Improving Patient Outcomes by Improving Interhospital Transfer. An Argument for Guided Transfer

Transcription

1 Improving Patient Outcomes by Improving Interhospital Transfer An Argument for Guided Transfer Theodore J. Iwashyna, MD, PhD University of Michigan Ann Arbor VA Center for Clinical Management Research umich. edu

2 Presenter Disclosure Information Presenter: Theodore J. Iwashyna, MD, PhD Title: Improving Patient Outcomes by Improving Patient Transfer FINANCIAL DISCLOSURE No relevant financial relationships exist. ACKNOWLEDGMENTS NIH K08 HL and VA IIR Thanks to the organizers for this kind invitation. The views presented here do not represent those of the U.S. Department of Veterans Affairs or the U.S.

3 Ontario s Strengths & Challenges Vast landmass, much of it with very low population density Differentiated systems of care essential to allowing both local hospitals and centers of excellence for unusual / expensive conditions Sophisticated transfer

4 The Argument Interhospital transfers is usually discussed as moving patient s to a higher level of care. But, in fact, that higher level has wide variation in its quality. If we had hospital- and condition- specific quality measures, one could send the patient to the best hospital, not just a higher level. In the U.S., those data are now publicly available. In Canada, they could be.

5 The Rest of the Talk Foundation: Transfers are safe and common Opportunity: Variation within a level of care Problem: Rarely recognized choice Proven Solutions: None Possibilities: Guided Transfers Potential Problems and Their Magnitude

6 Transfers Used to Be Scary Barry & Ralston (1994) Arch Dis Childhood.

7 Transfers Now Relatively Ontario: Critical events uncommon, 5.1% (981 of 19,228 urgent aeromedical) Philadelphia: No major adverse events (n=190) Scotland: no change in SMR for transfers (n=583) Singh (2009) CMAJ; Seymour (2008) Crit Care; Hughes (2010) Br J Anesthesia.

8 Long Distances are Singh (2009) CMAJ; Iwashyna (2009) Med Care.

9 Ontario s Strengths & Challenges Many hospitals have more than one big hospital within feasible transfer distance.

10 Scales et al (2011) JAMA efigure 3. The Opportunity: Ubiquitous Variability in

11 Opportunity: Wide Variation Even Among Same Tier Canadian Risk-Adjusted 30-Day Mortality Regional Variation: AMI : 6.5% to 13.8% Stroke: 10.4% to 28.0% Krumholz (2007) Health Affairs; CIHI (2010) Health Indicators 2010; Authors tabulation from

12 Can We Pick a Better Given that hospitals of seemingly similar capabilities vary widely in their processes and outcomes If one can direct patients to the hospital with better outcomes And those differences are causal, not measurement error Then guiding transfers to the hospital with the best outcomes can offer meaningful benefits. Hospital?

13 Can We Pick a Better Given that hospitals of seemingly similar capabilities vary widely in their processes and outcomes If one can direct patients to the hospital with better outcomes And those differences are causal, not measurement error Then guiding transfers to the hospital with the best outcomes can offer meaningful benefits. Hospital?

14 Can We Pick a Better Hospital? In the U.S. & Italy: data showing that little value is placed on hospital quality in determining destination. In Canada: not yet studied.

15 Guided Transfer: A For many conditions, high-quality risk-adjusted outcome data are or could be available. Guided transfer: Obtain reliable hospital outcome data; Default to choosing the destination hospital based on the condition-specific best published data Unless there is a strong patient-specific reason Iwashyna (2011) Curr Op Crit Care.

16 Proven Solutions: None There is a fair amount of data showing that systems of transfer can improve care. I am unaware of any gold-standard data proving that guided transfer based on hospital outcome data will improve outcomes.

17 Potential Impact: AMI Iwashyna (2010) Circ: CVOQ.

18 Potential Concerns High-quality condition-specific risk-adjusted outcomes data are not yet available for most Canadian hospitals. My patient does not want to go there. But the differences are not statistically significant! What about outcomes other than mortality? My patient is different than those included in the quality metrics. Won t this overwhelm the best hospitals? Or, the best hospitals won t accept my patient in a timely

19 Recap and Questions Foundation: Transfers are safe and common Opportunity: Variation within a level of care Problem: Rarely recognized choice Proven Solutions: None Guided Transfers: Recognize Choice of Destination Hospital is a Therapeutic Choice, Which Evidence of Relative Effectiveness Can Guide. Potential Problems and Their Magnitude Thank you for your attention I welcome your questions and comments. My is tiwashyn@umich.edu

20 Extra Slides Follow

21 Finalyson(1999) Med Care. My Patient Does Not Want to Go There Patient preferences matter. Patients often have preference for both outcomes and preferences of care. Physicians can substantially influence the degree of patient concern about

22 Potential Concerns But the differences are not statistically significant! True, but since the measures are unbiased, will still do better on average if select hospital with lower mortality. Further, is p<0.05 really the right standard of proof for this decision?

23 Outcomes Other Then Mortality Non-mortality outcomes are important We routinely lack non-mortality data on drugs and other therapeutics, yet make informed decisions anyway But, other measures (e.g. CLABSI rates, patient satisfaction) may not be well correlated with mortality, and so an informed trade-off may need to be made

24 Schweickert (2009) Lancet. My Patients are Different Problem 1: there are no good outcomes measures for patients with severe sepsis. Problem 2: my patient would have been excluded from the published measures. Solution: use the same reasonable clinical extrapolation so common in the rest of medicine.

25 Widespread adoption of guided transfer would likely require improved prioritization at receiving hospitals. This will be a problem for late adopters. Wunsch (2011) AJRCCM; Halpern (2004) CCM; Howell (2011) Curr Op Crit Care. Overwhelming Hospitals Many patients in receiving hospital ICUs are not receiving treatments that could only be delivered in an ICU. There has been a dramatic increase in ICU capacity over last two decades.

26 Opportunity: Wide Krumholz (2007) Health Affairs.

27 Iwashyna (2010) Circ: CVOQ. Frequent Aberrancy: Transfers for admitted AMI patients from nonrevascularization to revascularization hospitals 6%: to nearest hospital, also the best 45.8%: bypassed closer hospital to go to better hospital 36.8%: bypassed better hospital to go to farther hospital with worse 30-day outcomes

28 Bosk (2011) Med Care. Current Behavior: An In many hospitals, transfer for critically ill patients is an organizational routine In qualitative interviews, there was almost no discussion of picking a particular hospital based on that hospital s quality Transfer relationships are almost monogamous and highly stable for at least a decade in Medicare data But maybe some hospitals are as different as

29 Choosing a Specific What is the right metaphor for use with choosing a hospital to transfer a patient? Are all hospitals of a given class like different versions of an aspirin? Or all hospitals of a given class more like different versions of an antibiotic?

30 CMS (2011) HospitalCompare. U.S. Pneumonia Example Data from 3 close hospitals in Philadelphia There is a 1.5% absolute mortality difference Default could be to select hospital with 9.3% mortality for pneumonia transfers

31 Agenda Foundation: Transfers are safe and common Opportunity: Variation within a level of care Problem: Rarely recognized choice Proven Solutions: None Possibilities: Guided Transfers Potential Problems and Their Magnitude

32 Agenda Foundation: Transfers are safe and common Opportunity: Variation within a level of care Problem: Rarely recognized choice Proven Solutions: None Possibilities: Guided Transfers Potential Problems and Their Magnitude

33 Agenda Foundation: Transfers are safe and common Opportunity: Variation within a level of care Problem: Rarely recognized choice Proven Solutions: None Possibilities: Guided Transfers Potential Problems and Their Magnitude

34 Agenda Foundation: Transfers are safe and common Opportunity: Variation within a level of care Problem: Rarely recognized choice Proven Solutions: None Possibilities: Guided Transfers Potential Problems and Their Magnitude

35 Agenda Foundation: Transfers are safe and common Opportunity: Variation within a level of care Problem: Rarely recognized choice Proven Solutions: None Possibilities: Guided Transfers Potential Problems and Their Magnitude