Analysing and interpreting routinely collected data on sharps injuries in assessing preventative actions

Transcription

1 Occupational Medicine 2004;54: DOI: /occmed/kqh041 Analysing and interpreting routinely collected data on sharps injuries in assessing preventative actions G. Moens 1,2, G. Mylle 1, K. Johannik 1, R. Van Hoof 2 and G. Helsen 1 Background Sharps injuries (SI) occur frequently in hospitals and are a risk for exposure to bloodborne pathogens. During the 1990s, the safety service of a university general hospital introduced, in collaboration with the occupational health service, specific measures to reduce the number of SI. Aim The aim of this study was to assess the occurrence and evolution of SI during this period and to evaluate the effectiveness of the preventative measures taken, making use of routinely collected data. Methods In a retrospective study, we analysed the number of SI recorded from 1990 to The study population was all employees at risk of SI. Because the introduction of intensive preventative measures dates from 1996, an effect on the incidence of SI can be expected from To assess this effect, mean incidence rates for and for were compared. Results Conclusion Key words In the study period, a total of 4230 SI were recorded. The global SI incidence rate decreased from 33.4 SI per 100 occupied beds per year in to 30.1 in (P < 0.01). In the same period, among nurses a decrease in incidence rate from 17.2 to 12.7 SI per 100 person-years was noted (P < ) and for the hotel service from 4.8 to 3.7 (not significant). Although this study has various restraints, these results suggest that intensive preventative actions, in combination with technological advances, may have contributed to a drop of 67 SI cases per year. Hospital personnel; needle stick accident; occupational injury prevention; sharps injury. Received 31 March 2003 Revised 27 August 2003 Accepted 26 January 2004 Introduction Sharps injuries (SI) occur frequently in hospitals. They constitute more than one-third of all accidents reported. The most important pathogens which can be transferred through SI are the hepatitis B virus (HBV), 1 Department of Research & Development, External Service for Prevention and Protection at Work (IDEWE), Leuven, Belgium. 2 Department of Occupational and Insurance Medicine, Faculty of Medicine, Catholic University of Leuven, Leuven, Belgium. Correspondence to: Professor Guido Moens, IDEWE, Interleuvenlaan 58, B-3001 Leuven-Heverlee, Belgium. guido.moens@idewe.be hepatitis C virus (HCV) [1] and the human immunodeficiency virus (HIV) [2,3]. For HCV and HIV, no vaccines or immunoglobulins are available, as distinct from HBV. Prevention of SI therefore deserves attention [1 6]. From 1990 on, the safety service of a university general hospital introduced, in collaboration with the occupational health service, the use of a vacuum system for blood sampling and needle containers to reduce the number of needle-stick accidents. In 1994, needleless infusion systems were introduced at two hospital services [6] and as of 1995, intensive training sessions were organized in Occupational Medicine, Vol. 54 No. 4 Society of Occupational Medicine 2004; all rights reserved 245

2 246 OCCUPATIONAL MEDICINE the use of Universal Precautions [4]. In 1996 the preventative measures were included in the year action plan and a multidisciplinary working group was started, consisting of occupational health and safety professionals, hospital hygienists, employees belonging to various occupational categories and to the board of directors. At the same time, a randomly selected sample of 10% of the personnel completed a questionnaire (response rate of 88%) in order: (i) to assess the shortcomings and needs in prevention of SI; (ii) to adjust the prevention policy; and (iii) to assess the need for extra training. A brochure of guidelines was distributed in all departments in Furthermore, since 1996, needle containers have been routinely inspected during the health and safety audits by the occupational physician for recapped needles. In case of major safety violations (e.g. needles sticking out of a refuse sack), the head of the department was asked to inform the employees. Although studies about the effectiveness of such measures are scarce, Richard et al. [7] reported a drastic decline in SI after the introduction of sharp containers and of education programmes. The aim of this study was to assess the occurrence and evolution of SI between 1990 and 1997 and to evaluate the effectiveness of the preventative measures taken, making use of routinely collected data. Study population and methods The study population in this retrospective registration study consisted of all employees at potential risk of SI in the period : nursing personnel, physicians, dentists, maintenance personnel, laboratory workers and students. The SI among nurses, physicians and the employees of maintenance, laundry and kitchen (the hotel service ), will be discussed in detail. The collection of SI was done using first-aid records, present in each department. From these records, each SI with blood contact was entered into a computer database. The following variables were collected: year of SI, name of the employee, the department in which the SI occurred, the nature and description of the accident, an eventual infection of the patient and, if known, the person responsible for the accident. Also date of birth, seniority, profession and the status of vaccination were noted. SI were defined as follows: any contact with infected body fluid through needles, sharp instruments and blood spatters on mucus or impaired skin. Potentially infected fluids were: blood and other fluids which visibly contained blood such as cerebrospinal, synovial, pleural, peritoneal, pericardial, amniotic fluids, semen and vaginal secretions. Faeces, nasal secretions, sputum, sweat, tears, urine and vomit were not considered and neither was blood contact on unimpaired skin or needle accidents with sterile needles or materials. The SI incidence rate is expressed per 100 occupied beds, or if denominator data were available, per 100 person-years. Because intensive preventative measures were only introduced in the year action plan in 1996, a major effect on the SI incidence can only be expected since Because the number of data points was too small to perform a reliable trend analysis, the mean SI incidence rate for the period has been compared to that of The difference in incidence rates was tested using two-sided u-test or z-test (standard normal approximation of the Poisson probability distribution) [8]. No statistical testing was performed if the number of SI was <10. Results In Table 1, the incidence rates from 1990 to 1997 for the total personnel, for nurses and for hotel services personnel are presented. During a total of 4230 SI were recorded in this hospital. Over the years a decreasing incidence rate was seen, especially among nurses and hotel services personnel. For all employees the incidence rate per 100 beds was statistically significantly different between (incidence rate = 33.4) and the period (incidence rate = 30.1; z-test, P < 0.01). From 1990 to 1997, nurses reported 3017 SI or 71.3% of the total SI in this period. The average rate in was 17.2 and was statistically significantly different from 12.7, the average in (z-test, P < ). In Table 1 it is seen that among nurses numbers of SI overall show an absolute decrease between 1990 and 1997 (apart from the years ). From 1990 to 1994, there was a halt on recruitment for budgetary reasons, but from 1995 on approximately 400 new full time equivalent (FTE) nurses were hired. To take this change into account, the rates were preferred for comparison. The incidence rate is indeed calculated to take account of the number of person-years of exposure. Thus, when the FTEs increase, the number of person-years similarly increases. Thus, an increase in absolute numbers of SI can be expected if the risk remains stable because more persons are exposed. Hotel services personnel reported 4.7% of the total number of SI. The incidence rate of 3.7 in was not significantly different from 4.8 per 100 person-years in (P >0.05). The incidence rate for physicians could only be calculated in 1994 [6], 1995 and 1997 because the number of person-years was only available for those years. The number of SI among physicians in was 167, which was 10.5% of the total. The incidence rate of 4.4 per 100 person-years in was not significantly different from 6.1 in 1997 (z-test, P > 0.05).

3 G. MOENS ET AL.: SHARPS INJURIES AND PREVENTATIVE ACTIONS 247 Table 1. and incidence rate per 100 occupied beds per year for all employees and incidence rate per 100 person-years for nursing and hotel personnel in the period Year All employees a Nurses Hotel personnel 100 occupied beds b 100 person-years 100 person-years Total a The number of person-years was not available for the total personnel. b The mean occupation rate of beds was 81.9% and varied between 80.1 and 83.5%. Table 2. and incidence rate per 100 person-years for nurses according to department in the period and in Department 100 person-years 100 person-years Operating room Emergency unit Internal diseases Outpatient Surgery Geriatrics Pediatrics/maternity Psychiatry Oncology Intensive care units Function testing/radiology Mixed departments Other Total In Table 2, the incidence rates among nursing personnel are analysed according to the department, and the departments were sorted from highest to lowest 1997 incidence. Because person-year data were not always available, research periods have been divided differently: versus A decrease was noted in the operating room, internal diseases, outpatient, geriatrics, oncology, the intensive care units and mixed and other departments. However, none of these differences were significant. The same conclusion could be made for the departments in which an increase was noted, with the exception of one significant result in the department of paediatrics/maternity (increase in incidence from 9.4 to 15.7 per 100 person-years). The reason for this significant increase is not obvious. The description of the type of SI is given in Table 3 for and for The strong decrease (almost 50%) of the proportion of unspecified SI is probably due to the organization of training sessions and the brochures distributed following the 1996 survey. In these sessions and brochures the importance of giving a precise description when reporting a SI was stressed. Because of this decrease, the proportion of all other types has therefore increased, except for SI through splinters and instruments. Because relative frequencies depend on the total number of SI, the proportion of a certain type can have increased whereas the incidence rate has decreased. This illustrates that conclusions about risks cannot be based on relative frequencies alone. A further analysis of the relative distribution of the type of SI was made for nurses, physicians and hotel services in 1997 (not shown in Table 3). This distribution is clearly

4 248 OCCUPATIONAL MEDICINE Table 3. Absolute and relative frequencies of the nature of SI in and in among the employees having reported an SI Description dependent on the tasks performed by each group. Among nurses for instance, the SI occurring most frequently were attributed to blood sampling (23.2%), parenteral injections (19.7%) and infusion therapy (14.0%), whereas for physicians these were due to the handling of surgical instruments (34.9%) and for the hotel services to the collection of garbage (69.2%). Discussion Number of SI % Number of SI Not specified Recapping Parenteral injection Infusion therapy Blood sampling Cleaning up Collecting garbage Splinters Instruments Other Total Our study shows a significant decrease in SI following a specific prevention campaign introduced from The incidence rates among nurses decreased significantly and decreases (not statistically significant) were seen in a number of departments. Our findings showed that the nature of the SI was strongly influenced by the actions performed by the occupational group, such as blood sampling, parenteral injections and infusion therapy for nurses. Physicians were most at risk when handling surgical instruments. Among the hotel service personnel, most accidents occurred while collecting garbage. A survey showed that 6% of physicians and nurses still make errors in disposing of sharps [9]. Comparing our results to other studies, the large variation in SI rates is prominent, ranging from 130 cases per 100 person-years in Taiwan [10] to 5.3 in the USA [11]. This could be due to differences in reporting behaviour, in the calculation of rates, but especially because in most studies no information is available about many other determinants such as the type of patient, the type of pathology, the organization of care, the type of devices used, the presence of prevention programmes, etc. However, in a study in 58 US hospitals, a mean % incidence rate of 33 SI per 100 occupied beds was found [12], which is comparable to our results. Our finding that nurses are at greatest risk and that most SI happen in operating rooms, patient rooms and emergency units is confirmed by other studies [13,14]. Since probably not every SI has been recorded, the real SI incidence rate will be higher than the value reported. The survey [9] showed that only 43% of all employees recorded each SI in Even higher figures were reported in a study in Taiwan, where it was shown that 85% of doctors did not register the injury [15]. Major limitations of our study have to be mentioned. As already stated, reporting bias could have played a role because data has been collected based on self-reports. However, reporting bias in this case could be considered to attenuate any real decrease because the higher awareness as a consequence of the preventative actions is expected to increase the reporting of SI. Further, the design was a retrospective before after comparison and no control group was available. Thus, we cannot control for other possible factors which could have influenced the SI decrease (confounding), such as changes in the type of patients and pathologies, in the organization of care, in the characteristics of practices used (e.g. invasiveness), etc. Such confounders are difficult to account for in a retrospective study. Finally, the choice of the before and after period could be criticized. This choice however was made a priori and was defined by the starting year of the large and intensive preventative actions. For a reliable trend analysis, the number of data points was too small and data points were not available for every year. Another interesting research path which could not be investigated retrospectively was the follow-up of the serology among the injured personnel. However, studies following up injured personnel are very scarce and the number of documented infections after injuries appears to be very low [13,16]. Summarizing, we found that the global SI incidence in this university general hospital showed a significant decrease over eight registration years. It seems likely that the intensive prevention programme at least partially influenced this decrease. Advances in technology for the safe use and disposal of sharp instruments may also have played a role, suggesting that a combination of factors is responsible for the decrease in SI. The effect of these factors for all employees could be estimated at a reduction of 67 SI per year and for nurses of approximately 5 SI per 100 person-years. This study also clearly demonstrates the difficulties in using retrospective registration data in assessing preventative interventions. However, sustained efforts to repeatedly evaluate and adjust prevention programs are necessary to enhance effectiveness in the long run. Therefore the evaluative aim should be incorporated in

5 G. MOENS ET AL.: SHARPS INJURIES AND PREVENTATIVE ACTIONS 249 the registration system, with a more complete and standardized collection of relevant data, especially prevention-related characteristics and confounders. Following a registration feasibility study during 2002 in the same hospital, the Belgian National Institute of Public Health Louis Pasteur, will offer such a standardized registration system from 2004 to all Belgian general hospitals [17]. It is expected that these data will constitute a valuable basis for monitoring prevention of SI in Belgian hospitals. Acknowledgements The authors wish to thank the safety service of the University Hospital Leuven, the medical teams which collaborated in the study, as well as the documentation centre and the executive secretariat of IDEWE for the necessary support. They would also like to thank A. De Schryver (IDEWE) and C. Verbeek (IDEWE) for their critical comments and appreciated suggestions. References 1. Shapiro CN. Occupational risk of infection with hepatitis B and hepatitis C virus. Surg Clin North Am 1995;75: Chamberland ME, Ciesielski CA, Howard RJ, et al. Occupational risk of infection with human immunodeficiency virus. Surg Clin North Am 1995;75: Aitken C, Jeffries DJ. Nosocomial spread of viral disease. Clin Microbiol Rev 2001; 14: Porta C, Handelman E. A new approach to needlestick injuries among health care workers. OSHA compliance issues. Appl Occup Environ Hyg 1999;14: Gyawali P, Rice PS, Tilzey AJ. Exposure to blood borne viruses and the hepatitis B vaccination status among healthcare workers in inner London. Occup Environ Med 1998;55: Helsen G. Prikongevallen in een ziekenhuis: analyse met het oog op preventie [Needle stick accidents in a hospital: analysis in view of prevention]. MSc dissertation, Department of Occupational Medicine, Catholic University of Leuven, 1995 [in Dutch]. 7. Richard VS, Kenneth J, Ramaprabha P, Kirupakaran H, Chandy GM. Impact of introduction of sharp containers and of education programmes on the pattern of needle stick injuries in a tertiary care centre in India. J Hosp Infect 2001;47: Kirkwood BR. Essentials of Medical Statistics. London: Blackwell Scientific, Helsen G. Enquête: prikongevallen in het UZ [Survey: Needle Stick Accidents in a University Hospital]. Leuven: IDEWE, 1996 [in Dutch]. 10. Guo YL, Shiao J, Chuang YC, Huang KY. Needle stick and sharps injuries among health care workers in Taiwan. Epidemiol Infect 1999;122: Beekmann SE, Vaughn TE, McCoy KD, et al. Hospital bloodborne pathogens programs: program characteristics and blood and body fluid exposure rates. Infect Control Hosp Epidemiol 2001;22: Jagger J. A.E.P.: a new opportunity to make the health care work place safer. Adv Exposure Prevent 1994;1: Pettit LL, Gee SQ, Begue RE. Epidemiology of sharp object injuries in children s hospital. Pediatr Infect Dis J 1997;16: Heptonstall J, Turnbull S, Henderson D, Morgan D, Harling K, Scott G. Sharps injury: a review of controversial areas in the management of sharps accidents. J Hosp Infect 1999;43(Suppl.):S219 S Shiao JS, McLaws ML, Huang KY, Ko WC, Guo YL. Prevalence of nonreporting behavior of sharps injuries in Taiwanese health care workers. Am J Infect Control 1999;27: Baldo V, Floreani A, DalVecchio L, et al. Occupational risk of blood-borne viruses in health care workers: a 5 year surveillance program. Infect Control Hosp Epidemiol 2002;23: Institute of Public Health (IPH) Louis Pasteur. The Surveillance of Sharps Injuries in Belgian General Hospitals (Registration Protocol). Brussels: IPH, 2003.