Suppose your job is to treat mental illness in a community. You get a range of patients referred to you, and you have a range of resources available to treat them. A good way to allocate resources is to use a stepped-care model, right?
How it works
To see how it works, let’s simplify the stepped-care model to just three steps.
Each patient referred to you gets a cheap treatment first. It works for some, but not for others. So those others get a more expensive treatment. Again, it works for some but not for others, and those others get the most expensive treatment.
This gives you a way to allocate your range of resources. You can limit use of the expensive treatment to those patients who really need it, and you can tinker with the resource allocation on each step, matching resources to demand, so service managers have something to do, too. It makes everyone look busy.
By the way, the cheap treatment is usually referred to as “low-intensity” because it is not so resource-intensive as the other treatments. The term “low-intensity” does not refer to any therapeutic intensity (whatever that might mean), only to the cost. A simpler word for “not so resource intensive” is “cheap”.
But does this kind of model deliver value to your community and to the individual patients?
The most obvious problem with the model is that some patients get multiple treatments. Suppose a patient has a complex illness. This patient is going to be treated three times for the same illness. The first two treatments fail, because the complex illness really needs the most expensive treatment.
So this model fails to deliver value to the community, because it wastes treatment resources on treatments that don’t work, and wastes patients’ time while they receive those ineffective treatments. (This is not to say that cheap treatments don’t work at all, only that cheap treatments don’t work for everyone.)
The usual solution to this problem is to fast-track identified patients, who bypass one or more steps and get effective treatment immediately.
Fast-tracking introduces problems of its own, however, and these problems can outweigh any advantages. There’s a theoretical problem and a practical one.
The theoretical problem with having a fast track is that it causes the stepped-care model to break down. If you fast-track everyone who needs it, then you no longer have a stepped-care model at all. Instead, you have a triage model, where you assess every patient and allocate treatment accordingly.
And this leads to the practical problem. The triage model doesn’t work well in mental health because you cannot easily predict how patients will respond to treatment. So what typically happens in practice is that a small proportion of patients get fast-tracked based on the severity of their symptoms, not on the complexity of their illness.
Severe but simple cases end up getting expensive treatment, when they could have been treated more cheaply. Complex but only moderately severe cases get treatments that don’t work while they climb the steps.
Returning to the stepped-care model itself, another problem is its failure to address progressive illness.
The model assumes that once a patient is referred, the goalposts don’t move while you work through the stepped-care process. Unfortunately some patients are referred with illnesses that become more complex as time goes on.
Suppose a patient is referred with a condition that happens to require level-two treatment. Instead of getting level-two treatment immediately and leaving the system, the patient gets level-one treatment first, which doesn’t work. Meanwhile, the illness progresses. By the time the patient gets level-two treatment, the illness requires level-three treatment. So again, the treatment fails to address the illness. In some cases, by the time the patient gets level-three treatment, the illness is so complex that it cannot be treated.
The problem with progressive illnesses is made worse by waiting lists, and by merry-go-rounds and slides.
Resource limitations mean that treatments must be rationed in any practical system. The rationing method of choice in the NHS (which cannot use price) is to use a waiting list. Stepped-care models can incorporate a waiting list at each step, so those patients who have to climb all the steps can find that it takes them a very long time.
It’s not obvious how a waiting list achieves rationing. Suppose your community generates a thousand patients a year for you to treat. If you treat them all immediately when they are referred, then you need enough resources to treat a thousand patients a year. But if you operate a one-year waiting list, although every patient has to wait a year you still have to treat a thousand patients a year. The waiting list does not result in any resource savings.
What really happens is that when you force patients to wait a year for treatment, some of them just go away. This happens for various reasons. Some leave your community, dying of old age or by their own hand, or emigrating. Some buy external treatment (adding resources to the community’s overall mental health provision, even if not to your budget). Some come to accept their illness as permanent disability, and stop seeking treatment.
While all of these mechanisms make wait-list rationing an effective way for you to limit the resources you need for your service, they transfer the cost to your community and to the patients themselves.
You can see these mechanisms at work in your community’s suicide statistics, in its use of private healthcare and alternative medicine, in its need for both funded and voluntary support services for the permanently ill, and in the prevalence of self-medication with alcohol, tobacco and other non-prescription drugs.
Merry-go-rounds and slides
Allied to wait-list rationing, merry-go-rounds and slides have the same effect on patients, making them wait to move from step to step, and encouraging some to go away without ever receiving effective treatment.
A merry-go-round occurs when a patient is re-referred to the same treatment step. A typical example involves the use of SSRIs. A patient arrives in primary care with complex symptoms that include depression. Depression, you say? Try an SSRI!
The patient returns to say that nothing has changed. Give it time, you say. These drugs need time to take effect.
The patient returns to say that nothing has changed. Let’s try another SSRI, you say. These drugs are not all exactly the same.
And so it goes on. The SSRI merry-go-round can easily take a year or two.
Another example is fake CBT. A patient with a complex illness is recommended a self-help “CBT” book and tries to treat the symptoms. When this fails, the patient is referred to a computer program that uses the same methods and fails in the same way. Then a group led by someone who is not really a psychotherapist fails again. Then one-to-one therapy using the same techniques fails again. Applying the same treatment again and again with different packaging keeps the patient on the same step in the model.
A slide is a back-referral from a higher step to a lower step. For example, a patient has been treated in primary care (step 1) for a year with no improvement, and waited a year for a psychiatric assessment (step 2). The psychiatrist recommends continued treatment in primary care (back to step 1).
To summarize, even if a stepped-care model is implemented in an ideal way, it wastes resources by giving some patients ineffective treatments, and it can fail to treat progressive illness fast enough.
Typical implentations are less than ideal, when they incorporate wait-list rationing with merry-go-rounds and slides to prevent patients from moving from step to step. These methods create cost in the service by multiplying the number of ineffective treatments delivered, and they create cost in the community by increasing the amount of support needed for the long-term or permanently ill.
To combat these problems, the steps in a stepped-care model need to implement fast failure. Fast failure means that a patient stays with an ineffective treatment for as short a time as possible — only long enough to evaluate whether the treatment looks likely to work.
The system needs to have rules that prevent merry-go-rounds and slides. Thus, you can make a decision about whether to try a self-help “CBT” book, a computer program, a group or an inexperienced therapist, but you cannot try all four in succession if they are simply the same technique repackaged. You have to try just one of them, and only for long enough to make a judgement about whether it is going to help.
Cost savings realized by avoiding multiple ineffective treatments can then be used to make wait-list rationing less necessary, freeing up support resources in the community. And those support resources can then be focused on patients whose illnesses really are long term and resistant to treatment, so that you are not running a merry-go-round on the topmost step.
I started this post by writing, “Suppose your job is to treat mental illness in a community.” But now suppose it is not. Suppose you are a patient in a stepped-care model. You can implement some of these ideas for yourself, recognizing the typical failings in the model and doing your best to correct them.
- Don’t try a failed treatment longer than necessary before moving on.
- Don’t accept a failed treatment for a second time, even if it’s packaged differently.
- Don’t accept back-referral to an earlier step where you have been before.
- Don’t accept permanent disability (or emigration or death) when you have a treatable illness.
- Do take action if your illness is getting worse faster than the treatment model is treating it.
Applying corrections like these might take some effort. You might want to seek a mental health advocate or even a specialist lawyer to represent you. You might have to get involved in formal complaints, and even the courts. But good, effective treatment is available even in some of the worst stepped-care models — it’s just that not everybody receives it automatically.