Tuesday, December 21, 2010

A Negative Palliative Care Trial?


There has been a lot of good news about palliative care in the last year. Most notably is one well designed randomized control trial (RTC) finding improved survival and quality of life for stage IV lung cancer patients receiving outpatient palliative care consultations. This line of evidence has been particularly important for those of us attempting to grow palliative care programs both in the inpatient and outpatient arena. But, what should we do when the results of a RTC of inpatient palliative care consultations come up negative? Should we ignore these findings, dispute them, or acquiesce? These are the questions that I am grappling with after reading a research letter in the Archives of Internal Medicine by Dr. Steven Pantilat and colleagues.

The study in question is a randomized, prospective, clinical trial meant to evaluate the impact of a proactive palliative medicine consultation on the care of hospitalized chronically ill elders. The study population included 107 patients admitted to a large academic medical center who were 65 years or older with heart failure (51% of the subjects), cancer (22%), chronic obstructive pulmonary disease (20%), or cirrhosis (6%). The intervention group received palliative care consultations on enrollment and then every weekday of the resulting hospitalization. The authors, as best as I could tell, attempted to create an active control group, as the usual care group received a meet and greet from the palliative care physician and a packet of information on diet and exercise (I know – not very active of an active control group).

The outcomes of interest included the severity of pain, dyspnea, and anxiety, as assessed by a verbal numerical rating scale based on the symptom at its worst level in the past 24 hours. Patients also completed a telephone survey 2 weeks after discharge to reassess symptoms, rate physician and nursing care, and recall discussions of preferences (24% of patients ended up not completing the survey).

The good news is that symptoms scores improved from baseline. The bad news for inpatient palliative care consultations is that they improved for all subjects, during any of the assessment periods, no matter whether they were assigned to the intervention or the control group. There was also no difference between groups in the feeling that the "staff understood what they were going through" and that they "listened to their hopes, fears, and beliefs" (most respondents agreed with these statements).

The other interesting aspect of this trial is that only a minority of subjects reported that the physician discussed their "preferences for care (43%), their chance of surviving hospitalization (42%), or their religious beliefs (31%)", with no differences noted between the intervention or control group. This is despite the fact that palliative care consultations routinely assessed "psychosocial and spiritual needs, discussed treatment preferences, and consulted a pharmacist and chaplain as needed".

Why did this study show no improvement in symptoms with consultations? You can argue that it was small study with only 107 patients (there was no mention of power calculations), that it only occurred within one institution, and that it was a physician led team. Playing devils advocate though, this study is not much smaller than the Temel article on outpatient palliative care, which was also a single site and led by an MD and APN. You can also argue that they were automatically enrolled in this trial instead of being referred, making these patients look very different than the ones we see on a usual palliative care service. Lastly, you can argue that the reported outcomes are based on self-report and may not be at the heart of what palliative care is about (aligning treatments with preferences). 

So, in the end, we are left with a negative study of uncertain significance. Alex Smith wrote in a recent GeriPal comment that people disagree on how practice should change "because we interpret the evidence differently, and at the end of the day, our beliefs and hopes are still mixed in with our scientific understanding." I don't think this article will change my practice based on its limitations, although some would consider me biased.  Will it change yours (please comment)?

by: Eric Widera

5 comments:

Drew Rosielle MD said...

My reaction to this was mostly a shrug. No arguments with the authors but this is not anything like a trial of palliative care as actually practiced, and to me just underlines the very important question of what palliative specialists are supposed to be doing. While all the patients in this study needed to have 'palliative concerns' addressed, this does not mean that they needed or would benefit from specialty palliative care. Specialist palliative care is not for everyone: as underlined by this study (may not add anything to most patients) and professionally we'd all be bored and disillusioned. We've created this problem for ourselves to an extent - beating the 'palliative care for everyone drum' ...'upstream' etc. All of that true, but I firmly believe it does not mean specialist palliative care for everyone. The reason why Mass Gen RCT for lung cancer made an impact is b/c those patients were much sicker. Imagine the spectrum of 65 yo patients admitted with CHF - that's a lot more diverse group of patients (w/r/t prognosis, care planning needs) than patients with incurable lung cancer.

All of this fancy talk is not to say that I can say who should be seen by us specialists (and I should add I'm talking about palliative consultation here not the distinct-ish but very important specialist services delivered by hospice teams). It's not everyone (as this study reinforces; it's not even everyone who is close to death - many oncologists, primary docs, cardiologists, surgeons, etc. do just fine with straight-forward cases and the help of hospice services, etc.) and it seems to have a lot to do with 1) magnitude and complexity of patient symptom/psychoexistential/decision-making needs and 2) competence of the clinicians requesting us to get involved. Clearly there is a lot of variation there. I've thought about this a lot over the years and have recently given up on it, actually defining patient characteristics of who should be seen by us specialists, especially since #2 is so variable (even though yes yes I agree that many nonPC specialists overestimate their skills). Part of me wonders if what 'needs to be done' is to help all docs/clinicians understand and recognize and value what excellent communication, symptom control, psychosocial support, and care planning look like and to recognize when their patients aren't getting it and to refer appropriately. If you're the superstar oncologist who can do it all - swell! - if not, you should know when you're not meeting your patients needs and collaborate with teams who can.

Dan Matlock said...

Thanks Eric. I like Drew's comment. I felt a "shrug" as well. Also, the primary outcome of symptom control may not capture what we really do on the inpatient setting. I actually have a hard time controlling symptoms in the hospital (compared to the work I've done at a hospice) because the primary team and staff are often scared of the drugs/frequencies/doses that we recommend.

I think hospital based palliative care is really expert decision coaching. I'm actually working with some colleagues on an article arguing that inpatient palliative care should use outcomes from the decision sciences such as decision conflict, quality, and regret. I hypothesize that we'd see huge differences.

Eric Widera said...
This comment has been removed by the author.
Eric Widera said...

My take from reviewing the literature is that it is very hard to show symptom improvement with inpatient consults, especially in patients who are not targeted for having symptom distress. This is the second randomized control trial of inpatient palliative care teams that showed no significant improvement in symptoms. The first by Gade and colleagues (Journal of Palliative Medicine. 11 (2), 2008) also showed no improvement in symptoms although they did find that inpatient consultations improved satisfaction with care, decreased health care costs, decrease length of stay, increased advanced directives at index hospitalization discharge, and reduced ICU admissions on subsequent hospitalizations. This study also used as eligibility criteria a life-limiting diagnosis. As far as I know these are the only two RTCs on inpatient palliative care (am I missing one?)

So maybe, at least in regards to symptom control, the ability to show improvement is very much based on symptom needs. However, there may be a lot of other benefits that may come with a palliative care consult that are not captured by this current RTC like those listed by Gade.

Tim Cousounis said...

I'm astounded. Not by the finding that the palliative consult did not make a difference in symptom control. Rather, by the finding that fewer than half of the subjects reported that the physician discussed their "preferences for care", and that there were no differences noted between the intervention or control group. How could that possibly be? Doesn't the discussion about patient/family preferences lie at the heart of most(if not all)hospital inpatient palliative consultations?