review, we will focus on those digital interventions that facilitate and deliver health care services to patients with chronic conditions [7]. As patients directly interact with these interventions, they are the most relevant for the patient experience. The interventions can include remote appointments [8], mobile apps, wearable devices, and remote monitoring systems [9,10], electronic health records [11], and patients’ portals and online support groups [12].

Chronic conditions often require long-term management. For the patient, this can mean navigating services, interacting with health care professionals (HCPs), and adhering to treatments, which can all together create significant demands and load to the patient [13]. These demands can be referred to as treatment burden, which is defined as both the workload caused by the self-management of the condition and the impact the treatment has on the patient’s function and well-being [13]. In addition to the negative impact on patients’ well-being, high treatment burden is also associated with poor adherence and worse disease outcomes [14].

In this review, the terms treatment burden and burden of treatment are used interchangeably, consistent with previous literature. However, we draw specifically on the burden of treatment [14] theory, which provides a framework to explain the interactions between health care systems, patients, and their treatment burden [15]. This theoretical framework informed both our eligibility criteria and our interpretation of findings.

There are many ways digital interventions may affect treatment burden. Telehealth and self-management interventions may shift the responsibility of the care to the patient, causing additional demands [16,17], digital systems may be inaccessible or difficult to use [18], and patients may struggle with digital stress [19]. When properly designed, digital interventions can also make the health care processes simpler and more effective, reducing the treatment burden for patients. For example, remote care can reduce the time and costs needed for traveling to medical appointments [8], and self-management interventions can make self-management easier and more motivating [20,21].

Recent systematic reviews about the current digital interventions have shown limited efficacy and rather small benefits to patients [22,23]. Recent umbrella reviews show that most digital interventions in health care are mobile- or computer-based, the largest targeted condition group is mental illnesses, and most outcomes are focused on effectiveness [24-26]. The burden on the patient is often not considered or measured in the development and research of these interventions. On the other hand, the systematic reviews we have identified investigating the treatment burden on patients with chronic conditions [14,26-30] do not focus on digital interventions. We observe that there is a lack of systematic reviews combining these 3 concepts: treatment burden, digital interventions, and chronic conditions.

After the publication of the protocol for this review [27], 2 relevant reviews were published [28,29]. However, one of them was a scoping review, with a specific focus on telehealth [28], and the other was focused only on multimorbidity and the availability and benefits of digital tools [29]. The current systematic review will expand on these reviews by widening the scope to include all digital interventions that facilitate and deliver healthcare services to patients with a variety of chronic conditions. The summary of the related work is presented in Table 1.

The aim of this review is to evaluate the evidence of how digital interventions impact the treatment burden experienced by people with chronic conditions, and to explore how these effects vary across different conditions and intervention types. In this review, we use the term “impact” to refer to reported or perceived effects of digital interventions on treatment burden, rather than measured causal outcomes.

Based on the identified gap in research discussed above, we have defined 2 primary research questions (RQs):

To conduct this systematic review, we followed the University of Oulu ethics process as defined in the guidelines from the Ethics Committee of Human Sciences [35]. According to the guidelines, an ethics board review was not needed.

This systematic review used a convergent design for systematic mixed studies reviews [26] to synthesize published articles describing the impact of digital interventions on patients’ treatment burden. This review is reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) 2020 [36]. We registered the protocol on the International Prospective Register of Systematic Reviews (PROSPERO; CRD42023477605) and published the protocol describing our systematic review methodology [27].

The systematic search for articles was conducted, and screening was started before publishing the protocol and the detailed search strategy and inclusion criteria of this review are reported in the previously published protocol [27]. The search terms are presented in Multimedia Appendix 1. The search for articles published between January 1, 2013, and October 16, 2023, was conducted from bibliographical databases PubMed, Scopus, Web of Science, ACM, PubMed Central, and CINAHL. A supplementary search was conducted by hand searching the reference lists in systematic literature reviews and scoping reviews that were found during the literature searches. The search was done again on June 17, 2025, for papers published after October 16, 2023, to include more recent literature in the final results.

We included original publications written in English and accepted in peer-reviewed journals or conference proceedings published after 2013. Although the use of digital interventions in health care for chronic conditions extends further back [37], studies before 2013 mostly refer to research prototypes not available to the public. Since then, digital health technology has developed quickly. Therefore, we decided to restrict our search to studies published after 2013 to cover all relevant publications. Qualitative, quantitative, and mixed methods articles and conference proceedings are included, whereas reviews, protocols, and book chapters were excluded.

The eligibility criteria align with the PICO (population, intervention, comparison, and outcome) framework [38]. Population: studies were eligible if participants were patients with chronic conditions, their caregivers, or HCPs involved in their care. Studies focusing solely on mental health conditions were excluded. Intervention: only studies with digital intervention for treatment or management of chronic conditions were included. Comparison: studies were neither included nor excluded based on the comparator. Outcome: only studies where the outcome was a treatment burden for the patients were included.

We first imported the titles and abstracts of all the studies we retrieved from our search results to Covidence, a web-based screening tool [39], where duplicate records were removed. In the first stage of screening, 2 independent reviewers screened all titles and abstracts for inclusion. The Cohen κ of the consensus between reviewers (0.41 between reviewers MP and OK, and 0.59 between reviewers MP and PK) indicated moderate agreement. In the second stage, 2 reviewers screened and selected full-text papers for inclusion. The Cohen κ of the consensus in the second stage (0.32 between reviewers MP and OK, and 0.43 between reviewers MP and PK) indicated fair to moderate agreement. The Cohen κ score was calculated with the Covidence screening tool for all screened studies. Disagreements were resolved by discussion. For example, in a disagreement about whether the study included treatment burden outcomes or not, the reviewers checked the study together. The paper was included if reviewers were able to identify relevant outcomes.

A total of 36 articles were included in the data extraction stage of the initial search. From the data extraction stage, 1 study was moved back to full-text screening and removed as the intervention did not match the inclusion criteria. After data extraction, one study was removed due to a lack of participants. From the supplementary search, 5 articles were identified. One of these articles was a duplicate and therefore removed. From the second search in 2025, 8 articles were included. A total of 46 papers were included.

A data extraction form was designed using Covidence before starting the extraction. The predefined form included questions on population characteristics, study design, aims, intervention characteristics, and main results. Study designs were classified based on the tool from Grimes and Schulz [40]. The included studies covered a range of chronic conditions with varying levels of initial treatment burden, independent of the use of digital interventions. The initial treatment burden level of each condition was classified by the researchers who extracted the data (see Table 2). If explicitly stated by the original authors, the reported level of the burden was used for the classification. Otherwise, the treatment burden level was the burden to the patients. We considered a condition with high burden when the treatment is highly disruptive to the patient’s everyday life, involving more than daily medication and occasional check-ups. In cases where there was not enough information on the treatment burden level, or there were high variations between patients, we marked the treatment burden level as “Not classified.”

The data extraction was piloted beforehand; author MP extracted 5 articles, and authors PK and OK extracted 2 articles each. The authors then had a meeting to go through these extractions and suggest improvements to the data extraction table. Several adjustment needs were identified, and the data extraction form was adjusted accordingly. This included, for example, adding missing study design types and separate boxes for qualitative and quantitative results. Data extraction was performed independently by 2 reviewers. The researchers scheduled regular meetings to discuss the process. Most disagreements in consensus were resolved by reviewing the papers again. A disagreement about the treatment burden level of asthma had to be resolved during a meeting, which was decided to be low or medium after discussing together.

The quality assessment was conducted in parallel to the data extraction. The Critical Appraisal tools from the Joanna Briggs Institute were used [41]. Quality assessment results are presented in the appendix (see Multimedia Appendix 2). All studies reviewed with the tool for qualitative studies had a low risk of bias. For these studies, the only topic that caused concern for bias was the lack of self-reflection, which was present in 17 out of 38 studies. Some individual studies also had unclear reporting of ethical statements, theoretical statements, and the voices of participants. For studies reviewed with an analytical cross-sectional tool, we found 2 low-risk-of-bias studies. For randomized controlled trials, all studies were high-quality, but due to the nature of the intervention, blinding of personnel and participants was missing from these studies. For the studies reviewed with the tool for quasi-experimental studies, the experiments did not have comparison groups, which causes potential bias, but other aspects of the studies were rated high quality. Overall, while the analysis revealed some sources of bias, the reviewers agreed that the included studies had credible findings.

A convergent integrated approach was used for synthesis and integration [42]. In this approach, quantitative data is converted into qualitative data, and the qualitative and quantitative evidence are integrated and analyzed together [43]. We converted the quantitative findings into qualitative descriptions by narratively interpreting the effects reported in the quantitative studies. Examples of this conversion process are presented in Table S2 in Multimedia Appendix 3. After the conversion, both qualitative data and converted quantitative data were presented in a single dataset. With NVivo software (Lumivero), the dataset was coded line by line. An inductive coding approach was used [44]. Coded data were arranged into categories based on the digital intervention type and the reported impact on treatment burden. Examples of coding are presented in Table S3 in Multimedia Appendix 3. The coding and categorization were conducted by the first author.

The included studies used different types of digital interventions. We have categorized these interventions into 4 groups based on the components included in the interventions and the type of intervention. These groups are telehealth, informational resources, self-management tools, and facilitated tools. These categories are presented in Figure 2.

Telehealth refers to remote appointment portals and other tools for contacting HCPs. Informational resources cover all mobile- or web-based informational resources and decision aids provided to the patient. The self-management tools category includes all web- and mobile-based tools that the patient uses for self-management of the disease, without the help of a care team. The facilitated tools category includes the tools where an HCP or a care team is facilitating the use of the tool (for example, remote monitoring, patient portals, and digital care pathways). Some of these categories overlap. For example, digital care pathways often include many different types of components, such as remote appointment portals and informational resources.

To answer RQ1, we will present the impacts on treatment burden for each category. For the informational resources, self-management tools, and facilitated tools categories, we will discuss the impact on the conditions with high treatment burden and conditions with medium or low treatment burden. For the telehealth category, there were no studies about conditions with medium or low treatment burden, so the results will only cover conditions with high treatment burden and multimorbidity.

To answer RQ2, we categorized all the identified impacts into 4 main ways digital interventions can support patients with their treatment burden. These include providing support for improving access to health care, self-management, improving user experience, and addressing relevant concerns. Next, we will go through these categories in more detail.

This review investigated the impact of digital health care on the treatment burden of patients with chronic conditions. Digital interventions appear to impact on the treatment burden of people with chronic illness, for those whose illness generate high to medium burden the impact was positive on all technology categories, and some negative impact related mostly to time-consuming tools, accessibility issues, and difficulties of using tools The impact in people with low and medium treatment burden related to their chronic illness was mostly positive in other technology categories, but self-management tools had negative impact on the treatment burden, as they were often considered unnecessary and added responsibility to the patients.

We found that informational support had very little negative impact and many positive impacts. This may imply that it is mostly a suitable way to provide information to patients, and there are no major concerns in terms of treatment burden. Telehealth and facilitated tools seem to divide opinions, and self-management tools without support from HCPs had a relatively more negative impact.

Overall, digital interventions can help patients with their treatment burden by supporting self-management, improving access to health care, improving patients’ experience, and addressing relevant concerns.

Our findings were consistent with previous literature [8,20,21] in terms of finding the main positive impacts on supporting self-management, supporting communication with HCPs, and saving time and travel. We also found issues with accessibility and difficulty of using the tools and increased responsibility to patients, as in the previous literature on the negative impact of digital interventions [16-18].

The Burden of Treatment theory framework [15] about the relationships between patients’ treatment burden and the health care system in general is also consistent with our findings about digital interventions. The quality of digital interventions and the support available from HCPs directly impact the treatment burden experienced by the patient. Digital interventions have the potential to both reduce and increase the patient’s health care workload, which in turn affects the patient’s capacity to take responsibility for their care.

This is a novel review that covers the impact of digital health care on the treatment burden of patients. Our novel findings were related to the differences in impact on different types of digital interventions. We found that there are negative impacts for the group of patients who have conditions with low or medium treatment burden using self-management tools. This may be caused by people with higher treatment burdens having more needs for support and therefore gaining more benefit from interventions supporting the treatment. Self-management tools are also more problematic than facilitated tools in terms of patients’ treatment burden, as the sole responsibility is on the patient, and professional help is not available.

We have found novel findings demonstrating how treatment burden is a relevant concept for future digital health care research. Our findings show how the initial treatment burden of the patient’s condition and the type of digital intervention both have an impact on the benefits and harms of using the intervention and, therefore, may impact the overall success and usage of the intervention.

This review also has some limitations. First, our search was restricted to studies published in English. This may have excluded relevant papers from developing regions where English is not the primary language for scientific publication. Although we attempted to conduct thorough searches from 6 different databases, the included studies were mostly from Europe and North America, and from countries with high economic and digitalization levels. A recent scoping review found that digital interventions are harder to integrate in low-resource communities, and the experience of treatment burden can be different for marginalized groups [28]. Patients of low socioeconomic status are less likely to be offered digital interventions [91], and often live in areas that have poor internet infrastructure, causing more barriers and difficulties accessing digital interventions [92]. There is a need for future research focusing on the equity and treatment burden experience of structurally marginalized groups and patients living in low-resource areas. Programs should be developed to respond to the needs of these patient groups and address the accessibility issues.

Second, most of the included studies were investigating conditions with high burden or multimorbidity. We included a variety of conditions in the search terms, but especially the papers about telehealth and informational resources were missing research on conditions with low or medium treatment burden. Therefore, our findings about telehealth and informational resources may not apply to this patient population.

We found that some intervention types have more concerns than others, and some are not suitable for all patient groups. This can be useful information for the designers of digital interventions to design better interventions. HCPs can also target suitable interventions for the right patient populations. We have also identified a research gap concerning conditions with low or medium treatment burden. Treatment burden is also a relevant concept for these groups of patients, and future research should address this gap.

The results of this study indicate that digital interventions mostly support patients with chronic conditions with their treatment burden. The main benefits are support with self-management, informational support, and easier ways to contact HCPs. The main concerns were accessibility issues, time-consuming tools, and causing fear and anxiety. Future research about conditions with low or medium initial treatment burden would be needed to better understand the impact of digital solutions on patients’ treatment burden.