I'll take my chances

Suppose a patient presents at their doctor with an array of symptoms that the doctor subsequently diagnoses and treats. The patient's condition deteriorates further, and so he consults with other doctors and it turns out he has stage four cancer, which is incurable. For argument’s sake, medical expert opinion shows favourable evidence illustrating that when the patient first consulted a doctor, he was in the early stages of the cancer, which subsequently progressed to stage four due to the lack of treatment. The patient can illustrate that the initial doctor was negligent, however he cannot prove (with any certainty) that, in his individual circumstances even with the appropriate treatment, his cancer would not have progressed as far, or at what speed it would have progressed. Furthermore, the patient cannot show with any certainty what position he would have been in had he been diagnosed correctly. Is it fair that the patient should have no claim in our law?

Taking a chance

"[The doctrine of the loss of chance] permits a plaintiff to institute an action against a defendant for the loss of a chance of avoiding a result rather than merely for the result itself."

I am hard put to find a quote that encompasses the essence of the doctrine of a loss of chance quite as succinctly as the one above from the book The application of the doctrine of a loss of chance to recover in medical law by P van den Heever. In terms of this doctrine, a plaintiff is entitled to claim compensation that is proportional to the "proximate degree" by which a defendant negligently reduced or destroyed a chance he had in avoiding an injurious outcome, even where it cannot, on a balance of probabilities, be proved that the injurious outcome would not have resulted anyway. There has been sparse debate surrounding the introduction of the loss of chance doctrine into South African medical negligence law. This is surprising considering that it has the potential to do justice to both plaintiff and defendant, in that plaintiffs will be entitled to recover damages in cases where a loss of chance to avoid an injurious outcome occurred, however, a defendant's liability for damages will be in approximate proportion to their causal contribution to the injury.

Important legal considerations
According to J Neethling and JM Potgieter in Neethling-Potgieter-Visser Law of Delict our law follows an all-or-nothing approach in claims arising out of a conduct wrongfully and negligently caused. In terms of this approach, a plaintiff must claim for all damages already sustained or expected in the future, based on a single cause of action.

A cause of action plays an important role when considering the once-and-for-all rule, in that it determines from when a defendant becomes liable. It is generally accepted that a cause of action arises at the earliest stage when all the requirements for delictual liability are met and, insofar as damage is concerned, at the earliest stage when the first damage (even if not all the damage) occurred. The problem in our patient’s case is that causation is indeterminable due to the array of potential causes for his injurious result, all of which are hypothetical. One cannot so easily place a plaintiff in the position they would have been, when that position is not only unknown, but will never eventuate. In the normal course, the result may have in any event occurred, in which case the plaintiff had no chance at all.

As highlighted in the English case of Gregg v Scott, hypothetical events pose a problem in that they neither happened in the past, nor are they to happen in the future. In claims of this nature, a defendant's wrong prevents a result from ever materialising. According to Van den Heever, to achieve a just result in cases such as these, courts must endeavour to define a plaintiff's actionable damage in terms of the chance or opportunity lost, rather than by reference to the loss of the desired outcome (which was never within his or her control). 

Our court's approach thus far
Where exact quantification of damages is not possible, the court in the case of De Klerk v ABSA Bank Ltd, citing the English case of Allied Maples Group Ltd v Simmons & Simmons with approval, stated that differing standards of proof apply to cases where the issue is one of causation, or one of quantification.

The court in De Klerk went on to find that if a plaintiff is able to prove causatively that he had a real or substantial chance as opposed to a speculative one, then a court is entitled, in quantifying damages, to estimate the chances that will not require proof on a balance of probability.

De Klerk also quoted the case of Burger v Union National South British Insurance Company, a case which stipulated that where all available evidence provided by a plaintiff fails to alleviate the uncertainties present on the issue of quantum, a court is to do the best with the material available and, as a last resort, make an "informed guess". The court may not, in such a case, select from a range of possibilities the one least favourable to the plaintiff, merely because he bore the onus of proving the preferred possibility. Instead, it is recognisable in our law to have regard to relevant events that may occur, or relevant conditions that may arise in the future, even where it cannot be shown on a balance of probabilities that they will occur. The court provided the following illustrative example:

"If, for example, there is acceptable evidence that there is a 30 per cent chance that an injury to a leg will lead to an amputation, that possibility is not ignored because 30 per cent is less than 50 per cent and there is therefore no proved preponderance of probability that there will be an amputation. The contingency is allowed for by including in the damages a figure representing a percentage of that which would have been included if amputation had been a certainty. That is not a very satisfactory way of dealing with such difficulties, but not better way exists under our procedure."

Although our courts have dealt with future uncertainties, the loss of chance doctrine poses the even greater problem of evaluating a hypothetical future outcome that can never result, no matter what the probabilities. The De Klerk case illustrates that our courts have indeed applied the doctrine to cases of professional negligence and loss of commercial opportunity. Furthermore, damage has been assessed even where there is a less than 51% chance involved. Despite these developments, the doctrine cannot so easily be imported into cases of medical negligence, and the challenges facing an attempt to do so will now be examined.

Statistical limitations

Van den Heever highlights the problems the doctrine poses and the deficiencies of statistical evidence very efficiently in his introduction. By way of his illustrative example, the following can be noted: Where available, medical research can provide that, in a given case, a population's chances of recovery for a particular injury are 25:75. It is logical to conclude that someone who does not recover has not lost a 25% chance because, if he never would have fallen within the 25%, he has lost nothing at all. Such a plaintiff must show that he falls within the 25% and lost this chance due to the defendant's negligence. This poses the further problem that, if he could show he falls within the 25%, he would essentially have proved his case against the defendant and would be entitled to 100% of the damages and not 25%. The very nature of the problem in loss of chance cases is that the plaintiff is not able to prove whether he was one of the 25% versus one of the 75%.

Statistics are criticised for their potential to lead to inexact results that entail "instinctive guesswork". Arguments in favour of statistical evidence propose that, since statistics arise from previous experience of similar cases, they can provide a more accurate weighting on potential causes.

In an article by L Steynberg "Fair" mathematics in assessing delictual damages the learned author considers the anomaly that past and future losses are not treated equally in that the percentage proof of a past loss awards a plaintiff 100% damages, while the same percentage proof of a future loss awards a plaintiff merely that percentage of the damages. It is posed, quite rightly, that there may be no logical reason for such a distinction and that this is a result of the methods courts have used. Since the "winner takes all" approach cannot do justice between parties, perhaps either the principle of sharing losses or, alternatively, that judges use their discretion of fairness and justice, explains the treatment of future losses by courts.

Van den Heever considers various arguments on the role statistics play. In light thereof when dealing with medical negligence claims of such nature, we can be assisted by looking at the statistical probability of survival at the various stages of that particular cancer. This statistical probability would be informed by the long-term survival of previous patients with the same cancer. The pitfalls of applying a statistical chance of long-term survival based on past cases, is that this fails to consider the individual chance of the particular patient. The patient may have fallen in the percentage who would have survived, meaning the failure to diagnose deprived him of this chance, or he may have fallen in the percentage who would have died regardless, in which case he had no chance at all. This invites the argument that a statistical chance alone has no value unless and until it can be personalised.

In contrast, proponents of statistical chance argue that medical research is typically tailored to account for those who are matched, as far as possible, to the same physical and biological factors. For example, a particular form of cancer in people of the same gender, around the same age, all of whom maintain a certain level of fitness, are of a healthy weight, don't drink in excess and are non-smokers.

Furthermore, regardless of its all-or-nothing approach, accurate statistics enable reasonable fairness to patients who have chances nearer to zero and 100% of the predicted outcome. In respect of patients closer to zero, accurate statistics could enable an evaluation of a chance at various stages, so that a patient who maybe had only a 10% chance that was diminished to 5% can be said to have lost a 50% chance. Similarly, if the same patient's chances were diminished in their entirety, the plaintiff would have lost a 100% chance. Along these lines, Lord Nicholls in the English case of Gregg v Scott stated that "despite its imperfection, in practice statistical evidence of a diminution of perceived prospects will often be the nearest one can get to evidence of diminution of actual prospects in a particular case. When there is nothing better courts should be able to use these figures and give them such weight as is appropriate in the circumstances".

Where patients had a 50/50% chance of recovery, both parties would likely be happy with 50% of the full damages, as the plaintiff feels he has recovered at least something, while the defendant won't feel he has paid too much. 

When it comes to actually quantifying damages, Van den Heever considers various authorities of differing jurisdictions that place a different emphasis on what percentage is regarded as "speculative" versus "substantial", of which my preferred view is that of the learned author himself:

"In cases where the circumstances allow for the application of loss of chance, the principles with regard to quantification should be applied to both claimant and defendant equally. Should the chance, for example, be valued at 75 per cent, the claimant should be entitled only to 75 per cent of the damages he is able to prove."

It is this argument that can counter the problems the all-or-nothing rule poses to the loss of chance doctrine. Simply because statistics can aid plaintiffs in cases where chances are evaluated as closer to zero or 100%, their role would not promote fairness if used to impose adverse damages awards on defendants merely to protect plaintiffs who may have had legitimate claims, but courts choose not to entertain them because they are considered too remote. Neither, should the damages diminish the closer a plaintiff gets to a chance of 50%, even though the parties would consider damages in the region of 50% to be fair. Courts are meant to apply a uniform standard and perhaps statistics, and their accuracy, are not to be used to provide absolute results where a plaintiff's chance is remote, but rather a fair and proportionate result in all cases, where statistics can reflect a percentage chance that a plaintiff had.

A "chance" for South Africa

Considering that the loss of chance doctrine has been used in cases involving professional negligence and loss of commercial opportunity, I suggest that the time is ripe for our jurisdiction to consider the doctrine's applicability to cases of medical negligence. Van den Heever arguably accounts for most of the hurdles that would need to be overcome in his recommendation and conclusion.

Once again, the learned author emphasises that the doctrine provides justice to both parties as the burden of proof facing a plaintiff is alleviated, while a defendant's damages are limited. As endorsed by the court in De Klerk, in future hypothetical instances, causation is not established on a balance of probabilities, but on the court's assessment of the chances of the risk eventuating. Furthermore, a plaintiff can discharge the onus of proving a chance if it can be considered "real" or "substantial", rather than "speculative".

Van den Heever acknowledges that there are obstacles in assigning a value to a chance, however that they "are not insurmountable" and that policy considerations perhaps need to guide the setting of an objective measure. The learned author proposes that although hypothetical future events are notoriously difficult to assess, that is not a valid argument for awarding nothing. Furthermore, fears that acknowledging the doctrine will open the floodgates for highly speculative claims can be countered by the fact that other jurisdictions, which have recognised the doctrine, have managed to contain this aspect. Courts in these jurisdictions have been able to distinguish between "substantial" and "speculative" claims and our courts can seek guidance from them.

Finally, Van den Heever notes that common law jurisdictions, including South Africa, have clearly established that in professional negligence and commercial cases, questions on what would have occurred but for the negligence of a defendant, or future occurrences, can be decided on probabilities that are less than 50% with damages adjusted accordingly. The same cannot be said for the application of the doctrine in medical negligence cases. The learned author attributes this to a conflict between the courts not wishing to depart from traditional and entrenched principles of causation on the one hand, and the duty to compensate worthy claimants on the other. Van den Heever seems to allude to the fact that a departure from the all-or-nothing approach should be applied across the board, even to cases where previously a claimant would have recovered 100%, so that damages are proportional to the percentage established by the court in each case.


We should recognise that medical statistics can provide valuable guidance to a court in cases where a claim pertains largely to the statistical probabilities of a particular treatment regime proving to be successful or not.  The current approach being adopted fails to take this into account, with the result that full damages awards are imposed on defendants in all cases where a chance can be proved, where they would otherwise have been liable for nothing – thus these cases are treated as what is commonly referred to as "one-percenter cases".  Rather, in fairness to both parties, damages should be proportionate to the percentage at which a chance is valued by a court and that the very nature of "a chance" renders it illogical to suggest that only chances 51% or more should attract liability. While our courts have to date struggled to align the doctrine of chance with the Aquilian Action, strong arguments can be made that it could be taken into account when dealing with issues involving causation and, if not, then under the question of wrongfulness.

Download PDF Share Back To Listing
Loading data