Leading Transformational Change for Healthcare Quality It’s Time to Get it Right!

1. Leading Transformational Change for Healthcare QualityIt’s Time to Get it Right! Dale W. Bratzler, DO, MPH QIOSC Medical Director Oklahoma Foundation for Medical Quality July 19, 2006

2. RAND Study: Does this really apply to us?? Patients’ care often deficient, study says. Proper treatment given half the time. On average, doctors provide appropriate health care only half the time, a landmark study of adults in 12 U.S. metropolitan areas suggests. Medical errors corrode quality of healthcare system Medical Care Often Not Optimal Failure to Treat Patients Fully Spans Range of What Is Expected of Physicians and Nurses Study: U.S. Doctors are not following the guidelines for ordinary illnesses The American healthcare system, often touted as a cutting-edge leader in the world, suddenly finds itself mired in serious questions about the ability of its hospitals and doctors to deliver quality care to millions. .

3. “Almost one-third of Medicare spending for chronically ill unnecessary.” Go to: www.dartmouthatlas.org Look up your hospital

5. Overall Health Care in U.S. (Rand) Breast cancer screening (65-69) Outpatient ABX for colds Hospital acquired infections Hospitalized patients injured through negligence Post-MI b-blockers Defects per million Airline baggage handling Detection & treatment of depression Adverse drug events Anesthesia-related fatality rate Best in Industry 1 (69%) 2 (31%) 3 (7%) 4 (.6%) 5 (.002%) 6 (.00003%) s level (% defects) Source: modified from C. Buck, GE

6. http://www.ahrq.gov/qual/nhqr05/nhqr05.pdf

10. In performance measurement, is an “A” good enough?

11. All-or-None MeasurementThe Appropriate Care Measure

12. Calculating Performance Measures3 Options • Item-by-item Measurement • Composite Measurement (Opportunities Model) • All-or-none Measurement (Appropriate Care Measure) Nolan T, Berwick DM. All-or-none measurement raises the bar on performance. JAMA. 2006;295:1168-70.

13. Item-by-Item Measurement • Commonly used by performance measurement systems • Performance for each measure reported as a percentage

14. Acute Myocardial InfarctionQuarter 3, 2005 National averages as reported to the QualityNet data warehouse.

15. Heart FailureQuarter 3, 2005 National averages as reported to the QualityNet data warehouse.

16. PneumoniaQuarter 3, 2005 National averages as reported to the QualityNet data warehouse.

17. Composite Measurement • Often used when patient-level data is not available • Computed average performance across all patients and a number of performance measures • Sum of numerators over the sum of denominators for all measures • Hospital level, state-level, national

20. Appropriate Care MeasureAll-or-none • The patient is the unit of analysis • What percentage of your patients received all indicated care (all-or-none)? • If the patient did not receive any care that was indicated, the case is not in the numerator

21. Appropriate Care Measure • Denominator • Number of patients eligible for any performance measure for the topic • Numerator • Number of patients in the denominator MINUS any patient that did not receive indicated care for any measure that they were eligible for

22. QIO Program Vision “Every person receives the right care every time”

23. Why Should we Aggregate Measures? • Individual measures provide a limited and overly optimistic assessment of quality of care at the patient level • An “A” may not be good enough • There is no “partial credit” in all or none measurement – you did everything right for the patient or you did not

24. Quality from the Patient’s PerspectiveAdvantages of ACM “Appropriate Care Measure” reflects the interests and likely desires of the patient or their family How many of your patients got all care that they should have?

25. Fosters a Systems Perspective of QualityAdvantages of ACM • Requires an institution to focus on all aspects of care rather than the individual measures • Those measures that are hardest to address will have a large impact on the ACM

26. More Sensitive Scale for Assessing ImprovementAdvantages of ACM • High scores on individual measures of care can dampen enthusiasm for continued improvement • Example – 5 independent measures each at 90% performance • ACM is 59%

27. Measures for the Appropriate Care Measure • Measures should be evidence-based and important for patient outcomes • Process measures are best suited

28. Appropriate Care MeasureNational Performance Heart Failure Acute MI Qtr. 1, 2002 Qtr. 4, 2004 Qtr. 3, 2005 Pneumonia

29. Focus on Problematic Measures • Antibiotic timing for pneumonia • PPV • LV function assessment and use of ACE/ARB medications

33. Kahn et al, 1990 • Setting: 297 hospitals in 5 states • Patients: adults • Design: retrospective record review • Outcome: mortality 30 days after admission • Findings: Care that included antibiotics within 4 hours associated with 5.4% lower mortality • Issues: Retrospective, limited adjustment, timing not focus JAMA. 1990; 264:1969-1973.

34. Meehan et al, 1997 • Setting: 3555 hospitals in US; 1994-95 • Patients: random sample 14,069 aged > 65y, pneumonia by principal ICD-9 code, admit DX, and CXR. CAP only. • Design: retrospective cohort, medical record review, severity adjusted • Outcome: mortality 30 d following admission JAMA. 1997; 278: 2080-84.

35. Meehan, et alFindings Antibiotic administration within 8 hours associated with lower 30-day mortality among all patients: Severity-adjusted odds ratio = 0.85; 95% confidence interval, 0.75 – 0.96. JAMA. 1997; 278: 2080-84.

36. Meehan, et alFindings When limited to patients without pre-hospital antibiotics (75%), association at 8 hours stronger: adjusted OR = 0.78; 95% CI, 0.67 – 0.89. JAMA. 1997; 278: 2080-84.

37. Houck PM, Bratzler DW, et al. 2004 • Setting: 3732 hospitals in US; 1998-99 (13371 patients) • Patients: random sample 18,209 aged>65y; principal ICD9 code 48X, 038.X or 518.81 plus 48X; admit DX, (+) CXR, CAP, immunocompetent. • Design: retrospective, medical record data, adjusted (PORT index, ICU, region, race, care process) Arch Intern Med. 2004; 164: 637-44.

38. Outcomes by Time to First Antibiotic Administration* Arch Intern Med. 2004; 164: 337-44. *Patients who were on antibiotics prior to admission are excluded from this analysis.

39. First Dose Timing and Outcomes Using multivariate logistic regression [the model included the timing of antibiotic first dose, PSI score, ICU admission, US census region, race/ethnicity, other processes of care (oxygenation assessment, performance of blood cultures, and antibiotic selection)]. Patients who were on antibiotics prior to admission are excluded from this analysis. (Houck PM, Bratzler DW, et al. Arch Intern Med. 2004.)

40. First Dose Timing and Outcomes Using multivariate logistic regression [the model included the timing of antibiotic first dose, PSI score, ICU admission, US census region, race/ethnicity, other processes of care (oxygenation assessment, performance of blood cultures, and antibiotic selection)]. Houck PM, Bratzler DW, et al. Arch Intern Med. 2004.

41. Houck PM, Bratzler DW, et alPatients on Pre-hospital Antibiotics Among patients with pre-hospital antibiotics (n=4,438), 4 hour deliverynot associated with 30-d mortality: adjusted OR = 1.18; 95% CI, 0.97-1.45 8 hours associated with worse mortality: adjusted OR = 1.78; 95% CI, 1.04 – 1.88. Arch Intern Med. 2004; 164: 337-44.

42. Is 4 hours better than 8 hours? • Antibiotic administration within 4 hours, as opposed to within 8 hours, was associated with: • 31.7%* reduction in mortality in the elderly. • 12.5%* reduction in length of stay exceeding the median in the non-elderly • 7.4%* reduction in length of stay exceeding the median in the elderly • Additional analyses showed that those receiving antibiotics in 4-6 hours were similar to those receiving antibiotics in 6-8 hours. Singer ME, Krishnaswamy J, Bonomo RA. [Abstract] The American Society for Microbiology's 45th Annual International Conference on Antimicrobial Agents and Chemotherapy (ICAAC™).

44. PN-2: Pneumococcal Vaccination • Numerator • Patients with pneumonia, age 65 and older, who were screened for pneumococcal vaccine status and were vaccinated prior to discharge, if indicated • Denominator • Pneumonia patients 65 years of age and older • Key Exclusions • Transfers from another acute care facility, no working diagnosis of pneumonia on admission, patients receiving comfort care only, expired in hospital, left AMA, discharged to hospice care, transferred to another hospital

45. Missed OpportunitiesBackground Fedson DS, Houck PM, Bratzler DW. Infect Control Hosp Epidemiol. 2000;21:692-699.

47. Heart Failure Guidelines • The only way to determine left ventricular function is to do and document an objective test • Randomized trials over many years have confirmed the mortality benefits and improvements in quality of life for patients treated with ACE inhibitors (ARBs Class IIa)

48. Public Reporting and Pay for Performance Note: CMS and HHS have not made final decisions regarding public reporting for FY 2007 and pay for performance.

49. Deficit Reduction Act “for fiscal year 2007 and each subsequent fiscal year …a hospital that does not submit, to the Secretary … data required to be submitted on measures selected…, the applicable percentage increase…for such fiscal year shall be reduced by 2.0 percentage points…”

50. Deficit Reduction Act “…The Secretary shall expand, beyond the measures specified … the set of measures that the Secretary determines to be appropriate for the measurement of the quality of care furnished by hospitals in inpatient settings…”