Skip to main content
Have a personal or library account? Click to login
The Contribution of Case Mix, Skill Mix and Care Processes to the Outcomes of Community Hospitals: A Population-Based Observational Study Cover

The Contribution of Case Mix, Skill Mix and Care Processes to the Outcomes of Community Hospitals: A Population-Based Observational Study

Open Access
|Jun 2021

Figures & Tables

Table 1

Characteristics of the 2329 study patients, overall and by community hospital, Emilia-Romagna, year 2017.

COMMUNITY HOSPITALPATIENTS% FEMALEMEAN AGEMEAN BARTHEL SCORE AT ADM.% HIP FRACTURESAVG. NUMBER OF CONDITIONS% ADM. FROM ACUTE HOSPITAL
N%
146520.066.581.418.735.72.092.9
2703.065.783.131.70.02.00.0
3241.070.884.738.14.21.737.5
429812.863.183.920.431.22.696.0
51777.661.680.623.114.71.868.9
62018.665.281.932.118.42.063.2
71235.354.583.737.24.93.553.7
8964.164.681.940.53.11.646.9
9331.466.784.447.90.01.921.2
101315.667.283.229.817.62.153.4
111767.663.182.233.235.82.390.9
121195.149.681.723.212.63.295.8
1341617.953.682.751.81.71.9100.0
All2329100.061.582.431.118.92.279.6
Table 2

Rates of Barthel scoring improvement ≥10, transfer to acute care hospital, and length of stay ≥20 days in 13 community hospitals of Emilia-Romagna, year 2017.

COMMUNITY HOSPITALBARTHEL SCORING IMPROVEMENT ≥ 10TRANSFER TO ACUTE CARE HOSPITALLENGTH OF STAY ≥ 20 DAYS
N%N%N%
134073.3439.215834.0
234.5811.44970.0
3523.8625.01458.3
419666.0289.415552.0
57945.7158.59453.1
612463.32010.013969.2
74338.43427.66754.5
85764.82020.85456.3
9515.639.11648.5
106755.41410.78564.9
1111667.42212.59956.3
123832.81613.45344.5
1317548.7337.96615.9
All124856.326211.2104945.0
Table 3

Results from Bayesian multilevel logistic regression analysis: association of patient case mix with the study outcomes, Emilia-Romagna, year 2017. Odds ratio (OR) estimate is not available (—) if the patient-level predictor was discarded in preliminary automated regression analyses; clinical conditions not shown in the table were also discarded in preliminary analyses.

PREDICTORBARTHEL SCORING IMPROVEMENT ≥10TRANSFER TO ACUTE CARE HOSPITALLENGTH OF STAY ≥20 DAYS
POSTERIOR OR95% CRED. INTERVALPOSTERIOR OR95% CRED. INTERVALPOSTERIOR OR95% CRED. INTERVAL
Sex (female vs. male)1.150.94–1.400.830.63–1.091.30*1.07–1.55
Age (1-year increase)0.95*0.94–0.960.990.97–1.0030.990.98–1.01
Barthel score at admission (1-point increase)1.02*1.01–1.020.99*0.98–0.9950.99*0.986–0.993
Squared-Barthel score at admission (1-point increase)0.999*0.999–0.999
Hip fracture1.65*1.24–2.160.700.45–1.011.42*1.09–1.78
Chronic kidney disease1.90*1.28–2.70
Metastatic cancer0.45*0.22–0.83
Cerebrovascular diseases0.69*0.54–0.861.45*1.16–1.80
Dementia0.61*0.47–0.79
Admission source (hospital vs.home/residential facility)1.46*1.08–1.93
Variance Partition Coefficient (VPC)0.2710.0620.147

[i] * Significant at the 5% level.

Figure 1

Results from Bayesian multilevel logistic regression models for (a) Barthel scoring improvement ≥10 points, (b) transfer to acute care hospital, and (c) length of stay ≥20 days: CH-specific odds ratios (ORs), unadjusted and adjusted for patient case mix, Emilia-Romagna, year 2017. Dashed line indicates the null value of 1 (no difference compared to the overall average).

Note: Posterior ORs of Barthel scoring improvement were adjusted for sex, age, squared Barthel score on admission, hip fracture, metastatic cancer, cerebrovascular diseases, dementia, and admission source. Posterior ORs of transfer to acute care were adjusted for sex, age, Barthel score on admission, hip fracture, and chronic kidney disease. Posterior ORs of longer period of care were adjusted for sex, age, Barthel score on admission, hip fracture, and cerebrovascular diseases. The “unadjusted” posterior ORs of Barthel improvement were controlled for squared Barthel score on admission.

Abbreviations: CrI, credible interval.

Table 4

Results from Bayesian multilevel logistic regression analysis: association of skill mix and care processes of community hospitals with the study outcomes, Emilia-Romagna, year 2017. Odds ratio (OR) estimates are adjusted for patient-level characteristics listed in Table 3.

PREDICTORBARTHEL SCORING IMPROVEMENT ≥10TRANSFER TO ACUTE CARE HOSPITALLENGTH OF STAY ≥20 DAYS
POSTERIOR OR95% CRED. INTERVALPOSTERIOR OR95% CRED. INTERVALPOSTERIOR OR95% CRED. INTERVAL
Medical support (specialist vs GP)1.660.28–6.710.790.45–1.310.510.15–1.22
Years of operation (≥2 vs <2)1.160.23–3.910.950.55–1.511.240.46–2.85
Nurses’ working hours/week/bed (>12 vs ≤12)0.940.14–3.721.200.70–2.001.390.47–3.26
Patients/year (≥150 vs <150)3.780.59–13.60.610.33–1.071.020.31–2.54
Variance Partition Coefficient (VPC)0.2850.0550.143
DOI: https://doi.org/10.5334/ijic.5566 | Journal eISSN: 1568-4156
Language: English
Submitted on: Jun 26, 2020
Accepted on: Jun 2, 2021
Published on: Jun 21, 2021
Published by: Ubiquity Press
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year

© 2021 Davide Pianori, Kadjo Yves Cedric Adja, Jacopo Lenzi, Giulia Pieri, Andrea Rossi, Chiara Reno, Maria Pia Fantini, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.