A Model to Improve Hospital-Based Palliative Care: The Palliative Care Redistribution Integrated System Model (PRISM)

To better address the current palliative care access problem, we propose a new model: “The Palliative care Redistribution Integrated Service Model (PRISM; Figure 1).” Using the industrial engineering principle of “task shifting,” this approach leverages disciplinary diversity and shifts specific activities from more specialized to less specialized members.⁹ In this way, PRISM integrates hospital-based interdisciplinary teams across all tiers of palliative care delivery.

PRISM sheds a tier-based approach in favor of flexible, skill-based verticals that span all physician and nonphysician providers. By dividing the original pyramid into three domains – physical, psychosocial, and spiritual – providers with various spheres of expertise may serve patients on multiple tiers. For example, a bedside nurse may perform basic psychosocial assessment consistent with his or her training, while physicians may focus on code status or prescribe antiemetics or low-dose opiate monotherapy – skills they have refined during medical school. Analogously, secondary palliative care may be delivered by any provider with more advanced skills in communication or symptom management. In this way, we expand the pool of clinicians available to provide palliative care to include nurses, hospitalists, oncologists, intensivists, social workers, and chaplains and also recognize the diversity of skill sets within and between disciplines. Thus, a hospitalist may clarify the goals of care but may ask a social worker trained in psychosocial assessment for assistance with difficult family dynamics or a chaplain for spiritual needs. Interdisciplinary teamwork and cross-disciplinary communication – hallmarks of palliative care – are encouraged and valued. Furthermore, if providers feel uncomfortable providing a certain type of care, they can ask for assistance from more experienced providers within their discipline or outside of it. In rare cases, the most complex patients may be referred to specialist palliative care teams.

Inherent within PRISM is a recognition that all providers must have a basic palliative care skillset obtained through educational initiatives.⁷ Yet focusing solely on training the workforce as a strategy has and will continue to miss the mark. Rather, structural changes to the means of providing care are also needed. Within hospitals, these changes often rely heavily on hospitalists due to their central position in care delivery. In this way, hospitalists are well primed to be the agents of change in this model.

Since many hospitalized patients have unrecognized and underserved palliative care needs, a formal approach to assessment is needed. Lin et al. proposed criteria for a “sentinel hospitalization,” marking a major illness or transition in high-risk patients necessitating palliative interventions.¹⁰ Similar screening criteria have been validated among hospitalized oncology patients¹¹ and in critical care.¹² While checklists have been shown to help identify hospitalized patients with palliative care needs,¹³ their implementation has been slow, presumably because they are burdensome for busy providers to complete.

Technological automation may be a solution to the checklist conundrum. For example, if palliative care screening criteria could be automatically extracted from electronic health records, scoring systems could trigger hospitalists to consider the goals of care discussions or engage an interdisciplinary care team to fulfill a variety of needs. Frameworks for such scoring systems already exist and are familiar to most hospitalists. For example, admission order sets routinely calculate the Padua or Caprini score to facilitate decision-making for prophylaxis of deep vein thrombosis. An admission order set that screens and prompts decision-making around palliative care needs is thus feasible. One example is a hard stop for entering code status in the admission order set; in turn, this hard stop could also trigger providers to complete a “check-box” palliative care screening checklist. Automatic extraction of certain data from the record – such as age, prior code status, recent hospitalizations, or mobility scores – could auto-populate to facilitate decision-making. In turn, measuring the influence of such tools on access to palliative care, workflow, and capacity will be important, as most tools may not have quality or value intended.¹⁴