Data collection comes to palliative care

Late to the game, it is figuring out the rules

Palliative care was only recognized as a specialty five years ago by the American College of Graduate Medical Education. Because of its newness, those working in the specialty are still learning how to effectively collect data and make use of the information once they have collected it. Even once they have decided that data collection is a good idea — and not all of those who work in the specialty are there yet — they often rely on measurements and tools that seem sensible, but lack evidence that they are actually appropriate1.

Amy Abernethy, MD, associate professor of medicine and director of the Duke Cancer Care Research Program at Duke University in Durham, NC, has worked on figuring out what data are of value in palliative care. In an article published this month in Current Oncology Reports2, she and her peers looked at what data are collected and how they can affect "quality, value, and research within a palliative care organization." Without data, she notes, the specialty can't demonstrate its value and won't survive.

A number of problems have to be addressed, she says. First, most of what is commonly collected in palliative care is from medical documentation, which Abernethy notes "isn't discreet and can't help you do the kind of predictive modeling you need to affect care. Plus, this is a distressing time, and data collection can be repulsive to both patient and family. I think, too, that those who go into palliative care aren't into data collection as much as they are into caring for their patients. They are just not number wonks."

While palliative care and hospice are good at using patient and family satisfaction tools to see how they are doing, says Abernethy, the environment is what she calls "data naïve." There are attempts by many organizations to improve things. Specifically noted in Abernethy's paper is the Center to Advance Palliative Care ( which has a variety of tools and articles available (for a list of suggested data points suggested for collection by CAPC, see box, below). Among the types of data to collect, Abernethy and her coauthors name process data such as patient demographics, where care is delivered, and referral sources; and outcomes data that quantify the impact of care delivered on patients and their families.

Suggested data points for palliative care

  • Patient demographics: age, gender, race/ethnicity
  • Consultation diagnosis
  • Referring service and/or referring MD
  • Admission date
  • Discharge date
  • Consultation date
  • Disposition: inpatient death vs. discharge
  • Consultation volume
  • Disease distribution: cancer vs. non-cancer
  • Location of consult
  • Age distribution
  • Consults by referring service or physician
  • Length of stay
  • Length of stay outliers: admission-consultation > 30 days or consultation-death >30 days
  • Origin of admission: direct to palliative care (hospice or non-hospice), ED, ICU, ward
  • Type of inpatient unit: fixed bed unit (average daily census, average % occupancy) or swing bed unit (average daily census)

Source: Center to Advance Palliative Care

Once you figure out what to collect, you have to determine the best way to collect it. "This is a poorly reimbursed and time-intensive specialty," Abernethy says. "You have to make people more likely to participate." Putting more data collection at the point of care and making the collection systems more streamlined and easier to use could encourage more providers to engage in data collection. "We can use electronic pens, iPads or other tablets, and find ways to run algorithms in the background so that if you put in some incorrect data, you can be prompted quickly to correct it so that you don't have to go back to the patient or family and repeat something."

One leg up that Abernethy has on others is that there is a consortium of palliative care organizations in North Carolina that have agreed upon what data to collect and how to expand on that if desired. "If someone wants to do a study on difficulty swallowing, they can add something to the electronic form we have already created. They don't have to create a complete new form," she says. The state now has a single large data pool, too, which can help the individual organizations monitor their quality and outcomes compared to their peers.

There is no question that those working in palliative care are dedicated, but Abernethy says even the most dedicated providers can learn from data how to be more effective. One example uncovered in North Carolina was that African-American patients were more likely to suffer from constipation than other patients. "We don't really know why — it could be that people are afraid of being culturally insensitive and so we treat them differently when asking questions. Or perhaps they have some cultural sensitivity around discussing it with us. Or it could be an issue about their medication." Just knowing, though, allows an organization to create initiatives to treat the problem.

The consortium also learned from data collection that they were taking care of a much larger number of heart failure patients than they knew. "Once we found that out, we figured that they might have some special needs we could address, that we might have some workforce issues to deal with to handle their needs. We know now to ask certain questions. Before we collected the data, though, we didn't know it. It wasn't something we noticed off hand."

There will undoubtedly be more pressure on palliative care to prove its worth using data as time goes on. Meanwhile, Abernethy says that the data they collect can be used to improve care for patients and also to market palliative care services. "They can say they think about evidence-based practice."

For more information on this topic, contact:

  • Amy P. Abernethy, MD, Associate Professor of Medicine, Division of Medical Oncology, Department of Medicine, Duke University School of Medicine and Director, Duke Cancer Care Research Program. Durham, NC. Telephone: (919) 668-0647. Email:


  1. Hanson LC, Scheunemann LP, Zimmerman S et al. The PEACE project review of clinical instruments for hospice and palliative care. Jour of Palliat Med 2010 Oct;13(10):1253-60.
  2. Kamal AH, Currow DC, Ritchie C et al. The Value of Data Collection within a Palliative Care Program. Curr Oncol Rep (2011) 13:308–315.