|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||||
|
|
|
|
|
|
|
|
|
|
|
Hons supervision
(note topics from 2004 are also available should students be interested...)Tentative Research topics for 2005
#1 Big Fat Lies? An Economic Analysis of Obesity Epidemics.
This project will critically examine the arguments, evidence and inferences
for the claims by NZ and other public Health authorities that obesity is
a major public health problem (http://www.newhealth.govt.nz/toolkits/obesity/executivesummary.htm).
You will be expected to organize your thinking as an economist about the
key multi-disciplinary issues in this controversial topic, starting with
two recent popular books by Campos and Gaesser. The theoretical perspective
that you will learn to apply to this debate is Bayesian on measurement and
inference (a philosophy of science developed in Donald Berry), an economic
perspective that reflects the rational agent models we used in decision making
under uncertainty in Econ 321, and the political economy of risk assessment
(as developed in Graham and Sunstein). Along the way you'll learn a great
deal about measurement and interpretation of health risks.
#2 Risk Aversion in the Lab: an operational subjective Bayesian Perspective.
This project will re-examine the econometric modelling, model selection, and
inferences about risk attitudes in lab settings from a Bayesian perspective.
We have access to two data sets from two recent AER papers on this topic,
Harrison et al and Holt and Laury. No prior experience using operational
subjective Bayesian inference is presumed, but a general interest in experimental
economics and in developing and assessing simple predictive models will be
an advantage.
Glenn W. Harrison, Eric Johnson, Melayne M. McInnes & E. Elisabet Rutström "Risk
Aversion and Incentive Effects: Comment" October 2004 (forthcoming AER)
Holt, Charles A., and Laury, Susan K., "Risk Aversion and Incentive Effects," American
Economic
Review, 92(5), December 2002, 1644-1655.
Donald Berry Statistics A Bayesian Perspective (1995)
Frank Lad Operational Subjective Bayesian Statistics (1996) (Ch 3 on coherent
statistical inference and ch 6 on proper scoring rules.)
#3 [tentatively] Measuring and assessing
the changing economic value of bone density scanning technologies
General research question: "How does the predictive value of alternative
types of diagnostic technologies for measuring bone density change as new treatment
options for osteoporosis are developed, and clinically and experimentally understood?" This
research project will most likely be be carried out with in collaboration with
Dr Nigel Gilchrist The
Canterbury Geriatric Medicine's Research Trust at Princess Margaret hospital
. The project will proceed along 3 dimensions:
1 Review the international literature on the (co) evolution of diagnostic and
treatment technologies for osteoporosis during the last 20 years to develop
a broad overview of changing received and dissenting professional opinion on
the genral research question.
2 Apply established theoretical models based on Bayesian inference (learning)
principles and patient/doctor preferences for assessing bone health risks to
the general research question, and compare these models to commonly used Frequentist
statistical techniques for assessing diagnostic value.
3 Review the current and recent professional and institutional guidelines and
practises within New Zealand in light of (1) and (2)
#1 Is randomization in clinical trials an economically efficient way of practicing evidence based medicine?
Using human beings as subjects in experiments that test drugs, medical devices or surgical interventions has long been a source of serious ethical concern. Oppressive political regimes have used "prisoners" as experimental subjects, but even liberal regimes tolerate and encourage (eg fund) suspect practices. One experiment involved black men with syphilis remaining untreated for years so experimenters could observe the natural course of the disease. Several decades ago women at Auckland's Women's Hospital unknowingly became subjects in a doctors research into cervical cancer, with fatal outcomes for some. A recent study claims that 17 out of 1400 participants in the development of AIDS vaccines in the US have developed HIV. It makes sense as an economist to ask hard questions about the kinds of "contracts" that arise between experimenters, potential subjects and alleged beneficiaries of experiments. There is a serious potential for inefficiency arising from incentive incompatibilities and asymmetries in information in this situation. This research project focuses on one particular contentious issue: randomizing people to treatments.
The basic research question is: "Is randomization in clinical trials an economically efficient way of practicing evidence based medicine?" The project will use the lens of economically efficient contracting under asymmetric information (Milgrom and Roberts Chapters 2,5-8), Besanko et al (chapters 14, 15 on agency, Chapters 2 and 3 on the boundaries of the firm) and Bayesian decision theory to review the debate, contrasting theories, and some historical evidence about randomized clinical trials (RCTs) . Part of your task in this project will be to understand and critically summarize the differences between Bayesian statistical inference for rational decision making and the frequentist (Fisher, Neyman Pearson) statistical methods of hypothesis testing that dominate health related research. You may not have any training in Bayesian inference but don't despair, we will develop the necessary ideas on the fly in the project. The challenge is conceptual, not computational, and the "level" is nicely presented in Howson and Urbach and in Berry and Strangl (1996). As economists interested in informational and incentive problems, the project will review theories of the role of randomization in experimental design. Kadane is a good place to start to appreciate the issues involved; you can find a pdf file (1mb) of his article here . The project will also review some historical evidence about the actual behavioral, strategic and risk management issues that arise in randomized clinical trials generally, via the British Medical Journal symposia on randomized trials, and specifically, via the Office of technology Assessment OTA report on medical, ethical and legal issues involved in developing AIDS vaccines.
Depending on the preferences of students the project can lean in the direction of being analytical and theoretical (eg, trying to model the impacts of randomization in a model of strategic interaction between patient, physician- agent, and political regulators) or in the direction of being more interpretative and institutional (are the institutional arrangements under which clinical trials for AIDS vaccines have been conducted efficiently solving the moral hazard and adverse selection problems involved in the tradeoff between good care and good information?)
Paul Milgrom and John Roberts Economics Organisation
and Management
Besanko, D, D. Dranove, M. Shanley, S. Schaeffer Economics of Strategy
3rd edition 2004
Joseph Kadane "Introduction" to Bayesian Methods and Ethics in
Clinical Trials (1996) Wiley pp 3-18 (available as pdf file on website)
Kenneth Schaffner "Ethically Optimizing Clinical Trials" Bayesian
Methods and Ethics in Clinical Trials (1996) Wiley 52-62
Joseph Kadane and Teddy Seidenfeld "Randomization in Bayesian Perspective" in Rethinking
the Foundations of statistics
Donald Berry and D Strangl "Bayesian Methods in Health Related Research" Bayesian
Biostatistics (1996) 3-66
David Spiegelhater Laurence Freedman Mahesh Palmer Bayesian Approaches to
Randomized trials" Bayesian Biostatistics (1996) 67-109
Deborah Ashby and Jane Hutton "Bayesian Epidemiology" Bayesian
Biostatistics (1996) 109-138
Colin Howson and Peter Urbach. Scientific reasoning : the Bayesian approach (2nd
edition)
OTA : US Office of Technology Assessment "Adverse Reactions to HIV Vaccines:
Medical, Ethical, and Legal Issues" September 1995 OTA-BP-H-163 (pdf
1Mb).
OTA The Impact of Randomized Clinical Trials on Health
Policy and Medical Practice: Background
Paper (Washington, D. C.: U.S. Congress, Office of Technology Assessment, OTABP-
H-22, August 1983).PDF
1.1MB
Ezekiel J Emanuel;David Wendler;Christine Grady ,"What makes clinical
research ethical?" JAMA (The Journal of the American Medical Association);
May 24/May 31, 2000, 283, 2701-2711
#2 An economic assessment of health risks from Hormone Replacement Therapy (HRT) for post menopausal women.
Millions of postmenopausal women around the world have been using hormone replacement therapy (HRT) for over half a century for a variety of health related reasons involving bones, reproductive organs, heart, lungs, skin as well as physical comfort and sexual functioning associated with ageing . In November 2002 New Zealand's medical authorities MARC (Medicines Adverse reactions Committee) announced changed guidelines stating that generally long term use of HRT involves risks that exceed benfits (a copy of Prescriber Update pdf 328kb from Medsafe http://www.medsafe.govt.nz/index.asp and related Lancet article pdf 128 kb is available at http://cantua.cantebrury.ac.nz/~fou10 ; the MOH site http://www.nzgg.org.nz/library/gl_complete/gynae_hrt/index.cfm contains NZ guidelines and updates.). These actions are in line with a worldwide re-assessment of the use of HRT after results of recent large scale clinical trials reported unanticipated risks JAMA editorial and JAMA article on the latest results on HRT (see the Women's Health Initiative website http://www.whi.org , especially the videocasts of a 2-day workshop for an excellent historical and scientific overview of knowledge about HRT).
Currently, complexity and confusion prevails amongst all major participants (scientists, practitioners, and consumer/patients) about whether to take or to continue to take HRT in light of this new evidence. Combining all the evidence, and examining all relevant alternatives, one has to ask: are the (uncertain) benefits of taking HRT worth the (uncertain) harms? There are two interrelated questions here: (1) What is known, and not known, about the balance of risks and harms for HRT? and (2) How best can diverse sources of uncertain knowledge, new and old, be summarized and communicated to women and their physician agents to enable them to make good decisions about use of HRT? Answers to the first question typically come from meta analyses, a term describing systematic attempts to combine information from different sources. Answers to the second question are in the form of treatment guidelines.
This project will review some of the key meta analyses that have led NZ authorities to change the guidelines for use of HRT from a Bayesian statistical inference standpoint, Stangl and Berry, and from an economic standpoint Milgrom and Roberts, and Philipson. Ultimately we want to answer the question whether these analyses summarize and convey complex information from heterogeneous sources in a way that helps rational consumers and their agents make sensible tradeoffs about health risks? Along the way we need to review theories and debates about methods in meta analyses, Bayesian or frequentist, for selecting studies, incorporating inconsistencies and quality differences in study design and outcomes, comparing and matching outcome and covariate measures, adjusting for publication bias, developing models and incorporating model uncertainty, and for dealing with missing data and rational patient/agent risk management behaviour in incomplete contracting environments.
Writing Group for the women's
health Initiative "Risks
and Benefits of estrogen Plus Progestin in Healthy Postmenopausal Women" JAMA
(Journal of the American Medical Association) July 17, 2002, vol 288, 321-333 full
text pdf . see related editiorial in same issue full
text
OTA (Office of Technology Assessment) "Effectiveness and Costs of Osteoporosis
Screening and Hormone Replacement Therapy, Vol. I: Cost-Effectiveness Analysis" August
1995 OTA-BP-H-160 (full
text pdf )
Tomas Philipson "Economic Epidemiology and Infectious Diseases" Chapter
3 Handbook of Health Economics Elsevier 1999
Paul Milgrom and John Roberts Economics Organisation and Management,
chapters 2,5,6
D K Stangl and Donald Berry Meta Analysis in medicine and health Policy Marcel
Dekker 2000
| jfhome | research | contact | links | econ223 | econ337 | econ338 |EconDept |