Module Takes PROs From Paper to Practice

Cancer care specialists routinely gather "labs" from patients and use that data to guide care. They may also compile quality of life (QoL) data and patient-reported outcomes (PROs) from patients. While acting on the former information may be fairly straightforward, figuring out what to do with the latter is more of a challenge -- how can PROs be applied in a practical way to manage symptom burden, and in a meaningful way to address patient concerns?

Houston-based clinicians devised an answer to that question in the form of the MD Anderson Symptom Inventory (MDASI)-Lung Cancer (MDASI-LC). As explained by Loretta A. Williams, PhD, APRN, of the University of Texas MD Anderson Cancer Center, Meagan S. Whisenant, PhD, APRN, of the University of Texas Health Science Center at Houston (UTHealth), and colleagues, this was a "customized module ... that measures the severity at its worst in the last 24 hours of 13 symptoms common to all patients with cancer (pain, fatigue, nausea, disturbed sleep, distress, shortness of breath, trouble remembering, lack of appetite, drowsiness, dry mouth, sadness, vomiting, and numbness and tingling) plus three symptoms common to patients with lung cancer (cough, constipation, and sore throat) and six items that represent symptom interference with daily functioning (general activity, mood, work, relations with others, walking, and enjoyment of life) during the past 24 hours."

The authors tested MDASI-LC with 40 patients with non-small cell lung cancer (NSCLC), more than half of whom had stage IV disease, who described their symptom experience in single qualitative interviews; content analysis was used to create a patient-reported outcome measure of NSCLC symptom burden.

"The main symptom-related themes found in the analysis were symptom occurrence, symptom severity, symptom interference with activities, and symptom management," the authors stated.

The patients described 32 symptoms, and half a dozen of those – with shortness of breath (57.5% of participants), fatigue (47.5%), and distress (42.5%) – cited by at least 20% of participants. "Clinicians should routinely and proactively monitor the presence and severity of these symptoms in NSCLC," the authors wrote.

Williams, Whisenant, and colleagues also found that the symptoms varied among the participants by the treatment modality received (i.e., chemotherapy versus radiation therapy) but not by stage of disease. Patients also described ways in which symptoms interfered with different aspects of daily functioning, including symptom-related interference with general activity, mood, work, walking, relations with others, and enjoyment of life.

The team concluded that the research established the content validity of a PRO instrument for measuring symptom burden specific to NSCLC, and that well-designed PRO measures are "essential for accurate symptom assessment and management to maintain patient functioning and for use as tools linked to clinical research studies."

In the following interview, Williams and Whisenant discuss their study in further detail, and offer guidance to lung cancer specialists on how to make the most of PROs.

Williams is associate professor in the Department of Symptom Research, Division of Internal Medicine, while Whisenant is assistant professor in the Department of Research at Cizik School of Nursing at UTHealth.

The paper states that "several PRO measures for routinely monitoring symptoms during routine care for NSCLC exist but may not provide symptom-related data to clinicians that are easily interpretable and clinically actionable." Would you say this is another example of the difficulty of translating clinical trial results into real-world practice?

Williams: The answer is yes. A problem is that many clinical trial measures are trying to get some really specific answer. So they may involve very complex scoring algorithms, or they measure complex concepts, such as QoL.

While QoL may be really important to a clinician, just knowing that a patient has a good or a poor QoL doesn't really tell the clinician anything useful in terms of answering the question "What do I need to do to improve that [QoL]?"

I think the scoring algorithms and the complex concepts are things that do not lead that clinician immediately to "'Oh, this is the problem. This is what I should do to fix this."

Is the MDASI-LC available for use for clinicians outside MD Anderson? Also, how would you recommend lung cancer clinicians use other PRO measures in a practical way?

Williams: The MDASI-LC is available; [clinicians] can contact us and get permission to use it. We are using it in the Thoracic Medical Oncology Clinic at MD Anderson.

Clinicians need to think about why they are asking a patient to complete a PRO and what data they want to get out of it that can be used immediately. Clinicians are becoming more aware of the importance of finding out about the patient experience, but they may not know how to pick a [PRO] measure.

They know the PRO measures they are familiar with from clinical trials. A PRO measure may have been included in a trial because the trial sponsor or the FDA wanted the information provided by that PRO to judge the efficacy of a treatment. Clinicians may feel that particular PRO measure is what they should use. But they don't really think about how they're going to apply the information from the PRO directly in their care; how it's going to be useful to them when seeing a single patient in a single encounter.

And sometimes what patients are concerned about may not align with what a PRO is focused on. We have another paper coming out [in which] we asked patients "If you're asked to fill out one of these questionnaires, what would you like to see?"

Whisenant: [That paper] is in press and is slated for publication in March of next year. We asked patients their preferences around these types of measures, and how they could be used. These were advanced cancer patients receiving treatment.

Overwhelmingly, the patients told us that the most important thing is that someone is going to look at and address their responses. So they didn't want to be asked about PROs when they're sitting in the waiting room or before they come in for a visit, and then not have their responses discussed in a direct manner or not have someone respond to them.

We really feel strongly that clinicians need to give a lot of thought to the measures they're using, and really think about the types of questions that are being asked on those measures; think about how they're going to interpret the responses in the context of a clinical practice. Have they thought through the way that they would respond to the information patients share? It's not just about asking the question; clinicians really need to think about how they are going to respond.

How would you recommend that clinicians get PRO questionnaires to patients and still have time to consider an appropriate response?

Whisenant: These types of questionnaires can be incorporated into a [patient portal]. They can be sent to patients electronically prior to a visit, or in the interim between visits when patients are in active treatment or between clinical visits.

And all of this can be done electronically. We collect this type of data electronically in our research studies. The MDASI-LC is particularly useful in a busy clinical setting because the items -- the way that the patients respond to the items on the questionnaire -- are very easy to interpret. All the questions are asked using the same format and wording, and the same response options are available for each item.

It's really easy for a clinician to familiarize themselves with the measure and not need to look at a full summary score for the measure, but rather look at each item to quickly identify which symptoms may be problematic for that patient.

Williams: We have the MDASI in Epic [electronic health system or EHR]. As Meagan said, it doesn't take a lot of extra time. Let's say a patient is reporting, on a 0-to-10 scale, a pain score of 8 and a lack of appetite at 7. You can immediately order a nutrition consult or a pain consult. If a patient is reporting moderate nausea, the nurse can ask "Are you taking your nausea medication?," and explain to the patient "If you notice that your nausea is at level 4, take your medication. Don't wait until it's worse."

Whisenant: We like to think about these responses the same way we would think about a lab value. An elevated potassium level would be flagged in a patient's [EHR]. A page or message might go out to their provider so that somebody can address the issue within a fairly quick time frame.

Items on the MDASI-LC could be similarly flagged so that an elevated severity of shortness of breath, or pain, or distress, or sadness could trigger an alert to a provider, and the response could be fairly quick.

[Clinicians] do a great job in responding to laboratory and imaging data. We don't do a great job of responding to PRO data. But the model is the same. We need to change the way that we're thinking about what this data is telling us about our patients and how we respond to it.

With regard to the "Symptoms Identified in Qualitative Interviews," were there any that came to a surprise as your group, either in their inclusion or where they landed numerically on the list?

Williams: No, but what was really surprising to us was that the symptoms on the MDASI-LC were the symptoms that were important to patients. The MDASI-LC was developed many years ago and the symptoms included were based on clinician expert opinion. Nowadays we develop these measures based on qualitative interviews [with patients], and we pick the symptoms from there.

Since this had not been done with the MDASI-LC, we decided we needed to do the qualitative interviewing with patients to make sure we weren't missing something. What surprised us was that weren't missing anything. What the patients told us [with their symptoms] was what we already had [in MDASI-LC]. We were quite shocked at that because it usually doesn't work that way in terms of what comes out of the patient interviews and what the clinician experts say.

Among the patients who declined to participate in the study, did any give specific reasons for passing on the study that you found anecdotally interesting?

Williams: Most patients, when we approach them, they say "Oh wow, this is great, yes, I'll talk to you." In general, if they decline it's more often than not a scheduling issue. A lot of patients come to MD Anderson from out of town, and they have all of their appointments within a day or two. So it may simply be that there's no time in between the different scheduled clinic visits to do an interview.

Very occasionally we'll run across the patient who will say "I'm just feeling really bad today; I don't think I can do it." But most of the time, even patients who are not feeling well still want to do it.

Whisenant: Sometimes we'll catch a patient when they're on day 2 or 3 following a chemotherapy infusion, and those tend to be rough days for patients. So we can also say "We'd really like to visit with you about your experience. Could we set up an appointment for next week?" We have ways of working around when patients aren't feeling well so that we can still capture their experience.

Williams: Absolutely, because what we do is ask them about the entirety of their experience, not just about "right now."

What's the next step for your group with the MDASI-LC?

Williams: We have a large registry trial at MD Anderson right now where MDASI-LC is being used. We are starting to try to compare the experiences of patients with lung cancer who also have had COVID versus those who have not. We do have an instrument for measuring COVID symptoms, so we're hoping to look at that.

Our lung cancer survivorship clinic is going to start using MDASI-LC soon. There are some clinicians in other institutions who already use it for routine patient monitoring. It's been around awhile. This study was more to make sure we had everything in it that we needed.

Read the study here and expert commentary about the clinical implications here.