Emergency Department Utilization Report to Decrease Visits by Pediatric Gastroenterology Patients

Study Design and Setting

The present study was an analysis of data that were prospectively collected and previously disseminated as a quality improvement project to reduce ED utilization by pediatric patients followed in a pediatric gastroenterology practice. The project was conducted at an academic, tertiary care referral hospital in a major urban city. This study was approved by the hospital’s institutional review board (protocol no. 2013P000538). The pediatric gastroenterology outpatient practice included 20 attending physicians, 3 nurse practitioners, 3 nurses, and 10 administrative staff who handled ∼20 000 outpatient visits annually.

Intervention

Starting in May 2013, pediatric gastroenterologists began receiving reports with rates of ED utilization by their patients. Two separate reports were distributed: (1) one report individualized to the physician (Fig 1); and (2) one report with all physician rates for the divisional leadership (Fig 2).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Example of a physician-level report. MRN, medical record number.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Example of a practice-level report.

The reports contained information on only those patients who were seen in the ED and discharged from the hospital, excluding those who required admission to the hospital or to observation. The reports included both all-cause ED visits and GI-related ED visits. For both all-cause and GI-related utilization, rates were calculated for ED visits during: (1) all hours of the day; (2) weekdays; and (3) office hours.

A GI-related ED visit was defined by using the primary diagnosis upon discharge from the ED. Based on the International Classification of Diseases, Ninth Revision (ICD-9) code of the primary diagnosis, a diagnosis categorization tool was applied that clusters patient diagnoses (Clinical Classifications Software for ICD-9, Agency for Healthcare Research and Quality, Rockville, MD). Two GI-related diagnostic groups were included: (1) digestive; and (2) liver and pancreas.

Individual physician reports included the medical record number, patient name, date of last office visit, ED diagnosis, and whether the patient was seen during a weekday (all hours) or during office hours. A visit was deemed as during office hours if the time of presentation was between Monday and Friday, 9:00 am to 5:00 pm.

Starting in May 2013, preliminary reports were given to the leadership of the Division of Pediatric Gastroenterology and the Quality Leadership of Pediatric Gastroenterology to examine and modify. After approval by pediatric gastroenterology leadership, individual physician reports were distributed to their physicians in August 2013. Feedback and comments on the reports were received over the next month from the individual physicians and integrated into future reports. Follow-up meetings were held and reports were presented during the months of March 2014, February 2015, and May 2015.

The ED utilization reports resulted in discussion and a cultural shift; that is, starting September 2013, patients with urgent complaints were seen and treated preferentially in the outpatient office instead of the ED. Patients calling in with issues that appeared to require an urgent visit for evaluation and management were triaged to a same-day office visit instead of being sent to the ED. This approach was in contrast to the previous process in which the ED was generally recommended for these cases. If the patient’s primary gastroenterologist was not available that day, the patient was seen by a physician or nurse practitioner colleague in the division. This change required the coordination and efforts of the entire staff in the outpatient office. The administrative staff ensured that all urgent-appearing calls were returned by the triage nurses or the patient’s primary gastroenterologist. The nurses and administrative staff were empowered to offer the family an office appointment that same day.

Study Population and Data Collection

For each report, all eligible ED visits were identified from hospital databases for a 3-month period immediately before disseminating the reports. Eligible visits concluded with discharge from the ED for a patient who had seen a pediatric gastroenterologist in the office within the previous 12 months.

Specific data elements included patient name, medical record number, name of the pediatric gastroenterologist, date of last visit to the pediatric gastroenterologist in the office, ED visit date and time, ED visit disposition location, primary diagnosis in the ED, and secondary diagnosis in the ED. An ED visit was attributed to a pediatric gastroenterologist if it met the following 3 criteria: (1) patient was seen in the ED during the 3-month period; (2) the patient had at least 1 gastroenterology office visit with the physician within the 12 months before the index ED visit; and (3) the rendering provider was the same as the billing provider. For example, if a patient was seen in the ED on December 10, 2013, for the visit to be attributable to the physician, he or she must have seen the patient at least once in the office between December 10, 2012, and December 10, 2013. If the patient saw multiple gastroenterologists during the 12-month period, then the attribution was to the gastroenterologist who had seen the patient most recently before the index ED visit. Any patients seen primarily by a nurse practitioner were excluded.

Outcome Measures

The primary outcome was the mean rate of ED visits for the division before and after the start of physician-level reporting. The following rates of ED visits were calculated for each physician, as well as for the entire pediatric gastroenterology division: (1) ED visits during all hours of the day for all-diagnoses; (2) ED visits during office hours for all-diagnoses; (3) ED visits during weekdays for all-diagnoses; (4) ED visits during all hours of the day for GI-related diagnoses; (5) ED visits during office hours for GI-related diagnoses; and (6) ED visits during weekdays for GI-related diagnoses.

Calculation of Rates

Rates for each 3-month period were calculated. The numerator was the total number of attributable ED visits for that category, and the denominator was the total number of office visits seen in the gastroenterology practice by the physician over a 12-month period. All rates were defined as per 1000 office visits.

Rate of ED Visits (per 1000) = (Number of Attributable ED Visits in 3 months)/(Number of Office Visits in 12 months) × 1000

Balance Measure

As a balance measure, we examined if the increased work from the process change would affect the number of visits by new patients to the outpatient practice. During the study period, there were no increases in daily staffing of administrative and clinical personnel. As such, the additional work required for the process change would be work beyond the regular routine of the office. In addition, because new patients typically require more time than follow-up patients, we hypothesized that if the intervention overstressed the office clinicians, there would then be a significant decrease in new patient visits.

Statistical Analysis

To analyze the effect of the intervention on rates of ED utilization, control charts with upper and lower control limits set at 3 sigma were used. Specifically, we used u charts because the denominator changed during the study period. For the balance measure, we used a c chart because the data were represented as absolute values. The mean rates were then analyzed by using Poisson regression to test for statistical significance of the rates. A P value <.05 was considered significant. SPSS version 22.0 (IBM SPSS Statistics, IBM Corporation, Armonk, NY) and Microsoft Excel 2016 (Microsoft, Redmond, WA) with the QI Macros Add-in (KnowWare International Inc, Denver, CO) were used for analyses.