Use of an Electronic Medical Record to Assess Patient-Reported Morbidity Following Ureteroscopy

Introduction

Ureteroscopy (URS) has emerged as the most frequently performed procedure for stones in the United States, surpassing shock wave lithotripsy (SWL). ^1,2 Like SWL, URS is performed on an outpatient basis and is widely utilized by most urologists. Particularly in ambulatory surgery centers, but also in hospital-based settings, discharge following URS occurs within hours of completing the procedure, and therefore, limited time is available to provide patients with detailed instructions regarding their expected postoperative course.

Patient-centered outcomes research is playing a growing role in decision-making that takes into consideration patient-related factors and preferences. ³ While established outcomes such as complication rates and need for unplanned emergency department (ED) visits associated with URS have been quantified by analyses of a variety of datasets, ^4,5 other indicators of patient concern or lack of well-being reflected in patient-initiated contact with healthcare providers has remained largely unexplored. With the widespread use of electronic health records (EHR), patient-initiated encounters in the form of ED visits, hospital admissions, telephone calls, and secure online communication through a portal in the electronic medical record (EMR) are readily accessible. Since initiating our EMR (Epic; Epic Systems Corporation, Verona, WI), we observed that patients undergoing URS generated a seemingly unusually high number of patient-initiated encounters, considering the “routine” nature of this procedure. We sought to evaluate and quantify patient-unplanned medical encounters following URS to identify common patient-reported postoperative problems that might be avoided with more robust patient education or changes in perioperative care. We also sought to determine if URS differed from another outpatient endoscopic procedure with regard to patient-initiated encounters.

Methods

Following IRB approval, we reviewed the records of 298 consecutive patients who underwent 314 URS procedures for renal and ureteral calculi between July 2013 and November 2014. All patients underwent extensive pre- and postoperative counseling by one of two surgeons (J.A.A. and M.S.P.) and received a printed brochure developed by us with a detailed description of the procedure and the expected postoperative course, as well as instructions detailing signs and symptoms that should prompt contact with a healthcare provider. All patients were discharged with the same postoperative medication regimen that included a narcotic analgesic, alpha blocker, anticholinergic, bladder analgesic (phenazopyridine), and stool softener. Postoperative antibiotics were prescribed for ≤24 hours in accordance with the American Urological Association Best Practice Policy Statement on Urologic Surgery Antimicrobial Prophylaxis unless special circumstances dictated otherwise. All patients received a follow-up phone call on postoperative day 1 from the day surgery nursing staff questioning them about their general well-being following surgery. If patients expressed concerns or problems, they were encouraged to contact their physician, and the nursing staff sent a message to the physician if they deemed the problem significant.

We extracted patient demographics, stone characteristics, and operative details from the patient record. In addition, we thoroughly reviewed the EHR to identify all telephone encounters, secure web-based patient-initiated encounters (MyChart messages; Epic Systems Corporation), ED visits, and hospital admissions following URS. Telephone calls were handled and documented in the EHR by a clinic nurse during business hours and by the resident on call after business hours. We categorized the reason for the telephone call, MyChart message, ED visit, and hospital admission. ED visits or admissions to outside facilities were documented if the patient shared that information.

To compare the frequency and type of encounters associated with another ambulatory endoscopic surgery, we also reviewed the records of 56 patients who underwent outpatient transurethral resection of a bladder tumor (TURBT) during a similar, but shorter time period, between January and June 2014. We chose TURBT for comparison because the procedure is typically performed on an outpatient basis for small- to moderate-sized tumors, involves instrumentation of the bladder, but not the upper tracts, differs from URS in that typically does not involve an indwelling ureteral stent, and is performed on a routine basis by most urologists.

Continuous and categorical variables were analyzed using Student's two-tailed t-test and Fisher's exact test, respectively. Significance was set at P < 0.05. We performed univariate (UVA) and multivariate (MVA) analyses to identify factors predictive of postoperative patient encounters. Variables on UVA meeting a threshold P-value of <0.15 were included in the MVA. To compare the outcome of medically related encounters between the TURBT and URS groups, a multivariable model incorporating the type of procedure and controlling for potentially confounding parameters was constructed. All analyses were conducted using JMP version 12.0 (SAS Institute, Cary, NC).

Results

We identified 298 consecutive patients who underwent 314 URS procedures for renal and/or ureteral calculi. URS patient demographics are shown in Table 1. Among the URS patients 70% of patients had prior stones and 33% underwent a prior URS for stone treatment. Of note, 28% (n = 98) of all URS procedures were bilateral, and in 90% of patients (n = 281), flexible URS alone or in combination with semirigid URS was utilized (Table 2). A ureteral access sheath (UAS) was used in 82% of cases, including 40 procedures with a 9.5/11.5 F UAS (13%), 21 procedures with a 10/12 F UAS (7%), and 197 procedures with a 12/14F UAS (63%).

We identified a total of 201 (67%) patients who were responsible for 443 encounters, including 331 telephone calls, 71 MyChart messages, and 38 ED visits. Among these encounters, 79.5% (352) were medically related, which included report of symptoms (267 encounters) and questions regarding medications (102 encounters), with some overlap between the two (Fig. 1). The remaining encounters were administrative or scheduling related. There were no significant differences in baseline characteristics between patients who made contact for medically related reasons and those who did not (data not shown). Among those with medically related encounters, 105 patients generated 198 encounters for flank/abdominal pain, which prompted early imaging studies in 16 patients. While 11 patients had abnormal studies (hydronephrosis in 7 and renal pelvis dilation only in 4), no patient had an obstructing residual fragment and all patients were managed conservatively and followed closely. At subsequent follow-up, imaging studies reverted to normal in all 16 patients, and no patient required intervention. A total of 34 patients were seen in the ED, with most encounters due to pain (61%), and 10 ED visits culminated in hospital admission (mean length of stay 0.7 ± 1.3 days).

OPEN IN VIEWER

FIG. 1.

Flowchart of encounters. Some patients had multiple reasons for each encounter.

On UVA, younger age, bilateral URS, larger UAS size, stones in multiple locations, and larger number of stones were all predictive of a symptom-related encounter (Table 3). On MVA, after adjusting for other parameters, only younger age and larger UAS size remained independent predictors of a symptom-related encounter, while bilateral URS and larger number of stones both trended toward statistical significance (Table 3).

We identified 56 consecutive patients undergoing outpatient TURBT to serve as our comparison group. Patient demographics are shown in Table 1. The two groups were comparable with regard to race and body mass index, while age, gender, and ASA were statistically significantly different in that the TURBT group was older, more predominantly male, and had higher ASA scores. Patients undergoing TURBT generated 64 patient encounters, 46 of which were medically related. The most common reported symptoms were hematuria and lower urinary tract symptoms, which contributed to 14 and 10 encounters, respectively.

On MVA, when adjusting for confounding variables, the type of procedure did not impact the likelihood of a symptom-related encounter (Table 4). However, URS patients had more pain-related encounters (45% vs. 11%), while TURBT patients had more hematuria-related encounters (22% vs. 11%). On subgroup analysis, patients undergoing URS were 2.5 times more likely to have pain-related encounters than patients undergoing TURBT, when controlling for confounding variables (Odds ratio [OR] 2.54, 95% confidence interval [CI] 1.08 to 7.04, P = 0.03).

Discussion

While historically SWL had been the most frequently utilized treatment modality for upper tract calculi, recent trends indicate that URS has surpassed SWL as first-line therapy for patients with renal and ureteral calculi. Case logs submitted to the American Board of Urology from 2003 to 2012 by candidates for certification and first and second recertification reflected temporal trends in the distribution of surgical stone procedures. Coincident with a decline in the proportion of SWL procedures from 54% in 2003 to 36.3% in 2012, URS procedures rose from 40.9% to 59.6%. The rise in the proportion of stone procedures comprising URS was seen in both the groups of younger (certification candidates) and older (first and second recertification candidates) urologists, while the increase was greater in the younger compared to the older cohort (49% vs. 37.5% increase, respectively). ⁶ Likewise, time series analysis using a single-payer universal healthcare insurance database from Ontario, Canada, demonstrated a reduction in SWL utilization, from 69% in 1991 to 34% in 2010, concomitant with an increase in URS from 25% to 59% during the same time period. ⁷

The reasons accounting for the reversal in utilization of SWL and URS are not completely clear, but may reflect advances in ureteroscope design and instrumentation that have expanded URS indications to include renal calculi and larger stones, ² while SWL technology stagnated for some time and SWL indications narrowed as better means of patient selection emerged. ^8,9 The choice of URS over SWL or PCNL for management of renal and ureteral calculi is largely physician driven, taking into account success, retreatment, and complication rates. ^{10 –14} However, despite few objective complications following URS, patient concerns prompt some patients to seek medical attention postoperatively. Tan et al. reviewed 1798 consecutive outpatient URS procedures and identified 70 immediate unplanned hospital admissions (3.9%) postprocedure. ⁴ On MVA, after controlling for confounding factors, they found that previous admission related to stones (OR 3.24, 95%CI 1.7–6.1, P < −0.001), history of psychiatric illness (OR 2.28, 95% CI 1.17–4.47, P = 0.016), and bilateral procedure (OR 2.88, 95% CI 1.19–6.99, P = 0.019) were independent predictors of unplanned hospital admission, and distal ureteral stone was associated with a reduced likelihood of unplanned hospital admission (OR 0.26, 95% CI 0.08–0.85, P = 0.26).

Scales et al. analyzed data from the Marketscan dataset and identified 93,523 privately insured individuals who underwent a procedure for stones between 2003 and 2011. ⁵ Unplanned ED visits or hospital admissions occurring within 30 days of the procedure occurred in 15% of URS and percutaneous nephrolithotomy (PCNL) patients and in 12% of SWL patients. In a multivariable logistic regression model, using PCNL as the referent, SWL was associated with a reduced rate of unplanned visits (odds ratio 0.74, 95% 0.65–0.94, P < 0.001), while URS showed comparable risk of unplanned visits (OR 0.95, 95% CI 0.84–1.07, P = 0.39). Thus, URS was as likely as more invasive PCNL and more likely than SWL to be associated with an unplanned visit. The most common reason for admission in both these studies was pain and concern for infection, but the need for invention other than pain management and/or antibiotic therapy was not quantified.

While post-URS ED visits or hospital admissions are relatively uncommon, the prevalence of patient concerns regarding their expected postoperative course has remained largely unexplored. Patient-centered outcomes are assuming an increasing role in medical decision-making, and paternalistic patient care is being supplanted by shared decision-making. ^15,16 As such, patient-reported quality of life, along with traditional outcomes of success and complication rates, is an increasingly important metric.

With the widespread use of EHR, by which documentation of unplanned patient encounters is readily accessible, the extent of postoperative patient morbidity, real or perceived, is more apparent. In our study, two-thirds of patients initiated some form of communication or contact with a healthcare provider following surgery, and 80% of the generated encounters involved a medical, as distinct from an administrative or scheduling, question or problem, most involving symptoms. While none of these encounters led to invasive intervention, they reflect a high level of anxiety, uncertainty, or discomfort following their procedure.

We attempted to identify patient characteristics or operative parameters that were associated with a higher likelihood of unplanned symptom-related encounters and found that only younger age and larger UAS size were independent predictors. Because a UAS was used in 82% of our patients, we were unable to determine if eliminating use of an access sheath would reduce postoperative encounters for symptoms. That larger UAS are more problematic than smaller UAS does provide a potentially modifiable factor that could be initiated to reduce postoperative problems. While we suspected that previous stent placement or URS might impact the likelihood of a symptom-related encounter either positively or negatively due to familiarity with the procedure, we found no such association. However, a quality-of-life study among 155 stone patients found on multivariate linear regression analysis that the number of ureteroscopies and stent placements, each independently predicted lower scores on three of the four mental domains in the in the SF-36™ survey. ¹⁷ Furthermore, because younger age was also associated with greater risk of postoperative patient encounters, this patient group may warrant more intensive preoperative counseling regarding expectations following surgery and might benefit from a different pharmacologic regimen.

We hypothesized that URS generated more patient-initiated encounters than TURBT, possibly due to additional upper tract manipulation and placement of a stent with its associated symptoms. Indeed, despite a comparable proportion of medically and symptom-related encounters, URS was more than 2.5 times more likely to generate a pain-related encounter than TURBT when controlling for confounding factors.

We acknowledge several limitations of our study. Due to the retrospective nature of our data collection, patient-initiated encounters that were not directed to our health system could have been missed. Furthermore, it is possible that some encounters were not diligently documented in the EMR and are therefore unaccounted for. In addition, the URS and TURBT patient groups were not completely comparable, as TURBT patients were older and more predominantly male. However, we were unable to identify a more comparable, commonly performed outpatient endoscopic procedure with which to compare URS, and to some degree, our multivariable model took these differences between groups into account.

Conclusions

While analysis of objective patient outcomes following URS demonstrates that the procedure is highly successful and has low complication and retreatment rates, we found that patient-perceived postoperative problems are disturbingly common. Furthermore, this finding is distinct from another commonly performed outpatient endoscopic procedure, TURBT. We may be able to reduce patient morbidity following the procedure by focusing on patients who are most likely to initiate contact with their provider due to perceived problems, for example, altering the postoperative medication regimen in younger patients and/or modifying the procedure by using smaller UAS. Prospective analysis of the nature of patient-perceived postoperative problems may provide further insight that can be used to improve patient tolerance of the procedure, and more extensive preoperative patient counseling may better inform patient decision-making.