Friday, November 2, 2018
Nope. We STILL Shouldn't Claim Prolonged Survival in Hospice and Palliative Care
by Drew Rosielle (@drosielle)A group of investigators from Tulane recently published a meta-analysis in Annals of Behavioral Medicine indicating that outpatient palliative care improves survival and quality of life in advanced cancer patients (free full-text available here, although I'm not sure if that's permanent).
Perhaps you'll remember in June of this year when I pleaded with our community to stop claiming that palliative care prolongs survival (my little Twitter rant about this starts here).
My basic plea was this:
Hospice and palliative care community, I'm calling for a moratorium on all blanket, unqualified claims that hospice and palliative care improve survival.I still, 100%, stand-by that plea as stated.
However, this meta-analysis does, I think, show the way forward for future investigators to continue to pool survival data across trials, as some of the trials show trends toward a survival benefit which aren't statistically significant and pooling data in a meta-analysis is reasonable. I just don't think this one really adds much to the discussion, unfortunately.
We've been able to for a while now, and still should, I think, continue to be able to make basically unqualified claims that hospice and palliative care programs improve the quality of life for patients living with serious illness, because the preponderance of the evidence continues to show that, and this meta-analysis also looked QOL outcomes too and showed, not surprisingly, broad and consistent improvements in QOL (I won't really talk about this in this post because, while awesome, it's also not really news).
So - great news! However we'll look through the metaanalysis and I hope you'll conclude with me that it's hardly the final story on this.
Nor does it change the concern that I have that our (over-) exuberant trumpeting of such outcomes as a community may not be in our best interest long-term. At the very least, I think we should continually be on the forefront of articulating the idea in any and all venues available to us, the desperately necessary moral position, that longevity alone is not the fundamental goal of medicine.
Living longer is great, as long as it's living well (whatever that may mean to someone), and we have to be one of the key groups of professionals clearly, proudly articulating the idea that living well is as important as living longer. We all know too well the physical and emotional/existential devastation that longevity-obsessed, technology & organ-focused medical 'care' brings to too many of our dying patients and their families. Helping those patients receive medical care which actually helps them is our gig, and the longevity bonus we may bring sometimes is of course totally swell, especially as it lowers barriers to us being able to see the patients we can actually help.
On to the meta-analysis. Basically the analysis confines itself to high-quality, randomized (either by patient or cluster-randomization) trials of specialist, outpatient palliative care services for patients with advanced solid tumors. There's a lot they looked at, but it basically is a meta-analysis of the 2 big Temel studies (NEJM 2010, JCO 2017), 2 of the Bakitas ENABLE studies (JAMA 2009 and JCO 2015) and Carla Zimmerman's 2014 Lancet study. (They sort of partially included 3 more preliminary and lower quality studies, but most of the headline findings here are restricted to the 5 larger, higher quality studies listed above.)
The Temel studies were basically looking at early (near the time of diagnosis), automatic, ambulatory palliative care for patients with metastatic lung (2010), or lung and GI cancers (2017) vs usual care. Both studies showed some QOL improvements, the 2010 study famously showed a survival benefit, the 2017 one did not. Both studies took place at Massachusetts General Hospital.
The Bakitas studies are a little complicated, but look at a package of interventions which include a visit with a specialist clinician (an APRN in 2009, unclear who they were in the 2015 study) and then regular phone calls for support and coaching around a variety of important concerns - life review, coping, end-of-life planning, etc. The 2009 study showed improvements in QOL and a non-statistically significant trend towards improvement in survival. The 2015 study showed improvements in longevity but not QOL. Perplexing, right?
The Zimmerman study is a Toronto cluster randomized (by different cancer clinics all within the same organization) trial of automatic clinic-based palliative care for advanced cancer patients which basically showed improvements in QOL (but not longevity that I know of).
So, to be clear, the meta-analysis involves 3 studies involving palliative care clinics within 2 organizations, and the 2 ENABLE studies involve a sort of intervention which, while scalable and probably beneficial to patients, is sort of its own thing, and not something most people would recognize as typical ambulatory palliative care services.
But, as it turns out the meta-analysis doesn't even really include all these studies in the actual primary analysis (they are included in the qol analysis and secondary analyses), because they chose survival at 1 year their primary outcome, and only the Bakitas studies and Temel 2010 have 1 year survival data available. I.e., Temel 2017 and Zimmerman 2014, neither of which showed a survival benefit, were even included in the analysis (and they also happened to be the largest of the 5 studies).
I.e., this is a meta-analysis of the famous Temel 2010 study, and the 2 Bakitas studies, one of which bizarrely showed a survival benefit and no QOL benefit (which, again, is interesting, and I honestly don't know what to make of it, but it's not the sort of trial I am going to make broad generalizations about, especially given that it doesn't involve a clinical model that is routinely used in ambulatory palliative care.) If you're interested, the pool survival benefit was 14.1% at 1 year (56.2% surviving vs 42.1% with usual care) in the Bakitas and Temel 2010 studies.
In their secondary analysis, which included survival at 3 months (and thus data from all 5 high quality trials), there was no survival benefit. There was a tiny benefit at 6 months (using data from all the trial except Zimmerman which, by the way, was the largest of the trials), and they looked at 9, 15, and 18 month survival but these basically were the same trials as the primary outcome trials.
So, to try to summarize all this, the actual headline finding here is that when you combine data of 2 of the studies to actually show a survival benefit with one which showed a non-statistically significant trend in survival benefit, you end up with a result which shows a statistically significant survival benefit.
This is why I don't think this paper should change how we talk about survival benefits from palliative care. It's still unclear, a lot of studies have shown no benefit, a few have, maybe we do some times, in narrow circumstances, but that's still not clearly been shown in any generalizable way, and that's about all yo
u can say about it. We can continue to say that we don't worsen survival, because only one study I know of has shown that, and we should continue to shout from the fucking mountaintops that we improve the quality of life of our very sick, very much suffering, seriously ill patient population. And I'm damn proud of that.
I do look forward to more meta-analyses in the future, and think it's a good approach to pool data from heterogenous trials to try to get some clarity on the possibility that there's a sort of 'generalized' survival benefit to palliative specialist services or not, and I'm glad these authors tried this approach, although I don't think it ends up adding much to the survival debate. So, onwards we go....
Drew Rosielle, MD is a palliative care physician at the University of Minnesota Health in Minnesota. He founded Pallimed in 2005. You can occasionally find him on Twitter at @drosielle. For more Pallimed posts by Drew click here.