Shaping Nursing Science and Improving Health: The Michigan Legacy

Short Paper: Dyspnea Assessment and Treatment at the End of Life

I completed my studies and dissertation at the University of Michigan School of Nursing (UMSN) in 2006. I studied in the biobehavior track and did cognate work in neurophysiology, the science of emotions, and neurological mechanisms for fear activation and behaviors. I was advised through the completion of my dissertation by Drs. Therrien (chair), Metzger, Algase, and Maren (psychology). I was stimulated to pursue doctoral study through my clinical work as a palliative care nurse practitioner. The field of palliative care lacks the evidence base of other more established fields of clinical care. My aim with my doctoral completion was to increase the evidence base for assessing and treating dyspnea among patients at the end of life. I am proud to report that in the nearly 10 years since I completed studies at UMSN, I have achieved this aim and continue to contribute translatable evidence for clinical practice. Details about my program of research follow.

Dyspnea, also known as breathlessness, is a subjective experience of breathing discomfort that consists of qualitatively distinct sensations that vary in intensity. Dyspnea can only be known from a person’s self-report. Eliciting a patient’s dyspnea self-report is the gold standard for assessment, yet decreased consciousness typifies the last phase of a terminal illness, leaving many patients unable to provide a dyspnea self-report. Thus I established a biobehavioral theoretical model that respiratory distress is the observed corollary to dyspnea based on the person’s display of physical behaviors (Campbell, 2008b).

Respiratory distress is the suffering that arises from an asphyxial threat, such as hypercarbia, hypoxemia, and inspiratory effort (Campbell, 2008b). The human respiratory system automatically produces pulmonary stress behaviors in reaction to an asphyxial threat. In addition, the amygdala is activated. Subcortical activation of the amygdala leads to a fear affective response with characteristic behaviors (see Figure 2-a).

Figure 1-a.
Respiratory distress model.

In my dissertation study, I undertook testing of this biobehavioral model. I observed 12 adult men and women experiencing naturally occurring dyspnea during a mechanical ventilator weaning trial in a medical intensive care unit. I videotaped the participants framed from the waist up and obtained continuous data from a capnograph/oximeter, including heart and respiratory rates, peripheral oxygen saturation, and end-tidal carbon dioxide. When the participants had been restored to comfortable ventilator-assisted breathing, I sought their reports about the emotions they experienced when short of breath through a display of pictures depicting strong emotions, one of which was fear. From these data, I established the behaviors associated with respiratory distress across cognitive states (Campbell, 2007).

Using the behaviors displayed in the previous study, I developed a Respiratory Distress Observation Scale (RDOS) and began a series of studies to establish RDOS reliability and validity. Our first RDOS psychometric study enrolled 210 participants in three groups: (a) patients with chronic obstructive pulmonary disease (COPD) undergoing pulmonary rehabilitation; (b) orthopedic surgery patients in the first 24 hours after surgery before analgesia administration; and (c) healthy volunteers. We established convergent validity with COPD patient self-reported dyspnea, discriminant validity with patients experiencing pain, and healthy volunteers. Scale reliability (α) was also established (Campbell, 2008a).

In our second psychometric study, we tested the RDOS with the intended population of terminally ill patients at risk for dyspnea and unable to provide a symptom self-report. We established construct validity; the RDOS is correlated with hypoxemia. We also established perfect interrater reliability with RDOS measurement by registered nurses (Campbell, Templin, & Walch, 2010). In that study, we also established the high frequency of inability to self-report dyspnea among this sample of patients who were near death (Campbell, Templin, & Walch, 2009).

Since its publication, the RDOS has been embraced by palliative care providers for clinical use to standardize patient assessment. Further, the RDOS has been translated into Dutch, French, and Chinese by investigators around the world for research purposes.

With a reliable, valid, objective measure of patient respiratory distress, I was able to begin a program of intervention studies. Oxygen is a routine intervention applied to patients with respiratory distress. From my clinical work, I hypothesized that patients near death, who often have decreased consciousness, could have oxygen withdrawn. I conducted a double-blind, repeated-measures observation of 32 terminally ill patients using the patients as their own control. We randomly provided oxygen, medical air, and no flow via nasal cannula in 10-minute intervals; we measured the RDOS at the end of each interval. We ascertained that most patients do not need oxygen when near death (Campbell, Yarandi, & Dove-Medows, 2013). Clinical practice changes are suggested from these findings.

Patients near death often have small amounts of retained pharyngeal secretions that resonate and make noise, sometimes referred to as “death rattle.” A number of studies have been done previously to determine whether there is distress among patients with death rattle. Also, studies have been performed to determine which antisecretory medication is most effective at reducing death rattle. No one had previously established whether the patient experiences distress from death rattle. I conducted the first study to establish that there is no patient respiratory distress when patients with and without death rattle were compared. This study supports that death rattle is a normal sound, and antisecretory agents that might produce adverse effects are not indicated (Campbell & Yarandi, 2013).

Meanwhile, clinical users of the RDOS have requested distress cut-point identification. Therefore, we conducted a receiver operating characteristic curve analysis of RDOS scores among a sample of self-reporting dyspneic patients stratified at none, mild, moderate, or severe levels of dyspnea (Campbell & Templin, 2015). We have a study under way with terminally ill patients to substantiate the cut point identified by self-reporting patients.

Critically ill patients undergoing terminal ventilator withdrawal are often unable to self-report dyspnea but are at high risk for experiencing respiratory distress if the withdrawal is not well conducted. I have developed a nurse-led, RDOS-guided, algorithmic approach to this common palliative care process in intensive care units. We completed a pilot study and established feasibility and acceptability and proof of concept that the algorithm is superior to usual care in ensuring patient comfort (Campbell, Yarandi, & Mendez, 2015). An R01 application was submitted in June 2015 for a stepped wedge cluster randomized controlled effectiveness trial; a very good score was obtained from the reviewing special emphasis panel, and funding is anticipated.

Other projects are under consideration for funding. We have opened a study to determine the incidence, prevalence, and trajectory of respiratory distress among patients referred for hospice care. Previous studies of dyspnea prevalence drop patients when they can no longer self-report, so we do not know whether respiratory distress accelerates, remains unchanged, or abates as death nears. I hypothesize that it abates, but no one has rigorously followed patients’ experiences until death.

Lastly, we have prepared a layperson’s version of the RDOS for use in the home hospice setting. We opened a pilot study in the summer of 2015 to test whether the RDOS-family and a bundle of distress interventions taught by the hospice nurse yields positive patient and family outcomes.

In summary, the foundation in biobehavioral science I received at UMSN has enabled me to launch a program of clinical research that is changing palliative care practice here and abroad.