Sunday, June 12, 2011

"If I've got 6 months to live, I want to know so I can party"

Journal of Supportive Oncology recently published the results of a pilot trial of an evidence-based decision aid for patients with metastatic cancer (free full-text available here, as always with JSO papers).

This was a small, single cancer center study of 27 patients with metastatic solid tumors (mean age 63 years, 56% African American, with a mixture of metastatic breast, colon, lung, and hormone refractory prostate cancers).  All patients at the center (it's not clear how the patients were identified - if this was a consecutive vs a convenience sample, etc.) who were potentially eligible were referred to the study after approval of their oncologist.  The primary oncologist or oncology nurse could decline allowing the patient being enrolled due to concerns the patient would have 'great emotional difficulty' if they received straight-forward prognostic information.  While they note they didn't measure this, they estimated about 10% of eligible patients were not approached at the request of their oncologist.

Patients were introduced to the study, and given screening questions about whether they actually wanted to hear full prognostic information (1/27 declined full information - they note that patient started the aid then stopped it) and a 'pretest' to determine their current estimations of having a chance of cure and treatment outcomes specific to their type of cancer.  They were then administered the decision aid and their knowledge about chances of cure and estimation of the effect of treatment was measured again.  Hopefulness was measured before and afterwards on a hope scale.  And yes, thankfully, this was IRB approved.

There is an example of the decision aid in the appendix available online.  For instance, the 'Lung Cancer 2nd Line Chemotherapy' aid gives information along the lines of:
[2nd line chemotherapy] improves the chance of being alive at 1 year by 18 out of a 100 people.  With chemotherapy, 37 of 100 people were alive at one year.  Without chemotherapy, 11 of 100 were alive.  ...In this setting, there is no chance of cure.  
It also gives information about side effects, effects on pain and quality of life, and suggestions patients should address advance directives, code status, values and care goals.   The image above is an example from the lung cancer aid.

All patients said they wanted full information (as above, one stopped mid-intervention; they note that that patient's hopefulness did not change).  Patients overwhelmingly overestimated the chances of cure of a patient with metastatic cancer (which, for the type of cancers discussed, which all the patients had, was zero):  52% (only 41% indicated accurately that their cancer was incurable).  Patients grossly overestimated their chances of the cancer shrinking by half, as well as estimated survival (2/3 initially said over 3 years).  The decision aid itself had some impact (the numbers here are too small to measure statistical significance):  most strongly in patient's understanding that their cancer is not curable (it went down to 30% after the intervention).  Generally, the group's time-based estimations became shorter (=more accurate) by a little.  Other misconceptions (chance of the cancer shrinking by half with chemotherapy, chance of cancer symptoms being improved) didn't seem to change with the intervention: they remained over-optimistic.

Both objectively (by the hope index) and subjectively (by the researcher's anecdotal reports), patients were not distressed, upset, and did not 'lose hope' by being told accurate prognostic information.  This is not to say that all patients were all smiles and peaches about the news.  The paper gives sampling of patients comments about the intervention, which are illuminating.  Many seemed grateful or neutral (already knew it).  Many seemed to acknowledge they were told prognoses that were shorter than what they were anticipating but were perfectly happy to ignore that information, and assume it doesn't apply to them ("There were some things I didn't know - I didn't know about the 1-2 years - I'm not going to accept it though - I'm planning on more.")  A few seemed upset ("'s a bummer") - after all they were given upsetting news - but as before this did not translate into changes in the hope index, etc.  And, of course, one expressed gratitude for gaining knowledge which will help them plan the timing of When To Party.   25 patients indicated afterwards they thought the information was helpful.

Reflections on this, ignoring for now discussions of the limitations of this small, uncontrolled pilot trial, etc:

  1. This is further (Do we need more? Yes I think so!) data supporting the long-held wisdom that we (clinicians, docs) generally have little power to take away hope from our patients, at least via sharing knowledge, information with our patients about what we expect.  It's really, really tough to do.  Hope, like religious faith, doesn't spring from a sober assessment of one's material being and surroundings. These patients were given flat out frank information: 'no chance of cure' and 30% of them came out of there counting on one nonetheless.  Anyway: let's beat the drum that routine, straight-forward prognostic disclosure should be the standard of care for all patients with advanced illness.  
  2. That point, plus, gee whiz, it's tough to get patients to understand what seems like very straight-forward information.  Never, ever, underestimate one's ability to mystify and fail to communicate effectively with our patients.  I've been a grown up palliative doc for 5 years now and still impress myself with my ability to mystify my patients.  So much so that I've begun to stop considering that some sort of 'screw up' and instead just assume it's happening and endeavor to double, triple check things which I think are particularly important.  
  3. Part of this makes me wonder what really matters - what should our outcome be because accurate numeric understanding is probably unobtainable for a substantial minority of our patients - should we care about this?  Ignoring patient characteristics (of which we have no control), what is it we say, do, share, disclose, don't disclose, etc with our patients which actually improves the quality of their care and the sanity of their medical decision making?  Data, and patient knowledge are some part of this (and clearly there is a wide and measurable gap there), but of course so are other things - relationship/trust, clinician willingness to not recommend treatments unlikely to benefit a patient.  Maybe it's patient belief about what will happen to them, and, like hope, we may have so little control over that.  Maybe of course it's variable, individual, and fluctuates over time, and a young doc's desire to Understand These Things Systematically needs to yield to the wisdom of sitting at the bedside, and listening to an individual patient's story.  
  4. I meant that, but nonetheless I'll move on from my embarrassing attempts to sound wise and ask - does anyone know of another study which involves, as part of the intervention, effectively telling patients who think they can get cured that their cancer is incurable?  I'm not concluding there haven't been other studies which do this, I just can't think of any.  Is this a milestone or been done before?  I know some of the authors of the study read the blog, so if you read this guys, let us know in the comments if this was at all challenging with the IRB or not.  The Coping With Cancer Study very studiously did not disclose prognostic information to patients per their protocol (as best I understand it):  patient's prognostic understanding was measured however.  One of the important reasons to do this sort of pilot intervention as publishable research, not just a QI project,  is to demonstrate in the pilot setting that this sort of intervention is safe.  The safety data from this, and hopefully follow-up studies, plus the 'effectiveness data' from the CWCS (ie prognostic awareness** is associated with improved EOL & bereavement outcomes) could be used to argue for a larger trial of prognostic disclosure in cancer patients.  

**That's an oversimplification of what the CWCS was measuring, but now is not the time to belabor that.

ResearchBlogging.orgSmith TJ, Dow LA, Virago EA, Khatcheressian J, Matsuyama R,; Lyckholm LJ (2011). A pilot trial of decision aids to give truthful prognostic and treatment information to chemotherapy patients with advanced cancer. The Journal of Supportive Oncology, 9 (2), 79-86 PMID: 21542415

Pallimed | Blogger Template adapted from Mash2 by Bloggermint