Data mining technology; hair removal and surgical site infection; cytobrush versus curette sampling; pain wraps - Evidence For Practice

AORN Journal,  Sept, 2003  by George Allen

• E-mail
• Print
• Link

Use of data mining tools in medication error analysis

Topics in health information management November 2002

Data mining and machine learning are artificial intelligence technologies that analyze data from different perspectives and summarize it into useful information. The objective of this study was to determine if data mining tools are useful in identifying underlying factors related to medication errors. (1) Incident reports (n = 2,700) from a large urban hospital during a five-year period were analyzed. Errors were classified in relation to the person responsible (ie, nurse, physician, pharmacist) and location where the error occurred (ie, general care, surgery, intensive care, emergency department). Outcome variables assessed were near misses and adverse drug events (ADE) associated with medication errors. Data mining and machine learning technologies were used to analyze these variables.

Findings. Data showed that 80.7% of the medication errors reached patients, and 19.3% were classified as near misses (ie, the error was caught before the medication was given to the patient). Data mining analysis accurately classified 72% of the medication error cases and revealed that the three most important factors in predicting an ADE were type of medication, proximal cause (ie, why), and place where the error occurred. The person responsible, proximal cause, and place were the factors associated with near misses. Additionally, analysis revealed that if procedure was not followed, errors reached patients 93% of the time if a nurse was the responsible person, 68% of the time if a pharmacist was the responsible person, and 75% of the time if a physician was the responsible person. Conversely, if a nurse was responsible and the medication error was the result of miscommunication, an ADE was likely to result 20% of the tune, compared to 55% of the time if a pharmacist was responsible for the medication error.

Clinical implications. This study revealed that data mining tools can identify patterns and relationships that are not easily identified using conventional data analysis. In searching for ways to improve patient safety, perioperative managers must consider the use of techniques such as data mining to identify hidden patterns and underlying factors related to medication errors and other patient safety issues so corrective actions can be developed and implemented effectively.

Hair removal and surgical site infection rates

Otolaryngology--head and neck surgery January 2003

Removal of hair from a surgical site is an identified risk factor for surgical site infection (SSI), and the risk increases when hair is removed by shaving. Despite this fact, hair removal by shaving is a common practice before surgery on the skull. This retrospective study was undertaken to evaluate SSI rates after skull base resections for which hair was not removed. (2) The SSI rate in 175 patients who underwent skull base resections without hair removal was assessed retrospectively and compared to published SSI rates in patients who had undergone similar procedures for which hair was removed.

Findings. The overall SSI rate in the group that did not have hair removed was 1.1% (ie, two SSIs in 175 procedures). This infection rate was equivalent to published SSI rates for similar procedures for which hair was removed.

Clinical implications. The finding of no statistically significant difference in SSI rates between procedures for which hair was removed and procedures for which hair was not removed before skull base surgery is clinically relevant. These results strongly suggest that skull base surgical procedures without hair removal do not carry additional risk for SSI, and they potentially could have a positive impact on the psychosocial well-being of the patient. Perioperative nurses can urge surgeons and other clinicians to reevaluate commonly practiced procedures (ie, sacred cows) including removal of hair before surgical procedures on the skull.

Sleeved cytobrush versus endocervical curette sampling

Obstetrics & gynecology March 2003

Although false negative and false positive results plague the endocervical curette test, it continues to be used by many health care providers in the evaluation of women with abnormal cervical cytology. This randomized study was undertaken to compare endocervical specimens obtained using an endocervical curette to those obtained using a sleeved cytobrush. (3) Both endocervical curette and cytobrush sampling were performed on 62 patients undergoing cervical conization or hysterectomy.

The sampling order was randomized. In one group, cytobrush sampling was performed first, and then an endocervical curette sampling was performed before surgery. In the other group, an endocervical curette sampling was performed first, and then a cytobrush sampling was performed before surgery. Specimens were assessed for adequacy for diagnosis, the presence or absence of neoplasm, and the degree of neoplasm if present. Descriptive statistics were used to analyze differences between the two groups.