How Much Does It Hurt?

Northwestern's David Cella is leading a vast effort at better measuring pain, fatigue and outcomes. It may change the way researchers run clinical trials.
Author:
Publish date:

Sometimes, it's simple to measure the effects of medical treatment: A patient recovers from the flu or she doesn't. Cancer cells disappear or they return.

Yet many conversations between doctor and patient yield less-than-scientific answers — intangible descriptions, such as the amount of pain a person is in or how much a drug amplifies his drowsiness. Because researchers employ incongruous tools to measure the unmeasurable, a group of clinical studies devoted to a disease might seem like they were written in different languages.

To address this problem, David Cella wants to create the Rosetta Stone of medical symptoms. The chair of the department of medical social sciences at Northwestern University's medical school is leading PROMIS, a government-funded program aimed at standardizing the way patients talk to doctors about their bodies and minds. It has the ambitious goal of re-engineering the way clinical research is performed in the United States.

The PROMIS initiative began after officials at the National Institutes of Health, like many health experts, saw the limits of using varied methodologies to gauge the same symptoms, as it's difficult to compare the results of studies that employ different assessment tests.

"They've grown frustrated or weary of seeing fatigue measured in hundreds of different ways or pain measured in hundreds of different ways," Cella says.

In some cases, test results also fail to capture changes that clinicians observe among their patients. This is partly because some measures are too basic to yield detailed information. The standard way to measure pain, for instance, is just to ask a patient their level of pain on a scale from zero to 10, with little regard for whether they're able to perform everyday tasks.

"Only recently have we begun to recognize that it's important not only to document pain, but functional status," says Aaron Gilson, a director of the pain and policy studies group at the University of Wisconsin School of Medicine and Public Health.

In 2004, NIH started funding the Patient-Reported Outcomes Measurement Information System (PROMIS), a cooperative group of scientists at six research sites, with Cella in charge of a statistical coordinating center at Northwestern. They're mapping out ways to assess a range of well-being areas, from anger to social support.

To create the right vocabulary, they've had to fundamentally rethink the way medicine measures symptoms, assembling 7,000 questions from literature reviews and expert recommendations, and then tossing out overlapping concepts. They winnowed the final list down to a few hundred queries that provide the maximum amount of valid information in the smallest amount of time. (You can try a sample test by clicking on "CAT demo" here.)

They field-tested the battery among thousands of adults against the observations of clinicians, patient self-reports in traditional measures and objective outcomes like CT scans and survival. That helped determine where questions fit along the continuum of intensity for each perception. For example, the question, "How often were you too tired to take a bath or shower?" better gauges extreme fatigue levels than, "How often did you have enough energy to exercise strenuously?"

The group then created an adaptive software test that chooses each subsequent question based on a test-taker's prior responses. The final score, Cella believes, is an accurate, standardized measure of the way a person experiences a symptom — allowing researchers to compare him or her with other patients.

After two years of managing a web-based PROMIS system, this month Cella's team is releasing a standalone version of the software for research teams and hospitals to maintain in-house. Because it addresses users' privacy and technical concerns over the online system, he expects this overhauled version will soon double the number of hospitals and research teams using PROMIS to more than 2,000.

In the next few months, the project is opening up data from its first wave of questionnaires — about 20,000 patients' worth — so outside researchers can analyze it. They're also expanding testing protocols in a variety of domains, including arthritis and congestive heart failure, as well as new physical areas such as sleep/wake and cognitive functions.

To be sure, no system based on patient responses can accurately, objectively measure a feeling as abstract as human pain. Yet tech-heavy neurological research, which aims to do just that, remains years away from allowing scientists to understand specific brain states based solely on a computer image. For now, PROMIS seems to be the best way to quantify fuzzy emotions and sensations into a common scientific language.

Cella stresses that PROMIS isn't meant to replace all other tests, particularly not those targeted toward specific treatments or side effect. If it succeeds, though, it could become the lingua franca for interpreting everything from the effects of new drugs to the efficacy of addiction research.

Sign up for our free e-newsletter.

Are you on Facebook? Become our fan.

Follow us on Twitter.

Add our news to your site.

Related