The Cost-Effectiveness of Digital Health Interventions on the Management of Cardiovascular Diseases: Systematic Review

Background With the advancement in information technology and mobile internet, digital health interventions (DHIs) are improving the care of cardiovascular diseases (CVDs). The impact of DHIs on cost-effective management of CVDs has been examined using the decision analytic model–based health technology assessment approach. Objective The aim of this study was to perform a systematic review of the decision analytic model–based studies evaluating the cost-effectiveness of DHIs on the management of CVDs. Methods A literature review was conducted in Medline, Embase, Cumulative Index to Nursing and Allied Health Literature Complete, PsycINFO, Scopus, Web of Science, Center for Review and Dissemination, and Institute for IEEE Xplore between 2001 and 2018. Studies were included if the following criteria were met: (1) English articles, (2) DHIs that promoted or delivered clinical interventions and had an impact on patients’ cardiovascular conditions, (3) studies that were modeling works with health economic outcomes of DHIs for CVDs, (4) studies that had a comparative group for assessment, and (5) full economic evaluations including a cost-effectiveness analysis, cost-utility analysis, cost-benefit analysis, and cost-consequence analysis. The primary outcome collected was the cost-effectiveness of the DHIs, presented by incremental cost per additional quality-adjusted life year (QALY). The quality of each included study was evaluated using the Consolidated Health Economic Evaluation Reporting Standards. Results A total of 14 studies met the defined criteria and were included in the review. Among the included studies, heart failure (7/14, 50%) and stroke (4/14, 29%) were two of the most frequent CVDs that were managed by DHIs. A total of 9 (64%) studies were published between 2015 and 2018 and 5 (36%) published between 2011 and 2014. The time horizon was ≤1 year in 3 studies (21%), >1 year in 10 studies (71%), and 1 study (7%) did not declare the time frame. The types of devices or technologies used to deliver the health interventions were short message service (1/14, 7%), telephone support (1/14, 7%), mobile app (1/14, 7%), video conferencing system (5/14, 36%), digital transmission of physiologic data (telemonitoring; 5/14, 36%), and wearable medical device (1/14, 7%). The DHIs gained higher QALYs with cost saving in 43% (6/14) of studies and gained QALYs at a higher cost at acceptable incremental cost-effectiveness ratio (ICER) in 57% (8/14) of studies. The studies were classified as excellent (0/14, 0%), good (9/14, 64%), moderate (4/14, 29%), and low (1/14, 7%) quality. Conclusions This study is the first systematic review of decision analytic model–based cost-effectiveness analyses of DHIs in the management of CVDs. Most of the identified studies were published recently, and the majority of the studies were good quality cost-effectiveness analyses with an adequate duration of time frame. All the included studies found the DHIs to be cost-effective.


Background and Objectives
Provide an explicit statement of the broader context for the study. Present the study question and its relevance for health policy or practice decisions. 4 Target Population & subgroups Describe characteristics of the base case population and subgroups analysed, including why they were chosen. 5 Setting & Location State relevant aspects of the system(s) in which the decision(s) need(s) to be made. 6 Study Perspective Describe the perspective of the study and relate this to the costs being evaluated. 7 Comparators Describe the interventions or strategies being compared and state why they were chosen. 8 Time Horizon State the time horizon(s) over which costs and consequences are being evaluated and say why appropriate. 9 Discount Rate Report the choice of discount rate(s) used for costs and outcomes and say why appropriate. 10

Choice of Health Outcomes
Describe what outcomes were used as the measure(s) of benefit in the evaluation and their relevance for the type of analysis performed. 11

Measurement of Effectiveness
Single study-based estimates: Describe fully the design features of the single effectiveness study and why the single study was a sufficient source of clinical effectiveness data. Synthesis-based estimates: Describe fully the methods used for identification of included studies and synthesis of clinical effectiveness data. 12

Measurement & Valuation of Preferencebased Outcomes
If applicable, describe the population and methods used to elicit preferences for outcomes. associated with the alternative interventions. Describe primary or secondary research methods for valuing each resource item in terms of its unit cost. Describe any adjustments made to approximate to opportunity costs. Model-based economic evaluation: Describe approaches and data sources used to estimate resource use associated with model health states. Describe primary or secondary research methods for valuing each resource item in terms of its unit cost. Describe any adjustments made to approximate to opportunity costs. 14  For each intervention, report mean values for the main categories of estimated costs and outcomes of interest, as well as mean differences between the comparator groups. If applicable, report incremental cost-effectiveness ratios. 20 Characterizing Uncertainty Single study-based economic evaluation: Describe the effects of sampling uncertainty for the estimated incremental cost and incremental effectiveness parameters, together with the impact of methodological assumptions (such as discount rate, study perspective).
Model-based economic evaluation: Describe the effects on the results of uncertainty for all input parameters, and uncertainty related to the structure of the model and assumptions. 21 Characterizing Heterogeneity If applicable, report differences in costs, outcomes, or cost effectiveness that can be explained by variations between subgroups of patients with different baseline characteristics or other observed variability in effects that are not reducible by more information.

Measurement of Effectiveness
The effectiveness was retrieved from literature reviews, including the results from a randomized controlled trial and four meta-analysis. 1

Measurement & Valuation of Preference-based Outcomes
The baseline utility and the utility of 6-month text messages strategy were estimated from a health survey conducted in TEXT ME trial. The utility associated with a myocardial infarction event and stroke were retrieved from the literature review with referencing. The key simplifying assumption was that individuals could only have one of either an of myocardial infarction or a stroke, after which they moved to and then remained in the history of secondary event state until death.

Analytic Methods
Probabilistic sensitivity analysis with 1,000 Monte Carlo simulations was conducted and five scenario analysis were considered. 1

Study Parameters
Input values and ranges were specified. 1

Incremental Costs & Outcomes
TEXT ME was expected to lead to 563 fewer myocardial infarctions, 361 fewer strokes and 1143 additional QALYs, with an overall saving of $10.56 million for the health system over the patients' lifetimes. 1

Characterizing Uncertainty
Parameter uncertainty had little effect on the conclusion that TEXT ME was cost-effective, which was shown in a cost-effectiveness plane. TEXT ME was cost-saving in all the individual scenario analysis, whilst it was cost-effective when all scenario run simultaneously.

Measurement & Valuation of Preference-based Outcomes
The utility values for each NYHA class were retrieved from a manufacturer database and constructed using the Dutch utility weights. 1

Estimating Resources & Costs
The personnel-and hospital-related costs were from the Dutch health care costing manual. And the costs of the telemonitoring system were acquired from the manufacturer and adjusted in accordance with the market research. (3) The hospitalization costs were assumed to be treatment arm-independent, but NYHA classdependent. (4) The utility values were assumed to connect to the severity of the disease.

Analytic Methods
Probabilistic sensitivity analysis, threshold analysis, subgroup analysis, and scenario analysis were conducted. 1

Study Parameters
Input values and ranges were specified. 1

1
Characterizing Uncertainty The scenario including telenurse cost inputs in nurse telephone support yielded results that were slightly different from those from the scenario excluding this cost, when comparing all NYHA classes of severity.

1
The probabilistic sensitivity analysis suggested that there was a very low probability of home telemonitoring being cost-effective when nurse telephone support was available for the management of patients with chronic heart failure.

Characterizing Heterogeneity
Nurse telephone support dominated home telemonitoring, compared with usual care, in all NYHA classes except NYHA IV.

Measurement of Effectiveness
Authors mentioned the inputs were retrieved from literature reviews and statistics center. But it was unclear from which study each parameter estimate was obtained (each parameter was not referenced, respectively). 0

Measurement & Valuation of Preference-based Outcomes
The measurement of preference-based outcomes was not specified. 0

Estimating Resources & Costs
The costs of the tool worth 100 euros. The healthcare cost and non-sanitary cost were retrieved from the Ministry of Health, Social Policy and Equality. The assumption was not clearly specified. 0 Analytic Methods A univariate sensitivity analysis was carried out to assess the robustness of the results. 1

Study Parameters
Study parameters and the ranges were not specified. 0

Incremental Costs & Outcomes
The cost for introduction of the app was €19.012 per patient and that for no introduction was €28.315. The ICER was €9.303/QALY. However, the health outcome was not clearly specified and how to calculate the ICER was not clear. The measurement of effectiveness was not specified. 0

Measurement & Valuation of Preference-based Outcomes
The mother's utility was retrieved from a study with referencing but the utility of the newborn children was not clearly specified.

Estimating Resources & Costs
The estimated resources were presented but not clearly specified. 0

Currency, Price Date & Conversion
The currency, price date and conversion were not clearly specified. 0 Choice of Model Decision tree model 1

Assumptions
(1) The authors assumed that telemedicine had a 97% sensitivity and 96% specificity rate. (2) It was assumed that a further pregnancy with a normal outcome would occur, with a delay of one year in about 50% of women. (3) All 'standard-risk' women screened over 15 months could be reviewed in 11 working weeks. (4) It was assumed that each patient in the model would incur a cost to the health service. (5) It was assumed that on average each child would live to their expected lifetime. 1

Analytic Methods
Bootstrapping was used to stabilize the mean and to generate 95% confidence intervals around the mean value for the skewed cost and effectiveness data. One-way and probabilistic sensitivity analysis were also conducted. 1

Study Parameters
Input values and ranges were specified. 1

Incremental Costs & Outcomes
In the base-case assumption, the arm that all women received telemedicine screening was dominant. 1

Characterizing Uncertainty
The ICER of one-way sensitivity analysis were clearly presented. The probability of a screening strategy with telemedicine being cost-effective was nearly 100% at a willingness-to-pay (WTP) of £20,000 per QALY. Choice of Model Type of decision-analytic model was not specified. But a figure of the model was provided. 1

Assumptions
(1) It's assumed a telestroke system with 8 spokes (range 6-12), each of which had 12 telestroke consults per year (range 6-30), and 1 hub with 4 neurologists (range 3-5) rotating telestroke calls from either the office/hospital (1 shared hospitalbased telestroke unit) or the home (each with a home telestroke unit). (2) Patients with an modified Rankin Score (mRS) score (a 6-point disability scale) of 0 were assumed to be discharged to home, while those with scores between 1 and 5 could be discharged to home, a rehabilitation facility, or a nursing home. (3) In patients who were discharged to rehabilitation, the score was assumed to improve by 1 point at 90 days; the initial mRS score was assumed to not change in patients who were discharged either to home or to a skilled nursing facility.
1 Analytic Methods One-way sensitivity analysis and probabilistic sensitivity analysis were conducted. 1

Assumptions
(1) Acute ischemic stroke patients could only transfer from a less severe to a more severe health state or remain in the same health state at each cycle. (2) Stroke treatments between a telestroke network and no network differed only during the initial hospitalization for acute ischemic stroke, not after discharge from acute care. (3) Incremental effectiveness associated with treatments in a telestroke network only resulted from IV thrombolysis or endovascular stroke therapy during the initial hospitalization for the first-time acute ischemic stroke. (4) There was no difference in stroke-related mortality between patients with and without IV thrombolysis, and between patients with and without endovascular stroke therapy during hospitalization. (5) Rate of recurrent stroke was the same regardless of the treatment received during the initial hospitalization for acute ischemic stroke.
1 Analytic Methods One-way and two-way sensitivity analysis were conducted. 1

Study Parameters
The study parameters were specified. 1

Incremental Costs & Outcomes
Patients treated in a telestroke network incurred $1436 lower costs and gained 0.02 QALYs over a lifetime. 1

Characterizing Uncertainty
The one-way sensitivity analyses showed that the results were robust, with a telestroke network being the dominant strategy in all scenarios except when the spoke-to-hub transfer rate was varied. When varying the transfer rate from 0% to 100%, the model showed that the network remained a dominant strategy when the transfer rate increased to 60% and remained cost-effective with a WTP threshold of $50,000 per QALY when the transfer rate increased to 90%.

Measurement of Effectiveness
The effectiveness data was from a single trial without clear description of the trial. 0

Measurement & Valuation of Preference-based Outcomes
The utility values were estimated from literatures, with individual referencing. 1

Estimating Resources & Costs
The cost and reimbursement data were retrieved from encounter-level financial dataset. Medicare Severity Diagnosis-Related Groups 61-66 were also used. This data included emergency department, rehabilitation, and treatment for acute ischemic stroke with or without telestroke. The fixed cost of telestroke was retrieved from the local telestroke network.

Assumptions
(1) The model assumed that after presenting at the emergency department of this hospital, the patient 1 could receive tissue plasminogen activator or not. (2) Spoke facilities were assumed to be responsible for 0%, 50%, and 100% of implementation expenses. (3) The mRS scores based on discharge location were assumed as followed: home = 0-1, rehabilitation = 2-3, skilled nursing facility = 4-5, and death = 6.

Analytic Methods
One-way sensitivity analysis, probabilistic sensitivity analysis, and scenario analysis were conducted. Subgroup analysis for each National Institute of Health Stroke Scale severity category was also complete. 1

Study Parameters
Model inputs and ranges were adequately described. 1

Measurement & Valuation of Preference-based Outcomes
The utility for discharged HF patients was from literature reviews., with detailed approach for data retrieval.

Estimating Resources & Costs
Costs of interventions other than usual care were broken down into the costs of the device, monitoring cost, and medical care cost and were estimated using bottom-up costing methods for the National Health Service Foundation trusts for 250 HF patients. The inpatient admission cost for hospitalization was from National Health Service references costs.  Analytic Methods Scenario analyses, threshold analyses, and probabilistic analysis were considered. 1

Study Parameters
Model inputs were adequately described. 1

Incremental Costs & Outcomes
Compared with usual care, telemonitoring had an estimated ICER of £11,873/QALY, whereas structured telephone support with human-to-human contact had an ICER of £228,035/QALY against telemonitoring. structured telephone support with human-to-machine interface was dominated by usual care. 1

Characterizing Uncertainty
The cost-effectiveness acceptability curve was presented. The chance of telemonitoring being cost-1 effective at WTP of £20,000/QALY was 40%. The scenario analysis performed using higher costs of telemonitoring (£215/month) indicated telemonitoring was dominated by structured telephone support with human-to-human contact. Threshold analysis suggested that the monthly cost of telemonitoring has to be higher than £390 to have an ICER greater than £20,000/QALY against structured telephone support with human-to-human contact. (1) It was assumed that preventing a hospitalization prevented an inpatient and a two-month posthospitalization increase in mortality. (2) It was assumed the benefit of the CardioMems device would continue lifelong.
1 Analytic Methods One-way sensitivity analysis, probabilistic sensitivity analysis, and the subgroup analysis were conducted. 1

Study Parameters
The study parameters and ranges were specified. Analytic Methods One-way sensitivity analysis, probabilistic sensitivity analysis, and scenario analysis were conducted. 1

Study Parameters
The study parameters were specified.

Characterizing Uncertainty
The results of one-way sensitivity analysis did not dramatically change the results. In the scenario analysis including staff costs, the ICER was £22,342/ QALY. At the WTP of £20,000/ QALY, the probability of the device being cost-effective was 97.6%

Measurement & Valuation of Preference-based Outcomes
The measurement and valuation of preference-based outcomes were not specified. 0

Estimating Resources & Costs
Direct and indirect cost were both included. Each strategy was clearly specified with the estimated costs to be assigned. (1) It was assumed all patients received an ICD after the infection was cleared in the base-case analysis; (2) non-SCA mortality was assumed to be equal in all strategies; (3) patients who did not need ICD were assumed to have a reduced total mortality; (4) it was assumed that use of either a WCD or an ICD further impacted a patient's quality of life.
1 Analytic Methods Both one-way and two-way sensitivity analyses were conducted. 1

Study Parameters
The study parameters were not clearly specified. 0

Incremental Costs & Outcomes
The ICER of the WCD strategy was $20,300 per lifeyear or $26,436 per QALY compared to discharge home without a WCD. Discharge to a skilled nursing facility and in-hospital monitoring resulted in higher costs and worse clinical outcomes.

1
Characterizing Uncertainty The incremental cost-effectiveness ratio was as low as $15,392/QALY if the WCD successfully terminated 95% of SCA events and exceeded the $50,000/QALY WTP threshold if the efficacy was 69%. The WCD strategy remained cost-effective, assuming 5.6% 2-month SCA risk, as long as the time to reimplantation was at least 2 weeks. References: