The promise of Patient-Reported Outcomes: One Step Closer to Routine Care

This article refers to ‘Prognostic value of psychosocial factors for first and recurrent hospitalizations and mortality in heart failure patients: insights from the OPERA-HF study’ by I. Sokoreli et al., published in this issue on pages XXX. Heart failure (HF) is a condition characterized by a high symptom burden and is complicated by high rates of the dual outcomes of HF hospitalizations (HHF) and mortality.1 As the prevalence of HF continues to increase, so too will the societal burden.2 From a health care system perspective, reduction in morbidity and mortality, and more recently cost containment, have become critical objectives. Since nearly 80% of the cost of care for persons with HF occurs during hospital admission,3 efforts have been focused on this outcome.4 The ‘third wheel’ of patient-reported outcomes (PROs)—a spectrum that can be loosely grouped by measures of symptom and psychosocial status, functional status and general health perception—has received much less attention. The current study by Sokoreli et al.5 serves as a timely reminder that PROs can be measured easily using validated tools and used to strengthen our ability to predict adverse events. Sokoreli et al.5 evaluated data collected from 575 patients prospectively hospitalized for HF enrolled in the OPERA-HF observational study, who agreed to complete six validated psychosocial instruments to assess for depression and anxiety cognition, frailty and patient status of living alone. The cohort was followed for a median of 764 days, during which a total of 1600 events occurred, the majority (1041 or 65%) being subsequent events to the first. The scoring of each psychosocial factor was dichotomized. In addition to several clinical variables with known associations to repeat hospitalization, the presence of at least one of: moderate to severe depression, moderate to severe anxiety, or frailty was associated with an independent 1.8-fold increase in the risk of total hospitalization. The presence of cognitive impairment and status of living alone were associated with increased risk of the first recurrent event but not of subsequent events. Importantly, this study was conducted in a large series of patients with a high event rate, which is likely indicative of HF with reduced ejection fraction in real-world settings. They also found that older age, higher urea or creatinine and higher co-morbidity were associated with increased risk of future events. These findings are in keeping with previously published data and serve to increase their external validity. In previous reports, psychosocial factors have been shown to associate with mortality in patients with HF.6 In addition, other reports have shown association of frailty and depression with adverse outcomes in elderly populations, including time to first events in HF populations.7 The current study extends these findings to the hospitalized HF population for both time to first recurrent event as well as for all recurrent events, the latter diriving the majority of overall events, and thus, health care cost. Several limitations were noted for this study, including the exclusion of individuals with HF and preserved ejection fraction (HFpEF), a condition responsible for up to 50% of HF hospital admissions.2 Another important limitation was the enrolment in only one region of care using one language. The relatively low completion rate of 54% for four items and just over one third for all six items led to a smaller effective sample size. Only completed sets of psychosocial data were included in the analysis, whereas imputation for missing clinical variables (which occurred far less frequently) was allowed. The patient cohort only included those who were able to complete a questionnaire administered in English and patients with severe cognitive impairment were excluded. We are not told if questionnaire completion was observed or unobserved. We are unsure of the relative importance of each psychosocial element in prediction of readmissions since they were not reported in detail separately. Two major implications arise from this study. First, incorporation of PROs will significantly enhance the ability to predict repeat events following HHF. Most algorithms designed to predict HF mortality demonstrate a good but not great performance, with area under the curve (AUC) values ranging from 0.70 to 0.75.8, 9 Far fewer advances have been made in prediction tools for repeat hospitalization, where AUC ranges between 0.60 and 0.62.10-12 Clearly, better tools for risk adjustment of hospital readmission are needed, which could potentially incorporate PROs. This has not stopped the US Medicare-based Hospital Readmission Reduction Program (HRRP), which reduced payments to hospitals with increased risk-adjusted repeat hospitalization rates, from basing' decisions upon such a poorly performing algorithm.10 Potentially unintended consequences, such as a very small decrease in 30-day readmission rates following HHF, at the expense of an increase in 30-day mortality13, 14 are highlighted in a recent review.15 Pending further validation in other HF populations and clinical settings, PROs would be incorporated into future risk stratification tools. The second implication is that PROs are closely linked to clinical outcomes. This is, of course, in addition to the central role of PROs in helping clinicians to better understand their patients' needs and how to frame interactions. The result will be increased use of PROs in both research and clinical practice settings. Several immediate tasks lay ahead. The findings of Sokoreli et al. should be tested in other health care systems and languages, and in HF populations containing patients with HFpEF. Further studies to determine the optimal test inclusion and scoring should be undertaken. Quality of life and time trade-off tools should be incorporated into future risk tool assessment as should other social determinants of health. Translational studies should be performed to assess feasibility and optimize incorporation of these new measures into clinical practice. Although PROs are increasingly reported in clinical trials, funding agencies should make this a mandatory requirement for finding eligibility. While hospitalization or mortality are important to patients, so too are other outcomes. As such, emphasis on patient and family centered care has rightly characterized modern health care strategies. There is therefore an inherent need to understand the patients' perception of their illness in the context of the culture and value systems in which they live, and take into account their goals, expectations, standards and concerns. The measurement and documentation of these items require accurate data from PROs, and allows us to speak more directly to patients and their families. Similarly, the health care system in which we work suffers an increasing strain to deliver adequate resources for optimal care of HF, especially as treatment becomes more complex and costly. To do this, stakeholders in HF care must be able to listen to the concerns of our patients—all of them, including the psychosocial aspects of PROs. This can enable a better patient voice with increased potential for them to self-advocate for their needs and concerns. This is likely to occur to far greater effect than the current clinician efforts to lobby governments.16 The learning from the current paper forms one step towards that goal. Conflict of interest: none declared.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.