Abstract
Purpose:
The medication regimen complexity-intensive care unit (MRC-ICU) score was developed prior to the existence of COVID-19. The purpose of this study was to assess if MRC-ICU could predict in-hospital mortality in patients with COVID-19.
Methods:
A single-center, observational study was conducted from August 2020 to January 2021. The primary outcome of this study was the area under the receiver operating characteristic (AUROC) for in-hospital mortality for the 48-hour MRC-ICU. Age, sequential organ failure assessment (SOFA), and World Health Organization (WHO) COVID-19 Severity Classification were assessed. Logistic regression was performed to predict in-hospital mortality as well as WHO Severity Classification at 7 days.
Results:
A total of 149 patients were included. The median SOFA score was 8 (IQR 5-11) and median MRC-ICU score at 48 hours was 15 (IQR 7-21). The in-hospital mortality rate was 36% (n = 54). The AUROC for MRC-ICU was 0.71 (95% Confidence Interval (CI), 0.62-0.78) compared to 0.66 for age, 0.81 SOFA, and 0.72 for the WHO Severity Classification. In univariate analysis, age, SOFA, MRC-ICU, and WHO Severity Classification all demonstrated significant association with in-hospital mortality, while SOFA, MRC-ICU, and WHO Severity Classification demonstrated significant association with WHO Severity Classification at 7 days. In univariate analysis, all 4 characteristics showed significant association with mortality; however, only age and SOFA remained significant following multivariate analysis.
Conclusion:
In the first analysis of medication-related variables as a predictor of severity and in-hospital mortality in COVID-19, MRC-ICU demonstrated acceptable predictive ability as represented by AUROC; however, SOFA was the strongest predictor in both AUROC and regression analysis.
Keywords: critical care, medication regimen complexity, COVID-19, sequential organ failure assessment (SOFA), COVID severity score, quality improvement
Introduction
Mortality prediction models in the intensive care unit (ICU) are required for research and quality improvement as well as clinical prognostication.1,2 Traditional severity of illness scores (eg, sequential Organ Failure Assessment [SOFA]) are the standard for mortality prediction; however, a high-profile study demonstrated poor performance of SOFA to predict mortality in mechanically ventilated patients with coronavirus disease 2019 (COVID-19).3,4 A possibility exists that traditional predictors of mortality using severity of illness scores are not as relevant in a disease defined by severe organ dysfunction often limited to one system, as with COVID-19’s predominant effects on the lungs. 1 As mortality prediction remains a highly relevant topic for COVID-19, including medication-related variables may provide a broader, yet still objective, approach to enhancing meaningful predictions.5 -9
The medication regimen complexity-intensive care unit (MRC-ICU) Score has been proposed to characterize medication regimens and has been demonstrated to predict both patient outcomes, including mortality, and workload of those individuals charged with optimizing patient’s medication therapy.10,11 In preliminary studies, the MRC-ICU demonstrated correlation to illness severity scores (Acute Physiology and Chronic Health Evaluation [APACHE] II and SOFA), patient-centered outcomes (mortality, length of stay), and pharmacist activity (as defined by the number and intensity of medication interventions performed by critical care pharmacists).10,11 However, the MRC-ICU was developed prior to the advent of COVID-19.3 -7
The purpose of this study was to evaluate the relationship of MRC-ICU to in-hospital mortality in COVID-19. Additionally, MRC-ICU’s relationship to a novel scale proposed by the World Health Organization (WHO) to classify COVID-19 severity of illness based on clinical scenario (eg, room air vs invasive positive pressure ventilation) was explored. 8
Methods
This retrospective study was approved by the institutional review board. Data were collected between August 2020 and January 2021. Patients aged 18 or older with COVID-19 receiving therapy with dexamethasone at the time of study inclusion and admitted to the medical intensive care unit were included. Length of stay was defined as time from admission to discharge or death, whichever came first. The primary outcome was the area under the receiver operating characteristic (AUROC) of the MRC-ICU at 48 hours for in-hospital mortality. AUROC is a probability curve of true positive rate versus false positive rate that measures the degree of separability for a binary classification with values closer to 1 indicating higher performance.
The MRC-ICU is a validated 37-line item score that has been previously validated to calculate an individual patient’s MRC-ICU score at a given time point. To calculate, each medication is assigned a weighted value ranging from 1 to 3, and these values are summed to provide a total score. 11
MRC-ICU and relevant demographic data were evaluated through review of the electronic medical record. WHO Severity Classification was evaluated at 48 hours and 7 days. The AUROC was also assessed of MRC-ICU for mechanical ventilation at 7 days with patients expiring before 7 days excluded from the analysis. Binary logistic regression was applied to the primary outcome of in-hospital mortality for age, SOFA, MRC-ICU, and WHO Severity Classification.
Data points were determined to fit a non-parametric distribution due to small sample size, so continuous variables were reported as medians with interquartile range (IQR). Statistical significance was assessed using an alpha level of .05, and results were reported as odds ratio (OR) with 95% confidence intervals (CI). Statistical analyses were completed using IBM SPSS Statistics for Windows, Version 27.0 (Armonk, NY: IBM Corp).
Results
A total of 149 patients were included. The population was 51% male (n = 77) with a median age of 68 years (IQR 57-75). The median MRC-ICU score at 48 hours was 15 (IQR 7-21), and the in-hospital mortality rate was 36% (n = 54). At baseline, the median WHO COVID-19 Severity Classification was 5 (IQR 4-7), which is equivalent to non-invasive ventilation or high-flow oxygen, with 128 (86%) of the cohort requiring invasive positive pressure ventilation at some point in their hospital stay. Table 1 summarizes relevant demographic information.
Table 1.
Summary Demographics and Outcomes.
| Variable | (n = 149) |
|---|---|
| Age, median (IQR), y | 68 (57.5-75) |
| Sex, n (%), male | 77 (51) |
| BMI, median (IQR) | 30.4 (26.6-36.3) |
| Hospital length of stay, median (IQR), d | 12.8 (8.4-24.1) |
| Laboratory values, median (IQR) | |
| D-Dimer, ng/mL | 560.5 (319.5-1557.3) |
| CRP, mg/L | 12.1 (5-19.2) |
| Comorbidities, n (%) | |
| DM | 70 (46) |
| COPD | 16 (11) |
| CAD | 28 (18) |
| HTN | 97 (64) |
| Cancer | 25 (16) |
| Medications, n (%) | |
| Remdesivir | 59 (39) |
| Tocilizumab | 14 (9) |
| Outcomes a | |
| Overall mortality, n (%) | 54 (36) |
| WHO COVID-19 Severity Classification at 48 h (n = 149), median (IQR) | 6 (5-7) |
| WHO COVID-19 Severity Classification at 7 d (n = 139), median (IQR) | 6 (4-7) |
| WHO COVID-19 Severity Classification at 24 h | |
| 1 (Asymptomatic) + 2 (Mild limitation in activity) | 57 (38.2) |
| 3 (Hospitalized with mild to moderate disease on room air) | 4 (2.6) |
| 4 (Hospitalized with nasal cannula or facemask oxygen) | 11 (7.3) |
| 5 (Hospitalized with high-flow nasal cannula or noninvasive positive pressure ventilation) | 43 (28.9) |
| 6 (Hospitalized with intubation and mechanical ventilation) | 11 (7.3) |
| 7 (Hospitalized with intubation and mechanical ventilation and other signs of organ failure [hemodialysis, vasopressors, extracorporeal membrane oxygenation]) | 24 (16.1) |
| 8 (Death) | 0 (0) |
Note. BMI = body mass index; CRP = C-reactive protein; DM = diabetes mellitus; COPD = chronic obstructive pulmonary disease; CAD = coronary artery disease; HTN = hypertension.
Approximately, 7% (n = 10) of patients died prior to 14 days and had scores that reflected death throughout the remaining timeframes, while approximately 36% (n = 54) of patients were excluded from scores due to being discharged earlier than 14 days after hospital admission.
The AUROC of the MRC-ICU for in-hospital mortality was 0.71 (95% Confidence Interval [CI], 0.62-0.79). SOFA demonstrated the strongest prediction for in-hospital mortality with an AUROC of 0.81 (Table 2). The AUROC of the MRC-ICU for mechanical ventilation at 7 days was 0.85 (95% CI, 0.78-0.92), while the strongest prediction for mechanical ventilation at 7 days was WHO COVID-19 Severity Classification at 48 hours.
Table 2.
AUROC of MRC-ICU, SOFA, and WHO Category for In-Hospital Mortality and Mechanical Ventilation at 7 Days.
| Variable | AUROC | 95% CI | P-value |
|---|---|---|---|
| In-hospital mortality | |||
| Age | 0.66 | 0.56-0.75 | .001 |
| MRC-ICU | 0.71 | 0.63-0.79 | <.001 |
| SOFA | 0.81 | 0.74-0.89 | <.001 |
| WHO category | 0.72 | 0.63-0.81 | <.001 |
| Mechanical ventilation at 7 d | |||
| Age | 0.55 | 0.45-0.65 | .37 |
| MRC-ICU | 0.85 | 0.77-0.92 | <.001 |
| SOFA | 0.86 | 0.79-0.93 | <.001 |
| WHO category | 0.91 | 0.86-0.96 | <.001 |
Note. WHO = World Health Organization.
In univariate analysis, all 4 characteristics (age, SOFA, MRC-ICU, and WHO COVID-19 Severity) showed significant association with mortality. For mechanical ventilation at 7 days, SOFA, MRC-ICU, and WHO COVID-19 Severity Classification showed significant association. Here, every one-point increase in MRC-ICU was associated with 9% increase in mortality. In a model incorporating age, SOFA, MRC-ICU, and WHO COVID-19 Severity Classification, MRC-ICU, and WHO COVID-19 Severity Classification were insignificant while SOFA score and age remained significant predictors of mortality. Table 3 summarizes these results of the regression analyses for mortality. For the need of mechanical ventilation after 7 days, SOFA, MRC-ICU, and WHO COVID-19 were significant in univariate analysis, and following multivariate analysis, only WHO COVID-19 remained a significant predictor (OR 7.19, 95% CI 2.91-17.78, P < .001). Table 4 summarizes results of the regression analyses for mechanical ventilation.
Table 3.
Univariate and Multivariate Analysis of In-Hospital Mortality Prediction for COVID-19.
| Factor | Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P-value | OR | 95% CI | P-value | |
| Age, y | 1.04 | 1.01-1.07 | .003 | 1.05 | 1.01-1.09 | .01 |
| SOFA | 1.38 | 1.23-1.54 | <.001 | 1.36 | 1.15-1.60 | <.001 |
| MRC-ICU | 1.09 | 1.05-1.14 | <.001 | 0.981 | 0.91-1.06 | .627 |
| WHO COVID-19 Severity Classification | 2.06 | 1.48-2.88 | <.001 | 1.29 | 0.82-2.04 | .273 |
Note. OR = odds ratio; SOFA = sequential organ failure assessment; MRC-ICU = medication regimen complexity-intensive care unit.
Table 4.
Univariate and Multivariate Analysis of Mechanical Ventilation Prediction at 7 Days for COVID-19.
| Factor | Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P-value | OR | 95% CI | P-value | |
| Age, y | 1.01 | 0.99-1.04 | .381 | 1.02 | 0.98-1.07 | .374 |
| SOFA | 1.51 | 1.32-1.73 | <.001 | 1.11 | 0.87-1.42 | .416 |
| MRC-ICU | 1.21 | 1.13-1.29 | <.001 | 1.00 | 0.89-1.11 | .924 |
| WHO COVID-19 Severity Classification | 8.88 | 4.80-16.44 | <.001 | 7.19 | 2.91-17.78 | <.001 |
Note. OR = odds ratio; SOFA = sequential organ failure assessment; MRC-ICU = medication regimen complexity-intensive care unit.
Discussion
In the first evaluation of medication regimen complexity as a predictor for in-hospital mortality in patients with COVID-19, MRC-ICU demonstrated acceptable performance for mortality prediction. Interestingly, this evaluation is the second to suggest incorporation of medication related data may support mortality prediction. 9 The higher predictive value when adding medication related data compared to the published findings of SOFA may be because this combination of illness and treatment variables makes use of the interprofessional health care team’s collective assessment. 10 The medication regimen of a particular patient represents the result of the team’s assessment of the patient’s disease, illness severity, and overall treatment needs, and the possibility exists that by quantitating this more holistic assessment of a patient that results in a given medication regimen, a more useful predictive model can be developed.
Beyond mortality, the observed relationship between MRC-ICU and WHO COVID-19 Severity Classification at 48 hours as well as MRC-ICU with subsequent oxygenation status was an important finding for future investigation in larger, prospective studies. The WHO COVID-19 Severity Classification was originally proposed in February 2020 as a means to categorize patients with COVID-19 by severity of illness. Its ability to predict severe outcomes such as mechanical ventilation and organ support has since been investigated in a variety of large-scale studies. Because MRC-ICU and WHO COVID-19 both represent categorization based on treatments or interventions, they provide a different angle of evaluating severity of illness than traditional severity of illness scores that rely largely on laboratory parameters and vital signs. Investigation into how these standardized metrics can be used in tandem may improve the ability to make clinically relevant predictions.
Notably, SOFA demonstrated good performance as a mortality predictor, which differed significantly from Raschke et al, 1 where they observed relatively poor performance of SOFA for mortality prediction (with age alone outperforming SOFA). Increased performance of SOFA in this present cohort may be driven by the treatment effects of dexamethasone on respiratory failure: in this study, dexamethasone was the standard of care and was used 100% of the time compared to 36% in Raschke et al, 1 which potentially accounts for the significantly lower mortality rates observed in this study (59% vs 36% mortality). Additionally, the entire cohort in the study by Raschke et al was mechanically ventilated while only 86% of the present cohort required mechanical ventilation, further demonstrating the potential impact of dexamethasone on clinical course. Interestingly, SOFA score was computed using the worst values within the 48 hours before intubation in the previous study, while the present study evaluated the worst SOFA score within 48 hours of ICU admission. The differences in the time points used to calculate SOFA between the 2 studies likely may also reflect changing mortality risk throughout the patient’s clinical course and appears to impact the utility of SOFA score for mortality prediction.
This study has important limitations, including its single-center, retrospective design, which precludes external generalizability and causal inference. Moreover, small sample size can lead to model overfitting to spurious trends of this dataset, and external validation of any model is warranted.
Conclusion
MRC-ICU demonstrated acceptable predictive performance for in-hospital mortality and mechanical ventilation at 7 days in this small, single-center analysis of COVID-19 patients potentially supporting further investigation in larger, prospective evaluations.
Footnotes
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding through Agency of Healthcare Research and Quality for Dr Sikora provided through R21HS028485 and R01HS029009.
ORCID iDs: Aaron Chase 
https://orcid.org/0000-0002-5891-4492
Christy Cecil Forehand
https://orcid.org/0000-0001-7523-0490
Andrea Sikora
https://orcid.org/0000-0003-2020-0571
References
- 1. Raschke RA, Agarwal S, Rangan P, Heise CW, Curry SC. Discriminant accuracy of the SOFA score for determining the probable mortality of patients with COVID-19 pneumonia requiring mechanical ventilation. JAMA. 2021;325(14):1469-1470. doi: 10.1001/jama.2021.1545 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Newsome AS, Murray B, Smith SE, et al. Optimization of critical care pharmacy clinical services: a gap analysis approach. Am J Health Syst Pharm. 2021;78(22):2077-2085. doi: 10.1093/ajhp/zxab237 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Horby P, Lim WS, Emberson JR, et al. Dexamethasone in hospitalized patients with Covid-19. N Engl J Med. 2021;384(8):693-704. doi: 10.1056/NEJMoa2021436 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Gwynn ME, Poisson MO, Waller JL, Newsome AS. Development and validation of a medication regimen complexity scoring tool for critically ill patients. Am J Health Syst Pharm. 2019;76(Supplement_2):S34-S40. doi: 10.1093/ajhp/zxy054 [DOI] [PubMed] [Google Scholar]
- 5. Newsome AS, Anderson D, Gwynn ME, Waller JL. Characterization of changes in medication complexity using a modified scoring tool. Am J Health Syst Pharm. 2019;76(Supplement_4):S92-S95. doi: 10.1093/ajhp/zxz213 [DOI] [PubMed] [Google Scholar]
- 6. Olney WJ, Chase AM, Hannah SA, Smith SE, Newsome AS. Medication regimen complexity score as an indicator of fluid balance in critically ill patients. J Pharm Pract. 2022;35:573-579. doi: 10.1177/0897190021999792 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Newsome AS, Smith SE, Olney WJ, et al. Medication regimen complexity is associated with pharmacist interventions and drug-drug interactions: a use of the novel MRC-ICU scoring tool. JACCP. 2020;3(1):47-56. doi: 10.1002/jac5.1146 [DOI] [Google Scholar]
- 8. World Health Organization. COVID-19 therapeutic trial synopsis. https://www.who.int/publications/i/item/covid-19-therapeutic-trial-synopsis (2021, accessed 1 March 2021).
- 9. Al-Mamun MA, Brothers T, Newsome AS. Development of machine learning models to validate a medication regimen complexity scoring tool for critically ill patients. Ann Pharmacother. 2021;55(4):421-429. doi: 10.1177/1060028020959042 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Kleinpell R, Grabenkort WR, Boyle WA, Vines DL, Olsen KM. The society of critical care medicine at 50 years: interprofessional practice in critical care: looking back and forging ahead. Crit Care Med. 2021;49(12):2017-2032. doi: 10.1097/CCM.0000000000005276 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Sikora A, Ayyala D, Rech MA, et al. Impact of pharmacists to improve patient care in the critically ill: a large multicenter analysis using meaningful metrics with the Medication Regimen Complexity-ICU (MRC-ICU) score. Crit Care Med. 2022;50(9):1318-1328. doi: 10.1097/CCM.0000000000005585 [DOI] [PMC free article] [PubMed] [Google Scholar]