ProVent Score Predicts One-Year Mortality at Day Seven of Ventilation

Prognostic timing in prolonged mechanical ventilation

Patients requiring prolonged mechanical ventilation face significant risks of long-term morbidity and high mortality rates [1]. In the intensive care unit, identifying which patients are at the highest risk for poor outcomes is essential for guiding clinical interventions and aligning treatment with patient values [2]. Historical data suggest that communication regarding end-of-life preferences often occurs late in the hospital course, frequently leading to aggressive treatments that may not align with patient goals [3]. Furthermore, baseline characteristics such as frailty and chronic health status significantly influence these long-term trajectories [4]. While prognostic models exist to assist clinicians, their utility depends on their ability to provide accurate assessments early enough to influence the care plan [2]. A recent study evaluates whether a validated mortality score can be applied sooner than traditionally recommended to assist in these critical clinical decisions, potentially allowing physicians to initiate goals-of-care discussions earlier in the intensive care unit course.

Validating mortality risk in contemporary intensive care

The ProVent14 score was originally developed to estimate one-year mortality among patients receiving mechanical ventilation for at least 14 days. To determine if this tool remains accurate in modern practice and whether it can be utilized earlier in a patient's clinical course, researchers conducted a retrospective cohort study using the Medical Information Mart for Intensive Care (MIMIC-IV) database, a large repository of comprehensive clinical data spanning from 2008 to 2022. The study population included adult intensive care unit admissions receiving at least seven days of mechanical ventilation. To ensure the findings were specifically applicable to acute respiratory failure, the investigators excluded patients with chronic neuromuscular disease, extensive burns, or those already requiring chronic mechanical ventilation. The researchers constructed three non-mutually exclusive cohorts based on the duration of ventilation, grouping patients ventilated for at least 7 days, at least 10 days, and at least 14 days. The primary outcome was one-year mortality from the onset of prolonged ventilation. To calculate risk, the researchers assessed five specific clinical predictors from the original ProVent14 model: age, platelet count, vasopressor use, hemodialysis requirement, and trauma admission status. These predictors were evaluated at day 7, day 10, and day 14 to determine if the model could maintain its accuracy when applied earlier than the traditional two-week mark. This approach allowed for a direct comparison of the model's performance as the duration of life support progressed, answering a vital question for intensivists managing prolonged respiratory failure.

Predictive accuracy across early ventilation milestones

The study analyzed three distinct groups based on the duration of mechanical ventilation to determine if prognostic accuracy changed over time. The day-7 cohort included 1344 patients, the day-10 cohort included 634 patients, and the day-14 cohort included 249 patients. Across these groups, the researchers observed a high burden of long-term mortality, with one-year mortality ranging from 64% to 67% across all cohorts. This persistently high death rate underscores the clinical necessity of early and accurate prognostic tools for patients requiring extended life support. To determine how effectively the ProVent score could distinguish between patients who survived and those who did not, the researchers assessed model discrimination using the area under the receiver operating characteristic curve (AUC, a statistical measure where 1.0 represents perfect prediction and 0.5 represents random chance). The ProVent model demonstrated an AUC of 0.74 at day 7, an AUC of 0.72 at day 10, and an AUC of 0.71 at day 14. These results indicate that the score maintains a consistent ability to differentiate outcomes even when applied a full week earlier than the standard 14-day assessment period. Furthermore, the researchers evaluated the model's calibration, which measures how closely the predicted risk of death aligns with actual observed mortality rates, using the Hosmer-Lemeshow test. In this test, a higher P-value indicates that the model's predictions are well-calibrated to real-world data. Calibration was acceptable at day 7 with a Hosmer-Lemeshow P-value of 0.50, at day 10 with a P-value of 0.18, and at day 14 with a P-value of 0.21. These findings confirm that the ProVent score provides reliable risk estimates throughout the entire second week of mechanical ventilation.

Clinical implications for earlier goals-of-care discussions

The researchers found that the ProVent score maintains stable predictive performance in a contemporary intensive care unit population, confirming its utility despite significant shifts in critical care practices over the last decade. This reliability is underscored by the fact that model performance remained stable across different study periods within the 2008 to 2022 timeframe. For the practicing intensivist, these findings validate that five objective clinical markers (age, platelet count, vasopressor use, hemodialysis requirement, and trauma admission status) remain robust predictors of one-year mortality in the modern clinical environment. The observation of similar discrimination and calibration at days 7 and 10 suggests that prognostic assessment may be feasible earlier during mechanical ventilation than the traditional 14-day threshold. Specifically, the model achieved an AUC of 0.74 and a Hosmer-Lemeshow P-value of 0.50 at day 7, providing a statistical basis for risk stratification a full week sooner than previously established. Given that one-year mortality ranged from 64% to 67% across cohorts, the ability to utilize these objective markers at day 7 or day 10 allows physicians to initiate data-driven goals-of-care discussions earlier in the hospital course. This earlier prognostic window supports clinicians in aligning treatment intensity with a patient's long-term prognosis and personal values, potentially reducing the duration of non-beneficial life-sustaining interventions and improving the quality of end-of-life care.

References

1. Imam N, Citra T, Zuhriyah K, Nggorong WP, Padaallah AP. Inspiratory muscle training in intensive care unit patients: A Systematic review. Nursing and Health Sciences Journal (NHSJ). 2025. doi:10.53713/nhsj.v5i1.446

2. Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock 2021. Critical Care Medicine. 2021. doi:10.1097/ccm.0000000000005337

3. Connors AF, Dawson NV, Desbiens NA, et al. A controlled trial to improve care for seriously ill hospitalized patients. The study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). The SUPPORT Principal Investigators. JAMA. 1995. doi:10.1001/jama.274.20.1591

4. Muscedere J, Waters B, Varambally A, et al. The impact of frailty on intensive care unit outcomes: a systematic review and meta-analysis. Intensive Care Medicine. 2017. doi:10.1007/s00134-017-4867-0