Chronic diseases are prolonged conditions that normally do not improve with time and are rarely cured completely (5). They may cause premature deaths, decrease quality of life (QoL) of individuals, and have a negative economic impact on the individuals' families and society (Reference Patra, Popova, Rehm, Bondy, Flint and Giesbrecht36). According to the World Health Organization, noncommunicable diseases, such as cardiovascular diseases, diabetes, obesity, cancer, and respiratory diseases, currently account for 59 percent of the 57 million annual deaths and 46 percent of the global burden of disease (48). A 2007 study stated that the total costs of chronic disease in 2003 in the United States alone was US$1,324 trillion dollars (treatment expenditures = US$277 billion and lost productivity = US$1,047 trillion dollars) (Reference DeVol and Bedroussian15).
Advancements in treatment for chronic diseases have resulted in reduced length of hospital stay, and in some cases, the avoidance of hospital visits, so the demand for home care services has increased (Reference Romanow41). Healthcare providers can deliver home care services by visiting a patient's home or by using information and communication technology, also known as home telehealth. Home telehealth is a subset of telehealth that brings healthcare delivery to the home environment by connecting the patient with medical professionals. It is not intended to replace health professional care or visits, but rather to enhance the level of care (7).
The research objectives were to systematically review the current literature on the cost-effectiveness home telehealth and to provide a framework for economic evaluations of home telehealth to assist future studies. The economic review is limited to three chronic conditions that are the most studied with respect to home telehealth: diabetes, congestive heart failure (CHF), and chronic obstructive pulmonary disease (COPD). In developed countries, CHF is diagnosed in 1–2 percent of the general population (Reference Bundkirchen and Schwinger6), and an estimated more than 180 million people worldwide have diabetes (46), and approximately 210 million people globally have COPD (47). Usual care involves follow-up by a primary care physician or specialist after patient discharge from hospital.
METHODS
A protocol for the review of economic evaluations was written a priori and was followed in detail. A framework for economic evaluations of home telehealth programs is also described, followed by an example of the framework.
Literature Search Strategy
The following bibliographic databases were searched through the Ovid interface: Medline, Medline Daily Update, Medline In-Process & Other Non-Indexed Citations, BIOSIS Previews, and EMBASE. Parallel searches were run in PubMed, Cochrane Library, CRD Health Technology Assessment (HTA) database, CRD NHS Economic Evaluation Database (NHS EED), and Health Economic Evaluations Database (HEED). The search strategy included usual controlled vocabulary, such as the National Library of Medicine's MeSH (Medical Subject Headings), and keywords. The main search concept was home telehealth (including variations such as telehome care, home telecare, e-health, home telemedicine); methodological filters were applied to limit retrieval to cost analyses and other economic studies. OVID AutoAlerts and PubMed MyNCBI were set up to send monthly updates for new literature; monthly searches were also performed in Cochrane Library, HEED, and CRD. Results were limited to articles published from 1998 to 2008, and language restrictions were not imposed.
Selection Criteria
To be eligible, studies had to concern patients with at least one chronic disease and had to have home telehealth as the intervention compared with usual care. Economic evaluations, such as cost minimization analyses (CMA), cost-effectiveness analyses (CEA), cost utility analyses (CUA), or cost benefit analyses (CBA) were included. In addition, cost analyses were included if the assumption was made that home telehealth was at least as effective as usual care.
Selection Method
Two reviewers (D.C., K.C.) independently scanned the titles and abstracts that were identified through the electronic literature search. The full-text articles of citations deemed to be potentially relevant were reviewed using the selection criteria. The reviewers compared their selections of included studies and any disagreements were discussed and resolved by consensus.
Data Extraction Strategy
Data from each included study were extracted independently by two reviewers (D.C., K.C.), using a structured data extraction form. Country of origin, disease area, study design, patient population, comparators, resources measured, form of analysis, study perspective, time horizon, and study outcomes were noted. Any disagreements were discussed and resolved by consensus.
Quality Assessment
Study quality was assessed by one reviewer (D.C.) using a ten-point scale, which was developed based on criteria identified in three previous articles relating to the conduct of economic evaluation in telemedicine (Reference Mair, Haycox and Williams29;Reference McIntosh and Cairns31;Reference Whitten, Mair, Haycox, May, Williams and Hellmich45). The ten-point checklist is similar to that suggested by Drummond and colleagues (Reference Drummond, O'Brien, Stoddart and Torrance16) for assessing study quality for economic evaluations in general, but in this context is more specific to the requirements for studies in home telehealth. Questions were phrased for a yes/no answer, and the number of questions with a positive response was recorded for each study. This number should not be interpreted as a quality score, as the importance of each question is not equal. For example, the requirement to discount is limited and less important than the need that the data source be a study with high quality design. The questions and their description are found in Table 1.
Table 1. Quality Assessment Questionnaire
RESULTS
Quantity of Research Available
The literature search identified 1,567 studies. The total number of studies reviewed includes both duplicates (i.e., the same paper identified through different databases) and duplicate publications (i.e., the same studies reported in more than one publication). After a review of 150 articles, 22 studies were found to be relevant for inclusion in the economic review. Figure 1 presents the QUOROM flow diagram detailing the process of study selection.
Figure 1. Selection of included studies.
Study Characteristics
Country of origin. Most studies (n = 17) in our review were from the United States (Reference Benatar, Bondmass, Ghitelman and Avitall1;Reference Bondmass, Bolger, Castro and Avitall4;Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10–Reference Dansky, Palmer, Shea and Bowles12;Reference Galbreath, Krasuski and Smith20;Reference Heidenreich, Ruggerio and Massie23;Reference Jerant, Azari and Nesbitt25–Reference Laramee, Levinsky, Sargent, Ross and Callas27;Reference Myers, Grant, Lugn, Holbert and Kvedar32–Reference Noel and Vogel34;Reference Riegel, Carlson, Glaser and Romero38;Reference Riegel, Carlson, Kopp, LePetri, Glaser and Unger39;Reference Southard, Southard and Nuckolls43;Reference Vaccaro, Cherry, Harper and O'Connell44). Among the rest, one study was from Germany (Reference Biermann, Dietrich, Rihl and Standl3), one was from Spain (Reference Farrero, Escarrabill, Prats, Maderal and Manresa19), one was from Italy (Reference Maiolo, Mohamed, Fiorani and de Larenzo28), one was from the United Kingdom (Reference Mason, Young and New30), and one was from Canada (Reference Pare, Sicotte, St-Jules and Gauthier35).
Disease Area. Most selected studies (n = 12) focused on home telehealth for patients with CHF (Reference Benatar, Bondmass, Ghitelman and Avitall1;Reference Bondmass, Bolger, Castro and Avitall4;Reference Galbreath, Krasuski and Smith20;Reference Heidenreich, Ruggerio and Massie23;Reference Jerant, Azari and Nesbitt25;Reference Laramee, Levinsky, Sargent, Ross and Callas27;Reference Myers, Grant, Lugn, Holbert and Kvedar32;Reference Nobel and Norman33;Reference Riegel, Carlson, Glaser and Romero38;Reference Riegel, Carlson, Kopp, LePetri, Glaser and Unger39;Reference Southard, Southard and Nuckolls43;Reference Vaccaro, Cherry, Harper and O'Connell44). Five studies focused on patients with diabetes (Reference Biermann, Dietrich, Rihl and Standl3;Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10–Reference Dansky, Palmer, Shea and Bowles12;Reference Mason, Young and New30), and three focused on patients with COPD (Reference Farrero, Escarrabill, Prats, Maderal and Manresa19;Reference Maiolo, Mohamed, Fiorani and de Larenzo28;Reference Pare, Sicotte, St-Jules and Gauthier35). The two other studies included patients with a variety of diseases (all included CHF, diabetes or COPD) (Reference Johnston, Wheeler, Deuser and Sousa26;Reference Noel and Vogel34).
Patient Population. Studies that focused primarily on CHF imposed numerous inclusion criteria. One study (Reference Southard, Southard and Nuckolls43) allowed patients with coronary heart disease as well as those with heart failure. Three studies required patients to have moderate or severe CHF (Reference Myers, Grant, Lugn, Holbert and Kvedar32;Reference Nobel and Norman33;Reference Vaccaro, Cherry, Harper and O'Connell44). Six studies required either recent discharge from hospital or frequent emergency department visits (Reference Berger2;Reference Laramee, Levinsky, Sargent, Ross and Callas27;Reference Myers, Grant, Lugn, Holbert and Kvedar32;Reference Riegel, Carlson, Glaser and Romero38;Reference Riegel, Carlson, Kopp, LePetri, Glaser and Unger39;Reference Vaccaro, Cherry, Harper and O'Connell44). One study (Reference Jerant, Azari and Nesbitt25) required patients to be at least 40 years old, while another (Reference Riegel, Carlson, Glaser and Romero38) studied Hispanics only.
All five studies on diabetes included patients with different baseline criteria: adolescents (Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10), patients on intensive insulin therapy (Reference Biermann, Dietrich, Rihl and Standl3), indigent or economically disadvantaged patients with diabetes (Reference Cherry, Moffatt, Rodriguez and Dryden11), elderly patients with diabetes who were recently discharged from a hospital (Reference Dansky, Palmer, Shea and Bowles12), and patients with type 2 diabetes (Reference Mason, Young and New30).
Two of the three COPD studies (Reference Farrero, Escarrabill, Prats, Maderal and Manresa19;Reference Maiolo, Mohamed, Fiorani and de Larenzo28) involved patients who required long-term oxygen therapy, whereas the other COPD study (Reference Pare, Sicotte, St-Jules and Gauthier35) focused on patients with severe disease after discharge from their first hospitalization.
For studies that examined multiple chronic diseases, one (Reference Noel and Vogel34) required patients with at least three chronic or complex conditions and who received home care for at least 6 months. In the other study (Reference Johnston, Wheeler, Deuser and Sousa26), patients were eligible if they had any one of the following conditions: CHF, diabetes, COPD, stroke, cancer, or require wound care.
Comparators. Usual care varied in terms of whether it involved a specified organized home care or another support program (Reference Benatar, Bondmass, Ghitelman and Avitall1;Reference Bondmass, Bolger, Castro and Avitall4;Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10;Reference Dansky, Palmer, Shea and Bowles12;Reference Johnston, Wheeler, Deuser and Sousa26;Reference Myers, Grant, Lugn, Holbert and Kvedar32–Reference Pare, Sicotte, St-Jules and Gauthier35) or care as directed by the physician, which may or may not have included home care (Reference Biermann, Dietrich, Rihl and Standl3;Reference Cherry, Moffatt, Rodriguez and Dryden11;Reference Farrero, Escarrabill, Prats, Maderal and Manresa19;Reference Galbreath, Krasuski and Smith20;Reference Heidenreich, Ruggerio and Massie23;Reference Jerant, Azari and Nesbitt25;Reference Laramee, Levinsky, Sargent, Ross and Callas27;Reference Maiolo, Mohamed, Fiorani and de Larenzo28;Reference Mason, Young and New30;Reference Riegel, Carlson, Glaser and Romero38;Reference Riegel, Carlson, Kopp, LePetri, Glaser and Unger39;Reference Southard, Southard and Nuckolls43;Reference Vaccaro, Cherry, Harper and O'Connell44). The home telehealth interventions tended to be complex, often with more than one facet of patient management changed in the intervention arm of the study. All but one home telehealth strategy (Reference Mason, Young and New30) involved augmented interactions between the patient and a nurse. The majority of studies on home telehealth for patients with CHF involved regular telephone monitoring of patients by nurses (Reference Benatar, Bondmass, Ghitelman and Avitall1;Reference Bondmass, Bolger, Castro and Avitall4;Reference Galbreath, Krasuski and Smith20;Reference Jerant, Azari and Nesbitt25;Reference Laramee, Levinsky, Sargent, Ross and Callas27;Reference Riegel, Carlson, Glaser and Romero38;Reference Riegel, Carlson, Kopp, LePetri, Glaser and Unger39), with one also incorporating video interactions (Reference Jerant, Azari and Nesbitt25). Other studies involved a system whereby patient information was sent either by telephone (Reference Heidenreich, Ruggerio and Massie23) or through a computer (Reference Myers, Grant, Lugn, Holbert and Kvedar32;Reference Nobel and Norman33;Reference Sanchez, Harewood and Petersen42;Reference Southard, Southard and Nuckolls43) to the nursing staff to monitor the patient's condition. Three of the five diabetes studies (Reference Biermann, Dietrich, Rihl and Standl3;Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10;Reference Cherry, Moffatt, Rodriguez and Dryden11) evaluated different systems where patient data were relayed to a nurse, physician, or diabetes center followed by advice provided to the patient by telephone. One study (Reference Dansky, Palmer, Shea and Bowles12) evaluated a program of video visits between patients and home care nurses, and another study evaluated a call station managed by non-healthcare professionals (Reference Mason, Young and New30). Table 2 provides details of comparators in the included studies.
Table 2. Comparators of Included Studies
Form of Analysis. All selected studies were a cost-analysis, except for one which was a CUA (Reference Mason, Young and New30). In the CUA, the long-term costs and utilities were derived from a previous U.S. study and then transformed to the UK setting, with no further information provided on the costs included. None of the studies included a formal assessment of either utilities or the program impact on long-term outcomes.
Resources Included. Most studies involved an assessment of the costs associated with specific healthcare resources, such as hospitalizations, primary care, and emergency department visits and included costs required to set up a home telehealth program. The majority did not involve consideration of all healthcare resources, and only two (Reference Biermann, Dietrich, Rihl and Standl3;Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10) incorporated patient-based costs. Four studies did not report the healthcare resources included (Reference Heidenreich, Ruggerio and Massie23;Reference Mason, Young and New30;Reference Nobel and Norman33;Reference Noel and Vogel34). Seven studies failed to adequately incorporate the costs of home telehealth (Reference Cherry, Moffatt, Rodriguez and Dryden11;Reference Galbreath, Krasuski and Smith20;Reference Heidenreich, Ruggerio and Massie23;Reference Jerant, Azari and Nesbitt25;Reference Nobel and Norman33;Reference Riegel, Carlson, Glaser and Romero38;Reference Vaccaro, Cherry, Harper and O'Connell44).
Study Perspective. Study perspective was deduced as it was unspecified in the majority of studies. Two studies adopted a societal perspective (Reference Biermann, Dietrich, Rihl and Standl3;Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10). Ten studies (Reference Benatar, Bondmass, Ghitelman and Avitall1;Reference Bondmass, Bolger, Castro and Avitall4;Reference Heidenreich, Ruggerio and Massie23;Reference Jerant, Azari and Nesbitt25;Reference Myers, Grant, Lugn, Holbert and Kvedar32–Reference Noel and Vogel34;Reference Riegel, Carlson, Glaser and Romero38;Reference Riegel, Carlson, Kopp, LePetri, Glaser and Unger39;Reference Vaccaro, Cherry, Harper and O'Connell44), adopted the perspective of a specific health insurance provider. The remaining studies in the review adopted a healthcare system perspective.
Study Design. Fourteen studies were based on data from randomized controlled trials (RCTs) (Reference Benatar, Bondmass, Ghitelman and Avitall1;Reference Biermann, Dietrich, Rihl and Standl3;Reference Chase, Pearson, Wightman, Roberts, Oderberg and Garg10;Reference Dansky, Palmer, Shea and Bowles12;Reference DeBusk, Miller and Parker13;Reference Farrero, Escarrabill, Prats, Maderal and Manresa19;Reference Galbreath, Krasuski and Smith20;Reference Jerant, Azari and Nesbitt25;Reference Johnston, Wheeler, Deuser and Sousa26;Reference Mason, Young and New30;Reference Noel and Vogel34;Reference Riegel, Carlson, Glaser and Romero38;Reference Riegel, Carlson, Kopp, LePetri, Glaser and Unger39;Reference Southard, Southard and Nuckolls43). Four studies were based on data from a case-control study (Reference Myers, Grant, Lugn, Holbert and Kvedar32;Reference Nobel and Norman33;Reference Pare, Sicotte, St-Jules and Gauthier35;Reference Vaccaro, Cherry, Harper and O'Connell44), and four were based on data from a pre–post study (Reference Bondmass, Bolger, Castro and Avitall4;Reference Cherry, Moffatt, Rodriguez and Dryden11;Reference Heidenreich, Ruggerio and Massie23;Reference Maiolo, Mohamed, Fiorani and de Larenzo28).
Time Horizon. The time horizon across all studies ranged from 2 (Reference Dansky, Palmer, Shea and Bowles12;Reference Myers, Grant, Lugn, Holbert and Kvedar32) to 18 (Reference Galbreath, Krasuski and Smith20) months.
Quality Assessment. Based on the quality assessment questionnaire derived from published economic evaluations of home telehealth, most studies were considered to be of poor quality. Only seven studies had affirmative answers to more than half of the ten items on the checklist (Table 3). However, even the studies with a high proportion of affirmative responses had major methodological weaknesses. All studies had a positive response to question 5, which focused on whether the study compared a strategy of home telehealth with an alternative strategy. This item was an inclusion criterion for this review. Several studies also had a positive score for question 1 (appropriate study question; 17/22 studies) and question 3 (RCT or good quality observational study; 17/22 studies). The questions that had the least adherence to good quality were question 2 (appropriate study perspective; 3/22 studies), question 9 (marginal analysis; 4/22 studies), and question 10 (appropriate sensitivity analysis; 3/22 studies).
Table 3. Study Outcomes and Quality Assessment
Data Synthesis and Analysis
A formal meta-analysis was not feasible due to variations in the patient populations, study designs and interventions and comparators. Instead, the results in each selected study were reviewed and summarized qualitatively. In one CUA (Reference Mason, Young and New30), a diabetes call station managed by non-healthcare professionals was associated with higher costs leading an incremental cost per quality-adjusted life-years (QALYs) of £43,300. In another study, there was no difference in costs between telephone management by a nurse manager and usual care for patients with CHF (Reference Galbreath, Krasuski and Smith20). In the remaining twenty studies, the home telehealth strategies were found to lead to reductions in the costs of the healthcare resources included from a healthcare system or insurance provider perspective. Study outcomes for each selected study are described in Table 3.
Framework for Economic Evaluation
Focus. An analysis of the economic impact of home telehealth must focus on the incremental costs and health benefits associated with the application of the program to a population of patients, rather than to an individual patient. Such evaluations would allow an assessment of the impact of the program as a whole by including costs from both the patient and population levels. This would allow evaluation of costs based on different patient population sizes, as required for marginal analysis.
Study Perspective. Studies must specify and justify the perspective from which the home telehealth programs and health resource use are measured. Societal, healthcare system, third-party, and patient perspectives each have a unique focus that is reflected in the included costs.
Direct Costs to Be Included. Telehealth program specific costs: Cost of program administration, program delivery, and program capital costs (i.e., costs of technology required for the delivery of home telehealth) must be measured. The costs of technology used over a period of time must be amortized over the technology's useful life.
Home healthcare costs: These costs comprise the volume of nursing and other homecare visits and contacts.
Healthcare costs: All pertinent healthcare resources, such as family physician visits, specialist outpatient appointments, emergency department visits, hospitalizations, diagnostic tests and investigations, medications, and specialized equipment (including prostheses) must be calculated.
Patient borne costs: The potential financial impact on patients and their families and caregivers must be measured. Some items to consider are travel and other directly borne costs, medication costs, and family caregiver time.
Indirect Costs to be Included. Indirect costs are required for studies measuring costs from a societal perspective, but are more challenging to measure. Costs in this category may include the patient or caregiver's productivity losses as a result of disease management or travelling time to the patient's residence incurred by the healthcare provider.
Outcomes. An economic evaluation requires an assessment of incremental costs and incremental outcomes of each healthcare program evaluated. Thus, the majority of published studies are not economic evaluations of home telehealth and cannot assist in determining whether a treatment is justifiable based on the impact on costs and treatment outcomes. Many studies interpret a reduced use of healthcare resources as evidence of improved outcomes. Healthcare resources use may be limited due to fewer home healthcare contacts with home telehealth, meaning reduced frequency of access to other services and not necessarily a reduced need for these services. Studies should, therefore, have either clinical outcomes (which may be surrogate outcomes such as disease markers) or patient QoL. If home telehealth is more costly compared with usual care, then studies must take the form of either cost-effectiveness or cost utility analysis using these outcomes.
Quality of Life. A cost-utility analysis of home telehealth programs must include a formal assessment of patients' QoL to measure the QALY and an estimation of any short-term effects on utility due to the intervention.
Effectiveness. A cost-effectiveness analysis of home telehealth programs must consider collecting data on clinical outcomes associated with the particular disease studied, such as event rates and deaths. In many instances, an evaluation of home telehealth programs may have a short time horizon, which will not facilitate a demonstration of differences in long-term clinical outcomes. The studies should then consider inclusion of surrogate markers as indicators of no differences in clinical outcomes, such as glycemic control (HbA1c) for diabetes, forced expiratory volume in one second (FEV1) for COPD, and systolic blood pressure for CHF.
Study Population. A vast array of patients participate in a home telehealth program. Studies with diverse patient populations increase the external validity of their outcomes. If there is substantial heterogeneity due to the study populations, subgroup analyses based on patient characteristics may be carried out.
Comparator. The home telehealth program must be compared with an appropriate alternative strategy that represents the existing form of health service delivery for the study population. Some examples include regular home care visits by a healthcare provider or primary care visits or longer hospital stays.
Marginal analysis. In an economic evaluation, it is necessary to assess the marginal costs and benefits associated with health interventions that relate to the additional costs and outcomes associated with the interventions being considered. When evaluating home telehealth, it is generally required to evaluate the implementation of a program. Marginal analysis should focus on the additional costs and outcomes of program implementation and not necessarily on the additional cost and outcomes on a per patient basis. However, it would be necessary to conduct such analysis using alternative estimates of the number of participants in the program as the marginal costs per patient will be much lower than the forecasted average costs. Subsequently, average costs are likely to fall the more patients covered within the program.
Sensitivity Analysis. An economic evaluation must include at least one sensitivity analysis to determine the robustness of the study findings based on the assumptions made. Sensitivity analyses are done by varying the underlying assumptions over a possible range of possible values.
Application of the Economic Evaluation Framework
In the supplementary section, which can be viewed online at www.journals.cambridge.org/thc, an illustrative application of the framework is provided. The application relates to an evaluation of a telehomecare demonstrator project, “EMPcare@home”(40) conducted in New Brunswick, Canada. The evaluation of the telehealth demonstrator project was used to measure whether the combination of telehomecare, timely staff intervention and an enhanced patient education program produces a better QoL for patients, is accepted by patients and health professional and reduces the need for hospital care. The objective was to highlight how such analyses can be conducted and what additional data would be required.
DISCUSSION
The literature search identified twenty-two studies relevant for inclusion in the systematic review. No attempt was made to quantitatively synthesize the selected studies. Instead, data were summarized and appraised to identify common results and the related strengths and weaknesses.
Most studies in our economic review found home telehealth to be cost saving from healthcare system and insurance provider perspectives. Conclusions must be qualified as the quality of the studies in terms of economic evaluations was poor. The studies were also heterogeneous possibly due to diverse study populations, interventions, and the healthcare systems in which they are based, so it remains a challenge to make an informed decision on resource allocation. Several studies were published in 1999 (Reference Bondmass, Bolger, Castro and Avitall4;Reference Dansky, Palmer, Shea and Bowles12;Reference Myers, Grant, Lugn, Holbert and Kvedar32) or in 2000 (Reference Johnston, Wheeler, Deuser and Sousa26;Reference Noel and Vogel34). It is possible given the likely decrease in technology costs that telemedicine costs were higher in those studies compared with costs in the more recent studies, so the cost-effectiveness of the home telehealth program may be confounded by the study's publication year. Numerous studies in this review did not measure simultaneous changes in outcomes and costs, so they did not measure the cost-effectiveness of a home telehealth program. Finally, no home telehealth study in the published literature involved a formal assessment of health-related utilities or the impact of the program on long-term outcomes.
The economic evaluation framework presented advocates that an assessment of a home telehealth program include costs at both the patient and population levels to facilitate assessment based on different size populations. The sample analysis demonstrates how the economic evaluation framework of home telehealth could be implemented. There are limitations to how the economic evaluation can be done, given the absence of data on all healthcare resources, patient costs, quality of life, and clinical outcomes.
The critical appraisal checklist for economic evaluations by Drummond et al. helps readers to identify strengths and weaknesses of published health economic studies (Reference Drummond, O'Brien, Stoddart and Torrance16). The Canadian Agency for Drugs and Technologies' Guidelines for the Economic Evaluation of Health Technologies: Canada is further guidance for individuals conducting an economic evaluation, especially in the Canadian healthcare system context (8). Our framework is consistent with the above-mentioned guidelines and can serve as a guide for future studies that evaluate the economic impact of home telehealth programs. Mair et al. produced ten recommendations for the design of economic evaluations of telemedicine based on a quality assessment checklist designed for pharmacoeconomics studies (Reference Mair, Haycox and Williams29), and authors McIntosh and Cairns also developed a framework for economic evaluations of telemedicine (Reference McIntosh and Cairns31). The checklist by Mair et al. does not outline in great detail the types of outcomes and costs to consider in an economic evaluation of telemedicine. Moreover, McIntosh and Cairns's framework was limited to direct costs (e.g., equipment, medical personnel in telemedicine, and cost of treatment) and clinical and nonhealth outcomes (e.g., length of waiting, time to diagnosis and improved education and reassurance). Healthcare resource use, such as number of primary care visits to the general practitioner, was also mentioned (Reference McIntosh and Cairns31). The current framework presented is specific to home telehealth programs and can serve as a guide for all types of economic evaluations from various perspectives because it explicitly describes both the direct and indirect costs for consideration. In addition to the clinical and healthcare resource use outcomes, our framework also discusses the quality-of-life outcomes necessary for a CUA.
There are some limitations to our study. Some studies in the economic review had small sample sizes and a lack of information on patient characteristics, clinical outcomes, and study perspectives and, overall, were of poor quality. There are no published economic reviews specific to home telehealth on which to make an informed policy decision. Past systematic reviews on the cost-effectiveness of telemedicine interventions found the existing evidence on which to draw a conclusion to be very limited (Reference Mair, Haycox and Williams29;Reference Whitten, Mair, Haycox, May, Williams and Hellmich45). Hailey et al. also found high-quality studies assessing the clinical and economic evidence in telemedicine to be scarce (Reference Hailey, Ohinmaa and Roine21).
Future research should measure the economic impact of home telehealth programs on more diverse patient populations with chronic diseases to increase the external validity of their results and help identify those patients who can benefit most from home telehealth interventions and compare real-time versus asynchronous technologies to provide additional insight on the most effective disease management strategy for chronic diseases. Several studies interpreted decreased healthcare resource use as improved clinical outcomes. Studies should always include clinical outcomes, such as a disease marker or patient QoL to determine whether reduced use of health services is a result of limited access to health services versus the need for these services. Finally, a standardized approach to the evaluation of home telehealth should be developed to increase the quality of studies and amount of evidence available.
Home care is an instrumental part of the chronic disease management model, and home telehealth is an extension of healthcare delivery in a patient's home environment. Our economic review reported that home telehealth was cost saving from both the healthcare system and insurance provider perspectives, but the overall quality of the original research was low. A framework for economic evaluations in home telehealth was also presented as a guide to measure its economic impact, and an illustration demonstrated the constraints with conducting an economic evaluation with a lack of data on healthcare resources, patient costs, quality of life and clinical outcomes. If future studies adopted our proposed economic evaluation framework for home telehealth programs, then their relevance to healthcare decision making would be greatly improved.
SUPPLEMENTARY MATERIALS
Supplementary section and Supplementary Tables 1–3 (www.journals.cambridge.org/thc)
CONTACT INFORMATION
Julie Polisena, MSc (juliep@cadth.ca), Research Officer, HTA Directorate, Canadian Agency for Drugs and Technologies in Health, 600–865 Carling Avenue, Ottawa, Ontario K1S 5S8, Canada
Doug Coyle, PhD (dcoyle@uottawa.ca), Professor and Director of Graduate Studies, Department of Epidemiology & Community Medicine, University of Ottawa, 451 Smyth Road, Ottawa, Ontario K1H 8M5, Canada
Kathryn Coyle, MSc (kathrynogrady@hotmail.com), Associate, Coyle Consultancy, 29 Ella Street, Ottawa, Ontario K1S 2S3, Canada
Sarah McGill, MLIS (sarahm@cadth.ca), Information Specialist, Canadian Agency for Drugs and Technologies in Health, 600–865 Carling Avenue, Ottawa, Ontario K1S 5S8, Canada
