Predicting Prognosis and Addressing Goals of Care

March 27, 2009
dementia,Creatinine,ADLs,primary care physician,PCP,comorbidities,hospitalists, risk assessment,mortality,pneumonia,hospice,hospital medicine

As many of you are aware, a growing number of hospitalized patients are elderly patients with increasing numbers of comorbidities.

Clinical scenario: 84-year-old male with advanced dementia, HTN, CKD III, severe COPD (FeV1 750ml) on home O2 is bought to Emergency Department by family for evaluation of progressive weakness and decreased PO intake resulting in a 20lb weight loss and severe malnutrition. He is admitted to the Hospitalist service with a diagnosis of failure to thrive (FTT). His admission labs are remarkable for a Creatinine 3.2, Sodium 131, Hematocrit of 28, and an Albumin of 2.9. On further history, it was determined that the decreased PO intake was a result of loss of appetite (anorexia) rather than nausea, dysphagia, early satiety, etc. Given his severe dementia and weakness, he is dependent on his family for most ADLs (except for eating). On discussion with the family regarding goals of care, the son, who is also the health care proxy, stated that he had been meaning to discuss placement of a feeding tube with his father’s primary care physician (PCP).

As many of you are aware, a growing number of hospitalized patients are elderly patients with increasing numbers of comorbidities. While advances in medical technology have allowed many to enter into their golden years, there also seems to have been a cultural shift in the expectations patients hold from their medical providers. At times, the line between what is optimistic and what is unrealistic is blurred and may not be obvious to the patients, family, or even their long term PCPs. In fact, studies have shown that hospitalists or physicians who care for a patient during a short period of time, such as an acute hospitalization, are far more accurate in estimating prognosis than longstanding healthcare providers, such as PCPs. As hospitalists, we not only have an opportunity to provide medical care, but we are also able to evaluate the patient more broadly in terms of mental status, physical decline, and functional status. This places us in a unique position of having the most comprehensive, and thus, the most realistic idea of how well the patient will do after discharge. Furthermore, a discussion with the patient and family at this juncture may greatly impact their perceptions and influence critical decisions in the future regarding goals of care, preferences regarding code status, etc.

In the past, when PCPs cared for their own patients in the hospitalized setting, this was not so much of an issue. Additionally, residents are underexposed and in general feel uncomfortable approaching and carrying out such sensitive discussions. This lack of experience impacts their comfort and the likelihood of them being proactive in addressing prognosis as hospitalists. At first, I was hesitant to have such conversations with patients and families, believing that these intimate discussions were more appropriately addressed by PCPs. However, having spent much of my clinical time caring for elderly patients admitted to our Acute Geriatric Unit, I am realizing that opportunities to discuss prognosis and goals of care occur quite frequently. Furthermore, as we shift into an era where a more and more patients admitted to the hospital are being cared for by hospitalists, it makes the most sense to assume, or at least share, this responsibility. Not to mention, I have only received gratitude from patients and family after these meetings. This is because, more times than not, the patient and/or family has very limited insight into the severity of their illnesses, the utility of available treatment options, the effect of the present hospitalization on their prognosis, and their chance for survival after a cardiopulmonary arrest. In some cases, it has been appropriate to introduce the idea of palliative care and hospice, which a tremendous resource that is sadly under-utilized.

There are some cases where the prognosis is quite clear, especially in those who are doing very poorly or very well. Unfortunately, the majority of cases are not so clear, especially in the absence of illnesses such as terminal cancer, end-stage cardio-pulmonary-renal disease, brain death, septic shock, etc. While, as physicians, we can usually arrive at a reasonable clinical impression after consideration of the medical and functional factors, it would certainly be more reassuring to families if we had some data to support our recommendations. In my search for papers that reviewed methods of addressing prognosis and goals of care, I came across a study that actually objectified this assessment with a prognostic index, "Development and Validation of a Prognostic Index for 1-year Mortality in Older Adults After Hospitalization." (I have since heard this article quoted in several lectures involving prognosis in the geriatric population, so it must be as good as my first impression!). This article is short and worth a quick read (see link below) and provides a simple additive point system using data readily available at discharge. Following is a brief overview: It's a prospective study with 1495 patients, age 70 and older, discharged from hospital with 1 year follow up (1993-1997). Various factors were measured, including age, sex, dementia, ADL dependency, co-morbid conditions, length of stay, laboratory data. At the end, 6 independent risk factors were identified, and each is assigned a point value: male sex (1pts), number of dependent ADLs (1-4: 2pts; 5: 5pts), CHF (2pts), cancer (solitary: 3pts; metastatic 8pts), creatinine >3 (2pts), albumin (3.0-3.4: 1pt; <3.0: 2pts). After addition of the points, one year mortality estimates are 13% (0-1pt), 20% (2-3pts), 37% (4-6pts), 68% (>6pts). This was subsequently validated in another prospective study with 1427 patients with impressively similar findings. One thing to note is that after adjustment for functional status, age and dementia were not independent risk factors for mortality. Needless to say, this does not replace your clinical judgment, but can certainly be useful as an adjunctive tool in risk assessment and predicting 1 year mortality, which can further help in discussions of prognosis and mortality.

The above patient was hydrated and discharged to a skilled nursing facility for rehab with follow up with his primary care physician. He was readmitted 10 days later with a pneumonia, resulting in a prolonged hospital course of 9 days and even more functional decline. A family meeting was held to provide the son/wife with a clear understanding of the expected poor prognosis and the limited treatment options that would provide benefit. The discussion then was redirected to address the goals of care and the family was agreeable to hospice.

The second link below is an excellent article regarding hospice, also worth a quick read.

http://jama.ama-assn.org/cgi/content/full/285/23/2987 - Development and Validation of a Prognostic Index for 1-year Mortality in Older Adults After Hospitalization, JAMA. 2001 Jun 20;285(23):2987-94

http://www.annals.org/cgi/content/full/146/6/443 - "I'm Not Ready for Hospice": Strategies for Timely and Effective Hospice Discussions", Ann Intern Med. 2007 Mar 20;146(6):443-9.