Show me that you can divide the notes of a song;

But first, show me that you can discern

Between what can be divided

And what cannot.

An anonymous musical composition inspired by a classical Sanskrit poem.1


Two questions that often arise in disease litigation are whether a condition is divisible or indivisible and whether it is possible to apportion damages between causes or tortfeasors. In many cases, the answers are well-established and uncontentious. But in novel cases, and in cases of psychiatric injury in particular, they can be difficult to answer.

The purpose of this short article is to look at the applicable principles in such cases.

(Apportionment in mesothelioma cases and the decision in Barker v Corus (UK) plc [2006] 2 AC 572 are beyond the scope of this article.)

The Starting Point

The law’s general approach in this area was summarised by Lord Phillips of Worth Matravers in Sienkiewicz v Greif (UK) Ltd [2011] 2 AC 229 at [90]:

Where the disease is indivisible … a defendant who has tortiously contributed to the cause of the disease will be liable in full. Where the disease is divisible … the tortfeasor will be liable in respect of the share of the disease for which he is responsible.

(See also Williams v Bermuda Hospitals Board [2016] AC 888 at [31] per Lord Toulson.)

How, though, does the law determine whether a condition is divisible or indivisible? And is it possible to apportion damages even where a condition is indivisible?

To answer those questions, it is necessary to consider two key authorities: Rahman v Arearose Ltd [2001] QB 351 and BAE Systems (Operations) Ltd v Konczak [2018] ICR 1. 

The Key Authorities

In Rahman, the Claimant was assaulted in the course of his employment and suffered a serious facial injury. His employers were held liable in negligence for not putting in place sufficient protection. His injuries were negligently treated in hospital, causing him to become blind in one eye. He also developed a serious psychiatric illness. One of the issues before the Court of Appeal was whether his psychiatric illness and its consequences should be regarded as “the same damage” within the meaning of the Civil Liability (Contribution) Act 1978.

Laws LJ held that:

  • at common law, wrongdoers who together cause “a single indivisible injury” are each liable to compensate the Claimant for the whole of the injury suffered (see [17]); and
  • an injury is to be regarded as single and indivisible “where there is simply no rational basis for an objective apportionment of causative responsibility for [it]” (see [19]).

The expert psychiatric evidence in Rahman delineated the Claimant’s overall condition into four separate components (post-traumatic stress disorder, a severe depressive disorder, a specific phobia and enduring personality change) and, to some extent, was able to apportion the cause of those different components between the initial assault and the hospital’s negligence. Laws LJ held that in those circumstances the injury should be regarded as divisible, admittedly on a ‘rough-and-ready’ basis. The court accordingly approved the trial judge’s apportionment of 75% to the hospital and 25% to the employers.

BAE Systems was an employment case. The Defendant was responsible for acts of unlawful discrimination, giving rise to a claim for compensation for injury to feelings and psychiatric injury. The Claimant, however, had a history of stress and difficulties at work involving events for which the Defendant was not responsible. The question arose whether the Claimant’s psychiatric injury was divisible. 

Underhill LJ noted at [58] that there is often a well-recognised rational and objective basis for apportionment in industrial disease cases where an industrial injury has become worse as a result of exposures at work over a long period and where the Defendant is only responsible for negligent exposure during part of that period. In such cases, the court can make an assessment, however broad-brush, of the degree of disability attributable to exposure during the period for which a particular employer is responsible. This is usually done on a time-exposed basis.

Examples of such cases include:

  • noise-induced hearing loss (Thompson v Smiths Shiprepairers (North Shields) Ltd [1984] QB 405);
  • asbestosis (Holtby v Brigham & Cowan (Hull) Ltd [2000] ICR 1086); and
  • hand-arm vibration syndrome/vibration white finger (Allen v British Rail Engineering Ltd [2001] ICR 942). 

Underhill LJ went on to consider psychiatric injury caused by stress at work, citing Hale LJ’s “practical propositions” in Hatton v Sutherland [2002] ICR 613 at [43]. Proposition 15 says this:

Where the harm suffered has more than one cause, the employer should only pay for that proportion of the harm suffered which is attributable to his wrongdoing, unless the harm is truly indivisible. It is for the defendant to raise the question of apportionment …

Proposition 16 is this: 

The assessment of damages will take account of any pre-existing disorder or vulnerability and of the chance that the claimant would have succumbed to a stress-related disorder in any event.

These propositions were criticised by Smith LJ in Dickins v O2 plc [2009] IRLR 58 at [45]-[47]. She questioned whether proposition 15 could ever apply in practice, because psychiatric injury is always indivisible, and suggested that justice can be sufficiently achieved by the application of proposition 16 alone. Speaking extra-judicially in 2008, Smith LJ said this: 

I do not think that one can apportion damages for psychiatric injury. It seems to me that it is par excellence an indivisible injury.

Underhill LJ in BAE Systems considered that Smith LJ had overstated the difference between her position and that set out in Hatton. At [67], he reconciled the opposing views in this way:

As regards cases where all that is being said is that the claimant had a pre-existing vulnerability to psychiatric injury, I understand the guidance in Hatton to be that any reduction, or discount, should indeed be made by the application of proposition 16, as Smith LJ herself suggests, rather than by apportionment in accordance with proposition 15. And even in cases where there are “multiple extrinsic causes”, the court in Hatton says only that a “sensible attempt” should be made to apportion the harm between what is and is not attributable to the defendant’s wrong. It recognises that there may still be cases where the harm is “truly indivisible” and that in such cases apportionment would be wrong. There is thus no difference as to the applicable principle, which indeed is authoritatively stated in Rahman v Arearose Ltd [2001] QB 351. The difference is that Smith LJ believes that in the case of psychiatric injury the harm will always be indivisible, whereas the encouragement in Hatton to find a basis for apportionment where possible means that the court believed that the harm would be divisible at least sometimes.

The Correct Approach

Having considered the authorities as summarised above, Underhill LJ gave the following guidance in BAE Systems at [71]-[72] (emphasis added):

What is therefore required in any case of this character is that the tribunal should try to identify a rational basis on which the harm suffered can be apportioned between a part caused by the employer’s wrong and a part which is not so caused. I would emphasise, because the distinction is easily overlooked, that the exercise is concerned not with the divisibility of the causative contribution but with the divisibility of the harm. In other words, the question is whether the tribunal can identify, however broadly, a particular part of the suffering which is due to the wrong; not whether it can assess the degree to which the wrong caused the harm.

That distinction is easy enough to apply in the case of a straightforward physical injury. A broken leg is “indivisible”: if it was suffered as a result of two torts, each tortfeasor is liable for the whole, and any question of the relative degree of “causative potency” (or culpability) is relevant only to contribution under the 1978 Act. It is less easy in the case of psychiatric harm. The message of Hatton is that such harm may well be divisible. In Rahman the exercise was made easier by the fact (see para. 57 above) that the medical evidence distinguished between different elements in the claimant’s overall condition, and their causes, though even there it must be recognised that the attributions were both partial and approximate. In many, I suspect most, cases the tribunal will not have that degree of assistance. But it does not follow that no apportionment will be possible. It may, for example, be possible to conclude that a pre-existing illness, for which the employer is not responsible, has been materially aggravated by the wrong (in terms of severity of symptoms and/or duration), and to award compensation reflecting the extent of the aggravation. The most difficult type of case is that posited by Smith LJ in her article, and which she indeed treats, rightly or wrongly, as the most typical: that is where “the claimant will have cracked up quite suddenly, tipped over from being under stress into being ill”. On my understanding of Rahman and Hatton, even in that case the tribunal should seek to find a rational basis for distinguishing between a part of the illness which is due to the employer’s wrong and a part which is due to other causes; but whether that is possible will depend on the facts and the evidence. If there is no such basis, then the injury will indeed be, in Hale LJ’s words, “truly indivisible”, and principle requires that the claimant is compensated for the whole of the injury – though, importantly, if (as Smith LJ says will be typically the case) the claimant has a vulnerable personality, a discount may be required in accordance with proposition 16.

On the evidence in BAE Systems, and considering the claim’s convoluted procedural history, there was no rational basis for apportionment and the Defendant’s appeal was accordingly dismissed.

In a short concurring judgment, Irwin LJ said this at [92]-[93] (emphasis added):

As a matter of principle, and supporting the fundamental approach that compensation should never become windfall, where an injury is divisible, even if on a rough and ready approach to the division, recompense must be limited to the consequences of identified injury attributable to the tort in question. I further support the proposition that it will often be appropriate to look closely, particularly in a case where psychiatric injury proves indivisible, to establish whether the pre-existing state may not nevertheless demonstrate a high degree of vulnerability to, and the probability of, future injury: if not today, then tomorrow.

In my view, the problem exposed here, properly analysed, is not so much a problem of law as a problem of medicine or science. The territory between the non- pathological but sensitised and vulnerable individual and the person with a defined pathology constitutes highly debatable land. It should be closely and carefully mapped by the relevant experts, and it is imperative that they should bring to bear as much clinical and diagnostic precision as possible, paying close attention to one or both of the internationally recognised psychiatric diagnostic systems. In particular, it is necessary to consider whether a less serious but nevertheless established and defined disorder may not have been achieved before progression to the diagnostic end-state. In addition, it should be routine for the experts to assess the level of risk of crossing the borderland between non-pathology and pathology through some other stimulus than the tortious act or omission. It will be recognised that exercise is often difficult and uncertain, but it will often be possible to give such advice within reasonable parameters of time and to the level of probability. Such an exercise is necessary in order to address proposition 16 in Hatton.


It is suggested that Irwin LJ was right that the questions whether a condition is divisible or indivisible, and of the appropriate apportionment, are more questions of medicine and science than of law. The legal principles are clear, and in cases such as Rahman it was the expert evidence that enabled the court, in applying those principles, to apportion damages accordingly. By contrast, in BAE Systems the lack of evidence led to the court declining to make any apportionment.

The aim in cases of psychiatric injury – and, it is suggested, in novel cases more generally – is thus to establish, by expert evidence, a rational basis on which the harm suffered can be apportioned. Particularly at the boundaries of medical and scientific understanding, this may be easier said than done.

1 As quoted in The Gene: An Intimate History by Siddhartha Mukherjee (Random House, 2016), p. 485.