Prescribing antibiotics for perioperative prophylaxis in common surgical procedures presents an ideal target for educational intervention. In this situation, antibiotics are often used inappropriately, with consequent excess expense and risk of morbidity. We developed an educational intervention aimed at the choice and appropriate dosing of antibiotics for the prophylaxis of cesarean sections. Person-to-person educational messages targeted at authoritative senior department members were supplemented by brief reminders on a structured antibiotic order form. Time-series analyses were conducted on 34 months of antibiotic use data for 2,783 cesarean sections to estimate the trend of magnitude and significance of discontinuities associated with the start of the program. Prior to the intervention, 95% of sections receiving prophylaxis were given cefoxitin and 3% were given cefazolin. After the intervention, these proportions were reversed, with the shift in use occurring immediately after the intervention (p<.001). Two years after the intervention, virtually all patients undergoing cesarean sections who receive antibiotic prophylaxis are given cefazolin. Savings from this change alone accounted for over $26,000 each year, or $47.36 per patient-day of prophylaxis. Substantial changes in prescribing practices for routine procedures can be accomplished through the implementation of a coordinated educational program that enlists influential senior staff members in a department in which policy-making is highly centralized, coupled with a structured educational ordering system. Lasting improvements in clinical practices may be brought about by means that are noncoercive, inexpensive and well-accepted by medical staff.

There is a significant rate of percutaneous injury with needles during the care of patients with acquired immunodeficiency syndrome (AIDS). Following puncture injury, it is recommended that the source of the contaminating blood be checked, and if human immunodeficiency virus-type 1- (HIV-1)-seropositive, zidovudine prophylaxis be considered. As the source of contaminating blood may be unknown, we studied the detectability of HIV-1 antibody and circulating antigen (p24) in the residual blood from needles and pieces of glass at various intervals following exposure to blood. The residual volume of blood remaining in needles varied from 183 ±50 μ 1 for a 20 G needle to 7.8 ± 1 μ 1 for a 27 G needle, and the residual blood on small pieces of glass varied from 23 μ 1 for a piece weighing 558 mg to 2 μ 1 for a piece weighing 21 mg. Analysis of washed samples of residual blood from all 20 G through 26 G needles and from broken pieces of glass larger than 0.41 g that had been exposed to HIV-1-seropositive blood and left at room temperature for one hour, one day and one week resulted in positive tests for HIV-1 antibody by enzyme-linked immunosorbent assay (ELISA), immunofluorescence and Western blot assays. The circulating antigen was detected in residual blood of 20 G through 26 G needles, but not from contaminated pieces of glass. This technique could be applied to situations where a healthcare worker pricked him- or herself with a needle or with a piece of glass that had been contaminated with blood of unknown seroreactivity. If HIV-1 ELISA, immunofluorescence, Western blot and circulating antigen assays are negative, the individual can be reassured. Because only 0.4% of needlestick injuries with HIV-1-seropositive blood have resulted in seroconversion, there must be other factors, as yet unknown, that predispose to infection.

We employed a questionnaire to survey 334 registered nurses regarding their knowledge, perceptions and vaccine acceptance for Hepatitis B virus (HBV) infections. The study population was those persons working in areas considered at high-risk for contracting HBV. The nurses were asked to provide information regarding blood and body fluid exposures and the reporting of these exposures. The questionnaire was completed by 169 nurses (50.6%). Less than half of the respondents (42%) had been vaccinated. We found that 13 of 14 black nurses, compared to 85 of 154 nonblack nurses, had not been vaccinated (p < .01). No other demographic differences between recipients and nonrecipients were noted. Partially or totally incorrect answers regarding transmittal knowledge were given by 108 of 160 (68%) respondents. Reasons for failure to be vaccinated were varied. Fear of side effects of the vaccine, contracting AIDS or hepatitis from the vaccine, or doubt of efficacy of the vaccine were cited by 50 of 88 (58%) respondents as reasons for not being vaccinated. Inability to schedule an appointment or unawareness of the vaccine's availability were expressed by 23% and 17% of the individuals, respectively. Our data allow us to conclude that to improve vaccination compliance, the vaccine must be made more readily available, possibly through workplace on-site administration. Additionally, a concerted effort to educate our staff regarding HBV with particular emphasis on minorities is needed.

A surgeon-specific computer-generated monthly questionnaire was used to improve surveillance of surgical wound infections in outpatients as well as inpatients following discharge. From July 1988 through June 1989, 20,536 surgical procedures were performed at our medical center, of which 53% were for outpatients. The total wound infection rate was 0.63%: 0.13% in outpatients and 1.2% in inpatients (p<.005). Of the infected wounds, 20% were reported by the survey alone and would have gone undetected by conventional surveillance methods (71.4% of outpatient and 13.8% of inpatient wound infections). As a whole, clean and clean-contaminated wounds in outpatients were much less likely to become infected than those in inpatients. Wound cultures were not obtained in 85% of infections reported by the survey alone, and were less likely to be obtained in outpatients. The average time spent by the infection control department on the survey was approximately two hours per week.

Voluntary screening for the presence of human immunodeficiency virus (HIV) is recommended by the healthcare profession. The optimal settings to accomplish screening have not been established. We evaluated an admission HIV screening program in a large private hospital to assess advantages and disadvantages in this setting. In a three-month study period, 4,535 of 8,868 patients (51%) admitted to the hospital agreed to HIV testing. Serum specimens from 500 patients who refused testing were blindly, anonymously tested. The seroprevalence of the patients agreeing to (0.26%) and refusing (0.60%) testing was not statistically different (p = .12). There were 12 HIV cases discovered; ten (83%) of these were known to be in a high-risk group at the time of admission. Eighty-five percent of patients interviewed were in favor of this screening program. Difficulties associated with confidentiality or consent were not evident. Calculated charges of testing for each HIV case discovery was $14,550. There was no evidence that this screening program provided for a more effective infection control policy to prevent nosocomial HIV transmission. A hospital admission HIV screening program can be implemented can meet with favorable patient opinion and can detect previously unknown HIV-positive patients. Hospitals are an efficient and practical setting for HIV testing. The benefit of this program appears to be greater for the patient than hospital or healthcare worker. Cost-benefit analyses will identify optimal candidates to be screened in different hospital populations.

We developed an educational program that reported the rate of needle recapping to healthcare workers, in conjunction with emphasis on appropriate disposal procedures. Over 12 months, the rate of recapping needles used for venipuncture and for percutaneous medication injections fell from 61% to 16% (p<.0001). Over the same period, the recapping of needles used primarily for intravenous (IV) administration fell from 44% to 33% (p = .03). Re-evaluation of the rate of recapping eight months later showed a continuation of these lowered rates. Needlestick injuries were too few in number during the study period to detect any change accompanying the decreased recapping rate. We conclude that programs that report back to employees their rate of recapping can significantly reduce this activity in the disposal of needles used for venipuncture and for percutaneous medication injections. While such reporting may reduce the rate of recapping of needles used for IV administration, the effect is not nearly so marked. Modifications in design remain the most promising approach to preventing needlestick injuries from recapping needles used for IV administration.