There is an accompanying letter A Strategy for Dealing with Excessive Waiting Times.
Keywords: Health;
This note has a simple purpose: to demonstrate the gains from reducing waiting times are somewhat larger than they might at first seem: an economic evaluation of the benefits reducing waiting times is likely to suggest there are very high returns. Essentially this arises because while a shortening of waiting times may appear superficially to benefit just a few people – the numbers in the backlog which are treated – all the subsequent patients are benefited by the shortening of the waiting times. Thus there is a spectacular multiplier from reducing waiting times which makes the gains for the outlay to reduce the backlog far larger than they at first seem.
Background
To illustrate the principle, we use a paper Does Delay in Starting Treatment Affect the Outcomes of Radiotherapy? A Systematic Review by Jenny Huang, Lisa Barbera, Melissa Brouwers, George Browman, and William J. Mackillop. The abstract is attached as an appendix. It is a systematic review which includes 21 studies involving breast cancer patients
Although the paper also looked at head and neck cancer, this illustration focuses on radiotherapy and breast cancer. Pooling a large number of studies the meta-study compellingly shows show that the local recurrence rate (LRR) is higher if the patient receives treatment in 9 to 16 weeks compared to receiving treatment in the first 8 weeks. The study provides an estimate of the LRR of those treated in the first 8 weeks of 5.8 percent and those treated in the following eight weeks of 9.2 percent.
The gain is thus 3.4 percent, but at issue is 3.4 percent of what. It might seem that it is 3.4 percent of the backlog which a Waiting Times Strategy treats. In fact it is 3.4 percent of all the women who as a result of the elimination of the backlog are treated early. This, as we show in an example below, in a year is probably going to be 6.5 times the actual backlog, so that the gains in the first year are 6.5 times 3.4 percent of the backlog number, or 22.1 percent. If the backlog is eliminated for 5 years the gains are more than the backlog.
This report does not carry out a complete evaluation, but recurrence has two economic effects. First women are at increased risk of death and a lower quality of life if their primary treatment fails, and second there are additional treatment costs if they have to undergo “salvage” treatment which usually includes mastectomy, if still feasible, and is often accompanied by chemotherapy treatment. Both are reduced by eliminating the backlog.
Given the multiplier effect, it seems likely that the benefits relative to the costs are high.
Illustration
To show how the multiplier works, consider the following simplified scenario.
Suppose every 8 weeks there are 100 women diagnosed as suitable for treatment radiotherapy for their breast cancer.
We have two scenarios. In Scenario A they are treated within the 8 weeks, and in Scenario B they are treated in the 9 to 16 week period.
To simplify let us assume that in Scenario A they are treated in their 8th week and Scenario B in their 16th week. This means that there are 100 women on the waiting list in Scenario A and 200 in Scenario B. Essentially there is a backlog of 100 cases, which force all women to wait longer than the recommended 8 week treatment.
(Implicit in this is the assumption the waiting lists are stable, and not growing, which is reasonable in most cases over a year, although there is some seasonal variation.)
The total recurrences in Scenario A in one year is the total number of women who are recommended for treatment, which is 650 women (52/8 X 100), times the LRR of 5.8 percent or 37.7 on average.
The local recurrences in Scenario B in the year is the total number of women who are recommended for treatment, which is 650 women (52/8 X 100), times the LRR of 9.2 percent or 59.8 on average.
Thus the difference in local recurrences between Scenario A and Scenario B is 22.1.
Now how many additional treatments is it necessary to get from Scenario B to Scenario A? The answer is 100, because once those 100 additional treatments are done the waiting list will reduce from 200 to 100, that is from scenario B to Scenario A, and all the women can be treated within the 8 weeks.
Thus an outlay of 100 treatments gets a reduction of 22.1 recurrences within a year.
The choice of one year is arbitrary. Suppose, as a result of the removal of the backlog, it is possible to maintain the 8 week treatment target for 5 years. In that case treating the backlog of 100 would eliminate 5 time 22.1 = 110.5 recurrences, that is more than the original number of extra treatments.
This is not a magic. The core of the result is the backlog of 100 is putting 3250 women at additional risk over 5 years.
Economic Evaluation
There is not a sufficient data to do a detailed calculation, but some crude modelling was carried out on the effects of delaying treatment for breast cancer. It used a 10 percent p.a. (real) discount rate and the parameters from the meta-study. It still required an estimate of the benefit to cost ratio for the treatment, which not being known was parameterised. The population was assumed to be a steady state as in the above illustration. Other assumptions are conservative.
Even so the conclusions seems to be robust:
- If the treatment is worth doing, it is always better doing it earlier than later.
- Evaluating the treatment in terms of one cohort (of 8 weeks) is misleading. The social benefit of removing a backlog is forty times – and more – more valuable than the benefit to a single cohort. This arises because of the large number of women who benefit from removing the backlog. In effect for each women who is shifted from the backlog, at least another forty women benefit.
(The ratio of population net benefit to cohort net benefit was 43 if the Benefit to Cost Ratio was 1.1 rising to 62 if the benefit-to-cost-ratio was 3.6).