Propensity score methods for observational studies with clustered data: A review

Chang, Ting-Hsuan; Stuart, Elizabeth A.

doi:10.1002/sim.9437

Cited by 26 publications

(23 citation statements)

References 94 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Seventh, we were unable to obtain the vaccination status in the dataset. Finally, because information on hospital identification was not available from the dataset, hospital clustering was considered when calculating the propensity score, which might cause bias ( 42 ).…”

Section: Discussionmentioning

confidence: 99%

Intensive Care Unit versus High-dependency Care Unit for COVID-19 Patients with Invasive Mechanical Ventilation

Ohbe

Sasabuchi

Iwagami³

et al. 2023

Annals ATS

View full text Add to dashboard Cite

Rationale High-dependency care units (HDUs), also termed “intermediate care units”, “step-down units”, or “respiratory HDUs”, are areas in which degrees of patient care and costs are between those of the intensive care unit (ICU) and the general ward. In general, patients requiring mechanical ventilation are treated in the ICU rather than in the HDU, except for the use of HDU beds as surge capacity beds during a massive strain; however, the HDU, as well as ICU, are used as the standard care units for mechanically ventilated patients with coronavirus disease (COVID-19) in Japan. Objectives To assess the outcomes of patients with COVID-19 with invasive mechanical ventilation treated in the HDU versus those treated in the ICU. Methods In this retrospective cohort study, we used a multicenter inpatient database in Japan to identify mechanically ventilated patients with COVID-19 in the ICU or HDU on the start day of invasive mechanical ventilation from February 10, 2020, to November 30, 2021. The primary outcome was in-hospital mortality within 30 days from the start of the first invasive mechanical ventilation. Propensity score matching was performed to compare the outcomes of patients treated in the ICU with those treated in the HDU. Results Of 1,985 eligible patients with COVID-19 with invasive mechanical ventilation, 1,303 (66%) were treated in the ICU, and 682 (34%) were treated in the HDU on the start day of invasive mechanical ventilation. After propensity score matching, patients treated in the ICU had significantly lower in-hospital mortality within 30 days than those treated in the HDU (18.3% vs. 24.2%; risk difference, −5.8%; 95% confidence interval, −10.9% to −0.8%). Conclusions This multicenter observational study in Japan suggests that care for mechanically ventilated patients with COVID-19 in the ICU may significantly reduce in-hospital mortality within 30 days compared with care in the HDU. Establishing a critical care system that would allow patients with COVID-19 requiring ventilators to be treated in the ICU is desirable. Because this study was an observational study, our finding represents an association, not causation. Further studies of different critical care systems are warranted to confirm our findings.

show abstract

Section: Discussionmentioning

confidence: 99%

Intensive Care Unit versus High-dependency Care Unit for COVID-19 Patients with Invasive Mechanical Ventilation

Ohbe

Sasabuchi

Iwagami³

et al. 2023

Annals ATS

View full text Add to dashboard Cite

show abstract

“…Propensity score (PS) of oral antiviral nonusers, molnupiravir users, and nirmatrelvir/ritonavir users was estimated by multinomial logistic regression on 18 clinical characteristics, including age, sex, baseline date, cardiovascular disease, digestive disease, diabetes, malignant tumor, nervous system disease, respiratory disease, kidney disease, hemoglobin, white blood cell, platelets, creatinine, alanine aminotransferase, albumin, total bilirubin, and number of hospitalizations in the past years; the 7 geographic clusters of nursing home was included as a random intercept. 12 Average treatment effect weighting was used for those treated with nirmatrelvir/ritonavir. The balance of clinical characteristics between groups was assessed by absolute standardized mean difference, with a value of less than 0.1 indicating a good balance.…”

Section: Discussionmentioning

confidence: 99%

Clinical Outcomes Following Treatment for COVID-19 With Nirmatrelvir/Ritonavir and Molnupiravir Among Patients Living in Nursing Homes

et al. 2023

View full text Add to dashboard Cite

ImportanceOlder patients living in nursing homes are at very high risk of mortality after getting COVID-19.ObjectiveTo evaluate outcomes following oral antiviral treatment for COVID-19 among nonhospitalized older patients living in nursing homes.Design, Setting, and ParticipantsThis is a territory-wide, retrospective cohort study conducted between February 16 and March 31, 2022, with the last follow-up date on April 25, 2022. Participants were patients with COVID-19 living in nursing homes in Hong Kong. Data analysis was performed from May to June 2022.ExposuresMolnupiravir, nirmatrelvir/ritonavir, or no oral antiviral treatment.Main Outcomes and MeasuresThe primary outcome was hospitalization for COVID-19, and the secondary outcome was risk of inpatient disease progression (ie, admission to intensive care unit, use of invasive mechanical ventilation, and/or death).ResultsOf 14 617 patients (mean [SD] age, 84.8 [10.2] years; 8222 women [56.2%]), 8939 (61.2%) did not use oral antivirals, 5195 (35.5%) used molnupiravir, and 483 (3.3%) used nirmatrelvir/ritonavir. Compared with patients who did not use oral antivirals, those who used molnupiravir and nirmatrelvir/ritonavir were more likely to be female and less likely to have comorbid illnesses and hospitalization in the past year. At a median (IQR) follow-up of 30 (30-30) days, 6223 patients (42.6%) were hospitalized and 2307 patients (15.8%) experienced inpatient disease progression. After propensity score weighting, both molnupiravir and nirmatrelvir/ritonavir were associated with a reduced risk of hospitalization (molnupiravir, weighted hazard ratio [wHR], 0.46; 95% CI, 0.37-0.57; P &lt; .001; nirmatrelvir/ritonavir, wHR, 0.46; 95% CI, 0.32-0.65; P &lt; .001) and inpatient disease progression (molnupiravir, wHR, 0.35; 95% CI, 0.23-0.51; P &lt; .001; nirmatrelvir/ritonavir, wHR, 0.17; 95% CI, 0.06-0.44; P &lt; .001). Nirmatrelvir/ritonavir was comparable to molnupiravir in achieving better clinical outcomes (hospitalization, wHR, 1.00; 95% CI, 0.75-1.33; P = .99; inpatient disease progression, wHR, 0.49; 95% CI, 0.20-1.20; P = .12).Conclusions and RelevanceIn this retrospective cohort study, the use of oral antivirals to treat COVID-19 was associated with a reduced risk of hospitalization and inpatient disease progression among patients living in nursing homes. The findings of this study of nursing home residents could be reasonably extrapolated to other frail older patients living in the community.

show abstract

“…Propensity scores, which reflect the probability of being tested positive on preoperative COVID-19 screening, were calculated using random-intercept logistic regression models accounting for the multicentre design. 13 All the following variables collected a priori were forced into this non-parsimonious PS model: age class, chronic respiratory diseases, grade of surgery ( i.e. , major vs. non-major surgery), emergency surgery, ASA class ≥3, hypertension, type of anaesthesia, vaccination against COVID-19, and surgery for cancer.…”

Section: Methodsmentioning

confidence: 99%

Association of preoperative COVID-19 and postoperative respiratory morbidity during the Omicron epidemic wave: the DROMIS-22 multicentre prospective observational cohort study

Lapidus¹,

Garnier²,

Constantin³

et al. 2023

eClinicalMedicine

View full text Add to dashboard Cite

Propensity score methods for observational studies with clustered data: A review

Cited by 26 publications

References 94 publications

Intensive Care Unit versus High-dependency Care Unit for COVID-19 Patients with Invasive Mechanical Ventilation

Intensive Care Unit versus High-dependency Care Unit for COVID-19 Patients with Invasive Mechanical Ventilation

Clinical Outcomes Following Treatment for COVID-19 With Nirmatrelvir/Ritonavir and Molnupiravir Among Patients Living in Nursing Homes

Association of preoperative COVID-19 and postoperative respiratory morbidity during the Omicron epidemic wave: the DROMIS-22 multicentre prospective observational cohort study

Contact Info

Product

Resources

About