Incorrect Surgical Procedures Within and Outside of the Operating Room:  A Follow-up Report FREE

Decreasing the incidence of wrong-site surgery and invasive procedures is a challenge both in and outside of the operating room (OR).¹ Wrong-site surgery is the reviewable sentinel event most frequently reported to the Joint Commission² and has been estimated to occur at a rate of 0.09 to 4.5 per 10 000 cases.³ This 50-fold range in the estimated incidence reflects the variety of results reported in various settings and different methods for defining adverse events in this area. The rate of reported in-OR adverse events per year for the Veterans Health Administration (VHA) system is 0.4 adverse event for 10 000 cases. We recently reported on incorrect surgical procedures in VHA hospitals from January 2001 up to mid-2006 and recommended that the promotion of enhanced communication would decrease the incidence of surgical adverse events.¹ This study provides a follow-up report and an update on recent systemwide interventions.

We analyzed surgical adverse events and close calls reported in the VHA system to assess the current state of surgical safety in the VHA system. We searched the VA National Center for Patient Safety database for surgical adverse events and close calls. Only events pertaining to incorrect surgical procedures or incorrect invasive procedures inside and outside the OR were included.

All reported surgical adverse events in the VHA system are reviewed at the local facility. As part of this process, the patient safety managers in each facility rate the severity and probability of harm of the event using a Safety Assessment Code (SAC) Matrix.⁴ The SAC yields a rating from 1 to 3 for both actual events and close calls (in which significant harm could have been caused but was averted). Actual “SAC 3”–rated events have the highest severity of harm and all SAC 3 events, whether actual or potential, require a root cause analysis, while other lower-risk events can be reported and documented in a briefer safety report. Root cause analysis reports use a team to examine what happened, why it happened, and what should be done to prevent it from happening again. The data we considered were limited to information derived from root cause analysis and safety reports (similar to an incident report) submitted to the VA National Center for Patient Safety and were not abstracted from medical records.⁵

To improve communication and patient safety in the OR, the VHA developed the Medical Team Training (MTT) program and began pilot testing it in 2003. The program was implemented nationally between 2006 and 2009, targeting OR personnel at VHA medical centers with a surgical program.^6- 9 Medical Team Training requires 2 months of planning with a core change team, a 1-day face-to-face learning session where surgery is put on hold (and attendance by the OR team is mandated), and 12 months of follow-up and coaching. A key component of MTT is the emphasis of conducting preoperative briefings and postoperative debriefings, both guided by a checklist.⁹ One of the goals of the MTT program was to reduce the in-OR incorrect surgeries in VHA medical centers. The VHA continued to implement its directive for ensuring correct surgery both in and outside of the OR and continued with promoting safe surgery through its National Surgery program.

We included surgical events occurring between July 1, 2006, and December 31, 2009. Events occurring before July 1, 2006, were reported in a previous article.¹ Events were not limited to the OR and included incidents in procedure rooms, radiology suites, and the bedside.

Two of us (B.T.C. and N.E.) independently coded the reports into major categories such as type of event and body segment. The specific definitions for the coding have been previously described.¹ The first 20 cases were coded independently and then consensus was reached on items of disagreement to refine the codebook. The next 10 cases were coded independently, and coefficient κ was calculated (κ = 0.73). Thereafter, the cases were coded independently, results were compared, and consensus was reached on all items where coders had differences.

Two of us (P.D.M. and J.N.) coded the root causes into major categories such as education and communication. The first 35 cases were coded by consensus and in the process the root cause codebook was revised. Revisions in the codebook were addressed with previously coded cases as necessary. Eleven cases were then independently coded, the codebook was further revised, and consensus was reached on items that differed. Finally, 10 more cases were independently coded and coefficient κ was calculated (κ = 0.79). Thereafter, cases were coded independently. If additional root causes or contributing factors were identified that were not identified by the root cause team, these cases were discussed and consensus was reached on the coding.

A descriptive analysis was conducted. Results were compared with the previously published report.¹ We used the χ² test to examine the association between harm and type of event. To examine the decrease in the rate of actual SAC 3 events over time in VHA medical centers, we performed a trend analysis by modeling the count data of the number of SAC 3 surgical events using the Poisson distribution (the link function was the logarithm and surgical volume was the offset). The independent variable was calendar year (modeled as a continuous variable). Rate ratios and accompanying 95% confidence intervals were calculated to represent the strength of association between passage of time (year) and declining incorrect surgery rates, estimated using the Poisson regression. Using the Poisson distribution, we also studied the differences in incidence rate to investigate whether changes were statistically significant.

In general, a simple Poisson model has the general form: Log(Rate) = Intercept + Slope × Independent Variable.

In our trend analysis, Rate = Number of SAC 3 Events/Number of Total Surgeries in the OR.⁴ When comparing current and previous reports, Rate = Number of Reports/Number of Months.

All analyses were performed using SAS statistical software version 9.2 (PROC GENMOD; SAS Institute Inc, Cary, North Carolina). All reported P values are 2-sided, at a significance level of .05.

Our review produced 237 reported cases (Table 1). Of those cases, 101 were actual adverse events and 136 were “close calls,” 150 cases were in the OR, and 58 were out of the OR and for 29 cases, we were unable to determine if the case was in or out of the OR. The number and percentage of reported adverse events and close calls were similar to our previous report (Table 2) but the rate of actual adverse events per month and the severity of those events significantly diminished for in-OR events (Figure). The rate of reported adverse events decreased significantly from 3.21 reports per month to 2.4 reported adverse events per month (P = .02). At the same time, reported close calls increased from 1.97 reported close calls per month to 3.24 per month (P ≤ .001). For the 101 reports of adverse events, Table 3 displays the types of events sorted by service where the event occurred. Over the 42 months of data, Neurosurgery had 1.56 and Ophthalmology had 1.06 in-OR reported adverse events per 10 000 cases. No other specialty had at least 1 in-OR reported adverse event per 10 000 cases and the overall VHA in-OR rate for adverse events was 0.4 per 10 000 procedures.

Table 4 displays type of event by the location of the surgery (OR vs non-OR). The eTable displays the root causes associated with the 101 actual adverse events. (There were a total of 204 root causes reported for the 101 reported adverse events because events could have multiple root causes.) The most common root cause for adverse events was a lack of standardization of clinical processes (18%).

We assessed the association between types of events (such as wrong procedure and wrong patient) and clinical specialty with degree of the actual SAC score. The χ² analysis indicated that the event type most associated with a high SAC (indicating high harm) was “wrong procedure” (χ² = 5.30; P = .02). No other event types or specific service specialties were associated with higher SAC scores.

Trend analysis estimated that the rate of actual adverse events coded as SAC 3 events (highest harm) dropped by 14% (rate ratio, 0.86; 95% confidence interval, 0.75-0.97; P = .02) each year, roughly translating into a rate of drop of 0.17 event per 100 000 surgeries per year.

In this follow-up to our previous publication regarding incorrect surgical procedures within and outside the VHA OR,¹ we now report a decrease over time in reported adverse events where the actual severity and probability of harm was the highest (SAC 3 events).¹ This decrease spans from 2001 to 2009. Reported in-OR adverse events for the most recent year (2009) in this report included no deaths and only 1 case included the incorrect loss of a body part (a tooth). There are many possible reasons for this decrease. There has been an increased focus on OR safety in VHA medical centers. The implementation of the MTT program and improved OR team communication may have also contributed positively. The VHA MTT program has increased preoperative briefings and postoperative debriefings for surgical procedures, improved perceptions of teamwork and patient safety, and avoided undesirable events in the OR.^6- 12

Despite the overall decrease in patient harm, opportunities exist to further decrease the number of incorrect surgical and invasive procedures. Similar to our previous report, the top 3 specialties reporting adverse events in and out of the OR were Ophthalmology, Invasive Radiology, and Orthopedics. However, when in-OR–reported adverse events were analyzed as a rate, we found that Neurosurgery had the highest rate of reported adverse events, followed by Ophthalmology. The Neurosurgery adverse events in this report were spine cases. As of May 17, 2010, the VHA updated the Ensuring Correct Surgery and Invasive Procedure directive, adding greater specificity such as the attending surgeon must personally confirm the position of the spine marker, there must be read-back on implants, and the operative site must be marked prior to anesthesia providers performing regional nerve blocks.¹³

Ophthalmology continued to have challenges with wrong implants. In some situations, teams had the correct implant in the room before the procedure but also had several other lenses in the room that contributed to the risk for error. In other situations, they pulled the lens based on the incorrect patient's data so even though they verified what they thought was the correct lens during the time-out, the implant was pulled based on incorrect data. Orthopedics also had adverse events involving wrong implants and wrong side. However, the percentage of Orthopedic adverse events related to implants has gone from 46% (12 of 26) in the previous article to 23% (3 of 13) in the current article. Ophthalmology adverse events related to implants were still a challenge: 49% (22 of 45) in the previous report to 59% in the current report (13 of 22).

The current results also mirror the previous article in that 51% of adverse events (108) were in the OR and 49% (104) were non-OR.¹ Non-OR adverse events continue to be problematic with wrong patient/invasive radiology cases.

Overall, the most common root cause for incorrect surgery was “critical clinical processes not standardized.” These were situations in which a clinical process was left to the judgment of the clinician to accomplish. “Human factors problems,” the second most common root cause, were caused by problems with the human-machine interface, look-alike packaging of different implant components, and other problems with the environment or time pressures, distraction, or fatigue.

We are implementing several approaches to work with clinicians and leadership to prevent incorrect surgical events: we share deidentified lessons learned from the root cause analysis reports with the regional surgical leaders in collaboration with VHA surgical leadership. Preliminary and ongoing results associated with this study and the previous study¹ were also presented at annual meetings of the leaders of the VA National Surgical Quality Improvement Program (N.E.). Current plans and actions include continuing to share detailed lessons learned from root cause analyses, rapid notification of adverse events, policy changes as needed based on root cause analysis review, and additional focused MTT for sites as needed.

This study has limitations. First, these data are based on self-report and therefore we cannot be sure that all incorrect surgical adverse events were included. Some adverse events would also be addressed by an administrative board of investigation if there was a concern about intentionally unsafe acts such as blatant disregard for policies. Such events would most likely only have a brief safety report in this database, instead of a root cause analysis. We also do not have demographic data about patients since the root cause analysis and safety reports are deidentified. In addition, there were times the reports did not contain enough information to categorize the cases fully.

Despite these limitations, this report reveals an overall decrease in the number and severity of wrong-site surgery procedures in VHA medical centers. We must continue to improve. We believe that sharing lessons learned; implementing team training in non-OR settings (for example, those in interventional radiology and cardiac catheterization laboratories), which is now being done in the VHA medical centers; ensuring the use of standardized procedures (especially in the out-of-OR setting) and the use of a now-mandatory preoperative checklist with preoperative briefings; and ensuring that clinicians are following policies and procedures are appropriate steps based on this study and others that have been published on this topic.