Curb Your Enthusiasm: Definitions, Adaptation, and Expectations for Quality of Life in ICU Survivorship

What appears to be the first cohort study to follow intensive care unit (ICU) patients beyond hospital discharge was published in 1976 (1). A year after ICU admission, 73% of patients had died, 23% were at home, and 10 were still hospitalized. Like all scientists in a novel field, the authors of this study lacked a vocabulary for measuring their observations. They categorized survivors as “fully recovered,” “progressing to full recovery,” “partial recovery at best,” and having “no improvement from time of discharge,” but they never defined these terms.

Since that first study, more than 425 peer-reviewed papers have reported on the physical, cognitive, and mental health or quality of life of ICU survivors (2). The field has matured and developed terminology, such as post intensive care syndrome (3, 4), and research tools, such as core outcome measurement sets (5). Prospective studies have illuminated the persistent, and sometimes permanent, physical, cognitive, and mental health impairments experienced by ICU survivors (6–12). And funders like the National Heart, Lung, and Blood Institute and professional societies have declared improving long-term quality of life (QoL) for acute illness survivors to be a research priority (3, 13–16).

But so far, our collective work and enthusiasm have not improved QoL for ICU survivors. A 2018 Cochrane Systematic Review of follow-up services aimed at improving long-term outcomes of ICU survivors found no evidence that they improve health-related quality of life (HRQoL) (17). This year, a systematic review of nonpharmacologic interventions to prevent or mitigate adverse long-term outcomes among ICU survivors found that QoL was assessed more often than any other outcome (18). But in meta-analysis, exercise and physical rehabilitation programs, follow-up services, and psychosocial programs were not associated with improved QoL among survivors (18). Although absence of evidence does not mean such follow-up services conclusively do not work, the early results have not been promising. And yet, a growing number of academic hospitals are exploring opening specialty ICU survivor outpatient clinics offering exactly these sorts of interventions (19–22).

It is time to pause, review definitions, and reexamine our assumptions. In this essay, we first review how QoL is defined and measured and argue that assumptions about the relationship between functional recovery and QoL need to be empirically tested. Next, we review how QoL normally evolves after a change in health state and discuss how research into psychological adaptation during survivorship could help make future randomized trials more efficient. Finally, we propose that ICU providers are in a unique and powerful position to improve QoL by shaping patient and family expectations about survivorship.

Definitions: What We Talk about When We Talk about Quality of Life

The World Health Organization (WHO) defines QoL as “an individual’s perception of their position in life in the context of the culture and value systems in which they live, and in relation to their goals, expectations, standards and concerns” (23). This definition reflects that QoL is a subjective self-evaluation made by an individual—not a static or objective measure bestowed on that individual. The WHO’s International Classification of Functioning, Disability, and Health (ICF) framework (24) describes how physiological impairments can limit a person’s physical and cognitive abilities and restrict their ability to participate in usual activities of their home and work environments. Conceptual models of ICU survivorship extend the ICF to include QoL as a hypothesized outcome of participation restrictions (25) (Figure 1). Another framework proposed by Wilson and Cleary (26, 27) uses domains called biology and physiology, symptoms, functional status, health perception, and quality of life, but maintains that an individual’s QoL is strongly determined by their ability to perform valued tasks (see Figure E1 in the online supplement). Importantly, both frameworks assume that illness limits what people can do independently, and these limitations make people feel worse about their lives. This assumption is intuitive for most people (25, 28), but it needs to be rigorously tested. A study of more than 1,400 ICU survivors 1 year after discharge found that only mental health symptoms, not physical or cognitive functioning, were associated with whether survivors rated their current health state as acceptable (29).

Figure 1.

The relationship between the World Health Organization International Classification of Functioning, Disability, and Health (WHO ICF) framework, patient-reported quality of life, and response shift among intensive care unit survivors. Boxes contain WHO ICF domains. Dashed lines indicate modifying factors.

HRQoL is a related, but separate, term. According to the International Society for Quality of Life Research, HRQoL assesses “people’s level of ability, daily functioning, and ability to experience a fulfilling life” (30). In short, HRQoL measures ask: “Can you?” and QoL measures ask: “Do you care?” The nebulous link between what a person can do independently and their subjective well-being makes it essential for investigators to be specific about what we are measuring. For example, the 36-Item Short Form Health Survey (SF-36) physical component summary score is sometimes (mistakenly) called a QoL measure. The 10 questions in the physical component summary score ask if the respondent is limited in their ability to perform a series of physical actions, including walking, climbing stairs, bathing, carrying groceries, and working. But under the ICF framework, these are questions about participation restrictions, not QoL. In comparison, the WHO Quality of Life–BREF (WHOQOL-BREF) instrument (31) includes fundamentally different questions about the same activities, including “How satisfied are you with your ability to perform your daily living activities?” and “How satisfied are you with your capacity for work?” The WHOQOL-BREF questions capture an individual’s subjective perceptions in relation to their personal goals, expectations, and culture and apply to everyone regardless of whether they value carrying groceries. It is a measure of QoL.

To be clear, we are not advocating that clinical researchers stop using well-validated instruments like SF-36 or EQ-5D, especially because both instruments are core outcome measures for clinical research on survivors of acute respiratory failure (5). Rather, we are encouraging researchers to think critically and speak carefully about what these instruments are measuring, to purposefully measure what ICU survivors can do, and to measure QoL by using instruments that ask survivors for their subjective perception of their position in life. Doing so will facilitate comparisons with previous studies and generate data for exploring the relationship between HRQoL and QoL.

Response Shift: Are Trials Failing Because Survivors Are Succeeding?

Over time, ICU survivors may adapt to new impairments and report improvements in QoL that are not explained by changes in their physical or cognitive function. This phenomenon is known as “response shift,” (32, 33) a term that encompasses forms of psychological adaptation, including recalibration, reprioritization, and reconceptualization, that complicate the relationship between participation restriction and QoL (Figure 1). This phenomenon has been documented in patients with stroke (34), spinal cord injury (35), cancer (36), multiple sclerosis, and inflammatory bowel disease (37). Table 1 provides an example of response shift in a survivor of acute critical illness.

Table 1.

Illustration of positive response shift in an intensive care unit survivor

Case: Janice is a 58-yr-old woman with hypertension who developed pneumonia and acute respiratory distress syndrome. She was mechanically ventilated for 6 d and spent 2 wk in acute rehabilitation after hospital discharge. Before her hospitalization, she worked full time as a media relations specialist and was very physically active, jogging or swimming almost daily. Six months later she is no longer working, cannot run a mile, has short-term memory impairment, and reports that she is consistently too fatigued in the afternoon to go out for dinner or participate in evening social events. Despite these changes, she reports no change in her quality of life compared to before her illness.
Form of Response Shift	Definition	Janice’s Description in Conversation
Recalibration	A change in a person’s internal standards	“I walk a couple miles with my neighbor and her dog most days—as long as the weather’s good—and I’ve started doing yoga on Tuesday mornings. I’m very conscious of what I eat, too. That’s pretty good for someone nearing 60.”
Reprioritization	A change in personal values	“Being sick made me think about how I want to spend my time. The way friends and family rallied around me was just incredible. I decided I wouldn’t go back to work—retire early—even before I left the ICU. Next month I’m meeting childhood friends in Florida for a mini-reunion, and I’m going to my niece’s graduation in May. That’s the good stuff.”
Reconceptualization	A change in how quality of life is defined	“Last month I started volunteering at a women’s shelter, and some of these women, they’re handling so much, trying to keep their kids safe, to keep custody of them, find housing, work, everything… every time I’m there it drives home again just how lucky I am.”

Definition of abbreviation: ICU = intensive care unit.

Unfortunately, response shift is also sometimes treated as a measurement problem (38–40). Researchers in this camp see patient-reported QoL after adaptation as biased and seek to estimate a patient’s “true” unbiased QoL. From a statistical perspective, the measurement error approach treats patient-reported QoL as the sum of true quality of life (a latent variable) and an error term created by the patient’s psychological adaptation. We posit that the “true QoL” in the following equation is the QoL healthy people expect survivors to report: True QoL + Response Shift = Patient-reported QoL. Researchers advocating for this definition seek to estimate the response a patient would have given in a counterfactual world without psychological adaptation. The disparity between the expectations of healthy survivors and the reports of actual survivors stems from the disability paradox, whereby people experiencing chronic illness and disability are happier (on average) than healthy people predict they would be themselves under similar circumstances (41).

We discourage treating response shift as a form of measurement error, because doing so makes “true” QoL an objective value defined by an external observer, rather than by the survivor. We believe that each person must be empowered to define their own QoL. Therefore, we endorse the following model of response shift, which centers survivor perceptions: Physical Health + Response Shift = Patient-reported QoL = True QoL. One hypothesis about why previous trials of interventions to improve ICU survivors’ QoL failed is that substantial proportions of patients in both the treatment and control groups of these trials may have experienced positive response shift. These survivors are analogous to patients with preexisting immunity in a trial of disease prevention. Enrolling these immune individuals, who will adapt and thrive regardless of interventions, decreases a study’s power to detect existing effects (42). Studying adaptation in survivorship and identifying patient characteristics associated with response shift could help create more efficient clinical trials (43). A better understanding of trends in response shift would enable investigators to enroll ICU survivors who are least likely to report improvement in QoL (e.g., because of positive response shift) and instead focus on recruiting those survivors most likely to benefit from the study intervention.

Expectations: It’s Okay to Not Feel Okay

If we accept that true QoL is a survivor’s subjective perception, interventions to improve QoL should target expectations as well as psychological adaptation and physical and cognitive function. Research on the relationship between expectations and patient-reported outcomes draws heavily from marketing research on consumer satisfaction (44). Expectancy-discrepancy theory proposes that people are satisfied when outcomes meet or exceed their expectations and dissatisfied when expectations go unfulfilled (45, 46). Under assimilation-contrast theory (47), we evaluate outcomes using expectancy-discrepancy when the difference between expectations and reality is wide, but when outcomes are close to our expectations we revise our expectations to fit the truth.

Research on satisfaction with surgical procedures suggests these theories hold up in the context of high-stakes healthcare decisions (48). But satisfaction with provided health care is not the same as satisfaction with one’s “position in life in the context of the culture and value systems” (23). To move the needle on QoL, we should try talking to survivors about what their lives are likely to be like after discharge (49). Failing to do so in a standardized way may have contributed to the failure of previous trials if the intervention group had greater expectations for functional recovery. It is nearly impossible to blind ICU survivors to rehabilitation interventions. If survivors randomized to the intervention versus control group have higher expectations for their recovery, these survivors may be less likely to adapt to functional impairments and more likely to perceive their health and QoL as poor. Providing standard information to all enrolled patients regarding expected recovery trajectories might help create more comparable intervention and control groups.

There are more important reasons to set appropriate expectations than study design, though. Discussing possible outcomes is an essential part of ensuring patient safety after discharge. Although the focus of the critical care practitioner is, understandably, to ensure patients are physiologically suitable for discharge, waiting for symptoms of depression, anxiety, and post-traumatic stress to become obvious puts survivors at risk. Patients and their families leaving the hospital need to be debriefed and prepared before they are discharged. As author M.S.H., an ICU survivor, notes: “Unprepared survivors may suffer in silence if they’ve absorbed the message that feeling physically better should be ‘enough.’ This sentiment can lead to self-isolation, substance abuse, a delay in seeking treatment (if they ever do), and at worst suicide.”

Of course, it is very hard to predict how any individual will recover, and we are not advocating discouraging patients. But we are also not advocating unbridled optimism. Instead, we recommend disclosing and normalizing that previously healthy individuals may experience new long-lasting impairments through no fault of their own, regardless of participation in rehabilitation. This does not mean that patients are helpless. In fact, they have significant control over how they respond to the uncertainty before them. In short, we recommend honesty.

Research on aging provides clues about why awareness of the potential for future health impairments may be protective. A recent longitudinal study of community-dwelling adults in their 70s, 80s, and 90s found that adults who predicted (correctly) at enrollment that their health would decline, termed “realistic pessimism,” not only reported few symptoms of depression but lived longer than the elders who were incorrectly optimistic about their health (50). One hypothesis about the mechanism driving this association is that the realists are more likely to take preemptive actions that permitted healthy aging in place, including accessing preventative care, living near family members, and seeking in-home assistance. Disclosing and normalizing the experiences of ICU survivorship may foster similar adaptive behaviors.

Consider two survivors: one assuming a return to their baseline status and one who is fully aware that their recovery may be incomplete. When both find they have significantly less stamina despite completing all prescribed rehabilitation, how do they react? The survivor who assumed complete recovery is likely to feel disappointed, and they may feel betrayed or like they have failed their doctors. They are also unlikely to have secured long-term, logistical support. The patient who knew that recovery was uncertain is not surprised. They may even face adapting to their fatigue with a sense of determination and grit. The prepared survivor is also more likely to have marshalled community or financial resources and fostered relationships with friends and family who can provide long-term support. Although some people worry that informed survivors will be discouraged and less likely to adhere to prescribed rehabilitation, data suggest the opposite. Karnatovskaia and colleagues found that patients identified as having depressive, anxiety, and cognitive symptoms at ICU discharge were significantly more likely to participate in follow-up 3 months after hospital discharge (51). Furthermore, patients with pulmonary conditions and congestive heart failure who were informed of cognitive screening results before discharge had fewer readmissions and increased adherence to medical recommendations (52).

Disclosing uncertainty and fostering realistic expectations may seem counterintuitive. Self-help books like The Power of Positive Thinking (53), which preach positive affirmations and unrelenting optimism, have sold well in the United States for more than six decades. But honesty and bravely facing our uncertain futures are simple and inexpensive to try. Given our meager progress improving QoL for ICU survivors to date, the price of continuing to do the same things is no longer affordable.