The Impact of Variations in Skill Mix on Outcomes Jan Erik Askildsen Kamrul Islam Uni Reseach & University of Bergen MUNROS is funded by the European Commission FP7 programme. HEALTH-F3-2012-305467EC. Project co-ordinators Prof Bob Elliott. r.elliott@abdn.ac.uk and Prof Christine Bond. c.m.bond@abdn.ac.uk www.abdn.ac.uk/munros
Objective Impact of differences in skill mix on patient outcomes and resource use in hospitals Patient health outcomes as measured by Self-reported health condition 30 day condition specific mortality Patient satisfaction From last visit Health care resource use Length of stay Unplanned readmissions
Variables Dependent Variables: Mortality within 30 days EQ5D and Visual Analogue Scale Patient satisfaction Unplanned readmission within 28 days Hospital Length of Stay (LOS) Explanatory Variables: Skill mix Relative nurse involvement in tasks performed Demographic: Age, gender Disease Severity: DRG weight Co-morbidity Hospital characteristics
Data Sources Questionnaires: Eight countries Routine data and patients medical records 5 countries Register data at individual patient level Scotland, England, Norway
Results The impact of skill mix on outcomes within: Breast cancer (BC) Heart disease (HD) Diabetes T2 (T2D) Controlling for: Age, sex, comorbidities, hospital characteristics, country
Self Reported Health EQ5D Differences among countries DE NL IT CZ TR NO ENG -0.300-0.200-0.100 0.000 0.100 0.200 BC T2D HD Countries compared to Scotland Self reported health poorer in Turkey and Germany
Health EQ5D Impact of skill mix -0.009 BC skill mix 0.022 HD -0.010 T2D -0.020 0.000 0.020 0.040 Marginal effect Weak negative impact of skill mix on BC and T2D, positive for HD
Survival probability within 30 days after hospitalisation (data come from Patient Records/Routine data in 5 countries) -0.277 BC skill mix -0.424 HD -0.035 T2D -1.000-0.500 0.000 0.500 Marginal effect Weakly negative or insignificant impact of skill mix on 30 days survival probability
Survival probability within 30 days: Scotland (register data) Effect of Skill Mix on Probability of No Mortality (within 30 days of discharge) -0.00 0.05-0.15-0.10-0.05 0.00 0.05 0.10 0.15 Marginal Effect on Probability Type 2 Diabetes Heart Disease 95% confidence interval provided (n=1596 / 602 / 1507) Increased 30 days survival probability for HD patients in Scotland
Survival probability within 30 days: Norway (register data) -1.015 BC skill mix -0.468 HD -2.387 T2D -3.000-2.000-1.000 0.000 Marginal effect Impact of skill mix on 30 days survival probability tends to be negative in Norway
Patient Satisfaction: a composite indicator (data gathered from all countries through Patient Questionnaire) -0.120 BC skill mix 0.832 HD 0.668 T2D -0.500 0.000 0.500 1.000 1.500 Marginal effect Positive impact of skill mix on patient satisfaction for HD and T2D, insignificant for BC
Probability of no emergency readmission within 28 days (data come from Patient Records/Routine data) -0.277 BC skill mix -0.553 HD -0.130 T2D -1.000-0.500 0.000 0.500 Marginal effect Tends to reduction in probability for avoiding readmissions
Probability of no emergency readmission within 28 days: Scotland (register data) Effect of Skill Mix on Probability of No Emergency Readmission (within 28 days of discharge) -0.01-0.09-0.15-0.10-0.05 0.00 0.05 0.10 0.15 Marginal Effect on Probability Type 2 Diabetes Heart Disease 95% confidence interval provided (n=1596 / 602 / 1507) Reduced probability for avoiding HD emergency readmission in Scotland
Probability of no emergency readmission within 28 days: Norway (register data) -1.015 BC skill mix 0.265 HD -0.020 T2D -2.000-1.000 0.000 1.000 2.000 Marginal effect Reduced probability for avoiding BC emergency readmissions in Norway Increased probability for avoiding HD readmissions
Length of hospital stay (data come from Patient Records/Routine data) -0.321 BC skill mix 1.986 HD -0.130 T2D -1.000 0.000 1.000 2.000 3.000 Marginal effect Increased LOS for HD, insignificant for BC and T2D
Length of hospital stay: Scotland (register data) Effect of Skill Mix on Length of Inpatient Stay -0.69 0.13-0.23-1.5-1.0-0.5 0.0 0.5 1.0 1.5 Number of Days Breast Cancer Type 2 Diabetes Heart Disease 95% confidence interval provided (n=1596 / 602 / 1507) Weakly negative impact on LOS for BC in Scotland
Length of hospital stay: Norway (register data) 0.722 BC skill mix -0.538 HD -0.020 T2D -2.000-1.000 0.000 1.000 2.000 Marginal effect Skill mix reduces LOS for HD patients in Norway Increase for BC
Concluding remarks Main question: Does nurse involvement matter for outcomes? Health Resource use Health and patient satisfaction: Self-reported health Negative for BC and T2D, positive for HD Survival Negative impact of skill mx, but clearly positive for HD patients in Scotland Patient satisfaction Positive for HD and T2D, no impact for BC Resource use: Probability of avoiding readmissions: Negative impact of skill mix, but clearly positive for HD patients in Norway LOS Mixed outcomes LOS: Increase for HD in 6 countries, reduction in LOS for HD in Norway Tends to reduced LOS for BC Results are diagnosis and country sensitive