Home Telehealth for Chronic Disease Management: Selected Findings of a Narrative Synthesis

Composition of the Review Team

The review team included both researchers and “knowledge users,” including representatives of several British Columbia health authorities. These individuals were active in defining research priorities and providing consultation to the researchers. Their contributions to the review team ensured that research remained accessible and relevant to policy makers and others with direct involvement in healthcare planning and delivery.

Defining the Study Question

The term “chronic disease” encompasses a vast range of conditions. In order to reduce heterogeneity and increase the usefulness of this research to knowledge users, we limited our literature search to six chronic diseases selected by knowledge users in early meetings: asthma, cardiovascular disease, chronic obstructive pulmonary disease, diabetes, renal disease, and stroke. These conditions were identified as especially relevant to provincial health priorities. However, we recognize that our use of these diseases to represent chronic disease in its entirety may limit the applicability of our findings. The exclusion of arthritis, mental health disorders, and hypertension may be particularly controversial (here they were excluded except where comorbid with one of the six previously named conditions). The precise reasons for omitting these conditions from our analysis—a decision that entailed prolonged debate—varied. A common factor, however, was their perceived dissimilarity with the other target conditions in regard to management needs, pathophysiology, and/or structure of care delivery services.

For the purposes of this project, “telehealth” was defined as ‘the use of information and communications technologies to deliver health services and transmit health information over both long and short distances.” Examples of such information and communications technologies include videoconferencing, Web conferencing, text messaging, and digital image transmission. Consultation with our knowledge users led us to exclude interventions that used only stationary telephones.

Database Searches

The review team worked with a qualified and experienced health librarian to develop a search strategy. The EMBASE, Medline, CINAHL, Web of Science, and PAIS databases were searched for material relevant to the priority research questions and with a focus on one of the chronic diseases listed above. The search was built around three concepts: home telehealth, chronic disease, and economics. The exact search terms used differed in each database in order to take full advantage of variations in indexing methods. Searches excluded material in languages other than English or French and were restricted to the time period 2005–2010. Academic publications, white papers, and gray literature were all considered eligible for inclusion, provided an original research element was present. Systematic reviews were also included. Case studies, technology reviews, and articles without abstracts were excluded. Full details of the search strategy can be obtained from the authors upon request.

Critical Appraisal

The research team categorized studies according to the Oxford Centre for Evidence-Based Medicine (OCEBM) 2011 levels of evidence. (For a comprehensive overview of this system, please refer to Howick et al. ⁷ ) In this system, studies are classified as Level 1–5 according to their ability to provide “the likely best evidence.” ⁷ Level 1 and Level 5 studies were excluded from this review through our original screening process. All remaining studies were therefore Level 2 (individual randomized trial), Level 3 (nonrandomized controlled cohort/follow-up study), or Level 4 (case-series, case-control studies, or historically controlled studies).

The OCEBM Levels were used in conjunction with an instrument to assess quality of execution and reporting (Appendix 1). This instrument was developed by the research team but drew on the framework presented in Zaza et al., ⁸ although time and resource limitations necessitated a more abbreviated process. Sampling method, dropout rates, blinding, and interpretation of data were among the criteria considered. A study was downgraded one OCEBM level if it received a “weak” (≤3) execution/reporting score. This practice is sanctioned by the Oxford Level guidelines.

Data Extraction

In the first stage of screening, studies were classified as qualitative, quantitative/mixed methods, or systematic review. A different review process was used for each subgroup of results. Systematic reviews were retained for reference purposes but not subjected to systematic data extraction. Qualitative studies were reviewed for information on patient and provider experiences. This information was summarized through thematic analysis. A standardized coding sheet (Appendix 2) for use with articles on quantitative and mixed-methods studies was developed after preliminary review of the evidence. Information extracted included data on population characteristics, intervention design, clinical and economic outcomes, and patient and provider experiences.

Scope of Literature

In total, 34,353 references were retrieved by the literature search. After removal of duplicates, citations without abstracts, and documents with no apparent relevance to the research questions, 1,075 articles remained. One hundred eighty-six of these articles (representing 147 discrete studies) met our inclusion criteria and were reviewed in detail (Fig. 1). In this article, discussion is limited to the 137 quantitative and mixed-methods studies. All subsequent analyses (percentages, etc.) are based on this subset.

OPEN IN VIEWER

Fig. 1.

Search process.

Strength of the Evidence

Eighty-seven articles (64%) were classified as OCEBM Level 2, 16 as Level 3 (12%), and 34 as Level 4 (25%). Nineteen articles were subsequently downgraded a level for low execution/reporting scores. Most articles received an execution/reporting rating of strong (29 articles) or moderate (89 articles). The subgroups of studies of asthma, heart failure, and type 2 diabetes were consistently well executed and reported; the percentages of articles within these subgroups that were given strong or moderate ratings were 100%, 93%, and 90%, respectively.

Studies showed considerable variation in length and sample size. Duration ranged from 0.5 to 84 months, with 6- and 12-month durations the most common. Size of sample ranged from 7 to 6,184, with a mean of 268 and a median of 101. (Calculations are based on the number of participants included in the final analysis. When this number was not available [11 instances], the number enrolled was used. When multiple publications reported different outcomes for the same sample, the sample was counted once.) The largest studies were found in the type 2 diabetes and cardiovascular disease subgroups (mean sample sizes of 406 and 256, respectively).

Distribution by Chronic Disease

The majority of the literature focused on populations with type 2 diabetes or cardiovascular disease (particularly heart failure). These disease categories were further broken down when it became apparent that they included at least two subgroups of populations with distinct management needs (type 1 versus type 2 diabetes and heart failure versus coronary artery disease, respectively). There was a notable dearth of studies on renal disease patients (Fig. 2).

OPEN IN VIEWER

Fig. 2.

Articles retrieved. COPD, chronic obstructive pulmonary disease; CVD-CAD, cardiovascular disease (coronary artery disease); CVD-HF, cardiovascular disease (heart failure); T1 DM, type 1 diabetes mellitus; T2 DM, type 2 diabetes mellitus.

Outcomes

We collected data on patient outcomes, provider outcomes, and system outcomes. In this publication, we focus on patient outcomes: disease state, quality of life, and/or symptoms. Tables 1 –7 give a complete list of the 101 articles that reported these outcomes.

Limitations

This review had several limitations. One is its currency. Our initial search was limited to publications from 2005 to 2010 (inclusive). In July 2012, we conducted a final follow-up search for highly relevant studies published after 2010. This search was intended to fill gaps in the literature and to allow us to incorporate any major developments that had occurred since our initial searches. However, it was not as comprehensive as the initial searches, and the data extraction process was abbreviated. It is possible that a more thorough review of post-2010 material would have altered our findings, particularly given the speed at which home telehealth has developed. We observed a marked increase in the rate of publication. In our initial search, the number of publications that we accepted into our review was, on average, 34.5 per year searched. The follow-up search yielded an average of 86.5 publications per year. Although it is possible that this is due to improvements in indexing, or perhaps more widespread adoption of preferred terms, the simplest and most probable explanation is increased research.

Another limitation of this review is the fact that members of the review team worked independently to screen, code, and assess the quality of the articles retrieved. This limitation was mitigated to the extent possible: clearly defined criteria were used; reviewers consulted with each other on an ongoing basis to discuss questionable articles; and periodic exercises in double-coding were used to improved inter-rater consistency. However, this remains a limitation.

Finally, it is possible that not all relevant articles were included in this review. Although we strove to be as comprehensive as possible in our initial search, references without abstracts were ultimately excluded, as were those for which full text could not be retrieved.

Analysis

In the studies we reviewed, home telehealth was as or more effective than usual care in improving disease state, quality of life, and symptoms in patients with chronic disease. This was true despite wide variations in study design, target population, and technology of choice. Patients using telehealth experienced significantly greater improvements than control group patients in disease state, quality of life, or symptoms in 38.8% of comparisons from Level 2 studies and over 30% of comparisons from all studies.

However, we cannot confidently identify the circumstances under which such improvements are most likely to occur. We found relatively little research in this area; most studies focused not on how or why telehealth worked, but on whether it worked. The exceptions were found primarily in studies of diabetes, the chronic disease with the most extensive body of evidence, and particularly in publications associated with the 5-year IDEATel program. This may reflect the large initial sample sizes needed to achieve acceptable confidence intervals for subgroup and multiple regression analyses.

Researchers who have investigated factors influencing the success of telehealth interventions have considered such variables as uptake, ⁹³ the intensity of the intervention, ^61,78 and the moderating effects of sex, ¹⁹ baseline glycated hemoglobin, ^15,52 self-efficacy, ¹¹⁰ and depression. ¹¹¹ More extensive use of this type of analysis has the potential to yield both clinical and economic benefits. The results would enhance policy makers' ability to apply existing research by enabling them to make a more informed assessment of the generalizability of a study's results to their target population.

Another limitation of existing research is the frequent exclusion of patients with comorbidities, who comprise roughly 30% of all Canadian seniors (>65 years of age) with chronic conditions and account for 40% of senior healthcare use. The exclusion of these individuals from study populations may result in under- or overestimates of the benefits of home telehealth to individuals and the healthcare system. We also observed a high level of variation in the level of details that authors provided in their descriptions of “usual care”; in some articles, usual care was undefined. More thorough reporting of usual care also has the potential to enhance our understanding of the mechanisms by which home telehealth interventions are successful.