Bollettino epidemiologico nazionale

Invasive Cardiac Procedures for Ischemic Heart Diseas in the Veneto Region

Mario Saugo¹, Bortolo Martini², FedericaMichieletto³, Gianstefano Blengio⁴,  Sandro Caffi⁵ and Filippo Palumbo⁶

¹Servizio Epidemiologico ULSS Alto Vicentino, Thiene

²Unità Operativa Cardiologia, ULSS Alto Vicentino, Thiene

³Direzione Regionale per la Prevenzione, Venezia

⁴Centro Tematico di Epidemiologia Ambientale del Veneto, Bussolengo

⁵Direzione Generale, ULSS Alto Vicentino, Thiene

⁶Direzione per la Programmazione Socio-Sanitaria del Veneto, Venezia

The rapid spread of invasive cardiovascular procedures, especially in older age groups, and the need to evaluate the appropriateness of such procedures as well as their associated costs, have led to the development of routine monitoring systems that are population-based and are focused on not only the frequency of the diseases for which the techniques are considered appropriate as well as the rate of utilization of these techniques (needs-supply analysis). The objective of the current study is to use, in an exploratory manner, the health data bases currently available in the Local Health Authorities of the Veneto Region to estimate both the rate of access to invasive cardiology services as well as some of its determinants, with the goal of determining whether the use of invasive cardiac procedures can be correlated with indicators of health “needs” and of availability of such procedures.

The Health Information System and Information Technology Service of the Veneto region provided anonymous data from the computerized hospital discharge records for 2000. The conditions of interest (ischemic heart disease) was identified using ICD9-CM codes 410… as the primary diagnosis. Thus, admissions for unstable angina (411.1, 413.0, 413.1) and initial episodes of acute myocardial infarct. Invasive cardiac procedures were identified using the codes 88.55-88.57 (coronary arteriography), 36.01, 36.02, and 36.05 (transluminal coronary angioplasty; PTCA). The diagnoses and corresponding procedures were attributed to the Local Health Authority of residence of the patient. The presence of a hemodynampics laboratory in the Authority was certified by a panel of cardiologists in the context of another study (1).

For each Local Health Authority, direct standardization was used to evaluate the ratio of observed versus expected rates by sex and age for coronary arteriography and PTCA, and 95% confidence intervals were calculated using the Poisson distribution. The number of expected procedures in each Authority was calculated by attributing to each age strata of the Authority of residence the mean rate of procedures for 2000 for the region as a whole. To estimate the rate of hospitalization for acute myocardial infarct and unstable angina, rates were also standardized for each Authority using indirect standardization. 1995-1997 standardized mortality rates for myocardial infarctions for persons between 35 and 64 years of age were provided by the Regional Prevention Office of the Veneto Region.

During 2002, a total of 24,606 hospitalizations for ischemic cardiac disease (5773 with an initial hospitalization for acute myocardial infarct and 4773 with unstable angina) were identified. The rate of hospitalization per 100,000 was 312 for males and 163 for females; in those 25 to 74 years, the values were, respectively 336 and 115 per 100,0000.

A total of 7551 coronary arteriography and 3680 PTCA were performed. The age-specific rates per 100,000 residents are reported in figure 1. The mean rate of utilization per 100,000 residents was 167.4 for coronary arteriography and 81.6 for PTCA.

Figure 2 shows the rates standardized by age and sex for coronary arteriography for each Local Health Authority of residence. Of the 21 Authorities, 6 had values significantly above and 6 below the regional mean.

To examine possible determinants of the delivery of invasive cardiology procedures, a Poisson regression model was examined using 2 indicators of “need” (the rate of hospitalization for acute myocardial infarcts and the standardized rate of mortality for acute infarcts for 1995-1997 for persons 25-74 years of age) as well as the availability of a fully equipped hemodynamics laboratory as an indicator of available services.

The indicator of epidemiologic “burden” obtained from hospital discharge records is a weak predictor of the difference between the standardized rates of invasive procedures; in this ecologic analysis, mortality was instead inversely related with the rate of utilization (0.95 (95% CI 0.90-1.10) for coronary arteriography and 0,79 (95% CI 0.72-0.86) for PTCA). The availability of a hemodynamic laboratory within the Authority of residence, by contrast, explained greater access to invasive procedures (+22 for arteriography and +34 for PTCA).

This study, despite its definite limitations relative to the case definition, the enumeration of clinical events, the correctness of the ICD coding, and the ecologic nature of the study, nonetheless demonstrates that there are more invasive procedures done in those areas with a hemodynamics laboratory; this is not surprising in an expanding phase of technology. It would seem desirable to develop a network and diagnostic protocols at provincial or zonal level to help reduce potential disparities in access in those Authorities with fewer resources.

Editorial note

Carlo Saitto and Carla Ancona

Department of Epidemiology, Local Health Authority E, Rome

The use of hospital discharge records to evaluate health services has become a common tool in epidemiologic research in all countries where such systems have achieved an acceptable standard of quality.

Administrative data are collected outside the context of epidemiologic research and therefore do not represent a major cost for evaluative epidemiology units, which are often modestly funded. When administrative records are made available for research, the costs involved are generally limited to those involved with data exploration and analysis. The use of the hospital discharge records permits the evaluation of assistance in ordinary situations; analysis can include extended areas of the health care system and sometimes the entirety of services can be examined.

In the face of this apparent advantage, however, there is diffidence on the part of clinicians, and more generally, health care workers, toward observational studies based on administrative data. This attitude is due to the fact that the data suffer from a “constitutional” problem of internal validity. However, it also could be said that randomized clinical trials, which are regarded as the ‘gold standard’ in clinical medicine, suffer from the “constitutional” defect of lack of external validity.

Beyond the apparent exaggeration of this opinion, the problem with external validity seems to be of much less concern than that of internal validity; it is certainly true that the control for confounding, and thus internal validity, represents one of the fundamental problems of evaluation research on health care. For this reason, evaluation research is being combined with traditional statistical methods, including multivariate analysis techniques multilevel models, Bayesian methods, propensity scores, and instrumental variable analysis.

The solidness of the results produced through increasingly refined techniques for the control of confounding permits reliable evaluation of health services and represents a useful tool for health planning. In Italy, the computerized hospital discharge record system has been used in studies to evaluate surgical procedures linked to cardiovascular disease, including revascularization techniques (2) and invasive cardiology as reported in this number of the BEN.

These experiences confirm that a careful use of administrative data and an appropriate application of models that permit adjustment can provide useful information for the comparative evaluation of efficacy, for the description of availability of services, and for improving our understanding of the role of individual determinants in access to health services. Administrative data, however, represent an information source that has largely remained underutilized.

References

1. http://www.regione.emilia-romagna.it/agenziasan/aree/gov_clinico/

2. Agabiti N, Ancona C, Forastiere F, et al. Evaluating outcomes of hospital care following coronary artery bypass surgery in Rome, Italy. Eur J Cardiothorac Surg 2003; 23 (4): 599-606.