CIRCLIST - Cost-Benefit Analysis

Original document dated 30.Jul.2010.
Running update: 16.Jul.2013.
Author: Chris Eley.

A number of factors enter into this debate, primarily the identification and valuation of long-term benefits of circumcision both to an individual and to society at large. These need to be set against the up-front cost of prophylactic circumcision in a "Cost-Benefit Analysis".

The result of such a "CBA" enables an informed appraisal of the overall financial balance between prevention now and cure later, essential information both for Health Service Administrators and their policy-making masters.

A CBA is not (and is not intended to be) the sole determinant of whether infant circumcision should or should not be routine; it addresses only the financial trade-off issues. Culture, religion, ethics, morality, sexual pleasure and so on can be and should be debated separately from the cost issue - without any of those other factors imposing a veto on proper consideration of the financial implications.

Arguably, the CBA (being, relatively speaking, an objective analysis) should precede debate of the more subjective issues so as to provide a factual backdrop to those other discussions.

Not only is a CBA needed, it needs to be up-to-date. But do we have an up-to-date CBA? Apparently not. Two major areas of input to the analysis have changed drastically in the past few years.

The list of the long-term benefits of circumcision has grown considerably. We now need to include here not just the benefits to the individual male but also the benefits to their partner(s).

Schoen EJ et al (published, 2006) took into account benefits of circumcision resulting from the prevention of infant urinary tract infection, balanoposthitis, phimosis, HIV infection and penile cancer.

Morris, BJ, (pers. comm., 2010) suggests that the list of medical conditions now needs extending.

Combining the two approaches of Schoen and Morris suggests that we need to include:

Urinary tract infections ("UTIs") in infants
UTIs post-infancy through adulthood
Phimosis
Paraphimosis
HIV
High-risk HPV
HSV-2
Genital ulcer disease (GUD) and other STIs
Balanoposthitis
BXO
Penile cancer and possibly prostate cancer

Also including consideration of the impact of MC status on medical conditions in the female sexual partners:

Cervical dysplasia
Cervical cancer
HSV-2
GUD
Bacterial vaginosis
Other STIs, possibly including chlamydia and its clinical consequences (PID, infertility, ectopic pregnancies).

This is a formidable list and it has to be said that the extent of prophylaxis conferred by male circumcision (MC) is not yet fully understood in every scenario.

Why include the female? This is best answered by an example: Transmission rates of HPV, the virus that causes cervical cancer, are circumcision-related. Thus a proportion of the cost of treating the female’s condition needs to be allowed for when adding up the financial benefits of male circumcision, the 'proportion' in that instance being dependent on the differential in the transmission rate that is attributable to male circumcision.

There is another whole set of potential omissions to be considered - the social cost of sickness and premature death over and above the cost of treating the medical condition giving rise to it.

In the economic environment of a fully-integrated Welfare State, the following will need to be considered:

(a) Loss to the state of taxation on earned income if the patient is unable to remain fully and equally economically active (differential value from onset of the reduction through to normal retirement age, time-discounted);

(b) Additional Social Security support of the patient during life (absolute cost, time-discounted);

(c) Social Security support of the patient’s family after the patient’s premature death (differential cost with respect to normal life expectancy, time-discounted).

In insurance-based social settings, these same costs will arise but are distributed differently - often appearing in the form of insurance premiums to be paid by an individual or their employer. It needs to be carefully noted that the absence of a welfare system does not cause these costs to vanish. Failure to track them down and account for them results in a cost of non-circumcision being omitted from the CBA.

With so many issues having been omitted from previous CBAs, it is possible that the benefits side of the CBA may have been grossly under-estimated in the past. That said, a rush to include new issues in a revision should not lose sight of reality. Not only must it be established to normal standards of scientific proof that prophylaxis exists, but also the efficacy of the prophylaxis must be determined in each case. The degree of protection conferred by MC is rarely perfect and may vary with 'style' of circumcision, especially when transmission routes involving Langerhans cells are involved.

Calculating the value today of a benefit not received until many years hence involves an actuarial procedure that has been understood for several centuries. Put in simple terms, it involves a "Compound Interest" calculation done in reverse.

For the benefit of those not familiar with techniques of valuing the avoidance of future expenditure, here is a brief explanation:

Understanding "Compound Interest" involves nothing more than being able to answer this example question:

"If I put $100 into a savings account earning 5% per annum and do not withdraw the interest, how much money will I have in the account in five years time?"

From the first interest payment onwards, the capital sum isn’t just the $100 originally deposited. During the next interest-earning period, the "capital" is that original $100 PLUS the interest accrued to date. The interest, in the jargon, is "compounded".

At the end of the five years, the account will contain more than would have been the case had the second and subsequent interest payments been based only on the original $100. Substantially more when interest rates are high, only a little more when interest rates are low. But always more.

So what’s this got to do with the discounting of future expenditure? Suppose that in thirty years time you face a medical bill of $10,000. How much money do you need to set aside NOW in order that, when the invoice hits the doormat, you have funds in hand ready to pay?

Answering that question involves a calculation that is, in essence, a Compound Interest calculation done backwards. Well-established formulae and tables assist with the mathematics here; no new research is involved.

Why, in the realms of Cost Benefit Analysis (CBA) relating to male circumcision (MC), do we need this?

Put yourself in the position of a Health Administrator tasked with making the decision as to whether infant circumcision should or should not be funded by the healthcare provider. By the time the matter arrives on his office desk, a prophylactic effect has been proved to exist. But is the expenditure on infant circumcision "worth it" (in purely financial terms) viz-a-viz the risk of having to spend a greater sum on some cure or other at a later date?

The answer involves a considerable number of variables, including the local probability of suffering this or that medical condition as a result of not being circumcised and the local cost of treating it. The figures will vary from country to country, but the variables to be considered will not. Hence there appears to exist the possibility of a "universal skeleton spreadsheet".

One variable needed in every calculation relating to "deferred expenditure" is Interest Rate; this follows from the Compound Interest issue discussed above. Of late, the global economy has seen a marked fall in interest rates and the predictions are that the present low rates are likely to continue for some considerable time. Therefore we need to analyse the effect of this change on previous CBAs, with the objective of determining whether the conclusions have been invalidated in consequence.

Consider again that bill for $10,000 payable in thirty years time. According to Parry’s Valuation Tables (Present Value / No Tax):

At 0% interest rate, we need to set aside $ 10,000 now.
At 2% interest rate, we need to set aside $ 5,520.71 now.
At 4% interest rate, we need to set aside $ 3,083.19 now.
At 6% interest rate, we need to set aside $ 1,741.10 now.
At 8% interest rate, we need to set aside $ 993.77 now.
At 10% interest rate, we need to set aside $ 573.09 now.

So, at 10% interest rate, the amount we need to set aside is a mere 5.73% of the "cash stuffed under the mattress" situation. But 10% interest rates are a thing of the past and seem unlikely to return in the short or medium term. The idea that pushing expenditure into the far future reduces its cost near to zero is no longer valid.

In the real world we also have to take into account the dreaded concept of inflation. Medical treatment costed at $10,000 now won’t be available for $10,000 in thirty years time!

Predicting future inflation isn’t quite as daunting as it sounds. Taking a long-term average, stable economies tend to inflate by about 2% per annum. So in thirty years time today’s $10,000 treatment is likely to cost $18,114. In forty years, $22,080. Fifty years hence, $26,916.

Back to our example of $10,000 payable in 30 years time. Assuming 2% per annum inflation in the cost of the treatment, what then are today’s set-aside figures?

At 0% interest rate, we need to set aside $ 18,114 now.
At 2% interest rate, we need to set aside $ 10,000 now.
At 4% interest rate, we need to set aside $ 5,584.89 now.
At 6% interest rate, we need to set aside $ 3,153.83 now.
At 8% interest rate, we need to set aside $ 1,800.11 now.
At 10% interest rate, we need to set aside $ 1,038.10 now.

Note that all this works independent of the unit of currency. Pounds, Australian Dollars, Euros, US Dollars... The numbers remain the same.

Now pause and inspect the figures immediately above. Note that if interest rates remain low (as they are predicted to do), there is little if any saving to be made by deferring expenditure. If interest rates fall below the rate of inflation, deferral does not save money - deferral then costs money.

Thus one needs to be very suspicious as regards the ongoing validity of CBAs done at a time when "real" interest rates (actual rate minus inflation rate) were well to the positive side of zero. A simplistic reading of their conclusions, made without a re-run of the calculations relating to the discounting of future expenditure, will mislead in a way that spuriously favours deferred treatment over immediate prophylatic action.

--

Statistical Issues. [Updated 03.Aug.2010]

When designing a new CBA, it needs to be borne in mind that many of the input variables relating to the deferred expenditure option carry with them considerable degrees of uncertainty. Risk analysis techniques need to be adopted, typically "Monte Carlo simulation".

Waskett, J (Pers. Corr, 2010) has suggested that any form of analysis set up to re-evaluate the overall cost-benefit of infant MC should be trialled by using it to re-analyse Schoen’s data. Such an approach would give a preliminary indication of the gearing of change brought about by changed economic circumstances, giving a pointer to the extent to which the whole re-evaluation exercise is justified as well as testing the mathematical model.

Also in need of testing will be correct handling of differing rates of "herd immunity" arising from different take-up rates of infant circumcision. The two are not part of a linear, proportional relationship. The proposition here is that the prophylactic efficaciousness of any one circumcision varies, in public health terms, according to the rate of circumcision amongst the individual male’s peer group. For comparison, consider herd immunity in the context of needle-jab vaccination campaigns.

The mathematical model will also need to ensure that the population at risk is properly adjusted for deaths due to causes considered elsewhere within the scope of such a wide-ranging study. One is not at risk of manifesting Disease A if already deceased as a result of Disease B. Fatal diseases outside the scope of the study do not come into play here; being circumcision-neutral they are accounted for by the general figure for life expectancy. But as soon as one fatal disease is involved within the circumcision CBA, the risk of a person having to be treated for some other condition diminishes.

--

Spreadsheets produced by Futures Institute. [Item added 16.Jul.2013]

A number of visitors to this web page have asked why no reference is made to the Futures Institute Male Circumcision: Decision Makers’ Program Planning Tool compiled on behalf of USAID and WHO. Let’s put that right. Those preparing the Planning Tool were tasked with analysing only the cost-effectiveness of male circumcision as a means of curbing the HIV epidemic. That they did, with considerable success. You can see the results here:

Overview: http://futuresgroup.com/files/softwaremodels/MC_manual.pdf

There is also an Index to Spreadsheets. (See right-hand panel on the page)

The problem perceived by Circlist is that the model quantifies only the one benefit of HIV prophylaxis. That’s the same mistake as is often made by critics of circumcision. Circumcision is a multi-benefit procedure; the true gain only becomes apparent when all the benefits are quantified and included in the spreadsheets. Thus it is Circlist’s view that the Futures Institute mathematical model seriously understates the totality of the benefits.