Hospital data to improve the quality of care and patient safety in oncology

Symposium QUALITY AND SAFETY IN ONCOLOGY NURSING: INTERNATIONAL PERSPECTIVES Hospital data to improve the quality of care and patient safety in oncology Dr Jean-Marie Januel, PhD, MPH, RN MER 1, IUFRS, CHUV - UNIL

Summary To show the importance of routine data in measuring patient safety in hospital To describe the current state of the project to develop Patient Safety Indicators (PSI) at international level o Example of the PSI for postoperative pulmonary embolism and deep vein thrombosis Top establish the interest of PSI and their perspectives to evaluate care in oncology

From contemplative medicine to interventional medicine From the Antiquity to the end of Middle-age o Contemplative medicine in reference to the «Malade imaginaire» from Molière 1 Claude Bernard (1813-1878) o Physiological medicine and principles of interventional medicine o Increasing of iatrogenic risks 1 Shuster E. Lancet 1998; 351: 974-977.

Content of the quality of care Efficiency Patient centered Safe Timeless QUALITY Fair Efficient Crossing the Quality Chasm, IOM. 2001

Patient safety Care ability not to be iatrogenic (not to be harmful, not causing complications). To take measures to prevent the occurrence of hospital adverse events associated with care (HAE) Or to reduce hospitalization consequences in terms of complications 5

Epidemiology of HAE 10% of hospitalizations 43% avoidable International and ubiquitary issue o AE occurrence : 3% to 19% 1 o Variations due to the definition used to identify AE mainly 6 1 De Vries, et al. Qual Saf Health Care 2008;17:216-223

Scale of risks according usual exposures in the life Transfusion accident HAE (10% of hospitalizations) Incidence of road injury Aircraft accident Cancer incidence (total (incidence per fly) population) Chemical industry accident Ultra safe Risky Unsafe * De Vries, et al. Qual Saf Health Care 2008;17:216-223 7 10-6 10-5 10-4 10-3 10-2 10-1

Limits of ad hoc study Need of important resources (prevalence survey) Cannot be repeated with high frequency Small sample size Dependant of quality of patient record / agreement across observers to identify outcomes using chart review To use hospital routine data?

Example of indicators based on hospital routine data To use administrative data (diagnostic codes based on the international classification of diseases, ICD) to identify HAE potentially associated to health care (= outcome indicators) Indirect measurement of a selection of HAE

PSI algorithm 1 PSI = 1 HAE Algorithm of diagnostic codes PSI = Codes for secondary diagnoses (SD) corresponding to HAE clinical definition At risk population defined using DRG codes, diagnosis codes, and/or procedure codes McDonald K, Romano P, et al. AHRQ Publication No. 02-0038. Rockville, MD: Agency for Healthcare Research and Quality. 2002.

Precision of PSI measurement (criterion validity) Heterogeneous PPV according to PSI Chronology in assessing HAE (code for «present on admission») Version of the ICD that is used in countries Others factors that contribute to quality of coding for ICD data

PSI /1000 stays Robustness of PSI measurement (reliability) PSI /1000 stays 30 120 2005 H 25 2006 H 100 2005 20 2005 F 80 2006 15 2006 F 60 10 40 5 20 0 [18-40[ [40-65[ [65-75[ [75-85[ >= 85 ans 0 1 [2-4[ [4-8[ [8-15[ [15-22[ [22-29[ >= 29 PSI 12 - Categories of age PSI 12 - Lenght of stay (days) Januel J.M. et al. Série Sources et Méthodes. Ministère de la Santé, DREES, Etudes & Recherches, 2011.

Fréquence, N Citations of «PSI» in PubMed 50 40 30 20 10 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Année

PSI as a standard to assess health system performance In USA, Canada, Australia (program at national level) France: ongoing development of the PSI 12 (postoperative pulmonary embolism and deep vein thrombosis) in patients undergoing hip / knee arthroplasty (by the end of 2015)

International comparisons based on PSI IMeCCHI International Methodology Consortium for Coded Health Information - An independent consortium - Subgroup on PSI (CA, CH, F, GER, AUS, USA) - Adaptation of 15 PSI from ICD-9-CM to ICD-10 AHRQ - First project - 20 PSI (ICD-9-CM) OECD - HCQI Project - A sélection of PSI - ICD-9-CM / ICD-10 - International comparisons Januel et al. Rev Epidemiol Sante Publique 2011

PSI Interest for international comparisons Performance of health systems o Comparisons (benchmarking) o To learn from others To show differences o Health system organization o Practices of coding o Healthcare practices

International comparisons of the PSI 12 (postoperative pulmonary embolism [PE] / deep vein thrombosis [DVT]) To propose a new approach to perform comparative study using an evidence basedbenchmark (corresponding to the state of art practices in healthcare) To develop and to test a such approach to HAE that occur in postoperative (example based on postoperative PE/DVT)

A three steps study To establish an external benchmark to be used as a reference value for comparisons (evidence based practice) To compare the rate of PE/DVT that occurs in hospitalized patients undergoing hip replacement using hospital routine data (PSI measurement) across several countries To explore potential factors that could explain differences between countries (as confounding factors in comparisons interpretation = potential bias)

To choose a clinical benchmark = Evidence-Based Practice Hip arthroplasty % (95% CI) I² P Total LMWH (Observational + RCT) 0.58 (0.35-0.81) 51.8% 0.001 LMWH (Observational) 0.83 (0.19-1.48) 67.3% 0.230 LMWH (RCT) 0.51 (0.26-0.76) 45.4% 0.010 Direct inhibitor of IIa/Xa factors (EC) 0.31 (0.03-0.59) 32.8% 0.070 Indirect inhibitor of IIa/Xa factors (EC) 0.68 (0.26-0.97) 0.0% 0.380 TOTAL 0.53 (0.35-0.70) 49.4% <0.001 Januel et al. JAMA 2012

A cross-sectional study Hospitalized patients ( 18 yrs) undergoing hip arthroplasty 5 countries (Switzerland, France, Canada, New-Zealand, the U.S. State of the California) Patient Safety Indicator (PSI) 12 to measure postoperative PE/DVT (= venous thromboembolism, VTE) A priori confounding factors (stratification) o Length of stay o Number of coding secondary diagnoses o Procedure codes for ultrasound that was used to screen DVT systematically

PSI algorithm to identify postoperative PE/DVT 1,2 NUMERATOR DENOMINATOR Procedure codes N PSI Inclusions Inclusions Exclusions Exclusions 12 Postoperative Secondary Procedure codes Principal Stent in the PE / DVT diagnoses of for total or partial diagnosis of cava vena PE / DVT hip prosthesis PE/DVT (recurrent PE) 18 yrs MDC 14 1 Januel JM, et al. Rev Epidemiol Sante Publique 2011; 59: 341-350. 2 OECD Health Technical Report. N 19. DELSA/ELSA/WD/http 2008.

Hospital length of stay (LOS)

Hospital length of stay (LOS)

Hospital length of stay (LOS)

Number of coded secondary diagnoses 0.05 0.04 0.03 0.02 0.01 0.00 =<2 3 4 5 6 7 8 9 10 11+ Number of Second Diagnoses Coding Fields Switzerland France Canada New-Zealand California-US

Number of coded secondary diagnoses

DVT and Procedure codes of ultrasound (France) 3.0 2.5 2.0 1.5 1.0 0.5 0.0 VTE DVT (only) PE (only) DVT+PE Type of thromboembolic events stratified by quartile (Q) for systematic ultrasound Q1 Q2 Q3 Q4

Factors associated to ultrasound coded Quartiles 1 2 3 4 Proportion of stay with ultrasound by hospital, Median 0 1.43 5.35 46.25 [IQR, 25 th 75 th percentiles] [0 0] [1.95 2.02] [3.85 7.41] [18.18 85.55] % of stays with PE/DVT 0.88 0.68 0.94 2.55 (95% CI) (0.76 1.00) (0.58 0.77) (0.84 1.04) (2.31 2.79) Volume of hip arthroplasty by hospital, Median 82 155 137 140 [IQR, 25 th 75 th percentiles] [40 140] [97 228] [74 228] [72 238] Type of hospital Public hospital and assimilated, n (%) 503 (42.70) 269 (22.84) 277 (23.51) 129 (10.95) For profit private hospital, n (%) 354 (24.65) 181 (12.60) 377 (26.25) 524 (36.49) LOS, Median 13.18 12.31 12.02 11.73 [IQR, 25 th 75 th percentiles] [10.97 15.75] [10.34 14.36] [10.97 15.75] [10.97 15.75] Number of secondary diagnoses, Median 2.41 2.35 2.31 2.56 [IQR, 25 th 75 th percentiles] [2.31 2.51] [2.23 2.47] [2.21 2.40] [2.45 2.67]

PSI as standard metrics At international level o Standard (for comparisons) o Comparability, factors to explain differences potentially At national level o Improving healthcare quality remains the fundamental and principal objective o Complementarily between PSI and ad hoc study (± electronically patient records)

N Patient Safety Indicators (PSI) de la AHRQ (CIM-9-CM) IMeCCHI OECD 1. Complications of anesthesia X X 2. Lower DRG mortality - - 3. Decubitus ulcer X X 4. Failure to rescue - - 5. Body left during procedure X X 6. Iatrogenic pneumothorax X X 7. Central venous catheter bloodstream infection X X 8. Postoperative hip fracture X X 9. Postoperative hemorrhage or hematoma - - 10. Postoperative physiological and metabolic disorders X - 11. Postoperative respiratory failure - X 12. Postoperative pulmonary embolism and deep vein thrombosis X X 13. Postoperative septis X X 14. Abdominal-pelvic surgical wound dehiscence - - 15. Technical difficulty - laceration or accidentally puncture during care X X 16. Transfusion reaction X X 17. Birth Trauma - trauma in newborn X X 18. Obstetric trauma during a vaginal delivery (with instrument) X X 19. Obstetric trauma during a vaginal delivery (without instrument) X X 20. Obstetric trauma during a cesarean X X

Interest and perspectives in oncology PSI = HAE with avoidable criteria o Oncology patients are excluded of some AHRQ PSI (denominator) To refine interesting PSI for oncology care o Also include patients with cancer in the denominator o To develop new PSI (specific)

Jean-Marie.Januel@chuv.ch