When conducting any medical study, before and after results are used to measure the success of the treatment. Many times patients are asked to fill out surveys that measure their responses to treatment as the primary outcome or measure of results. A surprising thing has been noticed about back pain patients from before to after surgery. They have less pain and better function, but they rate their progress as worse than before treatment. They seem to see themselves as more disabled even though tests show they are stronger and more active. What’s going on here?
Social research scientists call this the response shift phenomena. The patient has adapted to the new level of ability and then his or her expectations change. It’s a bit like a moving goal post or using a shorter yardstick for measuring desired outcomes. Showing the benefit of treatment becomes a challenge because patients change their internal standards for how they view their pain or other symptoms. They may reassign importance of one symptom over another.
Take for example the patient with back and leg pain from degenerative disc disease who has surgery to remove the disc. After surgery, the leg pain is gone but the back pain is still there. Even though half the pain is gone, the patient rates pain-related disability as much worse than before surgery when there was back AND leg pain. This patient has had more than just one type of response-shift phenomena. Besides changing internal standards of pain, now he or she has reprioritized the relative value of pain severity. In other words, in the absence of leg pain, suddenly the same level of back pain is much worse and more disabling.
You can see how this could skew research results after treatment. The patient really got better as a result of the surgery. But self-reported ratings suggest treatment made the person worse. Measuring the value of the treatment is like trying to hit a moving target. What can researchers do to reflect the accurate benefit of treatment in patients who experience a shift in the meaning or impact that pain has had on their daily activities and function?
The first step was just recognizing this as an event that happens as patients adjust to a change in their health status. And when they took a look around, scientists discovered the response shift occurs in all medical research whether the health condition being studied is multiple sclerosis, diabetes, dental disorders, joint replacements, cancer, or back pain. The next step was to find ways to adjust for the response shift when analyzing data. Various models have been proposed over the past 10 years to help understand, explain, and measure this phenomenon.
Some scientists have taken a closer look at the factors that might influence how patients cope and adapt. These factors might change a patient’s frame of reference or the way quality of life is rated. How long the symptoms have been present might be a factor. Work status and level of income are two other possible factors affecting how and when patients shift in their perspectives about their health from before to after treatment. There may be personality traits involved. And it’s possible that even the type of treatment (e.g., exercise versus steroid injections) used affects how patients view their results.
Another group of scientists have evaluated the tests used to measure outcomes. Maybe there’s a better way to assign value or ratings to before and after results that incorporates or avoids the response shift phenomena. Scales used to measure pain are always subjective (patient opinion). Tests like the Oswestry Disability Index (ODI) commonly used to measure results don’t reflect when change is clinically significant. There are no normal standards for the ODI based on age or gender. This is a limitation of the test that should be noted.
The authors of this article suggest that spine researchers take a closer look at measuring outcomes — maybe even come up with a better way to evaluate pain that changes because of the response shift phenomena. Or perhaps re-evaluate other measures currently available (e.g., the Pain Impact Short-Form) to see if the response shift affects the ability of this tool to measure outcomes.
At the same time, there is a need to look at different types of patients and compare their response shifts. For example, Worker’s Compensation patients may have a very different response shift compared with non-Worker’s Comp patients. Or looking back at the example of the patient who had back and leg pain who rated results worse even though overall pain was improved. It’s possible that patients who have a partial cure like this are more likely to develop response shifts. It’s possible that data only has to be adjusted for the response shift phenomena in some (not all) patients. The patient who has a total cure (all symptoms gone completely) may rate their results very differently than those who get a partial recovery.
In summary, as a result of discovering that the response shift phenomena exist, response shift researchers are now looking for ways to detect this effect. They are considering all of the variables and factors discussed in this article. The goal is to find a reliable and sensitive way to measure response shift. Hopefully one tool can be developed that could be used with all patients regardless of diagnosis, age, gender, or treatment. That’s a pretty tall order. It’s more likely different detection methods will be discovered and/or developed that can be applied appropriately for each patient group.