Cost Benefit Analysis: Measuring costs

I have been reading the chapter called Cost Effectiveness and Cost Benefit Research in our latest reading (the one with the coffee stain). Something that sparked an interest in me was the difficulty in measuring direct costs of health care. The article states that “accountants have a difficult time figuring out how to allocate substantial overhead costs...” (p.87). After finishing my accounting degree I know this to be true. What we do is estimate the overhead as a percentage then allocate it out to the cost object. For example say the overhead is supervisors wages, we would give say 30% to emergency rooms, 35% to maternity, 15% to paediatrics, 10% to psychiatric, and 10% to geriatrics. These percentages could further be broken down into actual operations or treatments under each ward. Although this is a very crude explanation it points out that accountants rely on estimates. We will never know the correct amount to allocate; it will only ever be an estimate. Also, the reality is the accountant probably has little knowledge in what goes on in each ward, making it hard for them to determine the estimate in the first place. Both these factors result in a very arbitrary amount of so called direct costs. Accounting can be very easily manipulated within the rules. Once we factor in indirect costs (time lost from work, inconveniences) and opportunity costs, the final amounts of total costs can be completely different to the real total cost. This is important as a procedure that has a cost exceeding its benefit may not be implemented into Government health care spending. The first step in cost benefit analysis is to determine the costs, the most seemingly obvious being direct costs. But as mentioned above the direct cost is impossible to determine correctly. Therefore, if the easiest aspect of cost benefit analysis is in a way impossible, what hope is there for accurately determining the other costs and benefits?