The August 2010 publication of a randomized control trial (RCT) that showed that patients receiving early palliative care had better quality of life, lower resource use (costs), and longer life expectancy (~12 months for treatment v. ~9 months for controls of patients with stage IV lung cancer) garnered a lot of attention. Because this study was a RCT its internal validity is very high, meaning there is a great deal of certainty that the only difference between the treatment and control groups was an early consultation with the palliative care service upon diagnosis with stage IV non small-cell lung cancer (the treatment). Note that both the treatment (got early palliative care, and then what care followed) and the control group (did not have an early consultation) were eligible for any chemotherapy, surgery and/or radiation that was available and appropriate. The intervention was simply early exposure to palliative care, which more commonly comes very late in the disease process.
As strong as RCTs are in terms of internal validity, their external validity can be very weak. External validity means the degree to which the research findings can be expected to apply in other settings, and most importantly from a policy standpoint, the degree to which an intervention can be broadly applied in the health care system with some expectation that it will produce similar results to those found in the RCT. With a pharmaceutical, the delivery of the drug in the tested amount is fairly straightforward, though there can still be questions about external validity raised based on the composition of the study population as compared to the population to which you want to apply the new pharmaceutical (for example, if no children are in a trial, you don’t know if it works for children; similar issues are often raised about pregnant women and different racial groups).
When the randomized treatment is palliative care, it is far more complex than is the delivery of a drug because palliative care is a process that depends upon human interaction, especially since clarifying goals of care is important. My colleague at Duke James Tulsky along with Anthony Black at Washington has studied such interactions and communications and they can be taught, but most physicians are not skilled in difficult conversations that are a key part of palliative care. A study in the Journal of Palliative Medicine looks carefully at the components of the early palliative care consultation for persons with newly diagnosed lung cancer. From the abstract of the study describing early palliative care:
The median total time spent with patients for the initial visit was 55 minutes (range, 20–120). Consultations focused on symptom management (median, 20 minutes; range, 0–75), patient and family coping (median, 15 minutes; range, 0–78), and illness understanding and education (median, 10 minutes; range, 0–35).
There is quite a lot of variation in the amount of time spent on different aspects of the initial palliative care consultation, which is the treatment in this RCT. Variation in time spent doesn’t mean the care is bad, it is just not as straightforward as delivering a pharmaceutical, in large part because a key aspect of palliative care is listening to patient concerns and helping them to express and clarify their goals of care. In fact, variation implies it was patient-specific, which is a key point of palliative care. Lyle Fettig, a Palliative Medicine physician has a detailed post on this study and the initial RCT, delving into the components of an early palliative care consultation and how context-specific they can be.
during the first visit, focusing on immediate needs (both symptom and psychosocial), disease understanding, as well as trying to really get a good picture of the patient’s “organic” support all take precedence over code status discussions, etc, in most cases.
And he goes on to note the importance of the setting (outpatient palliative care consult v. inpatient); the RCT treatment was early outpatient-delivered palliative care:
Perhaps the inpatient consultation differs in that the hospitalized patient is already likely at a moment of crisis which allows for concrete discussion about wishes and goals in the context of that crisis. In other words, discussion about the patient’s understanding of their disease/acute illness (for instance, dealing with a malignant pleural effusion) seems like a more natural entree into goals exploration and preferences for life prolonging measures than a patient who is waiting around for their first cycle of carbo/taxol.
As I wrote recently, it is prime time for palliative care in the overall health care system because of the body of evidence that it benefits patients and may reduce costs. A big part of fulfilling the potential of what palliative care can do for patients and the overall health care system is working hard to define what constitutes palliative care in research studies, and then learning to translate the findings broadly and effectively.
To that end, one of the most exciting aspects of my research these days is being the health policy lead for the recently NIH-funded Palliative Care Research Cooperative headed by Amy Abernethy (Duke) and Jean Kuttner (Univ. of Denver) that will execute multi-site clinical trials of various palliative care topics. As with all clinical trials, a strength will remain internal validity to sort out the effect of the treatment being tested. However, we will build these trials with external validity in mind, and seek to translate the the findings as widely as possible, in a manner that can inform public policy discussions.
