Advance care planning and professional satisfaction from “doing the right thing”: Interviews with hospitalist chiefs

To the Editor:

In 2016, the Centers for Medicare and Medicaid Services (CMS) began to reimburse physicians for ACP conversations to stimulate advance care planning (ACP). Though research on ACP adoption predominantly focuses on the outpatient setting, these billing codes can be used in any care setting.^1–3 Our ongoing NIH-funded work seeks to characterize the influence of these billing codes on ACP rates in the acute care setting. This preliminary study sought to identify facilitators of and barriers to ACP in the acute care setting at a large national physician practice with an ongoing ACP QI initiative.

METHODS

We collaborated with a large national physician practice that employs more than 2,500 acute care providers serving over 300 hospitals in 41 states. This practice predominantly employs hospitalists. The ACP QI initiative includes but is not limited to: (1) training on ACP conversation facilitation, (2) compliance with CMS billing codes 99497 (16–30 minute ACP conversation) and 99498 (each additional 30 minute of ACP), (3) feedback on individual and team ACP rates, (4) nursing support to help identify patients who may benefit from ACP, (5) financial incentives of $20 per ACP conversation. This study is approved by the Dartmouth Committee for Protection of Human Subjects, Study 00031186.

We interviewed the leaders of hospitalist teams, ‘chiefs,’ to explore leadership and physician behavior around ACP practices. We purposively sampled chiefs from the highest and lowest terciles of ACP billing. We calculated ACP billing rates based upon the proportion of any patients over the age of 65 who had an ACP billing code entered by their discharging clinician, adjusted for diagnosis-related group (DRG), age, gender and season.

An administrative leader from the physician practice group contacted hospitalist chiefs to introduce the study. Study personnel followed up with interested participants and obtained consent for phone interviews. A multidisciplinary team developed the interview guide, which probed attitudes toward ACP. The interviews were conducted until theoretical saturation was reached, and then recorded and transcribed verbatim. Two investigators coded 30% of the transcripts and convened to develop consensus regarding emergent themes, which were brought to an adjudicator for review. One investigator coded all interviews using the finalized coding framework. Two investigators reviewed the final coded text to finalize themes.

RESULTS

Of 30 chiefs contacted, 15 agreed to participate (50%), most of which were from high ACP billing sites (10/15). Chiefs were on average 45 years old and graduated medical school 17 years prior (Table 1). Chiefs oversaw, on average, 15 hospitalists who cared for an average of 6631 patients per year with an average daily census of 92 patients. The chiefs worked at hospitals across the country (Table 2). The mean ACP rate for all hospitalist teams was 12.3% (range 2.86–48.4).

Table 1.

Exemplary quotations from interviews

Theme	Example
System-level factors
Billing	“Billing is our least favorite part of the day at the end of the day.”
Time	“The biggest barrier is time. Time for documentation. We still have this ACP discussion with the family because it’s the right thing to do.. but the documentation is a big challenge. I wish It’s as easy as just checking four boxes and saying alright, I’ve had an ACP discussion…If they want people to do the right thing.”
Compensation	“For the amount of time you spend, the amount of compensation you get is not worth it. IF I do a history and physical, let’s say, in 15 minutes, and I get $250 or whatever for the RVUs, but I have to spend 16 minutes doing ACP codes and I get $20 or $40…the compensation is not worth it.”
Opinion on CMS billing requirements	“Well they’re supposed to be 16 minutes long, for instance. 16 to 30 minutes for the initial billing episode. And I think that’s off-putting this on people, because if you have some facility with these conversations, they usually don’t take that long. So, you’re either fudging your time or you may just choose not to and just include it in part of your usual documentation.”
Physician-level factors
Comfort	“Initially there were [hospitalists] who were uncomfortable with it. They didn’t feel comfortable having the discussion. There’s no question when you have this discussion with patients they think ‘why are you asking me this?’ They think they’re going to die or something…and that’s clearly not the case. It’s a team’s comfort to be able to explain to the family at the bedside why we are having these conversations. It doesn’t mean we are expecting a poor outcome.”
Experience	“You see your own mortality in a different way when you’re older…You know, as a physician you are kind of trying to fix. And you see a role coming out of residency as a fixer. There’s a challenge, and you have to deal with it. As you get older and you face more and more difficult situations, you learn inevitably that you can’t fix everything. And you just have to do your best to get the patients and the families prepared for the journey they have ahead. You’re not going to be able to fix everything. But you learn that making somebody comfortable towards their end, and releasing their anxiety, is also very, very important.”
Transient relationship	“Our time, you know…that we have with somebody, is short. You have to be able to connect right away, and you know, get to it–maybe not day one. But day two, day three, to get where you need to go. And for some of that, you really have to have a lot of patience, sit down and spend time.”
Consultants	“The big barriers, I see, you know, is the patient has longstanding relationships with the pulmonologist or the hematologist-oncologist, and oftentimes especially patients with stage IV cancers... They’ve never had these discussions, or their doctor wants to do more chemotherapy… So, it can be a little awkward sometimes…And then they’ll say, ‘oh they haven’t tried this chemotherapy,’ so they want the functional status to improve, so let’s do everything we can. And when patients hear that they want everything to be done so they wouldn’t even consider DNR.”

Table 2.

Characteristics of interviewees

	Total (n=15)
Age, mean (SD), y	44.6 (5.5)
Male sex, n (%)	12 (80)
Race/Ethnicity, n (%)
American Indian or Alaska Native	0
Asian	7 (47)
Black or African American	1 (6)
Hispanic or Latino	0
Native Hawaiian or other Pacific Islander	0
White	7 (47)
Years since graduating medical school, mean y (SD)	17 (6)
Foreign medical graduate, n (%)	10 (71)
Internal medicine residency, n (%)	15 (100)
Fellowship completed, n (%)	2 (14.3)
Nephrology	1 (7.14)
Leadership	1 (7.14)
Years at Sound, mean (SD)	5 (3.4)
Annual volume, mean (SD)	6631 (4045)
Hospitalist service daily census, mean (SD)	92 (62)
Full time hospitalists, mean (SD)	15 (10)

Barriers to ACP were universal, including time and documentation burden (Table 1). Chiefs described hospitalists consumed with administrative responsibilities, so billing for ACP was not prioritized. Many described “burnout” with the increasing level of documentation required of physicians. Chiefs stated that the hospitalists believed in ACP but neglected the documentation necessary to bill. Some stated that the rules for ACP documentation were too stringent: if physicians had an appropriate ACP conversation but didn’t cover every item required for ACP billing, they could not bill. Some chiefs reported inadequate compensation for the sixteen-minute time requirement. They were reluctant to “stretch the conversation out” in order to bill. One of the chiefs said that hospitalists weren’t sufficiently incentivized. They could do a history and physical in fifteen minutes and earn $250 but earned $20 for a sixteen-minute ACP conversation. Overall, having the conversation and meeting the documentation requirements were treated as two separate issues.

Chiefs universally discussed how their hospitalists expressed discomfort performing ACP. They described avoidant behaviors, “burnout” and difficulty changing entrenched habits. Many mentioned the transient nature of hospitalists’ in patient’s life and illness, which created challenges in initiating ACP. This was particularly true for patients with chronic disease, who were described as having longstanding relationships with specialists. These specialists did not always “let the patient in on the secret” of the severity of their disease. As a result, patients were often unable to buy-in to a discussion about the severity of their illness because the physician they trusted was not communicating the same message.

Chiefs at most sites encouraged ACP practices by generating friendly competition among their hospitalists, for example, by sharing particularly high or low ACP billing rates of individual hospitalists during team meetings. They also sought to provide real-time performance feedback and emphasized the importance of leading by example. Chiefs at high billing sites in particular framed the importance of ACP around hospitalists’ professional identity and purpose. They emphasized the value of ACP to excellent patient care. They reported that once their hospitalists became more comfortable handling ACP conversations, they were able to appreciate its value and garner satisfaction because ACP was “the right thing to do.” In contrast, chiefs at low billing sites framed the importance of ACP around documentation and did not describe their hospitalists as having a moral impetus toward ACP.

COMMENT

In this preliminary qualitative work, we describe facilitators of and barriers to ACP billing adoption. Facilitators were related to leadership and friendly competition. Barriers were universally related to time, documentation, compensation, and comfort. Our findings provide insight into strategies for increasing ACP uptake among acute care physicians.

Interviewees emphasized the importance of ACP conversations. However, the ACP billing incentive was not the primary driver. Billing was not considered as important as performing ACP and added workload without sufficient compensation. It is possible, however, that there is not a price adequate to incentivize providers to bill for ACP. In our prior work, providing $100 financial incentives to primary care providers of Medicaid patients for performing ACP did not increase ACP, but provider-delivered patient financial incentives modestly increased ACP.⁴ This finding, taken with our interviews, may suggest that physicians are not driven as much by personal financial gain as by the wellbeing of their patients.

Chiefs with higher hospitalist ACP rates tended to frame ACP as a prosocial behavior, an observable action that benefits others regardless of the cost to oneself.⁵ Prosocial behaviors, which are shaped in large part by leadership, improve professional satisfaction and in turn engender commitment to the employer.^6,7 At high billing sites, positive reinforcement came from the chief and from the hospitalist’s own sense of clinical professional duty. In contrast, low billing site chiefs tended to frame the importance of ACP around documentation, an administrative duty. Sites with positive, prosocial reinforcement were able to more successfully overcome the universally reported barriers to ACP.

We used billing for ACP conversations to identify a heterogeneous group of sites. Sites may define and bill ACP conversations differently, leading to misspecification. Billed ACP conversations may be systematically different than non-billed ACP conversations. Nevertheless, ACP billing is a robust measure of ACP given the time requirement and effort taken to document. More chiefs from high billing sites agreed to be interviewed. Since we achieved thematic saturation, we do not believe this introduced bias. We only interviewed chiefs, not the hospitalists, and therefore do not have a firsthand account of hospitalist experiences.

In this study of ACP practices at a large acute care physician practice, we demonstrate that physicians have different attitudes toward billing for ACP and having ACP conversations. Further, leadership factors—mainly an ability to disseminate a broad, prosocial orientation toward ACP—may increase conversations and billing uptake. There is opportunity for QI interventions that specifically target leadership to increase ACP and billing uptake across institutions.

Table 3.

Characteristics of hospitals

	(n=15)
Geographic region
Southwest and West	1 (6.5)
Midwest	1 (6.5)
Atlantic	6 (40)
Pacific Northwest	3 (20)
Gulf	4 (27)
Mean ACP rate, %
Low billing tercile	4.5
High billing tercile	29
Enrolled in BPCI program, n (%)*	15 (100)

^*BPCI -- Centers for Medicare and Medicaid Services Bundled Payments for Care Improvement program