The Common Sense Guide to Improving Palliative Care > 2.5 Identifying and Testing Changes

Sick To Death book cover This extract from the online edition of The Common Sense Guide to Improving Palliative Care is used with permission.

Identifying and Testing Changes

With your aim and measures stated, your team needs to decide which changes to try. Of course, some interventions that you think will work may surprise you and behave otherwise. (That is why we test!) And remember that it is generally risky to implement a good idea on a large scale until you have it working well in one area with a few patients and staff.

How Do You Identify Which Changes to Make?

Do's and Don'ts

Good ideas are found everywhere; the challenge is to select the ones that are most worthwhile for you to try in your organization.

How Will I Know If an Idea Will Work for My System?

Measurement tells you what works and what does not.

Once you have identified changes to try, assess them on a small sample of patients, clients, or nursing home residents. Testing is a step-by-step process to help identify whether or not something is working. If the change does not have a good effect on your outcome, testing will give you some ideas on how to adjust it or whether to try something else. The four steps outlined below are the heart of the rapid-cycle QI model.

Step 1: Plan By now, you have a change you would like to try. You have to plan for the test, so consider the following:

Step 2: Do After you have planned your change, try it. Be sure to do the following:

Step 3: Study After implementing the change and collecting the data, get together with your team to analyze what happened, look at what the data say about the change, and summarize what you have learned from this test. Did your data and observations match what you had predicted for this change? If yes, what else happened? If not, then why not? Even failed tests teach teams a great deal about their care system.

Step 4: Act By now, you will know whether the change you tested is working. If it is, you can plan to implement it on a larger scale (e.g., with more patients or in another setting). However, if you find that the change did not work well on the small group, the team needs to evaluate what happened so as to get ideas on how to modify the change or whether to try something else. Either way, the cycle moves back to step one.

Keep trying until you find what works in your system. Generally, try many small changes, one at a time, rather than gambling on one large change. Let your mantra be: "We are going to give it a try; if it does not work, we will learn something from it."

Matching Aims, Measures, and Changes

You may find it surprisingly difficult to get an aim, a measure, and a change to fit together. It is important to work on this until it is right. A frequent error is mistaking a change for an aim. For example, a team might say their aim is that "within one month, every patient will have an assessment for symptoms within four hours of admission." At first, that sounds like a worthy aim, but it has a serious flaw: you do not actually care about this aim. The reason to want to assess symptoms is so that you can relieve them. That is the real aim. Doing the assessment quickly is an important change, one that might allow you to achieve the aim of having patients be comfortable, quickly; but it would not accomplish that on its own. So, incorporate the real goal of comfort into the team's aim statement, and make the changes serve that aim. Sometimes, the change is so closely tied to the aim that it hardly matters; but usually, using a change as the aim will be misleading down the road when you have accomplished the change, but the real goal has not been reached.

Usually, you can see the real aim by asking whether you would be satisfied achieving the aim and nothing else. If you had many assessments but no change in symptoms, you would not have succeeded in increasing the patient's comfort, so assessments alone are not the aim.

Usually, you will have one or two measures that directly track the aim, but you may also have a few others that track the changes (process measures). How good do your measures need to be? Certainly, they do not have to be perfect. If you have 10 units of effort to put into improvement, save one or two for measurement. Too much measurement will make you focus only on data collection, but too little will leave you without insight. The right amount of measurement is what it will take to convince a benevolent skeptic. This standard of the benevolent skeptic usually calls to mind some organizational leader you will have to convince, and that is fine. You will often find that you are your own most critical skeptic, so be benevolent.

Related topics: