Air particulate monitoring and geostatistical analysis. A new approach to risk assessment in clean room

Background A microbiological contamination in drugs preparation could increase hospital infection incidence and mortality rate. While most of recent studies are interested in stem cell or antineoplastic agent production, this work concerns a clean room commit in preparation of parenteral nutrition bag. This study aims to underline how classical statistical analysis and geostatistical iso-values maps can help to understand the potential hazard in drug preparation related to the environmental pollution. Methods The particulate count was carried out by a portable laser particle counter (0.3–25.0 mm sensitivity), according to the ISO 14644 and GMP. The monitoring was made from the beginning of a hospital clean room activity, twice a year for 3 years, in several conditions: ‘At rest’, ‘Operational’ and ‘Recovery Time’. Particles with diameter 0.5/m3 were sampled by probes positioned at 130 cm far from the ground, at bags filling height. Results Results from classical statistical analysis show a gradual increase of particles greater than 0.5 mm: at rest conditions particles average was 716 particle/m3 (95% CI 671.78–760.22) in the first monitoring (March 2006), but it reaches 29.228 particle/m3 (95% CI 27 395.73–31 060.27) in June 2008 and then it decreases to 1.174 particle/m3 (95% CI 1101.12– 1246.88) in the last control. Through spatial data modelling techniques (geostatistical analysis), maps of the particle’s diffusion expected values show that the highest number of particles is concentrated near the clean room entrance, while the working area maintains a lower concentration of particles (Indicative Goodness of Fit (IGF)=0.0667; K-order Rank = 1; Jacknife = 1.0016). Conclusions This study highlights the importance of a continuous environmental control to keep a clean room within highest security standards. Moreover geostatistical maps are of great importance because they allow to estimate the particles concentration over all monitoring area and not only where the samples are performed, other than showing clearly the particles distribution in the environment.