The Role of Statistical Science in Guiding Health Policy

1
The Role of Statistical Science in Guiding Health
Policy

• Dalene Stangl1
• Don Berry2
• Giovanni Parmigiani1
• 1Institute of Statistics and Decision Sciences
• 2MD Anderson Cancer Center

2
The Role of the Statistician in Policy Analysis
and Research

• Sir Claus Mosers, 1975 ASA meeting, the
• foremost responsibility (of the statistician) is
  to contribute to more enlightened and efficient
  decision making through the fullest possible
  exploitation of our skills in analyzing and
  interpreting the data.
• Dorothy Price, Director, National Center for
  Health Statistics, (1976),
• The American Statistician, The Role of
  Statistics in the Development of Health Care
  Policy
• . As in other areas of social policy, health
  statisticians and health data are increasingly
  expected to provide keys to rational decision
  making. To accomplish this goal, the
  statistician and decision maker need to interact
  to an increasing degree.
• John Tukey, 1976, Am. J. of Epidemiology
• those statisticians for whom opportunity and a
  natural bent combine to offer experience and the
  development of expertise ought, in the public
  interest, become as much policy makers as their
  roles allow.

3
Quote Commonalities

• All refer to importance of decision-making
• Recommend more involvement of statisticians
• All statements were made 20-25 years ago

4
Focus of past 25 years

• 1970s and 1980s
• Develop coordinated, systematic data base
• At NCHS we are searching for innovations to
  enhance data production with minimal increased
  demand on resources and to provide data in a
  timely fashion. We are being asked to produce
  more data, which is more relevant, with resources
  that are not growing commensurately. We are
  being asked to aid in the interpretation and
  analysis of the data as well.
• 1990s
• Statistics and Policy (B.D. Spencer ed., 1997)
• no mention of decision theory
• The statistical basis of public policy a
  paradigm shift is overdue (Lilford and
  Braunholtz, 1996)
• Bayesian methods superior to conventional
  methods.
• Primary advantages
• Utility functions were given one sentence
  reference

5
Typical Bayesian Solution

• Hospital Profiling
• Outcome ex. - mortality rate at time t (adjust
  case-mix)
• Classical - Z-scores
• Bayesian - Posterior probabilities of excess
  mortality
• Implicit rather than explicit decision analysis
• How is decision-making embedded in the analysis?
• choice of outcome
• time point
• relative performance measure versus national
  guideline of performance
• quality thresholds
• posterior tail areas
• Sufficient?

6
Why Insufficient?

• Needs explicit decision-theoretic framework
• Two proposals
• relinquish automatic constant utilities embedded
  in p-values and posterior probabilities
• present statistical output in ways that increase
  the possibility and probability of applying a
  wide diversity of utility functions

7
Prescriptive Perspective

• Making Health Policy Decisions Is Human
  Instinct Rational? Is Rational Choice Human.
  Paltiel and Stinnett, Chance, 1996
• Approach formal analysis from a prescriptive
  perspective, I.e. aim to provide decision-makers
  with information that can help them to make
  better choices but stop short of telling them
  what to do.
• By being forced to consider this issue
  explicitly, people may, whatever their final
  decision, benefit from scrutinizing and coming to
  grips with values to which they had previously
  given little thought.

8
Proposal 1 relinquish automatic constant
utilities embedded in p-values and posterior
probabilities

• threshold, D
• two decisions
• d0 accept - sufficient quality
• d1 reject

9

• d1 is better than d0 if qgtD
• d0 is better than d1 if qgtD
• U(d,q) measures the worth/utility of d when the
  uncertain value is q

10
Constant and Linear Utility
11
Compromise Utility
12
Loss of declaring sufficiency
13
Uncertainty in q

• Calculate expected loss
• ?L(q)p(q)?q
• Declare sufficient quality iff negative
• Balances cost/benefits in a simple, comprehensive
  way

14
Specifying Loss

• Not the statisticians loss function
• Statistician can help decision-maker articulate
  value judgements in a way that allows coherent
  procedure

15
Extensions

• other loss functions
• multivariate outcomes - Tan and Smith (1998)
• prior elicitation

16
Proposal 2 Predictive Distributions for General
Outcomes

• Hospital Profiling
• excessive mortality at time t as measured by 1.5
  x median mortality across all hospitals
• predictive survival curves across time
• Advantages
• allow diversity of utility functions
• metric upon which values are easily understood
• incorporate QUALYs
• Disadvantages
• requires event times
• harder to model event times than dichotomous
  outcomes

17
Other Thoughts

• Meta-Analysis in Medicine and Health Policy
  (Stoto)
• Attitudes of Policy World
• Graduate Education
• Attitudes of Statisticians

18
References

• Berger, J.O. and Delampady, M. (1987). Testing
  precise hypotheses (with discussion).
  Statistical Science 2 317-352.
• Berger, J.O. and Sellke, T. (1987). Testing a
  point-null hypothesis the irreconcilability of
  p-values and evidence ( with discussion). J.
  Amer. Statistica Assoc. 82112-139.
• Berger. R.L. and Hsu, J.C. (1996).
  Bioequivalence trials, intersection-union tests
  and equivalence confidence sets (with
  discussion). Statist. Sci. 11283-319.
• Bernardo, J.M. and Smith, A.F.M. (1994).
  Bayesian Theory. Wiley. Chichester.
• Casella, G. and Berger, R.L. (1987). Reconciling
  Bayesian and frequentist evidence in the
  one-sided testing problem. J. Amer. Statist.
  Assoc. 82 106-111.
• Cochran, W.G. (1976). The role of statistics in
  national health policy decisions. American
  Journal of Epidemiology 104(4)370-379.
• Lilford, R.J. and Thornton J.D. (1992). Decision
  logic in medical practice. Journal of the Royal
  Collegeof Physicians of London, 26(4)400-412.
• Lilford, R.J. and Braudholtz, D. (1996). The
  statistical basis of public policy a paradigm
  shift is overdue. British Medical Journal
  313(7057)603-607.
• Lindley, D.V. (1985) Making Decisions. Wiley,
  Chichester.
• Lindley, D.V. (1997) The choice of sample size
  (with discussion). The Statistician 46 129-166.
• Lindley, D.V. (1998) Decision Analysis and
  Bioequivalence Trials. Statistical Science
  13(2) 136-141.
• Lindley, D.V. and Singpurwalla, N.D. (1991). On
  the evidence needed to reach agreed action
  between adversaries, with application to
  acceptance sampling. J. Amer. Statist. Assoc.
  86 933-937.

19
References continued...

• Normand, S., Glickman, M., Gatsonis, C. (1997).
  Statistical methods for profiling providers of
  medical care issues and applications. J. Amer.
  Statist. Assoc. 92(439)803-814.
• Paltiel, A.D. and Stinnett A.A. (1996). Making
  health policy decisions Is human instinct
  rational? Is rational choice human? Chance
  9(2)34-39.
• Rice, D. (1977). The role of statistics in the
  development of health care policy. The American
  Statistician 31(3)101-106.
• Tan S.B. and Smith, A.F.M. (1998). Exploratory
  thoughts on clinical trials with utilities.
  Statistics in Medicine 172771-2791.
• Tukey, J.W. (1976). Discussion of Role of
  statistics in national health policy decisions.
  American Journal of Epidemiology 104(4)380-385.