HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Agabiti, N. Articles by Perucci, C. A. Search for Related Content PubMed PubMed Citation Articles by Agabiti, N. Articles by Perucci, C. A. Related Collections Cardiac - other Coronary disease

Eur J Cardiothorac Surg 2003;23:599-606
© 2003 Elsevier Science NL

Evaluating outcomes of hospital care following coronary artery bypass surgery in Rome, Italy

^a Agency for Public Health, Lazio Region, Via di S. Constanza 53, 00198 Rome, Italy
^b Department of Epidemiology, ASL RM/E, Rome, Italy

Received 21 June 2002; received in revised form 17 December 2002; accepted 20 December 2002.

^* Corresponding author. Tel.: +39-06-830-60476; fax: +39-06-8306-0463
e-mail: outcome{at}asplazio.it

	Abstract

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

 
Objective: Monitoring health outcomes across hospitals has become a growing interest as a potential means to promote quality of care, but in Italy it is at the beginning stage. We aimed at comparing the performance of different cardiac surgery units and testing the utility of routinely collected data in this respect. Methods: From the Lazio region hospital information system (HIS), we selected a cohort of 1603 individuals (84% males; mean age 63 years, SD 8) residing in Rome (2,685,890 inhabitants), who underwent isolated coronary artery bypass surgery (CABG, ICD-9-CM code: 36.1) in the period 1996–97 in seven major cardiac surgery units in the city. They were identified as A, B, C (teaching), D and E (non-teaching) units. Information on vital status at 30 days after the CABG surgery was obtained through an automatic record linkage with the Municipal Registry of Rome. The association between cardiac surgery units and outcome was evaluated through logistic regression taking into account the following a priori risk factors in different models: gender, age, socio-economic status, type of ischaemic heart disease and comorbidities. Results: The overall mortality was 5.4% (range 2.1–11.4%). Statistically significant predictors of outcome included: age (OR=7.5 for age70 vs. 35–49 years), acute myocardial infarction (OR=32.7 vs. acute–subacute forms/angina), chronic myocardial ischaemia (OR=4.2 vs. acute–subacute forms/angina), other heart diseases (OR=4.8), chronic renal disease (OR=16.0) and peripheral arterial disease (OR=2.9). Statistically significant variability in mortality was observed across hospitals; taking hospital A as reference, hospital D showed the highest risk (OR=5.7, 95% CI=1.9–17.3, in the fully adjusted model). Conclusions: We suggest that a true variation in quality of care play a role in the observed differences across hospitals, although chance and inaccurately measured risk factors cannot be excluded. Despite some limitations, the HIS is a valid tool for screening cardiac surgery units with poor performance.

Key Words: Coronary artery bypass surgery • Discharge abstract data • Thirty-day mortality

	1. Introduction

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

 
In the last few decades, quality assessment in health care has become a growing interest for policymakers, administrators and clinicians, each having different perspectives [1]. As cardiac surgery is more costly than other specialities and has a moderately high in-hospital mortality, special attention has been devoted to the quality of care provided by cardiac surgical centres. With respect to this, in-hospital mortality after coronary artery bypass surgery (CABG) has been studied extensively, methods for profiling providers have been developed, and several risk-adjustment models have been proposed to take into account differences in patient severity of disease [2,3]. In many countries, Cardiac Surgery Registries have been developed since the 1980s to monitor outcomes over time and across providers or surgeons, and to evaluate the efficacy of programmes aimed at improving the quality and appropriateness of cardiac surgical procedures [4,5]. In the UK, where since 1977 cardiac surgeons voluntarily submit their annual figures to an anonymous dataset mainly to assess their own practice, the issue of disseminating outcome data in cardiac surgical practice has been undertaken as a new challenge by the National Health Service [6].

Several studies in many countries have documented wide inter-surgeon, inter-hospital and inter-regional variations in patient mortality after CABG that persist despite statistical adjustment for differences in patient case-mix, i.e. demographic and clinical characteristics [7,8]. However, even if adjustment for patients' illness severity is well performed, it is not clearly established whether high mortality rates mean poor quality. It has also been suggested that hospital league tables for mortality following heart surgery can be of limited value because of the possible year to year random variations in death rate even when case-mix does not change [9], and mortality rates have been criticised as a valid tool in assessing quality of health care [10].

The Italian National Health System (NHS) provides universal coverage for all 57 million citizens. As in many European countries, health policy has been dealing with important choices aimed at improving health care quality and containing costs. Within a national research framework on the themes of effectiveness and equity in hospital care, we evaluated mortality at 30 days following CABG in five public hospitals in Rome, taking into account individual characteristics as reported in the regional hospital information system (HIS) data set, and using an empirical method to evaluate patients' illness severity.

	2. Materials and methods

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

 
2.1. Subjects
From the Lazio HIS, we defined a cohort of 1603 patients (age 35 years) residing in Rome (2,685,890 inhabitants), who underwent CABG (ICD-9-CM code: 36.1) during 1996–1997 in the five specialised hospitals in the city. We previously excluded patients who underwent major operations on heart and/or arteries other than CABG during the same admission to analyse isolated CABG patients. Personal identification number, gender, age, residential address, up to four diagnoses and up to four surgical procedures (ICD-9 codes), admission and discharge date were available from the HIS database. Information on census tracts (CTs) of residence and vital status 30 days after CABG surgery of each subject was obtained through an automatic record linkage with the Municipal Registry of Rome. For each CT (average population: 480 inhabitants) a small area socio-economic index (SEI; four classes, I being the lowest social class and IV the highest) was derived from selected census variables including educational level, occupation, dwelling ownership, family size, people/room density [11]. Patients were then attributed the SEI value relevant to the CT of residence. Hospitals were identified as A, B, C (teaching), D and E (non-teaching).

2.2. Risk factors
We analysed the diagnoses reported in the discharge abstracts (coded according to ICD-9) to define the following variables as a priori risk factors:

1. Type of ischaemic heart disease: acute–subacute forms/angina pectoris, chronic forms/old myocardial infarction, acute myocardial infarction, mixed forms (defined when more than one type was reported).
   
2. Comorbidities in the index admission: defined as chronic conditions other than ischaemic heart disease existing before the patient's admission to the hospital – other heart diseases, diabetes, hypertension, chronic obstructive respiratory disease (COPD), chronic renal diseases, peripheral arterial diseases (including cerebrovascular diseases). Comorbidities were defined adapting the ICD-9 codes assigned by Deyo et al. to the Charlson's comorbidity index [12] to obtain an illness severity definition more appropriate to the perioperative setting [13]. We also prepared a list of death-related conditions (e.g. cardiac arrest), and complications (acute conditions that probably arose after the CABG operation, distinguished in local and systemic), and excluded them when defining illness severity.
   
3. Comorbidities both in the index and in the previous hospital admissions: we also examined the comorbidities reported in previous hospital admissions (up to 12 months before CABG surgery) obtained by an automatic record linkage procedure within the HIS database to provide a better definition of the patient's illness severity. A patient was defined as having a comorbidity when this was reported in the index hospital admission and/or in a previous hospital admission.
   

2.3. Data analysis
Risk factors distribution and their association with the outcome were first explored by univariate analysis. Logistic regression analysis was then performed in order to evaluate the relationship between hospital of care and risk of 30-day mortality, after adjusting for potential confounders (ORs and 95% CI). Backward stepwise regression was used to discard the set of independent variables that did not add to the performance of the model (P>0.20) [14]. We applied different models to estimate better the effect of hospitals on mortality taking into account various potential confounders. The following variables were included in the models:

Model 1 (restricted model): demographic variables (gender, age, SEI);

Model 2: demographic variables, type of ischaemic heart disease, comorbidities (type);

Model 3 (full model): demographic variables, type of ischaemic heart disease, comorbidities (both in the index and in the previous admissions).

The area under the receiver operator characteristic (ROC) curve was estimated as a measure of the overall predictive ability of each model [15].

All statistical analyses were conducted using the statistical program STATA 5.0.

	3. Results

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

 
3.1. Characteristics of the study subjects
Most of the patients (84%) were males; mean age was 63.2 years (SD=8.5). A higher proportion of lower social class patients were treated in public non-teaching hospitals (hospitals D 27%, E 25%) compared to other hospitals. A similar pattern was observed for females and older patients. Overall, the most frequent type of ischaemic disease was chronic forms (64%), while acute myocardial infarction was reported in 2.3% of the cases. A total of 1227 subjects (65%) had at least one episode of care during the previous 12 months. Prevalence of reported comorbidities increased when including previous admissions (40% for the index admissions, 63% for index plus previous admissions). Hospital D showed a high rate of comorbidities (60%, Table 1) and the highest value of mean length of stay (43 days, Table 2). Complication rates also varied across hospitals (range: from 0.9 in Hospital A to 9.1% in Hospital C).

	4. Discussion

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

Random variation is one of the critical points when examining differences in mortality data. It has been observed that the validity of mortality rates depend on the availability of sufficiently large numbers, to minimise the potential effects of chance [9]. Studies on the impact of random variation on hospital mortality rates have led to conflicting results that may reflect differences in diagnosis that were studied and in sampling size [10]. Time periods greater than 1 year have been proposed as a better tool when comparing hospital mortality data [16]. In addition, under the assumption of a perfect risk-adjustment in simulated analysis, it has been shown that 56–82% of the observed mortality difference between high and low outlier hospitals could be attributed to random variation [17]. In general, making quantitative comparisons between institutions is a complex problem on the statistic point of view and caution is needed when screening and interpreting small differences in hospital performances [18]. In our study, as far as hospital D is considered, random variation seems to play a minor role, as shown by the high statistical significance of the crude and adjusted ORs.

Although mortality data have been extensively used in assessing quality of care, a number of limitations have been noted. When a hospital's observed mortality rate is so much greater than expected, the hospital is considered as a high outlier and is presumed to be delivering poor quality care. However, the validity of this assumption has not been clarified [17]. Previous studies have found weak relationships between risk-adjusted hospital mortality rates and other independent measures of hospital quality, and it has been proposed that optimum measurement of quality of care should take into account several outcomes and process of care indicators [19]. However, this criticism is generally based on the possible use of mortality data to compare hospital performance for different medical conditions or general surgery [17], while cardiac surgery, in particular CABG surgery, represents an exception. We used 30-day mortality after CABG as the outcome measure, which is considered to fairly reflect the institutional habits concerning surgical treatment and postoperative patient care and has been widely used as a valid tool to judge quality of surgical centres [4,5].

The quantification of illness severity represents a major problem in our case as in many studies comparing health outcomes, and the validity of administrative data in this respect has been questioned [20]. We derived information on illness severity from the ICD-9 diagnosis and procedure codes and we cannot completely exclude different patterns of accuracy and completeness of coding across hospitals. HIS in Lazio region has been implemented in 1994 and specific training programs for personnel (physicians and administrators) are periodically realised in each hospital to improve quality of coding on the basis of regional ICD-9-CM guidelines [21]. The regional health authority routinely checks data quality from all hospitals in the region and, in general, data registration has improved along years. A ‘re-abstract study’ of a random sample of 395 medical records for patients who had bypass operations in 1997–98 was conducted as part of a larger validation study to discharge abstract data in our region¹. Accuracy of CABG procedure code (ICD-9-CM 36.1) was high (confirmation rate: 98%); comorbidities showed high levels of specificity (i.e. diabetes 99%, hypertension 97%, COPD 99%) and of positive predictive value (i.e. diabetes 96%, hypertension 93%, COPD 85%) but lower levels of sensitivity (i.e. diabetes 50%, hypertension 33%, COPD 31%); similar results were found for surgical complications (specificity 98%, positive predictive value 58%, sensitivity 26%). Quality of coding tended to vary across hospitals leading to a possible undifferential misclassification of comorbidity status.

Many studies have suggested the critical value of chronic comorbidities in predicting health care outcomes, and an underreporting of comorbidities in administrative data compared to clinical data, has been reported, which differs across conditions and type of comorbidities [22]. Confirming other studies based on claims data, in our dataset, chronic conditions such as diabetes, hypertension and COPD tend to have a paradoxical protective role [20]. We tried to reduce this bias linking records under study with previous admissions to take more information about comorbidities, and the paradoxical protective effect vanished. Linking data sets have the advantage to reduce differences in coding patterns across hospitals, but it can also lower the impact on outcome of the current comorbidity status given the observed small decrease of the positive associations of some comorbidities when using both past and present information. In addition, little is known about the impact of comorbidities underreporting. In our study, the inclusion of a more detailed definition of comorbidities had no important impact on hospital ranking. Similarly, information available from previous hospital admissions yielded only small improvement in the performance of models in a recent study aimed at comparing two comorbidity risk-adjustment models [23].

Risk-adjustment is an essential tool when making comparisons among providers. Several risk-adjustment methods have been developed in the area of CABG surgery based both on clinical data and discharge abstracts and a debate about their reliability and application of different methodologies has recently started [22]. Discharge abstract data are commonly used to evaluate hospital quality for various surgical procedures and medical conditions and have important advantages, as they are readily available, inexpensive to acquire, computer-readable and typically encompass large populations. However, because of the limited insight into the patients' clinical conditions, they have often been judged inadequate, compared to clinical data sets in capturing illness severity. We had no information about clinical variables potentially related to outcomes after CABG, however, this problem could of limited importance in our contest of hospital profiling where we had to balance the importance of taking into account all potential confounders and the need of running sufficiently parsimonious models. The situation is different in the clinical setting where exhaustive information of physiological and clinical variables is essential to generate a prognostic score and to predict an individual patient outcome. Although in our study a possible residual confounding from unmeasured risk factors cannot be excluded, it is unlikely that it completely explains the magnitude of the observed differences. Moreover, it has been observed that different methods, either data sources – clinical or discharge abstract – produce generally comparable hospital rankings for CABG surgery, and severity-adjustment alone cannot be considered sufficient to fully interpret quality differences across hospitals [22]. More recently, the impact on provider profiling after CABG of different risk-adjustment models has been evaluated, and it has been found that the hospital risk-adjusted bypass surgery mortality rating is consistent regardless of the risk-adjustment model applied [24]. In our study, we did not apply commercial software to attribute individual severity indices, but empirically attempted to distinguish comorbidities from complications or death-related conditions, as suggested by other investigators, to improve the clinical validity of our models and control for potential confounders [25]. Old age, female gender, presence of comorbidities and concomitant surgery were risk factors for early mortality after CABG surgery, confirming previous findings [3].

A known side effect of risk-adjustment is to reduce the precision of comparisons. Our results suggest that the excess in crude mortality observed in hospitals C and E is mainly due to their dealing with more severe patients than hospital A. Although the residual excess in risk observed after adjustment could deserve further investigations, we hypothesise differences in unknown aspects of current care across the five hospitals in Rome. Our use of administrative data did not allow to examine all the technical factors which may have contributed to the different outcomes, however, differences in complication rates across hospitals made us confident in the hypothesis of different quality of care.

In conclusion, measuring quality of care is a priority in Italy and further research is needed in order to define the data required for valid hospital profiling and to diffuse this information efficiently. In this respect, discharge abstract data seem a valid tool for screening hospital performance as long as completely and accurately recorded. Public health researchers and clinicians should collaborate in developing new methods to better understand the causes that result in inter-institutional variations of health outcomes.

	Acknowledgments

 
This work has been partially funded by the Italian National Health Service – Progetto ‘Efficacia ed equità dell'assistenza ospedaliera: pubblicizzazione ed informazione ai cittadini’ (Ministero della Sanità-Dipartimento della Programmazione ‘Programmi speciali’ – Art. 12, comma 2, lett b), del D.Lgs. 502/92).

	Footnotes

 
Part of the manuscript was presented at: International Symposium ‘Risk Stratification in Cardiac Surgery for Quality Assurance and Accreditation’, Turin (Italy), January 16, 1999 and Fourth International Conference ‘Strategic Issues in Health Care Management’, University of St. Andrews, March 20–April 1, 2000.

¹ La valutazione della qualità della compilazione e codifica della scheda di dimissione ospedaliera nel Lazio. http://www.asplazio.it/asp_online/att_ospedaliera/sio/sio_altre_pubblicazioni.php.

	Appendix A

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

 
List of diagnoses and procedures with their ICD-9 codes presented in Table A1.

	Appendix B. Editorial comment

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

Received 18 December 2002; accepted 20 December 2002

Keywords: Coronary artery bypass surgery; Discharge abstract data; 30-Day mortality

There is no doubt that the publication of performance data of different cardiac surgery units is an effort toward an improvement of the quality of care to better serve the final consumer (i.e. the patient). Quality comparison is in fact mostly intended to stimulate internal quality improvement efforts by physicians and health care providers but is also used as a mean to control and contain costs. Nonetheless, several implications are related to such reports and interpretation of the data are often not univocal and subject to debate. The responsibilities of such reports are important because they can affect future decisions making at several levels. While these studies are expect to positively influence the delivered quality of care as demonstrated in some of the pioneering studies in this field [1], on the other hand they could negatively influence the decision whether or not operating on sicker or elderly patients, the restriction to surgical activities for resident in teaching hospitals, or the decrease of operating privileges for less experienced, low-volume surgeons. Health care providers of good performing hospitals would probably use such information to attract more patients or use it as a benchmark for obtaining more resources and financial support. Conversely, publications of data of poor performing hospitals would probably have a negative impact on citizens confidence and on the professional reputation of the people involved. In this respect this type of studies needs to be very accurate: the exactness and completeness of the collected data, the quality control of the collected data, the quality control of the study database, the definitions of co-morbidities and the risk stratification are among the most critical aspects that should be explored thoroughly. After the reports are published, it is difficult that the reader would go into such technical details but his attention would rather be focused on the outcome analysis.

Crude mortality rate often served as indicator of quality of care but it is of limited value without knowing the risk profile of the patients [2,3]. Non-risk stratified data are incorrect and misleading and their publication should strongly be avoided. Analysis of the EuroSCORE database [4] on 19030, European patients showed that a great number of cardiac factors contribute to surgical mortality. In the absence of any risk factor, the baseline mortality for isolated CABG was found to decrease from 3.4 to 0.4% implicating that the absolute risk of the processes and structure of care is almost absent. Therefore, the profound impact of patients risk profile on the outcome should guide any study aimed at predicting mortality or at assessing quality of care.

In the context of monitoring quality of care, Agabiti et al. in this issue of the Journal report an elegant study on the use of an hospital information system (HIS) for screening cardiac surgery units with different performances. Discharge abstracts were used to collect data as they are readily available, have a low cost, and cover a great number of patients.

This study has the merit of being one of the first attempt in Italy to demonstrate the inherent potentiality of administrative data. It is also evident how the authors have used at their best the data that were available. However, HIS data have not been validated and might not be sufficiently sound as the quality of the discharge abstract data might vary from the different centres [5]. The fact that the percentage of the reported complications does not correlate with the hospital stay is unusual and is probably an example of under- or over-reporting. Therefore, complications rate should not be used to validate the hypothesis of different quality of care. Furthermore, because of the source of the data, some of the risk factors that are usually present in most of the clinical data sets which served for risk-adjusted outcomes programs [6–10] could not be considered. In particular, the authors did not include the impact of some preoperative risk factors among others like previous cardiac surgery, neurological dysfunction, cardiogenic shock, NYHA status, ejection fraction or operative factors like urgent or emergent operation that have proved to be strongly linked with cardiac surgical mortality. Finally, their risk model was derived from a relatively small group of patients and reapplied to the same population without validation. It is already well evident that inclusion of some risk factors (models 2 and 3) significantly reduced the difference among hospital A and the other hospitals. Would these differences be further modified by including all risk factors? A final consideration would be the appropriate timing of such reports. Apart from verifying the methodology of the HIS as a screening tool, what is the value of publishing data 5 years after their collection? At this stage the published data might not serve as a stimulus for improving the quality of delivered care. In fact, health providers seems to be the most responsive to the performance data. A recent survey on reporting systems evaluated in peer review publication [11] demonstrated that patients and individual corporation rarely seek for such information and seldom use it or trust it while physicians are often sceptical and do not usually discuss it with their patients. Conversely, hospitals were found to implement process improvements especially if located in competitive environments, and used the performance data for ameliorating coding, educating staff and improving outcomes.

Quality control studies are also needed in response to the cost-containment issues that is of growing importance independently from the type of health care financing system. Nowadays decision making on technology and resource as well as human force utilization are no longer in the hands of the physicians. Non-medical decisions could therefore affect the quality of care delivered by a single or by a group of physicians. Indicators of good quality of care should therefore be clearly described so that the additional cost of a given procedure could be appropriately defended. Similarly, expensive and ineffective interventions could be more readily avoided. In this light the important role of the physician in guiding any cardiac outcomes reporting programs should be underlined. A recent pilot study demonstrated how a physician governed cardiac outcome registry was successful in ‘balancing the informational needs of the medical community with those of regulatory bodies’ [12].

There is an increased need for reliable information about the quality of delivered care. Despite some criticisms these data should be actively pursued for their continuous positive impact on internal quality control analysis. However, an even strenuous effort should be directed in assuring that the methods and the data used to generate such reports were independently evaluated and validated.

	References

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion Appendix A Appendix B. Editorial comment References

 

1. Schneider E.C., Riehl V., Courte-Wienecke S., Eddy D.M., Sennet C. Enhancing performance measurement. NCQA's road map for a health information framework. J Am Med Assoc 1999;282:1184-1190.[Abstract/Free Full Text]
2. Baretti R., Pannek N., Knecht J.P., Krabatsch T., Hubler S., Hetzer R. Risk stratification scores for predicting mortality in coronary artery bypass surgery. Thorac Cardiovasc Surg 2002;50:237-246.[CrossRef][Medline]
3. Nashef S.A., Roques F., Hammill B.G., Peterson E.D., Michel P., Grover F.L., Wyse R.K., Ferguson T.B., EurSCORE Project Group. Validation of European system for cardiac operative risk evaluation (EuroSCORE) in North American cardiac surgery. Eur J Cardiothorac Surg 2002;22:101-105.[Abstract/Free Full Text]
4. Ferguson T.B., Bradley G.H., Peterson E.D., DeLong E.R., Grover F.L., for the STS National Database Committee. A decade of change – risk profiles and outcomes for isolated coronary artery bypass grafting procedures, 1990–1999: a report from the STS National Database Committee and the Duke Clinical Research Institute. Ann Thorac Surg 2002;73:480-490.[Abstract/Free Full Text]
5. Al-Ruzzeh S., George S., Bustami M., Nakamura K., Khan S., Yacoub M., Amrani M. The early clinical and angiographic outcome of sequential coronary artery bypass grafting with the off-pump technique. J Thorac Cardiovasc Surg 2002;123:525-530.[Abstract/Free Full Text]
6. Ramsay S. Cardiac surgeons in England face publication of outcomes data. Lancet 2002;359:329.[Medline]
7. Vargas A., Doliszny K., Herlitz J., Karlsson T., McGovern P., Brandrup-Vognsen G., Luepker R.V. Characteristics and outcomes among patients undergoing coronary artery bypass grafting in Western Sweden and Minneapolis-St Paul, Minnesota. Am Heart J 2001;142:1080-1087.[CrossRef][Medline]
8. Ghali W.A., Quan H., Brant R. Coronary artery bypass grafting in Canada: hospital mortality rates, 1992–1995. Can Med Assoc J 1998;159:926-930.[Abstract]
9. Polioniecki J., Valencia O., Littlejohns P. Cumulative risk adjusted mortality chart for detecting changes in death rate: observational study of heart surgery. Br Med J 1998;316:1697-1700.[Abstract/Free Full Text]
10. Rosenthal G.E., Shah A., Way L.E., Harper D.L. Variations in standardized hospital mortality rates for six common medical diagnoses. Implications for profiling hospital quality. Med Care 1998;36:955-964.[CrossRef][Medline]
11. Michelozzi P., Perucci C., Forastiere F., Fusco D., Ancona C., Dell'Orco V. Inequality in health: socioeconomic differentials in mortality in Rome, 1990–95. J Epidemiol Community Health 1999;53:687-693.[Abstract]
12. Deyo R.A., Cherkin D.C., Ciol M.A. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613-619.[CrossRef][Medline]
13. Ancona C., Agabiti N., Forastiere F., Arcà M., Fusco D., Ferro S., Perucci C.A. Coronary artery bypass graft surgery: socioeconomic inequalities in access and in 30 day mortality. A population-based study in Rome, Italy. J Epidemiol Community Health 2000;54:930-935.[Abstract/Free Full Text]
14. Greenland S. Modeling and variable selection in epidemiologic analysis (commentary). Am J Public Health 1989;79:340-349.[Abstract/Free Full Text]
15. Hanley J.A., McNeil B.J. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29-36.[Abstract/Free Full Text]
16. Marshall G., Shroyer L.W., Grover F.L., Hammermeister K.E. Time series monitors of outcomes. A new dimension for measuring quality of care. Med Care 1998;36:348-356.[CrossRef][Medline]
17. Thomas J.W., Hofer T.P. Accuracy of risk-adjusted mortality rate as a measure of hospital quality of care. Med Care 1999;37:83-92.[CrossRef][Medline]
18. Goldstein H., Spiegelhalter D.J. League tables and their limitations: statistical issues in comparisons of institutional performance. J R Stat Soc 1996;159:385-443.[CrossRef]
19. Hartz A.J., Kuhn E.M. Comparing hospitals that perform coronary artery bypass surgery: the effect of outcome measures and data sources. Am J Public Health 1994;84:1609-1614.[Abstract/Free Full Text]
20. Romano P.S., Mark D.H. Bias in the coding of hospital discharge data and its implications for quality assessment. Med Care 1994;32(1):81-90.[Medline]
21. Arcà M, Borgia P, Papini P, Perucci CA, Saitto C. La riorganizzazione del sistema informativo ospedaliero nel Lazio. Progetto Salute 1995 (DGR n. 9158 /93).
22. Iezzoni L.I. Risk adjustment for measuring healthcare outcomes, 2nd ed. Chicago, IL: Health Administration Press, 1997.
23. Stukenborg G.J., Wagner D.P., Connors A.F., Jr Comparison of the performance of two comorbidity measures, with and without information from prior hospitalizations. Med Care 2001;39:727-739.[CrossRef][Medline]
24. Peterson E.D., DeLong E.R., Muhlbaier L.H., Rosen A., Buell H.E., Kiefe C.I., Kresowik T.F. Challenges in comparing risk-adjusted bypass surgery mortality results. J Am Coll Cardiol 2000;36:2174-2184.[Abstract/Free Full Text]
25. Elixhauser A., Steiner C., Harris R., Coffey R.M. Comorbidity measures for use with administrative data. Med Care 1998;36:8-27.[CrossRef][Medline]

1. Hannan E.L., Kilburn H., Racz M., Shields E., Chassin M.R. Improving the outcomes of coronary artery bypass surgery in New York State. J Am Med Assoc 1994;271:761-766.[Abstract]
2. Greenfiled S., Aronow H.U., Elashoff R.M., Watanabe D. Flaws in mortality data: the hazards of ignoring comorbid disease. J Am Med Assoc 1988;260:2253-2255.[Abstract]
3. Green J., Winfeld N., Sharkey P., Passman L.J. The importance of the severity of illness in assessing hospital mortality. J Am Med Assoc 1990;263:241-246.[Abstract]
4. Roques F., Nashef S.A.M., Gauducheau M.E., de Vincentis C., Baudet E., Cortina J., David M., Faichney A., Gabrielle F., Gams E., Harjula A., Jones M.T., Pinna Pintor P., Salamon R., Thulin L. Risk factors and outcome in European cardiac surgery: analysis of the EuroSCORE multinational database of 19030 patients. Eur J Cardiothorac Surg 1999;15:816-823.
5. Iezzoni L.I. Using risk-adjusted outcomes to assess clinical practice. An overview of issues pertaining to risk adjustment. Ann Thorac Surg 1994;58:1822-1826.[Abstract]
6. Parsonnet V., Dean D., Bernstein A.D. A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 1989;79(Suppl. I):3-12.
7. Higgins T.L., Estefanous F.G., Loop F.D., Beck G.J., Blun J.M., Paranandi L. Stratification of morbidity and mortality outcome by preoperative risk factors in coronary artery bypass patients. J Am Med Assoc 1992;267:2344-2348.[Abstract]
8. O'Connor G.T., Plume S.K., Olmstead, et al. Multivariate prediction of in-hospital mortality associated with coronary artery bypass graft surgery. Circulation 1992;85:2110-2118.[Abstract/Free Full Text]
9. Hannan E.L., Kilburn H., Jr, O'Donnell J.F. Adult open heart surgery in New York state: an analysis of risk factors and hospital mortality rates. J Am Med Assoc 1990;264:2768-2774.[Abstract]
10. Hammermeister K.E., Johnson R., Marshall G., Grover F. Continuous assessment and improvement in quality of care. A model from the Department of Veterans Affairs Cardiac Surgery. Ann Surg 1994;219:281-290.[Medline]
11. Marshall M.N., Shekelle P.G., Leatherman S., Brook R.H. The public release of performance data. What do we expect to gain? A review of the evidence. J Am Med Assoc 2000:1866-1874.
12. Goss R.J., Whitten R.W., Phillips R.C., Johnston G.G., Hofer B.O., Mansfield P.B., Tidwell S.L., Spertus J.A., LoGerfo J.P. Washington state's model of physician leadership in cardiac outcomes reporting. Ann Thorac Surg 2000;70:671-695.[Abstract/Free Full Text]

This article has been cited by other articles:

				D. Jegger, J.-P. Revelly, J. Horisberger, L. K. von Segesser, and P. Ruchat Establishing an association between a peri-operative perfusion score system (PerfSCORE) and post-operative patient morbidity/mortality during CPB cardiac surgery. Perfusion, September 1, 2007; 22(5): 311 - 316. [Abstract] [PDF]

				S. Picciotto, F. Forastiere, M. Stafoggia, D. D'Ippoliti, C. Ancona, and C. A Perucci Associations of area based deprivation status and individual educational attainment with incidence, treatment, and prognosis of first coronary event in Rome, Italy J. Epidemiol. Community Health, January 1, 2006; 60(1): 37 - 43. [Abstract] [Full Text] [PDF]

				F. Seccareccia, C. A. Perucci, P. D'Errigo, M. Arca, D. Fusco, S. Rosato, D. Greco, and on behalf of the research group of the Italian CAB The Italian CABG Outcome Study: short-term outcomes in patients with coronary artery bypass graft surgery Eur. J. Cardiothorac. Surg., January 1, 2006; 29(1): 56 - 62. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Agabiti, N. Articles by Perucci, C. A. Search for Related Content PubMed PubMed Citation Articles by Agabiti, N. Articles by Perucci, C. A. Related Collections Cardiac - other Coronary disease