Chest -- eLetters for Garland, 127 (6) 2151-2164

Sir,

we read with interest the two-part Critical Care Review 揑mproving the ICU� by Garland, in the June issue of this Journal ^1,2. The author raised an important question: whilst the assessment of the quality of care in ICU is mandatory in the modern healthcare system, we are not yet able to accurately accomplish this task. Lack of quality assessment hinders the performance improvement in the single ICU.

In Italy, long before an active interest of the healthcare system in this topic, questions on performance assessment and improvement have been raised by the intensivists themselves. The circumstances that ICUs in Italy are run nearly exclusively by Anesthesiologists and are mostly general (medical/surgical) units, mainly located in public hospitals, lead to a probably more homogeneous situation than in other countries. Early in 1991 an increasing number of ICU physicians agreed to collect and share data on organization, case mix and patients outcome as a basis for quality assessment and continuous education and established a nationwide network (GiViTI � Italian Group for the Evaluation of Interventions in Intensive Care Medicine).

As an example of the activity of this group we report on our findings in 2004: 136 ICUs out of the about 400 in force in Italy collected on an ad hoc software data on 43,426 consecutive critically ill patients. The data included demographics, admission diagnoses, comorbidities, reasons for admission, severity scores, major complications occurring during the stay, major procedures/interventions performed, ICU and hospital outcome. The GiViTI coordinating center (www.giviti.marionegri.it), founded by a mix of public and private unconditioned grants, produces an annual report, where a prediction model of hospital mortality is computed via logistic regression. A personalized report for each participating ICU is also provided every year. Here the units can find specific analyses to compare their own data with those of all the others and with themselves over time.

The GiViTI philosophy is based on three key concepts: 1) only extensive data collection allows adequate insight and analysis of ICU data; 2) only very sophisticated analysis of outcome data, adjusted for as many confounders as possible, allows acceptable benchmarking; 3) the comparison with comparable units is a very good approach for each unit to identify directions to improve performance. Together with this attempt to base benchmarking on true hospital outcome parameters, rather than on weaker indicators like error reporting or other surrogate endpoints, GiViTI is aware that in many important areas improvements are best achieved by running specific research projects. In this framework, large-scale, pragmatical, participated trials should be viewed not only as the gold standard to advance knowledge, but even as the most powerful educational tool to improve daily clinical practice ³. Observational studies on sedation ⁴, on infections, on costs and organization ^5,6, and a randomized trial on enteral versus parenteral nutrition ⁷ have been carried out by GiViTI following this philosophy of a strategic alliance between research and education.

The group is actually addressing on this multicenter basis the problem of infection surveillance and the difficult issue of 揹ying in ICU�. Meanwhile, we are working on the next most challenging step: to prepare an electronic patient record to be shared by the largest number of ICUs. This way we hope to largely extend the data acquisition, with low efforts for the single unit. The aim is to pioneer a new approach to benchmark ICUs based on the process of care, while pushing evidence based prescribing and error reduction.

Martin Langer MD�, Guido Bertolini MD*

on behalf of the GiViTI group

� II Servizio Anestesia e Rianimazione, Ospedale Policlinico San Matteo IRCCS, Pavia � Italy.

* GiViTI Coordinating Center, Istituto di Ricerche Farmacologiche "Mario Negri": Centro di Ricerche Cliniche Aldo e Cele Dacc�, Ranica (Bergamo) � Italy.

References

1牋牋牋牋� Garland A. Improving the ICU: Part 1. Chest 2005; 127:2151-2164

2牋牋牋牋� Garland A. Improving the ICU: Part 2. Chest 2005; 127:2165-2179

3牋牋牋牋� Bertolini G, Rossi C, Brazzi L, et al. Local clinical research and changes in local clinical practice. Anaesthesia 2003; 58:199-200

4牋牋牋牋� Bertolini G, Minelli C, Latronico N, et al. The use of analgesic drugs in postoperative patients: the neglected problem of pain control in intensive care units. An observational, prospective, multicenter study in 128 Italian intensive care units. Eur J Clin Pharmacol 2002; 58:73-77

5牋牋牋牋� Bertolini G, Rossi C, Brazzi L, et al. The relationship between labour cost per patient and the size of intensive care units: a multicentre prospective study. Intensive Care Med 2003; 29:2307-2311

6牋牋牋牋� Brazzi L, Bertolini G, Arrighi E, et al. Top-down costing: problems in determining staff costs in intensive care medicine. Intensive Care Med 2002; 28:1661-1663

7牋牋牋牋� Bertolini G, Iapichino G, Radrizzani D, et al. Early enteral immunonutrition in patients with severe sepsis: results of an interim analysis of a randomized multicentre clinical trial. Intensive Care Med 2003; 29:834-840