Context: Maximizing value in palliative care requires continued development and standardization of communication Quality Indicators.
Objectives: To describe the basic epidemiology of a newly-adopted patient-centered communication Quality Indicator for the palliative care inpatient setting among patients with advanced cancer.
Methods: Cross-sectionalG1 analysis of 207 advanced cancer patients who received palliative care consultation at two university medical centers in the United States. Participants completed the following Heard & Understood Quality Indicator the day before and the day following the initial palliative care consultation: “Over the past two days “24 hours” for the post-consultation version, how much have you felt heard and understood by the doctors, nurses and hospital staff? Completely/Quite a Bit/Moderately/Slightly/Not at All”. We categorized “Completely” as indicating ideal Quality. We considered whether ideal Quality differed by sex, age, race, education, financial security, religious affiliation, quality of life, emotional distress, physical distress, spiritual distress, comfort tradeoff preferences at end of life and patients’ one-year survival expectations.
Results: Approximately one-third indicated ideal Heard & Understood Quality before palliative care consultation. Younger age, financial insecurity, high emotional distress, unformed preferences for comfort-longevity tradeoffs at End of Life (EOL), and having no idea about the likelihood of living for one year were associated with less-than-ideal Quality. Among those with less-than-ideal Quality at baseline, 56% rated feeling more Heard & Understood the day following consultation. We observed the greatest pre-post improvement among people who had unformed EOL treatment preferences or no idea about their survival prognosis at baseline.
Conclusion: Two-thirds of patients felt incompletely heard and understood at the time of G2 referral to palliative care consultation and more than half improved following consultation. People with unformed EOL preferences or who had no idea about their survival prognosis improved the most.
Key Words: palliative care, quality measure, heard and understood, communication
Patient-centered communication is fundamental to the practice of palliative care1 and represents a key underlying mechanism by which palliative care improves quality of life among the seriously ill.2-12 Therefore, establishing meaningful indicators of patient-centered communication quality is essential to prepare for value-incentivized palliative care.13,14
Specific conceptual models of patient-centeredness vary somewhat, but all agree that high-quality communication should foster healthcare eG3 nvironments where people feel heard and understood.15,16 As such, the National Quality Measures Clearinghouse recently adopted a Heard & Understood field measure17 created by our team18 and endorsed by the American Academy of Palliative and Hospice Medicine/Hospice.14,18 initiative as a quality indicator for the inpatient palliative care setting.
Successful implementation of quality indicators requires basic epidemiological description within the desired target population to ultimately guide value benchmarking standards. This study focuses on one common palliative care population – hospitalized patients with advanced cancer – and informs benchmarking efforts in three important ways. First, we describe the prevalence and demographic characteristics of patients whose baseline Heard & Understood response values are already at the ideal level prior to palliative care consultation. Second, we explore the distribution of those ideal baseline demographics to identify any potential health inequities preceding palliative care involvement. Third, we describe the magnitude and distribution of improvement in Heard & Understood responses between baseline and the day following the initial palliative care consultation.
Overview: As described more fully elsewhere,19 the Palliative Care Communication Research Initiative is a multi-site observational cohort study involving hospitalized patients with advanced cancer referred for inpatient palliative care consultation. The cohort data includes patient, proxy and clinician self-report; direct observation & audio recording of palliative care consultation visits; and 6-months of health services utilization follow-up. This is a cross-sectional analysis of all 211 patientG4 -participants at study outset. Four participants were excluded due to missing response to the dependent variable resulting in a final N = 207.
Participants: The parent cohort study took place at two large academic medical centers in geographically distant areas of the United States: the University of Rochester Medical Center in the Northeast and the University of California San Francisco Medical Center on the West Coast. Hospitalized patients were eligible for this study if they meet the following criteria: age>21; English-speaking; diagnosed with a metastatic solid cancer; referred to palliative care for assistance with goals-of-care clarification or treatment decision-making; and able to consent for research themselves.
Dependent Measure: Despite the variability in some constructs of patient-centered communicationG5 , nearly all state of the art measures share a common core element: processes that enhance patients’ experience of being heard and understood.15,16 Therefore, we developed 19 the following PCCRI measure specifically for use in the field environment among seriously ill patients: “Over the past two days “24 hours” for the post-consult version, how much have you felt heard and understood by the doctors, nurses and hospital staff?” (Completely, Quite a Bit, Moderately, Slightly, Not at All). A two-day recall was used to ensure a time period of evaluation sufficient to establish a non-transient experience while simultaneously short enough to focus on current conditions prone to substantial evolution in the acute care setting of serious illness. The post-consultation item was shortened to reflect the time period of earliest potential impact by the palliative care consultation. Participants had the option to read the item and respond directly, or to have it read to them by study personnel and respond verbally; mode of administration did not affect response quality or distribution.18
Independent variables: Participants self-reported their gender, race, ethnicity (Hispanic/Latino), educational attainment, financial security, religious affiliations, quality of life, emotional distress, physical distress, spiritual distress, comfort tradeoff preference at end of life, patients’ one-year survival expectations and financial security. As reported previouslyI,6 , we implemented ordinal response options based on the well-validated I,7 approach of the Dartmouth COOP field measures for low burden, reliable, and clinically useful measurement in the busy clinical setting. Following one previously successful approach for avoiding ceiling effects in patient assessment of the quality of clinical interactionsI,8 , we replaced the highest response value from “extremely” to “completely.” All measures underwent pilot testing both individually and within the structure of the 18-item baseline questionnaire and are shown in Figure 1.
Analytic Approach: We describe the frequency and distribution of all study variables. We analyzed patient ratings of feeling Heard & Understood in two complementary ways. First, we identified the proportion of patients whose ratings represented a clinically-important threshold. Given the endemic upward bias in patient-ratings of healthcare communication-related phenomena,20 we categorized “Completely” to be indicative of high clinical quality as done in related work.20 We report observed proportion and 95% confidence intervals for this binary definition. Second, we considered central tendency of responses to the 5-item scale. We observed a left-shifted count distribution of responses with sufficient responses for each possible response category. Therefore, we calculated mean values and used Poisson distribution for any consideration of statistically significant differences among stratum estimates. We conducted the analyses above for both pre- and post-consultation assessment.
We evaluated confounding variables using multivariate regression (logistic for binary outcome and linear for mean difference) and an iterative model-building procedure guided by the impact of potential confounders on the magnitude of the observed crude associations.21 We defined potential confounders as those being associated with both independent and dependent measures.
This study was approved by the protection of human subjects review committees at the University of Vermont Medical Center, the University of Rochester Medical Center and the University of San Francisco Medical Center.
Two hundred seven patients participated in the study. Approximately half were women and the mean age was 61 years (SD=14 years). Fourteen percent self-identified as Black, 9% as Latino, and 75% as having a formal religious affiliation. Two-thirds of participants reported having insufficient income for housing, food, clothing and medications at least some of the time. Approximately one-third of participants reported feeling “Completely” heard and understood (72/207), one-third “Quite a Bit” (67/207) and one third less so (“Moderately”/”Slightly”/”Not at all” heard; 68/207) prior to palliative care consultation. We obtained a ninety-two percent follow-up rate (ie. questionnaire completion at one day post-consult). The most common reasons for non-completion included loss of capacity for item completion, discharge from hospital or death.
As shown in Table One, 1/3 of respondents endorsed feeling completely Heard & Understood prior to the palliative care consultation. In bivariate analysis, the following factors were associated with responses indicating ideal Heard & Understood Quality (feeling completely heard and understood): older age, financial security, low emotional distress, established some preference for EOL care, and patients’ one-year survival expectations. The magnitude of these associations persisted when controlling for age, gender, race, Hispanic/Latino ethnicity, educational attainment, financial security, religious affiliation, and clinician-rated G9 survival prognosis.
Among the 191/207 participants who completed the post-consult Heard & Understood assessment, slightly more than 4 in 10 indicated ideal Quality the day following the palliative care consultation (Prevalence: 42.7; 95% CI: 35.5, 49.5). Participants who either reported being unsure about their preferences for EOL treatment or who had no idea about their life expectancy (n=73) were more likely to indicate ideal Heard & Understood Quality post consultation compared to pre-consultation (38.4% v 22.9%, p=0.04). Among the 124 participants reporting less than ideal Quality at baseline, participants who did not endorse comfort-over-longevity preferences for EOL treatment, those with unformed opinions about their EOL treatment and those who had “no idea” about their 1-year survival prognosis demonstrated the most pre-post improvement in Heard & Understood ratings (see Table Two). All patient-level factors that were associated with ideal Quality at the pre-consultation assessment demonstrated substantially attenuated and statistically insignificant differences at post-consultation.
In this multi-site epidemiological study, we observed three key findings relating to the new Heard & Understood communication quality indicator for inpatient palliative care.
First, we found that only one in three hospitalized patients with advanced cancer indicated feeling completely Heard & Understood immediately prior to meeting with the palliative care team. This aligns with previous work finding patients’ experience in the acute care hospital lacks in overall communication quality.22,23
Second, we observed potential disparities in who felt completely Heard & Understood pre-consultation. (We refer to these differences as “disparities” because our measure is a reasonably direct measure of quality rather than a preferences-sensitive outcome.) In particular, we observed a graded relationship between lower levels of financial security and lower likelihood of ideal communication quality. Related work in oncology observes financial strain as having important, yet previously underestimated effects on patients’ treatment decision-making and quality of life in advanced cancer.24-26 Our findings suggest that financial strain is a barrier to patients feeling completely Heard & Understood by their healthcare team, possibly indicating a lack of clinicians’ attention to the financial impact of medical treatment in routine communication with seriously ill patients.G10
We also observed that patients’ who had unformed preferences for EOL treatment and those with “no idea” about their survival prognosis were least likely to feel completely Heard & Understood. Physicians too infrequently discuss life expectancy prognosis or wishes for EOL treatment with seriously ill patients.27-33 Our findings highlight the impact that poor endemic communication on end of life treatment has on acutely ill people feeling Heard & Understood.
Third, we observed substantial improvement in the patient ratings of feeling Heard & Understood by their healthcare team after consulting with palliative care. More than half of the patients who indicated less-than-ideal Heard & Understood Quality improved and all potential disparities identified pre-consultation resolved after consultation. These findings are important indicators that inpatient palliative care consultation improves the healthcare communication environment globally for patients — meaning that patients’ perceptions of the whole healthcare team improve. Although these data do not describe the clinical processes by which palliative care consultation promotes these effects, our close observations of palliative care teams in action suggests that both skilled and empathic direct communication during patient/family conversations as well as the derivate effects of sharing the outcome of these conversations with other hospital care teams may be partially responsible.
This study has important limitations. First, we only included observations from two university hospitals regarded as having highly skilled palliative care teams. Second, our measures do not evaluate why participants felt unheard. Future work is needed to better identify the clinical systems, cultures, physical structures and inter-personal relationships that foster better communication in hospital environments. Third, our measures evaluate communication by the whole health care team and do not evaluate communication of individual team members. Fourth, we only included English-speaking patients. Therefore, we were unable to assess the effects of primary language and language-related barriers on the feeling Heard & Understood.
Our findings reinforce the need to improve communication in serious illness, particularly during the crisis of hospitalization. Further, this study adds to the evidence that high-quality communication playG11 s a central role in the processes by which palliative care improves the lives of the seriously ill.
The authors report no conflict of interest.