Systematic review of model-based analyses reporting the cost-effectiveness and cost-utility of cardiovascular disease management programs

Abstract

Background:

The reported cost effectiveness of cardiovascular disease management programs (CVD-MPs) is highly variable, potentially leading to different funding decisions. This systematic review evaluates published modeled analyses to compare study methods and quality.

Methods:

Articles were included if an incremental cost-effectiveness ratio (ICER) or cost-utility ratio (ICUR) was reported, it is a multi-component intervention designed to manage or prevent a cardiovascular disease condition, and it addressed all domains specified in the American Heart Association Taxonomy for Disease Management. Nine articles (reporting 10 clinical outcomes) were included.

Results:

Eight cost-utility and two cost-effectiveness analyses targeted hypertension (n=4), coronary heart disease (n=2), coronary heart disease plus stoke (n=1), heart failure (n=2) and hyperlipidemia (n=1). Study perspectives included the healthcare system (n=5), societal and fund holders (n=1), a third party payer (n=3), or was not explicitly stated (n=1). All analyses were modeled based on interventions of one to two years’ duration. Time horizon ranged from two years (n=1), 10 years (n=1) and lifetime (n=8). Model structures included Markov model (n=8), ‘decision analytic models’ (n=1), or was not explicitly stated (n=1). Considerable variation was observed in clinical and economic assumptions and reporting practices. Of all ICERs/ICURs reported, including those of subgroups (n=16), four were above a US$50,000 acceptability threshold, six were below and six were dominant.

Conclusion:

The majority of CVD-MPs was reported to have favorable economic outcomes, but 25% were at unacceptably high cost for the outcomes. Use of standardized reporting tools should increase transparency and inform what drives the cost-effectiveness of CVD-MPs.

Keywords

Cardiovascular disease disease management cost effectiveness economic evaluation Markov model

Introduction

The concept of disease management started to spread in the US healthcare industry in the mid-1990s with growing expectations that a disease management intervention can be cost-saving when applied to chronic diseases.¹ Despite this expectation and the amount invested, the evidence on the ability of disease management to actually reduce or save healthcare costs without compromising the quality of care has been highly variable. The inconclusiveness is further compounded by considerable variation in economic evaluation methods, inadequate costing methods observed in economic evaluations in general,^2,3 and by the lack of a widely-accepted definition of disease management.⁴

Cardiovascular disease (CVD) is the largest contributor to the chronic disease cluster,⁵ and accounts for a substantial share of national health expenditures in the USA,⁶ the EU⁷ and Australia.⁸ Both CVD and its risk factors have often been the target of prevention and management efforts globally, and heart failure, coronary artery disease, hypertension and hyperlipidemia collectively accounted for 25% to 38% of the chronic conditions targeted by disease management interventions identified in systematic reviews^9,10 and meta-analyses.¹¹

Despite substantial investment made in recent decades, the cost-effectiveness of cardiovascular disease management programs (CVD-MPs) is inconsistent. Understanding under what circumstances CVD-MP is or is not cost-effective is an important determinant of efficient resource allocation. However, considerable methodological or reporting variation has hindered the assessment of what could have impacted the reported cost-effectiveness, let alone the reliable estimate of the cost-effectiveness of CVD-MPs.

Given the paucity of data that could collectively inform policy decisions and urgency to standardize cost analyses of these widespread and commonly applied programs, our objectives are to: 1) identify model-based economic evaluation of CVD-MPs and compare study characteristics, the methods used and cost-effectiveness reported across identified studies; 2) assess the degree of transparency against established methodological or reporting principles.

Methods

The Preferred Reporting System for Systematic Reviews and Meta-Analysis (PRISMA) strategy was followed to ensure systematic selection of studies.¹² Electronic databases (Medline, CINAHL, the NHS Economic Evaluations Database (NHS-EED)) were searched from 1990 to May 2013. Studies published before 1990 were not considered, as it was not until 1995 that disease management appeared frequently in the medical literature.¹³ We used MeSH headings and keywords to identify modeled analyses of CVD-MPs reporting incremental cost-effectiveness ratio (ICER) or cost-utility ratio (ICUR). The complete search strategies for Medline Ovid, CINAHL and the NHS-EED are provided in online in Supplementary Material Appendices 1–3.

The Taxonomy for Disease Management,¹⁴ a system of classification developed by the American Heart Association (AHA) was used as the operational definition of a CVD-MP, which includes eight domains: patient population, recipient, intervention content, delivery personnel, method of communication, intensity and complexity, environment and outcome measures. Two reviewers independently screened the titles and abstracts of the identified citations according to exclusion and inclusion criteria (Table 1). Disagreements between reviewers were adjudicated by a third author. After the articles were excluded based on titles/abstracts, the full-texts of the remainder were reviewed using a screening algorithm (Table 2), namely, if an article met any of the conditions in Group A in the order listed, it was excluded. The remaining articles were further screened for the conditions in Groups B and C. The algorithm was used to ensure only one reason for exclusion was assigned per article.

Table 1.

Inclusion and exclusion criteria.

Inclusion criteria (if all of the following met)	Exclusion criteria (if any of the following met)
1. Intervention was designed to manage or prevent a CVD using a systematic approach to care and employing multiple component intervention.	6. Studies not published in English.
	7. Design/protocol paper, methodological paper, (systematic) review, meta-analysis, commentaries/editorial.
2. Intervention encompassed all eight domains specified by the AHA Taxonomy for Disease Management.¹⁴	8. ICER/ICUR was not calculated based on modeled parameters even if a modeling technique was used.
3. Intervention was targeted at a CVD condition in adults.	9. Key clinical endpoints were not a CVD condition (pregnancy-related CVD complications were excluded).
4. ICER/ICUR was reported using modeled analyses.	10. Insufficient information or effectiveness estimates not based on actual trial data (e.g. hypothetical intervention or summarized effect).

CVD: cardiovascular disease; ICER: incremental cost-effectiveness ratio; ICUR: incremental cost-utility ratio.

Table 2.

Screening algorithm.

A	Design/protocol only
	Methodological paper
	Review/meta-analysis
	No ICER/ICUR reported
	Not CVD focused
	Pregnancy-related CVD
B	Not disease management
	Single intervention component
C	Not model-based
	Hypothetical/insufficient information

CVD: cardiovascular disease; ICER: incremental cost-effectiveness ratio; ICUR: incremental cost-utility ratio.

Data extraction

A data collection table was developed a priori, pilot tested on three randomly-selected included articles and refined accordingly. Two reviewers extracted data on study characteristics, intervention characteristics and information relevant to economic evaluation.

Measures of cost-effectiveness (e.g. ICER, ICUR) were converted to 2013 US dollar values using a web-based tool¹⁵ that enables the conversion of costs across currencies (using purchasing power parities conversion rates) and years (using GDP deflator index values).¹⁶ Costs reported in international dollars were converted to the country’s purchasing power parity (PPP) exchange rate using the Penn World Table¹⁷ and then converted to 2013 US dollars. US dollars were used as a standard currency in preference to euros. The PPP of a euro varies widely across European countries making conversion to euros using PPP highly variable. US dollars avoids this problem.

Study quality assessment

Two assessment tools were used, namely, the Grading System for the Quality of Cost-Effectiveness Studies (Chiou’s system)¹⁸ and the Consolidated Health Economic Evaluation Reporting Standards (CHEERS checklist)¹⁹ by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR). The CHEERS checklist consolidated previous health economic evaluation guidelines into one reporting guideline. Chiou’s system is also used because a scoring tool does not assume that each criterion shares an equal level of importance as in a checklist, and this tool was previously used to assess economic evaluations of telehealth services²⁰ and drug-eluting stents for percutaneous coronary intervention.²¹

Two reviewers independently assessed study quality using the two tools. Any discrepancies between reviewers were resolved by discussion with a third author. The CHEERS checklist consists of 24 items scored dichotomously as having met the criteria (in full) or not. Chiou’s system includes 16 weighted criteria with score values ranging from 0 (lowest quality) to 100 (highest quality). If a subgroup analysis was not undertaken (Chiou’s item 4), that item was not applicable and the weights were rescaled excluding that factor. The items deemed to have partially met the criteria were scored as ‘0’ (Chiou’s system) and ‘No’ (CHEERS checklist) to avoid introducing subjectivity associated with assigning part scores.

Results

Of 379 articles identified, nine articles (reporting 10 analyses and a total of 16 ICERs/ICURs including those of subgroup analyses) were included^22-28 (Figure 1). Study characteristics of CVD-MPs are summarized in Supplementary Material Table 1. The targeted cardiovascular conditions were hypertension (n=4), coronary heart disease (n=2), coronary heart disease plus stoke (n=1), heart failure (n=2), and hyperlipidemia (n=1), with eight cost-utility and two cost-effectiveness analyses^22,28 reported. The clinical trials that provided the key data inputs for the modeled analyses were carried out in Austria²⁹ (n=1), the USA^22,27,30 (n=3), England²⁶ (n=1 article or 2 analyses), Ireland and Northern Ireland^23,24 (n=2), Argentina²⁸ (n=1) and Pakistan²⁵ (n=1). (Supplementary Material Table 1.)

Although ‘usual care’ was used as the comparator in all analyses, it was not consistently defined. Of note, the usual care described in three articles included some level of active involvement such as the provision of two 30-minute surveys,²¹ an action plan during a diabetes review with a diabetologist,²⁵ or a detailed disease management plan sent to the appropriate primary care physician, who was asked to implement it.²⁸ Hence, it is plausible that some elements of the CVD-MP in one study could match the level of the usual care in another. Supplementary Material Table 2 shows the key items assessed in the economic evaluation.

Choice of model

Markov models (n=8) and ‘decision analytic models’ (n=1)²² were used. One article²⁵ did not describe how the observed clinical data were converted into CVD disability-adjusted life-years (DALYs), or what parameters were used in the DALY and therefore in the probabilistic sensitivity analysis. Only three articles^23,28,29 justified the choice of model.

Economic perspective

Six analyses (five articles)^22-24,26,27 stated that their analysis was performed from the healthcare system perspective including one²⁴ that implied it took this perspective, one article reported both societal and fund holders perspectives²⁵ and three articles used a third party payer perspective.^28-30

Discount rates

Eight articles discounted both benefits and costs, and one article discounted costs only.²² Four articles used 3%–3.5%, four articles used 5% in their base case and one analysis used different rates for each, that is, 3% for cost, 5% for benefit³⁰ (Supplementary Material Table 2). Although all articles reported discount rates, only two articles^25,26 justified the choice of particular rates, that is, ‘reflecting common conventions’²⁶ and ‘in accordance with the ISPOR guidelines’,^25,31,32 respectively. Discount rates were varied in sensitivity analyses in all but one article.²² In all but four CVD-MPs,^25,26,29,30 results were shown to be sensitive to the choice of discount rate.

Time horizon

All analyses were modeled based on data from interventions of one to two years’ duration. The time horizon ranged from two years (n=1) ²⁵ to 10 years (n=1)³⁰ and lifetime (n=8).

Six^23-27,29 of the nine articles explicitly stated the assumed duration of intervention effect beyond the observed data.

Reporting of costs

The reported net difference in costs for CVD-MPs ranged from cost saving of $–9055 to $6015 in 2013 US$ (original: US$–8788 to US$4850) (Supplementary Material Table 2). The included costs appear to reflect the perspective taken except in one article, where perspective was not explicitly stated.²⁴ The description of ‘intervention cost’ was least comparable. One article²² reported the cost of developing the educational modules as a one-time research ‘sunk’ cost that would not be incurred again if implemented for future patient care, and hence it was not included in their costing. It is unclear whether the same practice was followed elsewhere, for example, whether it was part of ‘intervention setup’ or whether similar costs were simply not incurred in the other studies. Demarcation of fixed or variable costs was also inconsistent. One article²⁷ assessed the impact of the allocation of intervention costs between fixed and variable components; the effect of this was to increase the ICER from $43,650 (base-case) to $129,738 per quality-adjusted life-year (QALY) gained when all intervention costs were assumed to be variable.

Reporting of effectiveness

The analyses of heart failure management programs^27,29 referred to the mortality observed in a trial. The other seven articles mapped an intermediate outcome (CVD risk factor observed in a trial) to a definite health outcome such as to reflect mortality in life tables or the mortality rates from the Framingham Heart Study.^22,33 Methods used for this mapping included use of the Framingham risk equation,^23,24,26,28 and the United Kingdom Prospective Diabetes Study (UKPDS) Risk Engine.³⁰ Health outcomes were valued using SF-6D utility weights from the SF-36,^23,24,27 preference weights converted from Minnesota Living with Heart Failure Questionnaire,²⁹ previously published health state values,^26,30 assumptions due to ‘lack of evidence’,³⁰ and for the DALY conversion²⁵ no details were provided. Two articles did not use (dis)utility as they reported health outcomes in life-years (LYs) only.^22,28 Reported net differences in QALYs ranged from 0.0051 to 0.84; LY ranged from 0.03 to 0.18. (Supplementary Material Table 2.)

Cost-effectiveness of CVD-MPs

All authors of the reviewed articles cautiously concluded their CVD-MPs to be a potentially cost-effective alternative to usual care. Stated acceptability constraints ranged from US$50,000/QALY,^26,30 Canadian$50,000/QALY,²⁹ US$100,000/QALY,²⁷ €45,000/QALY, ^23,24 US$50,000/LY,²² three times GDP/LY ²⁸ to three times GDP per capita per CVD DALY ²⁵ (US$43,089 to US$122,694 in US$ 2013 values).

The ICERs/ICURs are shown both in 2013 US$ and original figures in parentheses as reported in articles (Supplementary Material Table 2). Of all ICERs/ICURs reported including those reported of subgroups (n=16), six ICERs/ICURs^{22,25,26,28,29} were below and four^22,27 were above US$50,000/QALY or US$50,000/LY, which is most frequently cited as an acceptable threshold, and six ICURs were shown to be dominant (cost-saving) in the base-case^23,29,30 as well as in subgroups²⁴ (compared in 2013 US$).

Study quality

All articles were of adequate quality: Chiou’s system (minimum 63, maximum 93; median 82); the CHEERS checklist (minimum 44%, maximum 87%; median 73%) although the CHEERS is not a formal scoring instrument.

Items 8, 9 and 13 of Chiou’s system were not satisfied most often across the studies. The requirement to justify the choice of discount rates (item 8) was met by only one article²⁶ out of nine (11%), and the choice of economic model (item 13) by three articles^{23, ²⁸, ²⁹} (33%). Six articles^22-24,28-30 (66%) reported resource quantities and unit costs separately (item 9). Chiou’s item 4 on subgroup analysis (pre-specification) was applicable in only two articles.^22,24

The criteria met the least in the CHEERS checklist were items 7, 9, 15 and 17. The requirements to describe and to justify the comparator (item 7) were met by four articles^23,24,28,30 out of nine (44%), the choice of a particular discount rate (item 9) by three^25,26,28 articles (33%), and the choice of model (item 15) by three^23,28,29 articles (33%). Analytic methods including approaches to validate the model and sensitivity analysis methods (item 17) were clearly described by five articles^23,24,27-29 (56%). Valuation of preference-based outcomes (item 12) was not applicable in two articles that reported cost per LY.^22,28

The results per study are provided in the Supplementary Material Table 3.

Discussion

In this review of published modeled economic analyses for CVD-MPs, of all ICERs/ICURs reported including those of subgroups (n=16), six ICURs^23,24,29,30 were shown to be ‘dominant’ (more effective and less costly), six ICERs^{22,25,26,28,29} were below, and four^22,27 were above a US$50,000 willingness to pay threshold (in 2013 US$).²⁸

The patient groups represented in the selected studies were community-based outpatients with one or two conditions to manage,^22-28,30 except one article²⁹ with those recently discharged from hospital and at high risk for re-hospitalization and death. There were two heart failure programs applied to relatively low-²⁷ and high-risk²⁹ patient groups, respectively; however, the type and scope of the intervention also greatly differed in each case. This makes it difficult to infer the extent to which the ICURs may have been attributable to patient or intervention characteristics.

Whether CVD-MPs are cost-effective when all the costs for program development and implementation were considered⁹ remains uncertain in our review. For future research, systematic reporting and standardization of intervention costs may be facilitated through the use of a tool to aid comprehensive cost estimates such as the Tools for Economic Analysis of Patient Management Interventions in Heart Failure.³⁴ Such effort will be complemented by an increased use of study quality assessment tools (e.g. the CHEERS) to further standardize reporting of economic evaluation methods, and thus to facilitate cost-effectiveness assessment of CVD-MPs.

While numerous factors affect cost-effectiveness, reporting practice is one area that can be standardized to improve the transparency of economic analyses regardless of other critical variables. Because a model could be framed to favor one intervention over another, or assumptions can be introduced without adequate justification,³⁵ use of a common reporting format and the justification for the choice made by a modeler, as emphasized in the study quality assessment tools, are crucial. If results were reported in a comparable and transparent manner, it would facilitate discerning the differences in cost-effectiveness ratios due to study methodology from those due to the characteristics of the intervention evaluated.³⁶

To clarify differences in the methods used across studies, we assessed how well they adhered to established methodological and reporting principles by using Chiou’s system and the CHEERS checklist. Although the quality of an analysis would ultimately hinge on whether clinical and modeling assumptions were reasonable, not simply whether authors adhered to procedural guidelines for reporting,³⁷ adherence to the guidelines is nevertheless the first step to increasing transparency. Our review revealed considerable variability in reporting itself, which hinders the assessment of methodological variation across studies and, ultimately, of value for money of CVD-MPs.

The variables to which the results reviewed appeared sensitive were intervention costs,²⁷,²² transition probabilities,^27,29 utility weights for health states,^27,30 the discount rate of health outcomes (utilities)³⁰ and time horizon.³⁰ Seven reviewed articles,^22-24,26-29 used a long time horizon (i.e. lifetime). This is particularly relevant to CVD-MP studies as their benefits due to reduced resource utilization and/or health gain after the incremental intervention cost is taken into account may not become evident in a short timeframe. Some suggest that it takes at least three to five years to identify ‘true’ program effectiveness because of time lags in reaching full implementation,³⁸ while other evidence indicates that evaluations of a shorter duration (<1 year) can overestimate intervention effects of >1 year.^39,40 This may be due to the difficulty in sustaining the intervention effects over time as risk factors can be ‘heavily influenced by an individual’s health behavior’.⁴¹ As a lifetime horizon is preferable for chronic conditions and modeling is often needed for data extrapolation, model assumptions regarding unobserved long-term intervention effects as well as the cost of sustaining them should be made explicit as they could significantly impact its cost-effectiveness. In two heart failure management programs,^27,29 the base case conservatively assumed mortality rates to be the same across groups post-trial. In a hypertension or lipid control program,²⁶ continuing provision of the intervention was assumed to be required to achieve its continued benefit. In the studies for coronary heart disease (CHD),^23,24 the effect was assumed to persist for four years post-trial, and thereafter was assumed to be the same across groups. Whether each assumption made was reasonable can be further judged as long as such choice and rationale are clearly stated.

Apart from those assessed in the sensitivity analyses, clinical practice patterns or the availability of alternative treatments or healthcare resources³⁶ are also known to affect cost-effectiveness. For example, an intervention can be highly clinically and cost effective in one context but highly ineffective when transferred to another context in which accessibility and quality of current clinical practice are much higher.⁴² This is of particular importance for a high-context intervention such as disease management, whose effectiveness depends on background clinical practice patterns (usual care) and how healthcare is delivered and financed in each locale. ‘Usual care’ could mask variation across jurisdictions and over time in the availability of new technologies or the level of substitution effects. Moreover, the content of ‘usual’ or ‘routine’ care is often not sufficiently described.⁴³ The same rigor should apply to the reporting of what constitutes usual care as well as CVD-MP, for ‘usual case’ could also be a multi-component ‘intervention’. Referencing to an instrument such as an instrument to measure the intensity and complexity of heart failure disease management program⁴⁴ would significantly improve comparability across reports.

We excluded articles if clinical effectiveness estimates were derived from meta-analysis or hypothetical CVD-MPs. That is, we only included modeled analyses whose source of effectiveness data was a single study for three reasons: i) using the AHA taxonomy as a guide for defining CVD-MP we were unable to fit hypothetical or meta-analyzed CVD-MPs into the taxonomy structure; ii) cost-effectiveness estimated from the summarized effect of CVD-MPs is difficult to interpret as each program widely differs in its intensity or intervention components; iii) our interest was in the assessment of CVD-MPs actually implemented. Some hypothetical programs appeared to assess the expected impacts of alternative interventions prior to actual development and implementation, and thus, rely more on synthesized assumptions and projections than those included in our review. Such studies were not considered as providing an empirical measure for technical efficiency and hence were excluded.

Our review has limitations. First, there is substantial heterogeneity across studies from three main sources: the range of CVD conditions studied, the components in the CVD-MP intervention and the approach to modeling (discussed above). Different target conditions were included because that enabled us to assess how modeling was used for both intermediate and definite outcomes; for example, in two secondary prevention programs of CVD^23,24 and a prevention program of CVD in patients with diabetes,³⁰ the clinical outcomes observed in the clinical trials were CVD risk factors that were then transformed into LY outcomes (LY, QALY, DALY). Only two^27,29 out of nine articles used mortality outcome actually observed in a trial. As for the intervention components, their scope or intensity was not described consistently or sufficiently. The only observable commonality was that all CVD-MPs incorporated patient education. Second, restricting to model-based analyses (by excluding trial-based analyses) inevitably reduces the number of eligible articles; hence no statistical analysis was involved and as a result our analyses are descriptive in nature. Although this was a trade-off between the increased comparability within modeled analyses and the diversity allowing different target conditions, such comparability enabled us to assess methodological and reporting variations independent of the variation of trial-based versus model-based approaches.

Finally, although publication bias may be present, its extent is hard to gauge, given the lack of study registration requirements for model-based analyses as would be required for clinical trials (no audit trail). Although a funnel plot is widely used as a test for publication bias, it has been suggested that asymmetry of the funnel plot does not accurately predict publication bias.⁴⁵ True heterogeneity among studies itself can cause asymmetry,⁴⁶ which is highly probable in CVD-MPs where patient and intervention characteristics substantially vary across studies.

Conclusion

Although CVD-MPs appear to be cost-effective based on published model-based analyses, it is not possible to estimate the extent of the influence by the reporting variation across studies, which itself may have masked the variation due to methodological approaches and modeling assumptions. Transparency should be increased through standardization tools to enable a more robust determination of what truly drives the long-term cost-effectiveness of CVD-MPs.

Footnotes

Acknowledgements

Ms Jodie Vickery is acknowledged for assisting study selection. Mr Shamesh Naidoo is acknowledged for his initial work on this topic. JAW was supported by a research fellowship from the Queensland Government Department of Employment, Economic Development and Innovation, Queensland Health and Griffith University.

Conflict of interest

None declared.

Funding

This work was supported by a National Health and Medical Research Council of Australia program grant (grant 519823).

Implications for practice

Heterogeneity around the intervention and/or patient characteristics will remain obscured where substantial variations in reporting and methodological approaches affect cost-effectiveness. The use of practical standardization tools will increase transparency and third-party reproducibility.

Any improved comparability across studies will subsequently inform what type of intervention works for which patient population, enabling targeted approaches in clinical practice and resource allocation.

References

Nolte

McKee

. Caring for people with chronic conditions: A health system perspective. Maidenhead, UK: Open University Press/McGraw Hill Education, 2008.

Graves

Walker

Raine

. Cost data for individual patients included in clinical studies: No amount of statistical analysis can compensate for inadequate costing methods. Health Econ 2002; 11: 735–739.

Halliday

Darba

. Cost data assessment in multinational economic evaluations: Some theory and review of published studies. Appl Health Econ Health Policy 2003; 2: 149–155.

Conklin

Nolte

. Disease management evaluation (A comprehensive review of current state of the art). The RAND Corporation, 2010. Santa Monica.

Institute of Medicine of the National Academies. Promoting cardiovascular health in the developing world: A critical challenge to achieve global health. Washington, DC: The National Academies Press, 2010.

Trogdon

Finkelstein

Nwaise

. The economic burden of chronic cardiovascular disease for major insurers. Health Promot Pract 2007; 8: 234–242.

Nichols

Townsend

Luengo-Fernandez

. European Cardiovascular Disease Statistics 2012. Brussels: European Heart Network; Sophia Antipolis: European Society of Cardiology, 2012.

Australian Institute of Health and Welfare. Australia’s health 2012. Canberra: Australian Institute of Health and Welfare, 2012.

Ofman

Badamgarav

Henning

. Does disease management improve clinical and economic outcomes in patients with chronic diseases? A systematic review. Am J Med 2004; 117: 182–192.

10.

Zwar

Harris

Griffiths

. A systematic review of chronic disease management. Research Centre for Primary Health Care and Equity, School of Public Health and Community Medicine, UNSW, 2006.

11.

Weingarten

Henning

Badamgarav

. Interventions used in disease management programmes for patients with chronic illness-which ones work? Meta-analysis of published reports. BMJ 2002; 325: 925.

12.

Moher

Liberati

Tetzlaff

. Preferred Reporting Items for Systematic reviews and Meta-Analyses: The PRISMA statement. PLoS Med 2009; 6: e1000097.

13.

Norris

Glasgow

Engelgau

. Chronic disease management: A definition and systematic approach to component interventions. Dis Manage Health Outcomes 2003; 11: 477–488.

14.

Krumholz

Currie

Riegel

. A taxonomy for disease management: A scientific statement from the American Heart Association Disease Management Taxonomy Writing Group. Circulation 2006; 114: 1432–1445.

15.

Campbell and Cochrane Economics Methods Group and the Evidence for Policy and Practice Information and Coordinating Centre. CCEMG – EPPI-Centre Cost Converter (v.1.3), http://eppi.ioe.ac.uk/costconversion/default.aspx) (accessed August 2013) 2010.

16.

Shemilt

Thomas

Morciano

. A web-based tool for adjusting costs to a specific target currency and price year. Evidence Policy 2010; 6: 51–59.

17.

Heston

Summers

Aten

. Penn World Table Version 7.1, Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania, Nov 2012. https://pwt.sas.upenn.edu/ (accessed May 2013).

18.

Chiou

Hay

Wallace

. Development and validation of a grading system for the quality of cost-effectiveness studies. Med Care 2003; 41: 32–44.

19.

Husereau

Drummond

Petrou

. Consolidated Health Economic Evaluation Reporting Standards (CHEERS) – Explanation and elaboration: A report of the ISPOR Health Economic Evaluations Publication Guidelines Task Force. Value Health 2013; 16: 231–250.

20.

Wade

Karnon

Elshaug

. A systematic review of economic analyses of telehealth services using real time video communication. BMC Health Serv Res 2010; 10: 233.

21.

Ligthart

Vlemmix

Dendukuri

. The cost-effectiveness of drug-eluting stents: A systematic review. CMAJ 2007; 176: 199–205.

22.

Datta

Oddone

Olsen

. Economic analysis of a tailored behavioral intervention to improve blood pressure control for primary care patients. Am Heart J 2010; 160: 257–263.

23.

Gillespie

O’Shea

Murphy

. The cost-effectiveness of the SPHERE intervention for the secondary prevention of coronary heart disease. Int J Technol Assess Health Care 2010; 26: 263–271.

24.

Gillespie

O’Shea

Murphy

. Relative cost effectiveness of the SPHERE intervention in selected patient subgroups with existing coronary heart disease. Eur J Health Econ 2012; 13: 429–443.

25.

Jafar

Islam

Bux

. Cost-effectiveness of community-based strategies for blood pressure control in a low-income developing country: Findings from a cluster-randomized, factorial-controlled trial. Circulation 2011; 124: 1615–1625.

26.

Mason

Freemantle

Gibson

. Specialist nurse-led clinics to improve control of hypertension and hyperlipidemia in diabetes: Economic analysis of the SPLINT trial. Diabetes Care 2005; 28: 40–46.

27.

Miller

Randolph

Forkner

. Long-term cost-effectiveness of disease management in systolic heart failure. Med Decis Making 2009; 29: 325–333.

28.

Perman

Rossi

Waisman

. Cost-effectiveness of a hypertension management programme in an elderly population: A Markov model. Cost Eff Resour Alloc 2011; 9: 4.

29.

Moertl

Steiner

Coyle

. Cost-utility analysis of nt-probnp-guided multidisciplinary care in chronic heart failure. Int J Technol Assess Health Care 2013; 29: 3–11.

30.

Shah

. A Markov model of the cost-effectiveness of pharmacist care for diabetes in prevention of cardiovascular diseases: Evidence from Kaiser Permanente Northern California. J Manag Care Pharm 2013; 19: 102–114.

31.

Hay

Smeeding

Carroll

. Good research practices for measuring drug costs in cost effectiveness analyses: Issues and recommendations: The ISPOR Drug Cost Task Force report – Part I. Value Health 2010; 13: 3–7.

32.

Ramsey

Willke

Briggs

. Good research practices for cost-effectiveness analysis alongside clinical trials: The ISPOR RCT-CEA Task Force report. Value Health 2005; 8: 521–533.

33.

Franco

Peeters

Bonneux

. Blood pressure in adulthood and life expectancy with cardiovascular disease in men and women: Life course analysis. Hypertension 2005; 46: 280–286.

34.

Reed

Kamble

. Introduction of the Tools for Economic Analysis of Patient Management Interventions in Heart Failure Costing Tool: A user-friendly spreadsheet program to estimate costs of providing patient-centered interventions. Circ Cardiovasc Qual Outcomes 2012; 5: 113–119.

35.

Buxton

Drummond

Van Hout

. Modelling in economic evaluation: An unavoidable fact of life. Health Econ 1997; 6: 217–227.

36.

Drummond

Sculpher

Torrance

. Methods for the economic evaluation of health care programmes. 2005. 3rd edition. New York: Oxford University Press, 2005.

37.

Neumann

Greenberg

Olchanski

. Growth and quality of the cost-utility literature, 1976–2001. Value Health 2005; 8: 3–9.

38.

Serxner

Baker

Gold

. Guidelines for analysis of economic return from health management programs. Am J Health Promot 2006; 20(Suppl.): 1–17.

39.

Elissen

Duimel-Peeters

Spreeuwenberg

. Evaluation of disease management in The Netherlands. In: Nolte

Hinrichs

(eds) DISMEVAL. Developing and validating disease management evaluation methods for European healthcare systems. Final report. Santa Monica, CA: RAND Corporation and DISMEVAL Consortium, 2012.

40.

Linden

Adler-Milstein

. Medicare disease management in policy context. Health Care Financ Rev 2008; 29: 1–11.

41.

Nolte

Conklin

Adams

. Evaluating chronic disease management. RAND Corporation, 2012. Santa Monica.

42.

Pamela

Frances

Alan

. Systematic review of cost effectiveness studies of telemedicine interventions. BMJ 2002; 324: 1434–1437.

43.

Clark

Savard

Thompson

. What is the strength of evidence for heart failure disease-management programs? J Am Coll Cardiol 2009; 54: 397–401.

44.

Riegel

Lee

Sochalski

. Developing an instrument to measure heart failure disease management program intensity and complexity. Circ Cardiovasc Qual Outcomes 2010; 3: 324–330.

45.

Lau

Ioannidis

Terrin

. The case of the misleading funnel plot. BMJ 2006; 333: 597–600.

46.

Terrin

Schmid

Lau

. In an empirical evaluation of the funnel plot, researchers could not visually identify publication bias. J Clin Epidemiol 2005; 58: 894–901.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB