Novel Use of the Epic Electronic Medical Record Platform to Identify Lost Ureteral Stents

Introduction

Ureteral stents are commonly utilized in urology for drainage of the upper urinary tract and stent insertion is one of the most common urologic procedures. ¹ Ureteral stents are deployed in the management of many urologic conditions including nephrolithiasis, ureteral strictures, ureteral reconstruction, and malignant ureteral obstruction. ²

Unfortunately, ureteral stents can be forgotten, possibly because of lack of patient education, poor patient compliance, or other factors such as lack of health insurance. ³ Prolonged dwell time of ureteral stents carries the risk of serious complications such as obstruction leading to renal failure, infection, ⁴ and even death. ⁵ Also, forgotten stents may become encrusted and fragmented, necessitating additional and often more invasive surgeries for their removal. ^{6 –9} Timely follow-up is essential to minimize and prevent morbidity and mortality associated with retained ureteral stents.

A number of systems have been developed to track stent placement and confirm timely removal. Card-based registries and logbooks that require manual entry and review ^10,11 have been replaced by a number of electronic systems, including computer-based systems located within the patient medical record, ¹² and now more recently, proprietary cloud-based applications that are accessible on smart phones and the web. ^13,14 Although these efforts are intended to improve patient safety, they are still dependent on manual entry of data raising the issue of adequate capture, and require the integration of new applications rather than utilizing health care systems already in place.

Epic is an electronic medical record (EMR) vendor with a commanding market share among EMR systems in the United States. In addition to patient laboratories, imaging, and chart documentation, Epic allows for the point-of-care registration of intraoperative implants including ureteral stents. Within our institution, these data are collected prospectively as part of an established workflow to help with billing, inventory, and quality control. The registry may be queried to identify ureteral stents that were placed in patients and are at risk for loss to follow-up.

In this study, we describe our experience with novel use of Epic to prospectively detect ureteral stents with dwell times >90 days to identify patients who are at risk for loss to follow-up.

Methods

Yale-New Haven Health System first went live with the Epic EMR in 2013. The information collected in the EMR is extracted to a data warehouse that can later be queried.

Our hospital protocol is that all surgical implants, including ureteral stents, are logged in the EMR at the time of placement. While in the operating room, the operating room nurse scans the packaging of the ureteral stent and this automatically populates the patient's chart with the ureteral stent implant. The log accounts for the product name, lot number, and expiration date through a bar code, in addition to the anatomic location of implantation (including laterality), date of service, and implanting surgeon. Subsequent implant extraction (stent removal) can also be done using the EMR through this same system.

As part of a quality initiative to decrease lost ureteral stents, our department initiated a program to use this available information to identify stents that are past their intended dwell time. Starting in January 2015, a dedicated urology nurse requested and reviewed monthly reports from Epic that identified patients with ureteral stents placed >90 days prior without documentation of removal. EMR was reviewed for documented stent removal or persistence (e.g., within subsequent office visits notes, operative notes, pathology specimens, and abdominal imaging reports). If the stent was believed to be beyond intended removal date, the nurse would alert the surgeon of record, and if a retained stent was suspected, the nurse would then contact the patient to arrange for appropriate follow-up for definitive management.

The use of the Epic system by the Yale-New Haven Health System contributes to the development of a large EMR warehouse, Helix. Helix contains more than four million patient records that can be applied to research. The Joint Data Analytics Team (JDAT) handles clinical and research analytics and reports across the health system and the School of Medicine. Comprising >60 analysts, JDAT centralizes and coordinates all data analytics and supports Helix. Through the use of Helix and the JDAT team, we are able to query the EHR for patient data.

In this study, we used Helix and the JDAT team to query Epic to identify risk factors for retained stents as a secondary objective. We retrospectively identified all patients with ureteral stents placed in our hospital system from January 2015 to August 2017. Patients were classified according to age, gender, race, surgical urgency (elective vs emergent), native speaking language, comorbidities, length of stay, and home distance (by ZIP code) to the hospital. We also examined length of time stent was retained.

ICD10-CM diagnosis codes were used to identify medical diagnoses in inpatient and ambulatory surgery encounters. CPT codes were used to identify inpatient and ambulatory surgery encounters during which a stent was placed.

Clinical characteristics for patients with timely stent removal were compared with those lost to follow-up using a t-test or chi-squared distribution test. Multivariate logistic regression was used to predict risk of lost to follow-up. All statistical analyses were performed using STATA 9.0 (College Station, TX).

Results

We identified 1788 patients who underwent ureteral stent placement over the study period. Sixteen patients (0.9%) with stents in place >90 days were deemed lost to follow-up (Table 1).

Of the patients lost to follow-up, nine patients (56%) were contacted and followed up for stent removal. Four (25%) were unreachable by telephone and did not respond to certified mail correspondence. Absence of a stent was confirmed subsequently in one of these patients on abdominal imaging performed for nonurologic indications, suggesting stent removal at an outside hospital. Three patients (19%) refused definitive intervention (and stent removal) because of advanced comorbidities and poor short-term prognosis.

There was no difference in patient age, gender, or home proximity to the hospital for risk of failure to follow-up. Patients who subsequently failed to follow-up were more likely to have had stents placed in an emergent (obstructing stone with a fever, ureteral obstruction with acute kidney injury, etc.) rather than a scheduled elective setting (63% vs 30%, p = 0.004).

Using multivariate logistic regression (Table 2), stents placed in an emergent setting (odds ratio [OR] 3.5, p = 0.018) and black race (OR 4.03, p = 0.018) were independent predictors of failure to follow-up. The median dwell time for explanted stents was 15 days vs 165 days among those patients lost to follow-up.

Discussion

Ureteral stents are used commonly in urology. The most pressing acute adverse event associated with an indwelling stent is renal colic and pain. When stents are lost to follow-up, patients run the risk of significant encrustation, recurrent infection, ureter–arterial fistula with hemorrhage, and obstruction with compromised renal function. Removal may necessitate repeat procedures over multiple hospital visits, entailing a significant financial burden. The financial cost of these procedures exceeds the cost of timely stent removal. ¹⁵

There is also medicolegal accountability for the removal of ureteral stents. In one study, authors looked a negligence claims against urologists from 1995 to 2009 from the National Health Service Litigation Authority. They found that ureteral stent negligence claims accounted for 23 claims over this period, which was the largest number of negligence claims filed. ¹⁶

Lynch and colleagues looked at stent tracking systems based on patient registries. The authors implemented an electronic stent register (ESR) and a stent extraction reminder facility (SERF) that operates within the hospital's EMR. Stents are registered and a maximum stent life is determined, SERF automatically interrogates the ESR and identifies any stent that is overdue. Clinicians then received emails to ensure stent removal of their patients. This study also looked into bar-code technology to improve outcomes in these patients. ¹² The problem with this system is that it can burden the provider with emails and there is little oversight of the system itself. Secondarily it is specific to one EMR.

As already mentioned, there are proprietary smartphone applications that track ureteral stents made by Boston Scientific. ¹³ Although the smartphone platform is easy to use, it does require an additional step of inputting the patient and the stent into an application raising the limitation of compliance. Also, the application only tracks its own brand (Boston Scientific) of stent limiting the generalizability if multiple manufacturers or non-Boston Scientific stents are used in a hospital system. In addition, some patients and providers may not want to disclose patient information or practice patterns to a commercial entity.

Another study looked at wristbands as a way to mark patients as having an internal ureteral stent. ¹⁷ The authors point toward the visible reminder as a means of reminding patients to have their stents removed. This method as expected is cumbersome to patients and wristbands can easily be removed and rely on the patients to be complaint with maintaining their wristband.

Medical institutions track the implantation of foreign bodies through the use of an EMR. This information is readily available to clinicians and can be easily extracted to improve stent retrieval without the use of a new system. This concept, therefore, has applicability to any hospital system.

Despite our use of an EMR to track stents, several patients in our series were lost to follow-up. In 2013, Divakaruni and coworkers ¹⁸ studied characteristics of stent patients who were lost to follow-up and found that men and those without insurance were more likely to be lost to follow-up. Access to health care is invariable dependent on insurance or other payer status. Our records did have some payer data available; however, insurance status could not be reliably confirmed at that time of stent placement or during the follow-up period to allow a reliable assessment of this variable as a risk factor to prolonged stent dwell times. It bears mentioning that all 16 patients who required recall for stent removal did have insurance and so it is unlikely that insurance status was a significant factor in our experience.

In our series, patients who were stented on an emergent basis were more likely to be lost to follow-up. Emergency stents can be placed outside of regular hours and in patients with significant active medical problems such as infection or renal failure. Nearly two-third of patients who were lost to follow-up were stented in an emergent setting. The etiology of the additional risk observed in the emergent setting is unclear. It is possible that emergency patients have different comorbidity, resources, or access to care. It is possible that emergency patients received different care or counseling than scheduled patients, such as inadequate instructions regarding poststent follow-up. Three of our “lost” patients were found but refused follow-up and definitive management to allow for stent removal because of significant comorbidities, opting for a palliative indication for stent. The change to palliation raises the possibility that patient goals of care may not have been adequately considered in the emergent setting, but this is not possible to determine retrospectively. Owing to the results of this study, we changed the way that follow-up is scheduled for patients who are stented in an emergent setting. In particular, we ensure that all patients who have a ureteral stent placed are contacted within a 2-day period of their stent placement to arrange for their follow-up appointment.

Although on average patients were older in the retained stent category, this result was not significantly different (p = 0.21). This result is similar to the finding of Divakaruni et al. The authors of this study did find that men were more likely to be lost to follow-up; our study, however, did not find a significant difference between men and women.

Our study also evaluated race and its association with being lost to follow-up. We found that black race was independently associated with being lost follow-up. The etiology of this association is unclear.

Lastly, it is the combination of both the EMR and having a dedicated urology nurse to review outliers that makes this system effective in identifying lost stents. Although stent implantation and explantation are an automated process, the lost stents or stents that were not explanted correctly were identified by a nurse.

This study has limitations worth mentioning. We could not study patient knowledge or understanding of their ureteral stents. Patient education remains an important concept in preventing ureteral stents from being retained, ¹⁹ and pre and postoperative counseling are integral in patient understanding. A standardization for patient discharge instructions may also be helpful for patients who have stents implanted.

Secondarily, we identified black race as an independent risk factor for failure to follow-up. The etiology of this association is unclear. Although we attempted to control for measurable potential risk factors such as insurance status and proximity to the hospital, there are likely other confounders. Haider and coworkers performed a comprehensive review of racial disparities and health outcomes and found that many of these studies like ours are retrospective in nature. The data for this analysis were collected for the purpose of billing rather than studying disparities and, therefore, may miss key data on patient race. ²⁰ There are patient, provider, and systemic characteristics that contribute to racial differences in outcomes, and if not selected for they cannot be studied.

Conclusions

Lost ureteral stents are a rare but impactful urologic issue. In this study, we demonstrate that data in the Epic EMR can be queried to find patients with lost ureteral stents. We also identified that patients stented emergently are at an increased risk for loss to follow-up—extra care should be taken in these situations to educate patients regarding their stent and the need for appropriate follow-up. There is need for further study, including an evaluation of the association between race and recurrence, an unsuspected finding in this small series.