The NHS funded Cancer Genetics Service in Wales (CGSW) offers a resource
intensive model of care, offering high risk patients multiple counselling sessions
and genetic testing. Given that there is no information upon the preferences of
patients for the manner in which the Welsh model of providing cancer genetic
services is delivered, or data on the associated costs, research question 1 was
developed. Research question 1, "What a re the attributes of cancer gene tics
services that are important to high risk patients (the patients spending the most
time in contact with the service and receiving most services i.e. genetic testing
and counselling)? and what would be the cost of providing the service to comply
with patient preferences?" Having examined the literature on eliciting patient
preferences the relatively new and experimental technique of discrete choice
modelling (DCM) was identified as the most appropriate one to use to elicit the
data required to answer research question 1. The health economics literature
revealed that no one had experimentally examined DCM 's underlying decision
theory principal of random utility theory (RUT) in conjunction with a DCM
exercise. To supplement the deficiency in the health economics literature, the
decision theory/psychology literature was accessed. The large body of literature
on utility theory revealed that the descriptive ability of utility theory was in
question. These findings resulted in research question 2, "Do respondents of
DCM questionnaires make choices in accordance with Random Utility Theory?"
Empirical Aims
Experiment
• Experimentally examine respondents of a DCM exercise by means of an
information manipulation to see if they are adhering to DCM's underlying decision theory principals of Random Utility Theory.
Patient survey
• Ascertain the aspects of cancer genetics services that are important to
patients, and present service configurations prioritised in terms of preferences accompanied by their costs ( cost-consequences analysis) for high risk patients.
Methods
Experiment
A repeated measures information manipulation experiment was conducted with a sample of 142 first year undergraduate students at the University of Wales,
Bangor.
Patient survey
Patient preferences were gathered by issuing self administered discrete choice
questionnaires to 30 high risk patients of the CGSW following their genetic risk
assessment. Costs were estimated for the Cardiff clinic of the CGSW by
administering a questionnaire to all staff, conducting an audit of clinic rooms and equipment and obtaining gross unit costs from the finance department.
Results & Conclusions
Experiment
The primary conclusion of the experiment was that respondents were not making choices in accordance with random utility theory. There is clearly a need to conduct further research into RUT as soon as possible as until the uncertainty
relating to the descriptive validity of RUT is resolved DCM and other RUT based techniques are potentially invalid.
Patient survey
Given the results of the experiment, the results of the DCM survey with patients
of the CGSW must be interpreted with caution. Counselling by a genetics
associate accompanied by favourable levels of other attributes provided high
utility and also provided substantial cost savings. These findings support the use
of genetics associates for genetic counselling in response to the scarcity of
qualified consultant clinical geneticists. The savings obtained from such a
service configuration can be redirected to fund improvements in the service such
as more staff ( clinical and administrative) to reduce the waiting time between
receipt of referral and issuing patients with a risk assessment or be used in
relation to other attributes or completely different health services.