KEY POINTS

Question: Is there an association between unmet nonmedication discharge needs (UDNs) and early readmission or death among survivors of acute respiratory failure (ARF) who were discharged home?

Findings: In this prospective multicenter cohort study of 200 survivors of ARF who required an ICU stay, more than two-fifths of patients had one or more nonmedication need that remained unmet in the early postdischarge phase. UDN were not significantly associated with a composite of readmission or death within 90 days of hospital discharge.

Meaning: These results highlight the complex relationship between UDN and outcomes and identify areas that remain susceptible to care disruption.

Survivorship following acute respiratory failure (ARF) continues to improve despite an overall increase in disease severity and prevalence (1–4). Significant numbers of ARF survivors experience a constellation of “new” or “worsened” health states across multiple domains. This post-intensive care syndrome requires targeted support, follow-up, and interventions after discharge (5–10). Often, significant mismatch exists between healthcare needs identified during the discharge process and those that are met in the postdischarge phase (11). Although unmet discharge needs are predicted to be associated with worse outcomes, limited prospective data and empirical analyses exist exploring such associations. Detailed characterization of unmet needs is an important step in understanding and improving postacute care.

We conducted a multicenter prospective cohort study (Addressing Post-Intensive Care Syndrome [APICS-01]) to explore the association of unmet healthcare needs after hospital discharge and subsequent readmission or death within 90 days of hospital discharge among survivors of ARF who were discharged home (12). Discharge needs can be broadly classified into two categories: medication and nonmedication. Nonmedication needs can be further categorized into durable medical equipment (DME), home health services (HHS), and follow-up medical appointments (FUAs). The complex relationship between unmet medication needs after hospital discharge and outcomes from this cohort have been published earlier (13). Herein, we: 1) characterize the proportion of early nonmedication needs that go unmet, 2) analyze association between early UDNs and readmission or death within 3 months of discharge, and 3) identify predictors of UDNs among ARF survivors who were discharged home.

METHODS AND ANALYSIS Study Design

APICS-01 (NCT03738774) is a prospective, multicenter cohort study of ARF survivors conducted at six academic medical centers across United States. Background, rationale, and protocol for the study have been previously published (12). This article focuses on the analysis of early unmet nonmedication discharge needs (DME, HHS, and FUAs) and their association with a composite outcome of death and readmissions within 90 days of discharge from index hospitalization.

PATIENTS

Adult patients with ARF who required an ICU stay and were discharged home were eligible for recruitment in this prospective observational study. ARF was defined as greater than or equal to 24 consecutive hours of: 1) mechanical ventilation via an endotracheal tube, 2) noninvasive ventilation, and/or 3) high-flow nasal cannula oxygen with Fio2 greater than or equal to 0.5 and flow rate greater than or equal to 30 L/min. Major exclusions included patients receiving respiratory support solely for reasons such as airway protection, advanced dementia according to Informant Questionnaire on Cognitive Decline in the Elderly screening, or inability to complete telephone-based follow-up (e.g., homelessness).

Exposure

Unmet discharge needs were the primary exposure of interest and were ascertained via telephonic interview conducted 7–28 days after hospital discharge. Unmet needs, both medication and nonmedication were assessed using an instrument developed and modified according to multidisciplinary stakeholder input from participating centers. Details of the discharge need assessment questionnaire have been published earlier (12). This questionnaire was designed to capture the status of healthcare needs (met or unmet) at the time of exposure ascertainment.

The needs assessment questionnaire was administered by a centralized outcome assessment group within the data coordinating center. Assessments were performed as soon as possible within a time window of 7–28 days after discharge to allow sufficient time (i.e., at least 7 d) for discharge needs to be arranged. Both patients’ and caregivers’ responses to the questionnaire were accepted depending on who could most accurately provide the information. Status of healthcare needs at the time of interview was categorized as either met or unmet. Further, DME needs reported as delivered or in-process were considered met. HHS needs reported as either occurred or scheduled were considered met. FUA reported as occurred or scheduled were considered met. Reasons for unmet DME and HHS needs were categorized as either: cannot afford, not delivered/scheduled, patient declined/refused, cancelled by provider, has COVID-19, and other. The primary exposure variable was the proportion of UDN assessed at the time of interview, that is, if a patient had 10 UDNs and four were unmet, the proportion was 0.4.

Outcome

The primary outcome was a composite of hospital readmission or death within 90 days of discharge after the index hospitalization. Key secondary outcomes were the individual constituents of the primary outcome.

STATISTICAL ANALYSIS

Demographics, discharge needs, timing of the assessment, and the proportion of early UDN were characterized with descriptive statistics, with percentages rounded to whole numbers using default settings of the statistical software. The primary research question was whether early UDN are associated with readmission or death within 90 days of hospital discharge. Primary analysis compared patients with discharge needs dichotomized as above or below median of proportion unmet. To account for reverse causation and confounding by indication, we used propensity score (PS) adjustment to estimate the average effect in those treated, or exposed to UDN (14). Prespecified covariates for multivariable balancing models included patient demographics and baseline hospital characteristics. Covariate balance was assessed by comparing observed and adjusted covariates with standardized means differences and the Kolmogorov-Smirnov test. The primary outcome model used Poisson regression and included the PS as a weight and as covariates: number of needs, enrolling site, age, and sex. Death and readmission were also analyzed separately with the same approach. Robust ses for modified Poisson regression were used to calculate p values and 95% CIs.

In an exploratory analysis, we aimed to identify predictors of unmet nonmedication needs. This was done to identify determinants of care disruption, which could be potentially targeted to improve postacute care. To reduce number of variables, baseline characteristics and demographics were included in a random forest model. Continuous variables were normalized, whereas categorical variables were flagged with a value of one. Patient outcome was coded as 0 for no unmet needs and 1 for greater than or equal to unmet needs. Random forest hyperparameters were optimized with a grid search and model with the lowest out-of-bag estimate of error was selected. After training the decision trees on a random subset of patients, patients left out, or out-of-bag were presented to the decision trees for prediction. The out-of-bag error rate was computed as the number of errors that were occurred from the out-of-bag sample. The top four variables of importance with respect to the selected model’s predictive ability were identified. These variables were included in a Poisson regression model with presence or lack of unmet discharge needs as the dependent variable. Robust ses were used to calculate p values and 95% CI. Due to the exploratory nature of the analysis, p values were not adjusted for multiple comparisons.

A two-tailed significance threshold of 0.05 was used. All analyses were performed using R Statistical Package (Version 4.0.3; R Foundation for Statistical Computing, Vienna, Austria).

SAMPLE SIZE

Assuming 152 evaluable patients (accounting for estimated rates of in-hospital death and loss to follow-up), we estimated 80% power with two-sided α = 0.05 to detect an increase in the risk of the composite outcome from 30% for those in the lower unmet needs category to 53% for those in higher unmet needs category.

Ethical Considerations

APICS-01 was funded by the U.S. Department of Defense (DoD) and prospectively registered at ClinicalTrials.gov (NCT03738774). APICS-01 was approved and overseen by the central Institutional Review Board (IRB) at Vanderbilt University (IRB No. 181120; date: July 19, 2018) with further oversight by Human Research Protections Office of DoD. Informed consent was obtained from patients or their legally appointed representatives prior to enrollment. Study procedures were conducted in accordance with the Declaration of Helsinki.

RESULTS Patients

Of 5,728 patients with ARF meeting inclusion criteria between January 2019 and August 2020, 4,908 were excluded. Among 820 eligible participants, 249 were consented and 200 enrolled after accounting for further exclusions and withdrawal after consent. Due to one withdrawal and loss to follow-up in four others, the final analytic cohort included 195 patients for which both exposure and outcomes were known. Details of cohort derivation have been previously published (13).

Baseline characteristics of the analytic cohort (n = 195) are summarized in Table 1. Median age was 55 (interquartile range [IQR], 43–66), 104 (53%) were female, and 66 (34%) were people of color. The median hospital length of stay was 14 days (IQR, 9–21 d) and 33 (16.9%) were diagnosed with COVID-19 during admission. Most (94 [48.2%]) patients had private insurance, and 84 (43.1%) were working prior to admission. PS adjustment provided good covariate balance (eFig. S1, https://links.lww.com/CCM/H238).

TABLE 1. - Baseline Characteristics and Demographics of Analyzed Patients (n = 195) Attribute Central Tendency (Dispersion) Age, yr, median (IQR) 55 (43–66) Female sex, n (%) 104 (53.3) Race, n (%)  Black 48 (24.6)  Hispanic/Latinx 5 (2.6)  Non-Hispanic White 129 (66.2)  Other/multiple 13 (6.7) Insurance status, n (%)  Private and public 24 (12.3)  Private 94 (48.2)  Public 66 (33.8)  None 11 (5.6) Acute Physiology and Chronic Health Evaluation II severity of illness score, median (IQR) 20 (15–26) Clinical Frailty Scale, median (IQR) 3 (2–4) Multidimensional Scale Perceived Social Support, median (IQR) 72 (60–81) Functional Comorbidity Index, median, (IQR) 2 (1–3) Charlson Comorbidity Index, median, (IQR) 1 (0–3) Employment status prior to admission, n (%)  Working (full/part-time), looking for work, or in school 84 (43.1)  Unemployed, not looking for work 7 (3.6)  Sick leave 1 (0.5)  Retired 29 (14.9)  Receiving disability payments 32 (16.4)  Unknown 42 (21.5) Diagnosed with COVID-19 during admission, n (%) 33 (16.9) Hospital length of stay, d, median (IQR) 14 (9–21) Enrolled in ICU recovery program, n (%) 32 (16.4)

IQR = interquartile range.

The breakdown of nonmedication discharge needs was as follows: 118 (60.5%) patients were prescribed DME, 134 (68.7%) required HHS, and 189 (96.9%) needed at least one FUA with a specialist or allied health professional. Of the 195 patients analyzed, 98.4% (192/195) had at least one identified nonmedication need at hospital discharge (Fig. 1). Of patients, 48.2% (94/195) had at least one need in each of the three categories and 31.3% (61/195) were identified to have needs in two categories.

Figure 1.:

Patient-level Venn diagram of discharge needs ordered by type: durable medical equipment (DME), home health services (HHS), and follow-up medical appointments (FUAs).

Figure 2 displays the total frequency of discharge needs and proportion unmet. Of the 192 DME needs, 40 (20.8%) were unmet. There were 321 HHS needs, of which 118 (36.8%) were unmet. FUA had the highest frequency at 599 and the lowest proportion unmet at 96 (16.0%).

Figure 2.:

Total frequency of discharge needs by type and proportion unmet: durable medical equipment (DME), home health services (HHS), and follow-up medical appointments (FUAs).

The median (IQR) of DME, HHS, and FUA needs for the cohort were 2 (1–2), 2 (1–3), and 3 (2–4), respectively. Overall, there was a median of 5 (IQR 4–8) needs when combining the three types. Median proportion of unmet needs across three categories was 0 with IQR ranging from 0–15% for DME, 0–50% for HHS, 0–25% for FUA, and 0–20% when considering all three. Those with greater than or equal to 1 needs were therefore considered above the median proportion of unmet needs and those with no unmet needs were considered below median. Number of patients with one or more unmet needs for DME was 30 (25.4%), HHS 47 (35.1%), and FUA 69 (36.5%). Of the patients, 43.8% (84/192) had one or more unmet need at the time of follow-up interview (Table 2). Further characterization of the UDNs, status of delivery and reasons for remaining unmet at follow-up are shown in eTables S1–S5 (https://links.lww.com/CCM/H238).

TABLE 2. - Nonmedication Discharge Needs by Type and Proportion Unmet Discharge Need Type (n = 195) No. of Needs, Median (IQR) No.of Unmet Needs,Median (IQR) Proportion Unmet, %, Median (IQR) One or More Needs Unmet, n (%) All nonmedication needs, n = 192 (98%) 5 (4–8) 0 (0–1) 0 (0–20) 84 (44)  Durable medical equipment, n = 118 (61%) 2 (1–2) 0 (0–1) 0 (0–15) 30 (25)  Home health services, n = 134 (69%) 2 (1–3) 0 (0–1) 0 (0–50) 47 (35)  Follow-up appointments, n = 189 (97%) 3 (2–4) 0 (0–1) 0 (0–25) 69 (37)

IQR = interquartile range.

Notable unmet DME needs included shower chair (41.7%, 5/12), commode (50%, 6/12), oxygen (12.5%, 5/40), and noninvasive ventilation equipment (31.3%, 5/16). Twenty-three percent patients prescribed walkers, 12% with prescription infusion medications, and 33% of patients requiring vacuum-assisted closure devices reported having their needs unmet (eTable S1, https://links.lww.com/CCM/H238). Proportion of unmet HHS needs varied widely with the most common unmet needs being respiratory therapy services (100%, 2/2), social work consult (100%, 3/3), physical therapy (45.2%, 38/84), occupational therapy (49.2%, 31/63), speech-language pathology (53.9%, 7/13), and home health aide and infusion care (20%, 1/5) (eTable S3, https://links.lww.com/CCM/H238). Proportion of unmet physician FUAs ranged from 11.1% (1/11) for internal medicine specialists to 42.8% (3/7) of cardiac rehabilitation appointments (eTable S5, https://links.lww.com/CCM/H238). Notably, 57.1% (4/7) of speech-language pathology and 80% (4/5) of occupational therapy follow-ups remained unmet or unscheduled at the time of follow-up.

Of the 599 FUAs specified at discharge, there 173 (28.9%) where the patient never attended the appointment. Of those unattended appointments, 30.6% (53/173) were booked before discharge and 19.1% (33/173) were booked after discharge. There were 50.3% (87/173) where no appointment was booked at all. Of the 86 appointments that were booked where the patient did not attend the appointment, 26.7% appointments (23/86) were missed due to the patient being readmitted before the scheduled appointment.

Cost was not perceived as a major barrier to DME, except for chairlifts. COVID-19 status was the most common reason for HHS needs being unmet (eTable S4, https://links.lww.com/CCM/H238). 20.6% (13/63) of occupational therapy, 30.8% (4/13) speech-language pathology, 19.1% (16/84) physical therapy, and 100% (2/2) of unfilled respiratory therapy follow-ups were attributed to COVID-19 status of survivors.

Study Outcomes

Fifty-six patients experienced the composite outcome of death or readmission within 90 days of discharge from index hospitalization. The median (IQR) time to the primary outcome was 14 days (4–47 d); the median time to readmission was 12 days (6–32.8 d) from discharge; and time to death was 70.5 days (36.3–80 d) from discharge, when it occurred (Table 3). Study outcomes occurred during the follow-up time period 35 (17.9%) times, of which 23 (11.8%) occurred before the follow-up assessment was completed (eFig. S2, https://links.lww.com/CCM/H238).

TABLE 3. - Ninety-Day Outcomes for Primary Outcome and Constituents (n = 195) Primary Outcome and Constituents All, n = 195 Death or readmission before 3 mo, n (%) 56 (28.7)  Time to primary outcome among those achieving primary outcome, d, median (IQR) 14 (4–47) Death before 3 mo, n (%) 10 (5.1)  Time to death among decedents, d, median (IQR) 70.5 (36.3–80) Hospital readmission before 3 mo, n (%) 52 (26.7)  Time to readmission among those readmitted, d, median (IQR) 12 (6–32.8)

IQR = interquartile range.

Primary Analysis

We did not identify a statistically significant association between unmet needs and the primary outcome (risk ratio, 0.89; 0.51–1.57; p = 0.690). Analysis of the outcome constituents, death, and readmission demonstrated similar results. These estimates represent the risk of the outcome among patients with “any” unmet need compared with those with no unmet needs since the median proportion of unmet needs was 0 (Table 4).

TABLE 4. - Propensity-Adjusted Multivariate Regression Estimates for the Association of Unmet Nonmedication Discharge Needs With Primary Outcome and Constituents Outcome Risk Ratio (95% CI) p Primary outcome (death, readmission) 0.891 (0.507–1.568) 0.690  Readmission 0.831 (0.438–1.576) 0.570  Death 0.547 (0.122–2.454) 0.431

In the exploratory analysis, the best performing random forest had an out-of-bag estimate error rate of 33.8%. The four variables that had best predictive value were hospital length of stay, Acute Physiology and Chronic Health Evaluation II score, Multidimensional Scale Perceived Social Support, and age (eFig. S3, https://links.lww.com/CCM/H238). When these variables were included in a logistic regression model, hospital length of stay and age were statistically significant for unmet discharge needs, with odds ratio (95% CI) of 1.02 (1.01–1.03) and 1.02 (1.00–1.03), respectively (eTable S6, https://links.lww.com/CCM/H238).

DISCUSSION

Overwhelming majority (98.4%) of ARF survivors who went home had healthcare needs identified at discharge. FUA were the most common posthospital need, followed by HHS, and DME. Broken down into individual categories, the proportion of unmet needs varied widely, ranging from 0–15% for DME to 0–50% for HHS. More than two-fifths of ARF survivors had one or more unmet nonmedication needs in the early postdischarge phase. We did not find a statistically significant relationship between UDN and our prespecified composite outcome, which was driven primarily by readmissions rather than mortality.

In addition to highlighting the magnitude of UDN, our study identified specific areas, which appear to be particularly susceptible to mismatch between the discharge plans and the reality patients experience after discharge. For example, 31.3% of patients prescribed nocturnal noninvasive ventilation and 12.5% of those prescribed supplemental oxygen reported not receiving either. Although affordability did not appear to be a major driver of nonreceipt of DMEs, this may not necessarily apply to other populations where higher proportions are uninsured. Only 5.6% of our cohort reported having no insurance coverage. Nearly 50% of physical, occupational, or speech-language pathology services requested as part of the discharge plan remained unfilled at follow-up. Our enrollment period overlapped somewhat with COVID-19 waves in United States, so these lapses could be partly explained by COVID-19–associated strains on healthcare systems. More importantly, gaps may also be more reflective of the fragmented nature of care coordination and downstream consequences of chronic workforce shortages (15–17). Although COVID positivity emerged as a theme of unmet needs, its attributable role remains unclear because of a small proportion (16.9%) of COVID-positive patients. The follow-up APICS-COVID (NCT03738774) observational study is anticipated to shed light on the impact of COVID-19 on disruption of postdischarge care. Scheduling difficulties emerged as a theme for unmet FUAs; a substantial proportion of patients did not have prescribed follow-up scheduled “before” discharge. We hypothesize that an emphasis on scheduling FUA prior to discharge may represent a pragmatic step toward reducing missed follow-ups. Notably, primary care appointments were either met or scheduled 87% of the time. This relatively high proportion suggests, that within the current framework, primary care practitioners may be best positioned to be key champions of care coordination. Primary care visits may also present as a triaging point for reappraisal of needs, which may have been modified postdischarge. Advanced age and hospital length of stay were further identified as predictors of unmet discharge needs in our exploratory analysis. These findings highlight the need for careful attention to discharge planning for such patients.

Our results provide previously unexplored insights into nonmedication needs that remain unmet after hospital discharge following ARF. Although exploratory, these data could be used to identify high-yield process improvements to mitigate the mismatch between the healthcare prescribed and those delivered in the early postdischarge phase.

Our findings confirm significant gaps in care delivery for survivors of ARF. The burden of such gaps may be expected to be higher in resource-limited, smaller centers. Although specialized ICU recovery clinics continue to grow, ICU survivorship appears to outpace their adoption. These models have shown promising but not definitive evidence of efficacy (18–20). Institutions with established ICU clinics typically target a window of 4 weeks or later for initial visits (21–26). Our results indicate that substantial proportion of readmissions occur prior to the traditional window of follow-up. Taken in context, the current framework may be insensitive in detecting and therefore ineffective in addressing gaps in care as readmissions occur before initial visit. Enrollment in an ICU recovery program did not have a significant effect on unmet needs in our cohort, but this intervention was unevenly applied and therefore these results should be considered exploratory. Fragmentation in postdischarge care may be even more pronounced where no dedicated infrastructure exists and likely be further exacerbated due to the recent pandemic (27). Advancing follow-up closer to discharge and integrating follow-up with targeted support services may represent a pragmatic step toward reducing care fragmentation. Given the wide range of postdischarge needs, comprehensive patient-centered care models including community outreach, such as those used successfully for patients with heart failure, may serve as blueprints, which could be tailored to address the distinctive needs of ARF survivors (28,29). The modest success of a recently concluded single-center, multicompartment program among survivors of sepsis has demonstrated some promise but such approaches will need to be formally tested before widespread implementation (30).

STRENGTHS AND LIMITATIONS

Strengths of our study include inclusion of patients from six tertiary level medical centers across United States with high participant retention that is broadly representative of the current state of postacute care among survivors of ARF. Discharge needs were assessed and characterized with high granularity using an instrument developed with multidisciplinary input from all study sites. Our study has several limitations. First, median time to readmission in our cohort was 12 days: we hypothesize that early readmission may have led to inaccurate ascertainment of exposure given the follow-up window extended well beyond this period. Second, we measured compliance with discharge plans as the exposure; it is possible that discharge plans included nonoptimal components. The quality and proper utilization of services rendered (e.g., DME or HHS) was beyond the scope of our assessment. Our results are representative of tertiary level medical centers in United States and therefore may not necessarily be extrapolated to other systems. 5.6% of participants in our cohort were uninsured; hence, these findings may not apply to populations with higher proportion of uninsured, who may have additional barriers to access and care coordination. While we were able to exclude a large effect size, our sample size may have failed to note a more modest effect. Finally, although we used PS adjustment (14), residual confounding cannot be ruled out.

This prospective observational study provides novel and comprehensive insights into gaps in the healthcare continuum among survivors of ARF across six centers across United States. These results might be extrapolated to other conditions requiring ICU stay. Although we did not find a significant association between UDN and 3-month mortality and readmission, we identified areas where mismatches were common. Recognition of these vulnerable areas and identification of broad-based themes, which lead to such fragmentation could help streamline care coordination of ARF survivors.

In conclusion, our prospective multicenter study demonstrates that ARF survivors have significant unmet nonmedication healthcare needs in the immediate postdischarge period, and readmissions happen early. Current systems of healthcare coordination have significant gaps in addressing the needs of this rapidly growing vulnerable population and further research should focus on optimizing processes to mitigate these deficiencies.

ACKNOWLEDGMENTS

Addressing Post-Intensive Care Syndrome (APICS-01) Study Team: Elise Caraker, Sai Phani Sree Cherukuri, Naga Preethi Kadiri, Tejaswi Kalva, Mounica Koneru, Pooja Kota, Emma Maelian Lee, Mazin Ali Mahmoud, Albahi Malik, Roozbeh Nikooie, Darin Roberts, Sriharsha Singu, Parvaneh Vaziri, Katie Brown, Austin Daw, Mardee Merrill, Rilee Smith, Ellie Hirshberg, Jorie Butler, Margaret Hays, Rebecca Abel, Craig High, Emily Beck, Brent Armbruster, Darrin Applegate, Melissa Fergus, Naresh Kumar, Megan Roth, Susan Mogan, Rebecca Abel, André de Souza Licht, Isabel Londono, Julia Larson, Krystal Capers, Maria Karamourtopoulos, Benjamin Hoenig, Andrew Toksoz-Exley, Julia Crane, and Lauryn Tsai.

REFERENCES 1. Zimmerman JE, Kramer AA, Knaus WA: Changes in hospital mortality for United States intensive care unit admissions from 1988 to 2012. Crit Care. 2013; 17:R81 2. Erickson SE, Martin GS, Davis JL, et al.: Recent trends in acute lung injury mortality: 1996-2005. Crit Care Med. 2009; 37:1574–1579 3. Zambon M, Vincent JL: Mortality rates for patients with acute lung injury/ARDS have decreased over time. Chest. 2008; 133:1120–1127 4. Stefan MS, Shieh MS, Pekow PS, et al.: Epidemiology and outcomes of acute respiratory failure in the United States, 2001 to 2009: A national survey. J Hospital Med. 2013; 8:76–82 5. Brown SM, Bose S, Banner-Goodspeed V, et al.: Approaches to addressing post-intensive care syndrome among intensive care unit survivors. A narrative review. Ann Am Thorac Soc. 2019; 16:947–956 6. Elliott D, Davidson JE, Harvey MA, et al.: Exploring the scope of post-intensive care syndrome therapy and care: Engagement of non-critical care providers and survivors in a second stakeholders meeting. Crit Care Med. 2014; 42:2518–2526 7. Needham DM, Davidson J, Cohen H, et al.: Improving long-term outcomes after discharge from intensive care unit: Report from a stakeholders’ conference. Crit Care Med. 2012; 40:502–509 8. Svenningsen H, Langhorn L, Agard AS, et al.: Post-ICU symptoms, consequences, and follow-up: An integrative review. Nurs Crit Care. 2017; 22:212–220 9. Heyland DK, Groll D, Caeser M: Survivors of acute respiratory distress syndrome: Relationship between pulmonary dysfunction and long-term health-related quality of life. Crit Care Med. 2005; 33:1549–1556 10. Dowdy DW, Eid MP, Dennison CR, et al.: Quality of life after acute respiratory distress syndrome: A meta-analysis. Intensive Care Med. 2006; 32:1115–1124 11. Taylor SP, Chou SH, Sierra MF, et al.: Association between adherence to recommended care and outcomes for adult survivors of sepsis. Ann Am Thorac Soc. 2020; 17:89–97 12. Akhlaghi N, Needham DM, Bose S, et al.: Evaluating the association between unmet healthcare needs and subsequent clinical outcomes: Protocol for the Addressing Post-Intensive Care Syndrome-01 (APICS-01) multicentre cohort study. BMJ Open. 2020; 10:e040830 13. Brown SM, Dinglas VD, Akhlaghi N, et al.: Association between unmet medication needs after hospital discharge and readmission or death among acute respiratory failure survivors: The addressing post-intensive care syndrome (APICS-01) multicenter prospective cohort study. Crit Care. 2022; 26:1–11 14. Imai K, Ratkovic M: Covariate balancing propensity score. J R Stat Soc. 2014; 76:244–263 15. Landry MD, Hack LM, Coulson E, et al.: Workforce projections 2010-2020: Annual supply and demand forecasting models for physical therapists across the United States. Phys Ther. 2016; 96:71–80 16. Richards LG, Vallee C: Not just mortality and morbidity but also function: Opportunities and challenges for occupational therapy in the World Health Organization’s rehabilitation 2030 initiative. Am J Occup Ther. 2020; 74:7402070010p7402070011–7402070010p7402070016 17. U.S. Bureau of Labor Statistics: Speech-Language Pathologists. 2022. Available at: https://www.bls.gov/ooh/healthcare/speech-language-pathologists.htm#tab-6. Accessed April 22, 2022 18. Schofield-Robinson OJ, Lewis SR, Smith AF, et al.: Follow-up services for improving long-term outcomes in intensive care unit (ICU) survivors. Cochrane Database Syst Rev. 2018; 11:CD012701 19. Kuehn BM: Clinics aim to improve post-ICU recovery. JAMA. 2019; 321:1036–1038 20. Bloom SL, Stollings JL, Kirkpatrick O, et al.: Randomized clinical trial of an ICU recovery pilot program for survivors of critical illness. Crit Care Med. 2019; 47:1337–1345 21. Bakhru RN, Davidson JF, Bookstaver RE, et al.: Implementation of an ICU recovery clinic at a tertiary care academic center. Crit Care Explor. 2019; 1:e0034 22. Huggins EL, Bloom SL, Stollings JL, et al.: A clinic model: Post-intensive care syndrome and post-intensive care syndrome-family. AACN Adv Crit Care. 2016; 27:204–211 23. Teixeira C, Rosa RG: Post-intensive care outpatient clinic: Is it feasible and effective? A literature review. Rev Bras Ter Intensiva. 2018; 30:98–111 24. Ahmad MH, Teo SP: Post-intensive care syndrome. Ann Geriatr Med Res. 2021; 25:72–78 25. Sevin CM, Bloom SL, Jackson JC, et al.: Comprehensive care of ICU survivors: Development and implementation of an ICU recovery center. J Crit Care. 2018; 46:141–148 26. Martillo MA, Dangayach NS, Tabacof L, et al.: Postintensive care syndrome in survivors of critical illness related to coronavirus disease 2019: Cohort study from a New York City Critical Care Recovery Clinic. Crit Care Med. 2021; 49:1427–1438 27. Castro-Avila AC, Jefferson L, Dale V, et al.: Support and follow-up needs of patients discharged from intensive care after severe COVID-19: A mixed-methods study of the views of UK general practitioners and intensive care staff during the pandemic’s first wave. BMJ Open. 2021; 11:e048392 28. Halatchev IG, McDonald JR, Wu WC: A patient-centred, comprehensive model for the care for heart failure: The 360 degrees heart failure centre. Open heart. 2020; 7:e001221 29. Albert NM, Barnason S, Deswal A, et al.: Transitions of care in heart failure: A scientific statement from the American Heart Association. Circ Heart Fail. 2015; 8:384–409 30. Taylor SP, Murphy S, Rios A, et al.: Effect of a multicomponent sepsis transition and recovery program on mortality and readmissions after sepsis: The improving morbidity during post-acute care transitions for sepsis randomized clinical trial. Crit Care Med. 2022; 50:469–479