Your RCA: How do you select events for analysis?

Stave off sentinel events by evaluating lesser ones

By Patrice Spath, ART
Brown-Spath Associates
Forest Grove, OR

The 1999 performance improvement standards of the Joint Commission require facilities to conduct a root-cause analysis (RCA) for all sentinel events that result in patient death or serious permanent injury. Unexpected clinical occurrences also should be analyzed to determine which of these events represent significant "near-miss" situations. A significant near-miss situation is one that could have — but by chance or through timely intervention did not — result in patient death or injury. Other events that may signal the need for further analysis include:

    • a patient death or loss of function following a discharge "against medical advice"
    • an unsuccessful suicide attempt;
    • minor hemolysis following transfusion;
    • medication errors that result in the need for treatment with another drug or increased resource use;
    • adverse occurrences involving employees or visitors.

These less serious adverse events also may be considered sentinel events because they signal that something important is happening in the process of patient care. By actively looking for the root causes of these undesirable situations, organizations can obtain the information necessary for taking appropriate preventive actions.

The Joint Commission expects facilities to define the situations that require a formal RCA to be conducted. It is fairly easy to describe the unexpected events that result in patient death or permanent injury. The difficulty lies in determining what other types of sentinel or serious occurrences should trigger an RCA.

Any type of incident may foreshadow more serious events. That’s why an RCA is conducted in many high-risk technical industries for any unusual occurrence, regardless of severity or complexity. Health care facilities generally lack sufficient resources to conduct formal investigations of every "semi-serious" incident. Organizations must somehow identify those incidents that are manifestations of larger problems within the system of patient care and that, if corrected, will likely result in improved processes throughout the facility.

The Joint Commission expects organizations to conduct an RCA for all events that, by definition, require such analyses. The standards are silent on how these events are to be defined, so organizations have used a variety of methods. Following are the three most common strategies:

Incidents are categorized according to their severity level.

A cut-off point is established and RCA is conducted for all events that fall into the higher levels of severity. Shown in Figure 1 (see chart, inserted in this issue) is the incident scoring system used by the New York Patient Occurrence Reporting and Tracking System (NYPORTS). Many facilities, not just those in New York State, are adopting this classification system. Although there is some room for individual interpretation, the definitions are more objective than those of most other scoring systems. If this, or a similar, incident classification system is used, the facility’s leaders should define which events require an RCA. Data are gathered about those incidents not requiring analysis and used in the facility’s ongoing performance measurement activities.

• The quality manager or risk manager screens all potentially serious events and gathers additional details. The form shown in Figure 2 (see chart, inserted in this issue) can be used to solicit additional details about the event to be used in determining whether an RCA is warranted. This form is completed by the unit or department manager, or, in the case of incidents involving physicians, the medical staff department chairman. A leadership group usually including the chief executive officer, the nurse executive, the medical staff president or his or her designee, and the quality or risk manager reviews each potentially significant incident to determine if there is a process improvement opportunity. If actions have already been taken that are thought to be likely to prevent recurrence of a similar event, then no further analysis may be necessary. Otherwise, the leadership group may choose to initiate a formal RCA.

• Any incident that results in increased resource use — for example, initial or prolonged hospitalization or need for additional treatment — is reviewed by a leadership group similar to that described above. An incident screening checklist (see chart, p. 98), is used to determine the extent of the analysis to be conducted. All questions answered yes receive a pre-established number of points. The total possible score for an event is 100. The leadership group must decide what score is considered high enough to subject the occurrence to a formal RCA investigation. If you wish to use a screening tool such as this check sheet, the questions and corresponding weights can be adjusted according to your own organizational priorities.

Patient safety improvement cannot be achieved if organizations conduct RCAs only when harmful sentinel events occur. In many facilities, such events happen very rarely. That’s why the Joint Commission requires that "significant" incidents also be subjected to formal investigation. By evaluating less harmful events, caregivers can identify where additional safeguards are needed to further reduce the likelihood of a fatal sentinel event.

(Editor’s note: See the July issue of Hospital Peer Review for a blow-by-blow description of how to conduct a root-cause analysis. Next month’s Quality-Co$t Connection: How to involve the governing board in your organization’s commitment to patient safety improvement.)