1 / 33

Today ’ s Objectives: Data Type, Quality & Source

Health and Cost Data Inputs Training in Clinical Research DCEA Epi 213 Lecture 5 28 February 2013 James G Kahn, MD, MPH. Today ’ s Objectives: Data Type, Quality & Source. To Understand the General Issues in Gathering and Presenting Health and Cost Data Inputs

dallon
Download Presentation

Today ’ s Objectives: Data Type, Quality & Source

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Health and Cost Data InputsTraining in Clinical Research DCEA Epi 213Lecture 528 February 2013James G Kahn, MD, MPH

  2. Today’s Objectives:Data Type, Quality & Source • To Understand the General Issues in Gathering and Presenting Health and Cost Data Inputs • To Understand Data Sources and Synthesis Methods for Health and Cost Inputs • To Understand Common Criticisms Surrounding Health and Cost Data Inputs

  3. Levels of Evidence - Efficacy • Systematic Review of RCTs with homogeneity (meta analysis) • Systematic Review with some heterogeneity • Large RCTs • Small RCTs • Systematic Review of cohort studies • Individual cohort studies • Case control studies • Case series • Expert opinion Center for Evidence Based Medicine

  4. Hierarchy For Other Questions Oxford Centre for Evidence-Based Medicine. http://www.cebm.net/index.aspx?o=5653

  5. Best Estimates and Plausible Ranges 1. Best Estimate = Base Case = Baseline a. The most likely value for the input: the value in the center of the best available data. b. You can intentionally err in one direction to highlight the strength of your result. • 2. Plausible Range:similar to 95% (not 99.99%) C.I. • a. When using a single empirical data base, calculate a formal confidence interval (typically 95% CI). • b. With multiple sources, informal use of best range, based on available estimates and expert opinion.

  6. Health Inputs • Overview • Key questions • How to Find Inputs

  7. Health Inputs Overview 1. Health State Outcomes a. Relevant Outcome States, including side effects b. Probability Estimates 2. Health Preferences Weights a. Preference Weights for Outcomes b. Utilities, QALYs 3. Population Characteristics a. Relevant Population b. Disease Prevalence in Population

  8. Health States Key Questions: • What Are the Relevant Health States Over Time for the Disease Under Study? • When Do These States Occur, and How Long Do They Last? • What Are the Likely Side Effects or Other Unintended Consequences for Each Group? • For which health states are there credible estimates appropriate for your RQ and population?

  9. Example – Aneurysm Analysis For the aneurysm analysis, health outcomes were estimated from multiple sources: • Aneurysm rupture rates large cohort study • SAH case fatality meta-analysis • SAH disability medium cohort study, • meta-analysis • RR mortality with disability small cohort study • Surgical mortality, disability meta-analysis • RR rupture (= 0) expert opinion (informal)

  10. Preference Weights - Utilities a. Disease-specific Utilities b. Generic Utilities c. Key Questions: i. Do Estimates Exist for Your RQ & model? ii. Whose Perspective Are You Taking? iii. Disease-Specific Ratings vs. Health attributes iv. Community Ratings vs. Patient Ratings

  11. Population Characteristics a. Prevalence of the Disease b. Key Questions: i. What is the Relevant Disease Prevalence? ii. National, Representative Samples iii. Disease Surveillance vs. claims c. What Competing Risks Exist (Unrelated to RQ)? d. Are These Estimates Appropriate for Your RQ? How to adapt?

  12. Example: Population Characteristics - Aneurysm Analysis • Prevalence of disease - not needed, not a study about screening or an estimate of societal costs • All-cause mortality - very important because of the low risk of aneurysm rupture and hence the high risk of dying before rupture occurs • Estimated by age and sex from a data base maintained at CDC, available on the internet.

  13. Health—How to Find Inputs • Comprehensive Literature Review • Existing Databases • Primary Data Collection

  14. Comprehensive Literature Review • Employ Levels of evidence • Always consider relevance to your model. • Revise your model when the cost of the data your model requires is greater than the benefit from not revising the model.

  15. Commonly Used Health Data Sources a. Clinical Trials b. CMS/VA Databases c. Disease Registries d. Quality of Well-Being Index (QWB), Health Utilities Index (HUI) - www.healthutilities.com/overview.htm e. Disability/Distress Index, EuroQol Instrument

  16. Cost Inputs What costs to include: • Direct vs. time • Fixed vs. Variable

  17. Analytic Perspective

  18. Muennig 2008

  19. Looking for Direct Costs 1. Published Estimates 2. Resources Used x Cost per unit Used 3. Cost Data Bases

  20. Costs must be Updated • Adjust using the medical component of the Consumer Price Index (CPI), maintained by the Bureau of Labor Statistics (BLS) • Substitute particular unit costs with updated values for the same services • Whichever method is used, all costs must be denominated in a single currency adjusted to a single year (ref at the end)

  21. Unit Costs • Reimbursements • Billed Charges • Cost Accounting Systems • Price References

  22. Reimbursements a. Acceptable, especially if based on negotiated rates (“allowed”) • Medicare reimbursement for inpatient care is based on prices established for DRGs (Diagnostic Related Groups) • Excellent approach: RBRVS (resource-based relative value scale) used by Medicare for outpatient services • When deductibles and copayments are charged, these should be included when calculating cost for the societal perspective • In some databases, “allowed charges” summarize total reimbursement

  23. Billed Charges • Can be used cautiously • Must be adjusted with hospital department-specific cost-to-charge ratios • Even then, imperfect: a single cost-charge ratio is used for all services in a department

  24. Price (Unit cost) References • Drugs: “Pharmacy Red Book” of average wholesale prices • WHO CHOICE – health care services, inputs • Public estimates for health worker hourly wages, diagnostic and laboratory equipment, and even for most supplie

  25. Cost Data Bases • Record both resources and costs for specific diseases • Examples: • California Office of Statewide Health Planning & Development (OSHPD) hospital discharge data base • Medical Expenditure Panel Survey

  26. Fixed costs • Fixed costs do not vary with each unit of service provided; variable costs do. • If an intervention is unlikely to require new fixed costs, they can be omitted. • In the long run, there are no fixed costs.

  27. Time Costs • Opportunity cost (to the patient & possibly caregivers) of receiving an intervention • Opportunity cost = value of activities foregone = value of lost work & household productivity • What time is properly counted? • How is time counted & valued?

  28. What Time is Counted • Time required for the intervention is counted. Age- and gender-adjusted values from published tables • Time lost due to the illness (care, disability, early death) is not counted: the valuation of this time is captured in the utility assessment, and should not be double-counted • Note that “intervention” vs. “illness care” varies across research questions: HIV is “illness” for an HIV prevention program, but part of “intervention” in a CEA on antiretroviral therapy

  29. Example: Aneurysm Analysis Cost input Value (range) Source Clipping $25,150 (18,000-35,000) Cohort study – cost accounting system Moderate/severe disability $20,000/yr (13,000-30,000) Published estimate SAH hospitalization $47,000 ($33,000-$67,000) Cohort study – cost accounting system Discount rate 3% (0-5) CEA guidelines Time Costs Not Included - Could be based on the time for surgery and recovery. Assuming one month of lost time, at 10 hours/day and $10/hour = $3,000.

  30. 5-Example: Final Product - Aneurysm Analysis

  31. Health databases • www.pubmed.gov (or, UCSF link) • Google • Cochrane • Web of Science, EmBase, many others

  32. Other Information Sources • USA Prices: http://www.bls.gov/ http://research.stlouisfed.org/fred2/data/GDPCTPI.txt • Currency Conversion & global inflation: www.cia.gov/cia/publications/factbook/, also World Bank and others • Acronym Finders: http://www.acronymfinder.com/ http://www.acronymattic.com/

  33. Global Health Information Sources • WHO Statistical Information System: http://www.who.int/whosis/en/ • CHOosing Interventions that are CE http://who.int/choice/country/en/index.html • Unit cost for Health Services by Country: http://who.int/choice/country/en/index.html • CDC International Health Data Reference http://cdc.gov/nchs/data/misc/ihdrg2003.pdf http://www.tufts-nemc.org/cearegistry/index.html v

More Related