Nutrition and Coronary Heart Disease
The finding by Smith et al (1995) has two major implications on the efficacy of
nutrition interventions. First, it may be the case that one intervention is not effective but
multitudes of different interventions are. Such an approach would allow for the
different stages of change, as well as cater for the different types of individuals.
Secondly, any analysis of a single intervention becomes problematic, as its effectiveness
may be inter-related with other interventions.
The result of this may be that the highest effectiveness rate may be obtained when a
wide variety of different interventions are implemented.
Portfolio Investment
The use of a wide variety of interventions, as stated above, would concur with ‘portfolio
investment’ theory (Hawe and Shiell, 1995). This approach argues for hedging
investments to minimise risk.
Risk is defined as a lack of evidence on the expected returns of a high-risk intervention,
whereas an intervention for which there is good evidence is a low risk intervention.
Unfortunately, there is very little concrete evidence about the expected returns on
nutritional interventions. Hence, all interventions carry a certain level of risk. A second
form of risk is associated with the program itself. For example, there are good mass
media campaigns and there are bad ones. What differentiates a good campaign from
bad one is not widely known and hence carries risk.
A second reason for examining portfolio theory is that the interventions considered in
this study may have some level of interdependence (see section 5.4). Therefore, a
variety of different interventions may be more successful than any one intervention on
its own.
To invoke portfolio theory there is a need to examine the variance of the results of an
intervention. Such a method would bring the study into line with the classical finance
literature of retum∕variance objective functions6. Using the concepts of return∕variance,
6 See Varian (1990, p227-39) or Brealey and Myers (1991, ppl55-74).
CHERE Project Report Jl - November 1999
40