Patient-reported outcomes: Is this the missing link in patient-centered perioperative care?

Abstract

Patient-reported outcomes (PROs) have been increasingly recognized as valuable information for delivery of optimal perioperative care to high-risk surgical patients in recent years. However, progress from clinical research on PROs has not been widely adopted in routine patient care. This review discusses the current concepts and practice status regarding PROs and addresses the missing links from research to practice adoption to further improve patient's experiences and clinical outcomes in perioperative care. Insufficient empirical research on appropriate PROs and its methodologies, insufficient implementation research to solve the practical issues, and insufficient data collection methods and experiences on ePROs are also discussed. Future research agenda should focus on evidence-supported, PRO-based symptom monitoring systems for early diagnosis and management of impending compromised clinical outcomes.

Introduction

Patient-centered care in perioperative medicine emphasizes addressing the preferences, values, and concerns of surgical patients in an effort to improve the overall experience. Person-centered care is a holistic and integrative approach designed to maintain well-being and quality of life. A Patient-Reported Outcome (PRO) is defined as “any report of the status of a patient's health condition that comes directly from the patient, without interpretation of the patient's response by a clinician or anyone else.” [1] It is imperative that patients are followed closely after major surgery to not only evaluate their risk for organ-/system-specific complications but also to achieve baseline health status [2]. Patient-centered care goals include the identification and optimal care of high-risk surgical patients, fostering patient-centered decision-making throughout the surgical perioperative pathway, and reducing unwarranted variation in practice [3]. Many stakeholder groups (e.g., clinicians, patients, patient advocates, and those interested in quality assurance) have made a strong case for the inclusion of PROs in routine oncology practice [4,5], and there is now a general recognition that this patient-centered care and outcome measures need to be extended to all fields of medicine. Patient-reported symptoms, function, and more global quality of life potentially could play a critical role in the post-discharge care of patients, especially with the development of platforms that can track patients electronically at home with simple questionnaires delivered through the Internet. There are early reports with practical recommendations on how such a program can be used in surgical settings to improve patient experience with patient-centered outcomes [6].

Instruments (tools) for measuring PROs

A growing body of literature provides support for using PROs in both clinical research and practice. Guidance from the United States Food and Drug Administration (FDA) as well as professional societies, provide standard methods for establishing a PRO-based questionnaire that demonstrates the PRO's reliability, validity, and sensitivity to change [1]. In principle, a good PRO instrument should be brief, simple, and understandable to patients with diverse educational and cultural backgrounds. For example, the 0–10 numeric severity rating scale has been widely used to obtain reliable information directly from patients, because its anchors have been translated and adapted for different cultures, thereby minimizing possible misunderstanding [7]. The categorical responses (i.e., none, mild, moderate, severe) are used on many scales, including the PROMIS assessment method [8] and the PRO-based version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) [9]. The ideal PRO instrument should be validated for repeated use and should be “fit for purpose”—that is, it measures critical and targeted patient outcomes for the disease and the procedure, and has been shown to be sensitive to both acute and chronic conditions on symptoms and functioning in the patient population of interest. These qualities will facilitate acquisition of real-time data during the perioperative recovery phase, which can be viewed as a continuous process.

Finally, meaningful PRO assessments should be able to detect worsening of symptoms and decline in functioning postoperatively that can support clinical decision-making. Although PROs have been widely accepted in clinical research [10,11] and are endorsed by the FDA for use in drug labeling-claim trials [12], and although sufficient guidelines promote inclusion of PROs in clinical trials [13]; the introduction of PROs in perioperative patient care is under development and early in the journey for wider clinical implementation [14,15].

Concepts and progress for incorporating PROs in perioperative care

The meaningful patient-centered outcomes and PROs starts from the recognition of concepts. Patients who undergo major surgery experience an acute systemic inflammatory, neuroendocrine, and metabolic stress response to procedure-related tissue injury and to the medications used during the perioperative period. Collectively, this response often encompasses a cluster of organ-specific symptoms related to complications (e.g., head and neck pain, abdominal bloating from abdominal or pelvic surgery) and systemic symptoms (e.g., fatigue, poor appetite, drowsiness, disturbed sleep). Such morbidity may cause significant postoperative symptom burden along with functional impairment during recovery, leading to unscheduled clinic or emergency department visits. Therefore, despite the clear need to incorporate patient-centered outcomes into perioperative care, in current clinical practice, the need has been tempered by various practical difficulties [16].

The feasibility of using PROs in the perioperative setting may be influenced by difficulties with integrating PROs into clinical workflows and the electronic medical record. There is also a concern that PRO assessment is a time-consuming effort that adds little in terms of clinical benefit [5]. The length of time needed for patients to complete a multi-item PRO instrument has been an issue for both patients and clinicians. Furthermore, although validated PRO instruments are available, their utility and the meaning of PRO scores relevant to clinical outcomes such as hospital length of stay and 30-day postoperative complication rate, remain relatively unexplored. These challenges to PRO adoption might be met by providing empirically established methods for interpreting PRO measurements [17].

Nonetheless, significant progress has been made toward using PROs to support nonsurgical patient-centered care (systemic and radiation therapy) in oncology practice, and there is increasing recognition of the role and value of PROs in perioperative practice as well. Studies show that PRO-based measures of symptom severity and functional impairment are sensitive to the posttreatment recovery trajectory when administered frequently, and that PRO-based quality of life (QOL) measures can predict objective clinical outcomes such as surgical complications [16,18]. An observational study from a comprehensive cancer care center demonstrated that using the MD Anderson Symptom Inventory (MDASI) throughout postoperative recovery from thoracic surgery can be an effective strategy for evaluating functional recovery. There was a good correlation of PRO-based recovery profile to surgery type, preoperative performance status, and comorbid conditions [19]. In liver cancer surgery, Lillemoe and Aloia et al. used a similar tool and showed that enhanced recovery pathways with a focus on minimizing symptom burden and functional interference in the immediate postoperative period demonstrated earlier return of patients to their preoperative functioning level and to adjuvant cancer therapies [20, 21]. Beyond perioperative care, evidence from feasibility studies and clinical trials supports the use of electronic (e) PRO symptom monitoring to address the needs of cancer patients during chemotherapy, radiotherapy, and palliative care and during cancer survivorship [22-24]. A randomized study by Cleeland et al. showed that frequent symptom monitoring with alerts to clinicians when symptoms became moderate or severe reduced symptom severity during the 4 weeks after thoracic surgery [25].

The missing link: Translating PROs from clinical research to routine practice

Effective and timely communication with patients, and engaging and empowering them in their surgical journey is an essential component of patient-centered perioperative care [26]. However, the clinical utility of PRO monitoring in perioperative care has not yet been widely adopted. Currently, pain assessment is the only routinely used screening tool. Hence, the question is whether PROs may be a missing link from research to real-world practice in patient-centered perioperative care.

Although the number of studies using PROs in research has increased, PROs remain the domain of innovation and research protocols, as they are not typically used as outcome measures in routine practice. PROs have been widely used in oncology practice compared to most other specialties. Clinicians involved in multidisciplinary cancer care need precise, evidence-supported PRO-based monitoring pathways and guidance when distressing symptom management needs immediate attention or may be predictive of impending compromised clinical outcomes. To extend this practice to routine perioperative clinical care practice, these approaches should include (1) a workable methodology for using PRO measures to characterize postoperative recovery after complex procedures from the patient's perspective, focusing on symptom burden and functional impairment (especially once the patient is discharged from hospital) and (2) knowledge about how to overcome the barriers to adopt PROs in routine patient care.

Insufficient empirical research on the methodology of using PROs

There has been recent interest along with increasing published reports on patient-centered care for perioperative medicine. However, most of the recent reports on postoperative PRO measures are with low patient numbers, often from a single site without external validation, varied in their approach to missing data, and narrow in the range of domains and the diversity of patients studied [27]. Meaningful PROs should inform treatment teams about clinical outcomes (e.g., risk for complications, length of hospital stay, and return to planned treatment) in order to assess such issues as recovery to preoperative symptom levels or the effectiveness of enhanced recovery programs [14,28].

Although the general recommendation for using core PROs in clinical trials has been published [29], the research questions to be answered in perioperative care include which critical symptoms should be included for a specific disease or procedure, what symptom-severity score thresholds and timing should be used to trigger an alert to clinicians for triage [30], and what functional interference score should trigger action in order to prevent further delay of adjuvant cancer therapy. The knowledge gained from such methodological research could be clinically relevant and a critical component of routine surgical practice for clinicians, as it may help to determine the patient's ability to tolerate treatment and to resume additionally needed cancer therapies. Thus, integrating PROs into standard postoperative care has multiple advantages for monitoring and evaluating recovery from cancer surgery. Recent research utilizing supervised machine learning with PROs reported an improved, robust predictive model for postoperative outcomes [31].

Insufficient implementation research

Recognition of the potential benefits of incorporating PROs into the trajectory of care has been tempered by awareness of problems in implementing PROs [16]. Implementation requires not only sufficient research to provide PRO parameters that will be useful in routine patient care, but also the ability to transition from research to practice, which can take 10–20 years in a large population. A systematic review evaluating the use of PROs during recovery after adult cardiac surgery [27] illustrated similar challenges that are faced by cancer researchers [14]. Significant heterogeneity and methodological weaknesses including differing follow-up periods, lack of standardization in the frequency of measurement and the appropriate tools used to assess recovery, variations in the handling of missing data, assessment of limited domains, and the fact that most were single-site studies were the limitations cited. These weaknesses support concerns that the evidence base regarding wider adoption of postoperative patient-centered outcomes still needs to be strengthened to guide data-driven improvement in postoperative recovery.

Insufficient data collection methods

Electronically administered patient-reported outcomes (ePROs) can be used to evaluate symptomatic and functional recovery. The ePROs may also be used as a tracking system to monitor new onset of severe symptoms in the post-discharge phase of care and therefore be used as a rapid rescue tool to alert surgical teams for the need to contact patients under their care before overt clinical deterioration [14]. However, for wider adoption of this type of monitoring and activation of clinical response from the provider teams; there needs to be a demonstrable patient benefit.

Smartphones have grown in ubiquity, and with increasing computing power, they can play an ever-increasing role in patient-centered health care [32]. As a recognized game changer, electronic data collection could significantly facilitate the use of PROs in routine clinical practice [10,18]. In the past few years, a transition from earlier methods such as paper-and-pencil assessment at clinic visits to electronic data capture via telephone-computer interactive voice or web-based methods on computers or smart phones has facilitated frequent PRO assessments from patients at home. Many ePRO systems can provide actionable links to clinical care, such as summary reports in a patient's electronic medical record and real-time email alerts to providers when patients report acute needs [33]. These newer methods increase the feasibility of tracking patients after discharge repeatedly and with relatively little patient burden, while making these data available to clinicians in real-time.

One result of time-related concerns related to multi-item PRO assessments is low temporal granularity in measurement with assessment only at clinic visits. However, today's digital platforms may allow for high-frequency PRO measurements, as frequently as on a daily basis [34]. Leveraging such technology provides novel opportunities to obtain detailed insight into the recovery process. The methods could enhance perioperative care not only during hospitalization after surgery, when patients are highly symptomatic, but also after discharge, during the functional recovery period.

The expansion of electronic data-capture systems to be deployable on home phones or computers, smartphones, and tablet computers makes it much more feasible to collect patient reports outside the clinic in real time [18,35]. This allows patient status to be followed frequently throughout active treatment and even after treatment has been discontinued. As the feasibility of adopting ePRO capture increases, the relationship between the patient's perspective and postoperative clinical outcomes as associated with different diseases and procedures should be further examined.

Practice points: Current trends of using PROs in perioperative medicine

The successful implementation of PROs in perioperative care depends on the purpose of the assessment within a specific context, the ePRO delivery approach, and the burden on clinicians who must assess and interpret the scores. In this context, symptom screening with a brief, easy-to-administer PRO questionnaire is usually followed by a more comprehensive symptom history to identify patients who actually experience clinically relevant symptom burden. For instance, National Comprehensive Cancer Network (NCCN) guidelines suggest that symptoms such as pain, fatigue, and distress need additional clinical attention when rated ≥4 on a 0 to 10 scale [36-38]. However, a meta-analysis of papers revealed that clinical actions were taken for only one third of patients with moderate to severe pain (≥5 on a 0–10 scale) who were receiving anticancer treatment [39], even though an electronic PRO instrument was available for patients to complete at every clinic visit [40].

The contrast between the high rate of suprathreshold scores and low response rates suggests that routine PRO assessment in practice might potentially generate a high resource burden for clinicians and other health care providers. To relieve this “alert burden,” effective approaches (e.g., defining actionable thresholds according both to derived cut points and to clinical feasibility) will be needed to promote PRO integration into surgical practice [30].

Patients with cancer, for example, frequently report multiple severe symptoms for the first few weeks after major definitive surgery. Although the cost of clinician follow-up calls to patients is prohibitive, the growing literature on randomized clinical trials indicates that computer-assisted electronic symptom monitoring (1) provides a symptom benefit, even without threshold alerts to clinicians [41-43], (2) enhances patient-provider communications [44], (3) increases patients’ sense of control over their care [45], and (4) may lower costs by identifying problems early enough for rapid intervention.

The concept of “return to planned oncologic therapy” in oncology practice calls for effective solutions to ensure that multidisciplinary care is resumed on schedule after all types of major oncology surgery [21]. For example, a nationwide study found that more than half of patients with head and neck cancer were not able to return to adjuvant therapy 6 weeks after surgery, the timeframe promoted in NCCN guidelines as optimal for functional recovery with fewer adverse long-term clinical outcomes [46,47].

PRO-based strategies to address this concern, such as randomized studies with severe-symptom alerts that trigger clinician response to reduce postoperative symptoms and to prevent delay in further treatment due to worsening complications, have not been well studied. However, preliminary evidence with clinically significant data from symptom research in perioperative care supports the use of severity cut points to trigger triage and active symptom management in patients with lung cancer after thoracic surgery. Cleeland et al. [25] reported results from a randomized clinical trial showing that at-home symptom monitoring for 3 weeks after discharge, plus feedback to clinicians when any symptom exceeded a predefined threshold, was feasible and contributed to more-effective postoperative symptom control (Fig. 1). Patients (n = 79) rated symptoms twice weekly on the MDASI by using a computerized, telephone-based interactive voice response system. For the intervention arm, an email alert was sent to clinicians if ratings on the 0–10 scale reached a threshold of 5 for pain, distress, or disturbed sleep, or 3 for shortness of breath or constipation. The intervention group had a significant reduction in threshold events on selected symptoms when compared with the control group (19% vs. 8%).

Fig. 1.

Mean number of symptoms exceeding a predefined severity threshold, as reported over time by patients using an automated, telephone-based monitoring system. Reprinted from Journal of Clinical Oncology: Cleeland CS, Wang XS, Shi Q et al. Automated symptom alerts reduce postoperative symptom severity after cancer surgery: a randomized controlled clinical trial. Journal of clinical oncology: official journal of the American Society of Clinical Oncology. Copyright 2011, with permission from ASCO.

The European Commission's PAIN OUT registry is a good example of a collaborative institutional effort for collecting data (demographics, clinical, PROs) from patients undergoing surgery. This real-world database supports immediate individualized feedback and the effective management of perioperative pain [48].

• The successful implementation of PROs in perioperative care depends on the assessment context, the ePRO delivery platform, and the clinical utility by the providers.

• Effective approaches include actionable thresholds based on the derived cut points and clinical feasibility; targeted critical PROs; appropriate frequency and timing during the continuum of therapy based on the individual patient's goals of care.

Research agenda: Future needs and opportunities

The academic discussion of PRO application has illustrated the necessity of a standardized procedure that incorporates fit-for-purpose tool development, procedure-driven study design, and expert data analysis and interpretation. With such a process, PRO-defined perioperative outcomes can quantify the benefit obtained from implementing a new perioperative care pathway, such as an enhanced recovery pathway, and can provide tracking targets for routine perioperative care.

Specific funding from governmental and regulatory agencies are needed to support the routine implementation of PROs in perioperative patient care setting. Establishing solutions that better serve key stakeholders (in particular, patients and clinicians) by tailoring patient-centered care and evaluating procedure-specific perioperative care after major surgery will require support from academic societies and regulatory agencies as well as collaboration among investigators from the surgery, outcomes research, nursing, and anesthesiology fields.

It is worth noting that QOL research within the larger arena of clinical oncology research has a long history of focusing on various dimensions of the patient's perspective. Wilson and Cleary [49] established a classic conceptual model of PROs that well defines the differences among the terms “symptom,” “functional status,” and “heath perception” as they relate to overall QOL. This model indicates that when considering how the individual evaluates his or her functional status, it is equally important to measure symptom burden, which is an extremely important clinical consideration for perioperative care. Being free of symptoms (e.g., nausea, pain) and being independently mobile were identified both by patients and by health care professionals as the most important goals of postsurgical recovery [50]. Indeed, functional recovery after major cancer surgery dictates whether adjuvant therapy will be delayed or canceled, either of which can negatively affect long-term clinical outcomes [21]. Current research has explored the role of PRO-based measures of functioning in providing robust data even when patients are in remote locations or before they come to the clinic, as opposed to measuring objective signs in the postoperative follow-up clinic [51]. Thus, the discussion about the design of PRO assessments has progressed from assessing one dimension of recovery, such as physiological variables, to a multidimensional assessment of physical, emotional, functional, and cognitive function [11,16,50].

• Research on the development of PROs for clinical practice necessitates a standardized procedure that incorporates fit-for-purpose tool development, procedure-driven study design, and expert data analysis and interpretation.

• Scope of PROs should cover multidimensional concepts from symptoms burden, functional status, and individual health status to quality of life.

Summary

Further collaborative research that includes medical specialties, nursing, information technology support, and institutional policymakers is needed to investigate the challenges and barriers to adopting PROs in real-world perioperative care. Evidence of clinical benefit from adding a measure of multiple symptom and functional outcomes that provides the patient's perspective on their recovery trajectory should be available in the workflow pathway for specific diseases and procedures to improve perioperative care. To incorporate PRO-based methods for defining major symptom burden, a research plan should include a validated PRO instrument(s) that measures critical patient outcomes and that has been shown to be sensitive for detecting both acute and chronic treatment effects on symptoms and functioning in the patient population of interest (i.e., a fit-for-purpose measure). These methods can then be used to establish the utility of PROs as clinical outcomes in surgical care practice, such as for defining functional interference and recovery during perioperative care and predicting poor recovery.