Reports From the Field

Sustainability of Ambulatory Safety Event Reporting Improvement After Intervention Discontinuation



From Novant Health and Novant Health Medical Group, Winston-Salem, NC (Dr. Clegg and Mr. West), and Atrium Health, Charlotte, NC (Mr. Anderson).


  • Objective: An educational intervention stressing anonymous, voluntary safety event reporting together with monthly regular audit and feedback led to significantly increased reporting of safety events in a nonacademic, community practice setting during a 15-month intervention period. We assessed whether these increased reporting rates would be sustained during the 30-month period after the intervention was discontinued.
  • Methods: We reviewed all patient safety events reported in our ambulatory clinics for the period 2012–2016, and selected 6 clinics that comprised the intervention collaborative and 18 specialty- and size-matched clinics (1:3 match) that comprised the comparator group. To test the changes in safety event reporting (SER) rates between the intervention and postintervention periods for the intervention collaborative, interrupted time series analysis with a control group was performed.
  • Results: The SER rate peaked in the first month following the start of the intervention. Following discontinuation of regular auditing and feedback, reporting rates declined abruptly and reverted to baseline by 16 months post intervention.
  • Conclusion: It is likely that sustaining enhanced reporting rates requires ongoing audit and feedback to maintain a focus on event reporting.

Keywords: patient safety; safety event reporting; voluntary reporting system; risk management; ambulatory clinic.

We have previously shown that patient safety reporting rates for a 6-practice collaborative group in our non-academic community clinics increased 10-fold after we implemented an improvement initiative consisting of an initial education session followed by provision of monthly audit and written and in-person feedback [1]. The intervention was implemented for 15 months, and after discontinuation of the intervention we have continued to monitor reporting rates. Our objective was to assess whether the increased reporting rates observed in this collaborative during the intervention period would be sustained for 30 months following the intervention.


This study’s methods have been described in detail previously [1]. For this improvement initiative, we reviewed all patient safety events reported in our ambulatory clinics for the period 2012–2016. We identified 6 clinics, the intervention collaborative, in family medicine (n = 3), pediatrics (n = 2), and general surgery (n = 1), and 18 specialty- and size-matched clinics (1:3 match), the comparator group [1]. For the intervention collaborative only, we provided an initial 1-hour educational session on safety events with a listing of all safety event types, along with a 1-page reporting form for voluntary, anonymous submission, with use of the term “safety event” rather than “ error,” to support a nonpunitive culture. After the educational session, we provided monthly audit and written and in-person feedback with peer comparison data by clinic. Monthly audit and feedback continued throughout the intervention and was discontinued postintervention. For event reporting, in our inpatient and outpatient facilities we used VIncident (Verge Solutions, Mt. Pleasant, SC) for the period 2012–2015 and RL6: Risk (RL Solutions, Toronto, ON) for 2016.

The baseline period was 15 months (January 2012–March 2013), the intervention period was 15 months (April 2013–June 2014), and the postintervention period was 30 months (July 2014–December 2016). All 24 clinics were monitored for the 60-month period.

To test the changes in the rate of safety event reporting (SER) between the pre-intervention and postintervention periods and between the intervention and the postintervention periods, interrupted time series (ITS) analysis with a control group was performed using PROC AUTOREG in SAS Enterprise Guide 6.1 (SAS Institute Inc., Cary, NC). Because SER rates are reported monthly, ITS analysis was used to control for autocorrelation, nonstationary variance, seasonality, and trends [2,3].

Next Article: