Expert Panel Names Five High-Value Practices to Speed Sepsis Recovery

More than one million adults survive sepsis in the United States every year, yet the weeks and months that follow hospital discharge remain a dangerous and poorly mapped phase of care. A new expert panel study, published in the Journal of General Internal Medicine, has now produced one of the first structured frameworks for deciding which recovery practices hospitals should actually prioritize. Using the RAND/UCLA appropriateness method, a modified Delphi approach refined through iterative anonymous scoring and moderated discussion, a 15-member multidisciplinary panel rated 18 candidate care practices on validity, importance, and resource intensity. The result is a ranked shortlist of five practices judged both highly beneficial and only moderately expensive to deliver, alongside a cautionary list of eight practices whose systematic implementation the panel declined to endorse.

The clinical stakes are considerable. Sepsis, defined as life-threatening acute organ dysfunction caused by infection, leaves survivors vulnerable to new cognitive impairment, worsening functional limitation, and exacerbation of chronic conditions. As many as 40 percent of survivors are rehospitalized within three months, and up to 30 percent die within the following year. Recovery is often prolonged and demands complex multidisciplinary care. Yet the Surviving Sepsis Campaign guidelines that address peri-discharge management rest largely on low-certainty or ungradable evidence, and the few randomized trials of sepsis recovery have evaluated complex multi-component interventions, making it difficult to isolate which individual practices drive benefit. Health systems have therefore implemented recovery programs in a fragmented manner, often without a strong empirical basis for their choices.

To close that gap, the research team, working within the Optimizing Veteran Recovery from Sepsis (OVeR-Sepsis) study funded by VA Health Systems Research, assembled practices from the Surviving Sepsis Campaign guidelines, hospital discharge guidelines, prior literature, and interviews with Veterans Health Administration clinicians. The panel that evaluated them included experts in sepsis, critical care, hospital medicine, primary care, geriatrics, inpatient and outpatient nursing, physical therapy, quality improvement, and post-critical illness recovery, as well as three Veteran patient representatives. Panel meetings ran from January through March 2025. For each practice, panelists independently rated validity, importance, and resource intensity on a nine-point Likert scale using Slido software before discussing the evidence and rerating. Disagreement was assessed statistically using the InterPercentile Range Adjusted for Symmetry, and practices with a median validity score of 7 to 9 without disagreement were rated valid.

Of the 18 practices considered, 10 were rated valid, eight uncertain, and none not valid, with no disagreement in the final ratings. Five practices emerged as the highest-value targets: scheduling recommended follow-up visits and testing before discharge; screening for limitations in basic and instrumental activities of daily living and need for durable medical equipment; educating patients on sepsis itself; educating patients on common post-sepsis symptoms such as fatigue, shortness of breath, and brain fog; and counseling patients or caregivers on discharge medications and their indications. All five were rated highly important yet only moderately resource-intensive, a combination the authors say should guide systematic implementation, quality improvement, and resource allocation.

Several of the panel’s judgments departed from existing guidelines in technically revealing ways. Scheduling outpatient follow-up within two weeks of discharge, and requiring that such visits be in-person, were both downgraded to uncertain validity. Panelists argued that reflex scheduling of every sepsis survivor, regardless of need, could erode clinic access for patients with more urgent problems, and noted that no evidence establishes a specific follow-up interval as outcome-improving or in-person visits as superior to virtual care. The final position aligns with the 2021 Surviving Sepsis Campaign guidelines, which issued no recommendation on early follow-up due to insufficient evidence, and with AHRQ’s Project RED toolkit, which recommends scheduling follow-up before discharge but does not specify timing.

The panel was similarly skeptical of systematic screening for cognitive impairment, anxiety, depression, and post-traumatic stress disorder after sepsis, a notable divergence from prior consensus recommendations. Panelists acknowledged that these conditions are common among survivors but questioned whether proactive screening improves management. Lingering delirium, they observed, may resolve with time, and an early screen risks an inaccurate diagnosis of cognitive impairment that could distress the patient, persist in the medical record, and alter care without benefit. Anxiety and depression after acute medical illness may likewise resolve with supportive care, and annual mental health screening is already recommended by the US Preventive Services Task Force and embedded in VHA practice. The panel concluded that the benefits of additional post-sepsis screening may not outweigh potential harms.

Other practices earned valid ratings but at high resource cost, placing them in a second implementation tier. Physical therapy evaluation before discharge was rated valid and highly important, with panelists suggesting that the AM-PAC

The panel’s reasoning also illustrates how resource intensity was defined expansively rather than narrowly. When panelists scored a screening practice, they were instructed to count not only the cost of administering the screen itself but also the downstream work it triggers, including any confirmatory testing, referrals, or treatment that a positive result would set in motion. This framing matters because practices that appear inexpensive at the bedside can become substantially more burdensome once their full cascade of follow-up is accounted for. Conversely, practices like medication counseling require little more than clinician time and careful communication, which helps explain why they clustered in the favorable high-importance, moderate-resource quadrant of the panel’s final ratings.

The composition of the panel itself shaped these judgments, and the authors are candid about its limits. Fifteen voting members included eight clinicians from the Veterans Health Administration, four clinicians from outside that system, and three Veteran patient representatives drawn from the Ann Arbor VA Veterans Research Engagement Council. Panelists were affiliated with seven VHA facilities across seven geographic regions and four national program offices covering hospital medicine, quality and patient safety, primary care, and nursing. The study team participated in deliberations but did not vote. Notably, the panel included no social work representation, a gap the authors flag as potentially influential, since several of the practices under consideration—particularly those touching on social support and care coordination—depend heavily on social work expertise. The predominance of VHA clinicians also raises questions about how ratings might differ in settings with different staffing models, payment structures, or patient populations.

The deliberative process itself changed outcomes in measurable ways. Five practices shifted between the initial and final rounds of rating, with four moving from valid to uncertain and one moving in the opposite direction, from uncertain to valid. These reversals followed structured discussion of the evidence, during which panelists reviewed literature summaries circulated before each meeting and debated practices where initial scores showed disagreement or wide spread. The audio-recorded and transcribed discussions were later mined for illustrative quotes that document the panel’s rationale. This dynamic—where expert consensus can erode as well as build under scrutiny—underscores that the ratings reflect collective judgment about the state of evidence rather than settled proof of effectiveness. A rating of uncertain validity is not a claim that a practice is useless; it is a statement that the panel lacked confidence it reliably improves recovery.

That distinction has practical consequences for how health systems should read the results. The five prioritized practices share a common character: they are largely informational and organizational interventions that connect patients to appropriate care rather than diagnostic or therapeutic maneuvers in themselves. Scheduling follow-up before discharge removes a well-known failure point in transitions of care, since patients who leave the hospital without an appointment in hand frequently fail to establish one. Screening for limitations in basic and instrumental activities of daily living, and for durable medical equipment needs, addresses the functional decline that sepsis survivors commonly experience and that often goes unrecognized until it causes a crisis. Education about sepsis itself, including the signs and symptoms of recurrence, equips survivors to recognize deterioration early, while education about expected post-sepsis symptoms such as fatigue, shortness of breath, and brain fog can normalize the recovery trajectory and prompt appropriate care-seeking. Counseling on discharge medications and their indications targets the medication errors and unintentional discontinuations that frequently accompany complex post-hospital regimens.

The panel’s emphasis on patient-centered scheduling deserves particular attention. In discussing the highest-priority practice, panelists stressed that follow-up should be arranged in a way that accounts for patient preferences regarding modality and timing, rather than blindly assigning patients to whatever clinic slots happen to be open. This nuance separates the endorsed practice from the downgraded one: the problem was never with follow-up itself, which the panel rated valid and highly important, but with rigid, one-size-fits-all mandates about when and in what format it must occur. The heterogeneity of sepsis presentations—ranging from patients discharged after brief resuscitation to those with prolonged organ dysfunction—makes uniform timing rules difficult to defend.

For quality improvement leaders, the framework offers a defensible starting point for resource allocation decisions that have until now lacked empirical grounding. Because the panel explicitly scored resource intensity alongside importance, the results can be mapped directly onto implementation planning: the five high-importance, moderate-resource practices represent the clearest near-term targets, while the valid but resource-intensive practices, such as pre-discharge physical therapy evaluation, may warrant adoption where institutional capacity permits. The authors intend the ratings to inform management, quality improvement efforts, and resource allocation, and the methodology—transparent scales, statistical assessment of disagreement, and documented deliberation—allows other institutions to adapt the framework to their own constraints. As evidence from future trials accumulates, the panel’s judgments can be revisited, with uncertain practices either earning validation or fading from recommended practice.

Subject of Research: Best Practices for Optimizing Recovery from Sepsis; Results of an Expert Panel

Article Title: Best Practices for Optimizing Recovery from Sepsis; Results of an Expert Panel

Article References: Sussman, J. B., Hogan, C. K., Whittington, T., Akgün, K. M., Anderson, M. L., Duseja, R., Eaton, T. L., Falvey, J. R., Ganz, D. A., McPeake, J., Taylor, S. P., Pawlikowski, S., Piazza, K. M., Sullivan, S. C., Bernstein, S. J., & Prescott, H. C. (2026). Best Practices for Optimizing Recovery from Sepsis; Results of an Expert Panel. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10926-1

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10926-1

Keywords: Best, Practices, Optimizing, Recovery, Sepsis, Results, Expert, Panel, scientific research