Admitted critically patients from hospital general wards, either medical or surgical, will face them a situation of increased morbidity and mortality risks compared to hospitalized patients who were not admitted to intensive Care Units (ICUs). 

This ICU admitted Length of Stay (LOS) period may last from few hours to several months. Most critical care mortality risk assessment studies specify 28 days as a clinical meaningful studied period for mortality assessment. Also, overall, 28-day ICU mortality is commonly further subdivided into two mortality phases, early mortality for critically ill patients who died after ICU admission to up 14 days and late mortality for critically ill patients who died after that. ^[1-4]

To reduce early, late, or overall, critically ill patients, it is important to group admitted ICU patients based on their mortality risk to early prioritize cares and treatments in the global era of expensive managements and limited supplies. While ICUs are highly specialized equipped departments that developed for early intervention, close monitoring, and intensively caring and treating of acute diseases Many recognizable factors, either adjustable or nonadjustable, are associated with critically ill patients’ negative clinical impacts, including mortality. One of these wide variety factors, is pre-ICU admission days. ^[5-8]

Prolonged pre-intensive care unit (ICU) hospital stay length days was previously shown to be independently associated with poorer critically ill patients’ outcomes. This is probably due to accumulated nosocomial offenders’ exposure and gradual deterioration of physiological functions in hospital staying, and in more advance situations probably leading to critical care admission. Unstable or rapidly deteriorating clinical conditions for hospitalized patients demand ICU admission and because hospital staying days in the ICUs cause high-cost expenditures and affect patients, their families, and the country’s economy, targeting any adjustable prognosticator may be helpful for mitigation overall ICU admitted negative clinical impacts., especially in country of an increasing general and aging population likes our country. ^[9-12]

In order to improve an overall quality of our institutional medical care we investigated the impacts of pre-ICU admission days on the critically ill patients’ major clinical outcomes, including but not excluded to, ICU admission days, overall hospital stay days, early, late, and overall mortality risk.

Critically ill patients without a previous ICU admission and were transferred from our institutional wards, surgical and medical, to the Medical-Surgical ICU (MSICU) between Jan 2018 and May 2021, were retrospectively reviewed via the Electronic Medical System (Hakeem). Our MSICU is a 33-bed multi-disciplinary at King Hussein Medical Center (KHMC) of the Royal Medical Services (RMS) Hospitals, a major tertiary referral medical center in our country, Jordan. 

Critically ill patients who were below 18 years and whose studied variables were totally or partially missed were excluded from this study. All eligible studied critically ill patients were grouped by their hospital length of stay (HLOS) prior to ICU admission: HLOS <6 days (Group I) or HLOS ≥6 days (Group II). In addition to the recorded pre-ICU admission days, relevant admitted ICU patients’ clinical data were retrospectively retrieved. Collected data contained; demographics, diagnostics, hemodynamics, biochemical lab levels, and prognosis risk scores of Charlson Comorbidities Index (CCI) and Sequential Organ Failure Assessment (SOFA). Also, the patients Pre-ICU admission wards were further subdivided into medical and surgical wards, and the overall 28-day ICU mortality were further subcategorized into early mortality (≤14 days) and late mortality (>14 days). Serial serum albumin levels during the 1^st week of ICU admission were classified into; At admission level, 2^nd-3^rd day level, and 4^th-5^th day level. 

The Comparative variables across the two studied groups were analyzed by One Sample and Independent T-Tests or Chi Square Test to represent the analysis results as Mean±SD and Mean Difference±SEM or Number (Percentage and odd ratio (OD), respectively. The 28-day overall ICU mortality and the early ICU mortality versus late ICU mortality prognostic values of the pre-ICU HLOS prognosticator in all eligible studied critically ill patients were explored using receiver operating characteristics (ROC) method. The area under the ROC curves (AUROCs) of the pre-ICU HLOS were expressed as AUC (95% CI; Range). The optimal Cutoff operating point on each ROC curve was picked by investigating the highest youden’s index, and accordingly the sensitivity analysis results of sensitivities, specificities, accuracies, positive and negative predictive values, and negative likelihood ratios of the optimal pre-ICU HLOS regarding the early, late, and overall ICU mortality were reported in this study. All results were analyzed using SPSS version 20 (Statistical Package for the Social Sciences, Chicago, IL, U.S.A.) with p-value <0.05 as a level of significance.  

From 2155 adult and elderly admitted critically ill patients in our ICU department at KHMC, RMS, Amman, Jordan between Jan 2018 and May 2021, 1638 critically ill patients were finally included in this study with 517 non-eligible patients were excluded. The mean age of the whole study cohort was 51.88±16.22 years, and the Pre-ICU LOS<6 days Cohort were significantly older than the Pre-ICU LOS≥6 days Cohort (52.87±15.48 years versus 51.14±16.72 years, respectively, P-value=0.033). Insignificantly, males were distributed in the study in approximately 2.55: 1 ratio compared to female [1177 (71.9%) versus 461 (28.1%), respectively, p<0.05] in which 71.3% (501 critically ill men) and 28.7% (202 critically ill women) were belonged to the Pre-ICU LOS<6 days Cohort compared to 72.3% (676 critically ill men) and 27.7% (259 critically ill women) were belonged to the Pre-ICU LOS≥6 days Cohort. The mortality risk estimates for female relative to male was OD (F/M), 1.052 (95% CI; 0.847-1.308). 

An overall 28-day ICU mortality was detected in 1318 of all eligible studied critically ill patients with an overall incidence of 80.46% during an average of 14.08±4.06 days and   21.00±5.61 days of the ICU and overall hospital admission days, respectively, in which only the overall hospital LOS was significantly higher in the Pre-ICU LOS≥6 days Cohort compared to the Pre-ICU LOS<6 days Cohort [23.57±5.09 days vs 17.57±4.28 days, P-value=0.000] while  it was insignificantly higher in the Pre-ICU LOS≥6 days Cohort compared to the Pre-ICU LOS<6 days Cohort [14.18±4.02 days vs 13.96±4.11 days, P-value=0.279].

Comparatively, Pre-ICU LOS<6 days Cohort had significantly lower pre-ICU HLOS, early ICU mortality, late ICU mortality, and overall 28-day ICU mortality compared to the Pre-ICU LOS≥6 days Cohort [3.62±1.19 days, 172 (34.9%), 321 (65.1%), and 493 (70.1%) versus 9.40±3.18 days, 372 (45.1%), 453 (54.9%), and 825 (88.2%), respectively, P-value<0.005]. 

There were an overall 994 (60.7%) studied medical patients and 644 (39.3%) studied surgical patients which respectively distributed to 409 (58.2%) and 294 (41.8%) within the Pre-ICU LOS<6 days Cohort and 585 (62.6%) and 350 (37.4%) within the Pre-ICU LOS≥6 days Cohort. The mortality risk estimate for critically ill patient admitted from medical wards compared to surgical wards was OD (Med/Sur), 0.832 (95% CI; 0.681-1.017).

Serially, serum albumin levels (ALB) were only significantly higher in the Pre-ICU LOS<6 days Cohort compared to the Pre-ICU LOS≥6 days Cohort at 2^nd to 5^th day after ICU admission with Mean±SD of [2.13±0.43 g/dl and 2.31±0.75 g/dl versus 2.04±0.48 g/dl and 2.12±0.81 g/dl, respectively, P-value<0.05]. Although there were insignificant differences between the two studied cohorts (Cohort II versus Cohort I) regarding ∆SOFA [3.23±1.68 versus 3.08±1.79, respectively, P-value=0.083], Cohort II had significantly higher modified Shock Index (mSI) compared to Cohort I [1.62±0.26 bpm/mmHg versus 1.57±0.25 bpm/mmHg, respectively, P-value=0.001]. All analyzed comparative variables across the two studied cohorts are fully constructed in Table 1-2.

The areas under receiver operating characteristic (ROC) curves for the Pre-ICU Stay Days and its prognosticating utility for the 28-day overall ICU mortality and early ICU mortality versus late ICU mortality are illustrated in Figure 1-2. Our tested prognosticators had significantly AUROC with Area±SEM (95% CI; Range) of 0.694±0.015, (95% CI; 0.664-0.723) and 0.587±0.016 (95% CI; 0.556-0.619), respectively. The operating cut-off point, sensitivity (TPR), specificity (TNR), Youden index (YI), positive and nega­tive predictive values (PPV and NPV), negative likelihood ratio (NLR), and the accuracy index (AI) for the pre-ICU HLOS regarding the 28-day overall ICU mortality and early ICU mortality versus late ICU mortality are separately described in Table 3-4.

The present study includes two studied outcome cohorts, the Pre-ICU LOS<6 days Cohort (Cohort I) and the Pre-ICU LOS≥6 days Cohort (Cohort II) of admitted critically ill patients. The uniqueness of our study is primarily involved in its multi-studied negative major clinical outcomes among both medically and surgically admitted critically ill patients.

In contrast to many studies, our study didn’t show significant negative impact of the pre-ICU HLOS on the ICU LOS [13.96±4.11 days versus 14.18±4.02 days, respectively, P-value=0.279]. conversely, our tested prognosticator had significant negative impacts on all investigated mortality series, early ICU, late ICU, and overall, 28-day ICU mortality with Numbers (Percentages) of 372 (45.1%), 453 (54.9%), 825 (88.2%) versus 172 (34.9%), 321 (65.1%), and 493 (70.1%), respectively, P-value<0.05. ^[13-14]

The gap difference in albumin level across the two tested cohorts was positively and steadily increased, at least after ICU admission through 2^nd to 5^th day, with Mean difference±SEM of +0.094±0.023 g/dl and +0.192±0.039 g/dl for 2^nd-3^rd and 4^th-5^th albumin day levels, respectively, P-value<0.05.

Many studies address the negative impact correlations between pre-ICU admission days and critically ill patients’ major clinical outcomes. Most of these studies have invoked suboptimal care prior to ICU admission as the reason for poorer ICU admitted patients’ outcomes. Also, insufficient facilities and/or qualified/trained healthcare persons, failure to appreciate clinical urgency, and suboptimal clinical supervision and care have frequently suggested as factors in this negative impact correlations. In our study, we revealed that the admitted critically ill patients whose pre-ICU admitted days equal or exceed 6 days has a significantly higher baseline Sequential Organ Failure Assessment (SOFA) score ≥5 than the comparative cohort of admitted critically ill patients whose pre-ICU admitted days didn’t exceed 6 days [877 (93.8%) versus 610 (86.8%), respectively, P-value<0.05] with odd ratio of 2.305 (95% CI; 1.635-3.251). ^[15-17]

In conclusion, ICU admission from a general care floor, medically ICU admitted patients and surgically admitted ICU patients, after ≥6 days is associated with poorer clinical outcome as compared to earlier pre-ICU admission days. This study is limited by its retrospective design. A larger, multisite, and prospective study is needed to control for multiple confounders and to clarify the causations between our tested prognosticator and mortality. Despite these limitations, our conclusions may have an added value to the current excessively evolving controversial pieces of evidence, especially in critically ill cohorts.

Conflict of Interest: No

Funding: No funding sources

1. Kalish, Richard L., et al. "Costs of potential complications of care for major surgery patients." American Journal of Medical Quality 10.1 (1995): 48-54. DOI: 10.1177/0885713X9501000108. 

2. Collins, Tracie C., et al. "Risk factors for prolonged length of stay after major elective surgery." Annals of surgery 230.2 (1999): 251.

3. Kozyrskyj, Anita, Carolyn De Coster, and Philip St John. "Long stay patients in Winnipeg acute care hospitals." Healthcare Management Forum. Vol. 15. No. 4. No longer published by Elsevier, 2002. https://doi.org/10.1016/S0840-4704(10)60177-9 .

4. Singer, Adam J., et al. "The association between length of emergency department boarding and mortality." Academic Emergency Medicine 18.12 (2011): 1324-1329. https://doi.org/10.1111/j.1553-2712.2011.01236.x 

5. Schneider, Eric B., et al. "Patient readmission and mortality after surgery for hepato-pancreato-biliary malignancies." Journal of the American College of Surgeons 215.5 (2012): 607-615. https://doi.org/10.1016/j.jamcollsurg.2012.07.007 .

6. Jannasch, Olof, et al. "Nosocomial infections and microbiologic spectrum after major elective surgery of the pancreas, liver, stomach, and esophagus." Surgical infections 16.3 (2015): 338-345. https://doi.org/10.1089/sur.2013.248 .

7. Freitas, Alberto, et al. "Factors influencing hospital high length of stay outliers." BMC health services research 12 (2012): 1-10.

8. Osnabrugge RL, Speir AM, Head SJ, et al. Prediction of costs and length of stay in coronary artery bypass grafting. Ann Thorac Surg. 2014;98(4):1286–93 https://doi.org/10.1016/j.athoracsur.2014.05.073 .

9. Almashrafi, Ahmed, Mustafa Elmontsri, and Paul Aylin. "Systematic review of factors influencing length of stay in ICU after adult cardiac surgery." BMC health services research 16 (2016): 1-12.

10. Lukasiewicz, Adam M., et al. "Patient factors associated with 30-day morbidity, mortality, and length of stay after surgery for subdural hematoma: a study of the American College of Surgeons National Surgical Quality Improvement Program." Journal of neurosurgery 124.3 (2016): 760-766. https://doi.org/10.3171/2015.2.JNS142721 .

11. Matsui, Kunihiko, et al. "Comorbidity as a correlate of length of stay for hospitalized patients with acute chest pain." Journal of general internal medicine 11.5 (1996): 262-268.–

12. Polastri, Massimiliano, et al. "Preoperative physiotherapy in subjects with idiopathic pulmonary fibrosis qualified for lung transplantation: implications on hospital length of stay and clinical outcomes." Journal of thoracic disease 8.5 (2016): E264.https://doi.org/10.21037/jtd.2016.03.52 .

13. Earnest, Arul, Mark IC Chen, and Eillyne Seow. "Exploring if day and time of admission is associated with average length of stay among inpatients from a tertiary hospital in Singapore: an analytic study based on routine admission data." BMC Health Services Research 6 (2006): 1-8..

14. Nobili, Alessandro, et al. "Polypharmacy, length of hospital stay, and in-hospital mortality among elderly patients in internal medicine wards. The REPOSI study." European journal of clinical pharmacology 67 (2011): 507-519..

15. Zeng, Rui, et al. "Hypoalbuminemia predicts clinical outcome in patients with type B acute aortic dissection after endovascular therapy." The American journal of emergency medicine 34.8 (2016): 1369-1372. https://doi.org/10.1016/j.ajem.2016.03.067 .

16. Rivers, Emanuel P., and Tom Ahrens. "Improving outcomes for severe sepsis and septic shock: tools for early identification of at-risk patients and treatment protocol implementation." Critical care clinics 24.3 (2008): 1-47..

17. Jellinge, Marlene Ersgaard, et al. "Hypoalbuminemia is a strong predictor of 30-day all-cause mortality in acutely admitted medical patients: a prospective, observational, cohort study." PloS one 9.8 (2014): e105983. https://doi.org/10.1371/journal.pone.0105983 .