Comparison of eVisit Management of Urinary Symptoms and Urinary Tract Infections with Standard Care

Introduction

Telemedicine is a rapidly expanding health care delivery model. ¹ eVisits, using asynchronous communication with a health care provider via a secure online patient portal, offer an alternative to face-to-face (F2F) visits. ^{2 –5} While providing benefits of cost savings, convenience, time savings, and improved access, ^6,7 eVisits have been regarded with some skepticism. Concerns reported in the literature include that eVisits lead to increased health care utilization ⁸ and antibiotic overprescribing. ⁹

Urinary symptoms and urinary tract infections (UTIs) are frequent chief complaints for visits to primary care providers and emergency departments. It was noted in 2007 that UTIs in the United States accounted for 10.5 million ambulatory visits, which was 0.9% of all ambulatory visits, and almost 2 million emergency room visits. ^{10 –12} The cost of these infections in the United States is estimated at $3.5 billion per year. ¹¹ It is accepted that UTIs can be diagnosed by history alone, without physical examination or urine testing, ^13,14 and UTIs have been treated by non-F2F methods for years, primarily by registered nurse (RN)-administered phone protocols. Using RN phone protocol treatment for UTI has been well established in the literature. ^{15 –18} However, nursing protocols are not a billable service, but require nursing resources, and on occasion provider resources.

The current literature has limited published research looking at UTI evaluation and treatment via eVisits. Mehrotra et al. found that treatment of UTIs via eVisit had lower rates of testing and similar rates of follow-up but had higher rates of antibiotic prescribing when compared with F2F visits. ⁹

The purpose of this study was the following: to investigate antibiotic prescribing, clinical outcomes, and follow-up rates for the online assessment and treatment of urinary symptoms and UTIs compared with the established standard of both in-person F2F visits and RN phone protocols.

Methods

Mayo Clinic Rochester primary care empaneled patients have different F2F and non-F2F options for care of UTI symptoms. Non-F2F evaluation and management are available with asynchronous online text-based eVisits or phone call by nurse protocol. eVisits are patient-completed symptom-specific question sets sent to a pool of advanced practice providers (APPs) via a secure online patient portal, which are then reviewed and responded to by an APP within an hour (if submitted between 8 am and 11 pm) 365 days a year. The APP does not have clinical support to order urine testing for eVisit patients. The other option for non-F2F management of UTI is the RN phone protocol. Patients can call and request to be treated via a protocol by an RN for their UTI symptoms, or sometimes end up being treated via a protocol as part of the phone triage process. If the patient does not meet criteria for treatment via the RN phone protocol, the nurse will then triage the patient to a different endpoint for care. Unlike with eVisits, occasionally if the patient does not meet criteria for treatment with protocol, an APP or physician in the primary care setting will be consulted and order laboratory urine testing without a formal F2F appointment, and treatment options considered based on test results.

For F2F visits, patients can have an in-person evaluation at their primary care office or at Mayo Clinic Express Care (MCEC). MCEC is a retail clinic setting staffed by APPs offering walk-in appointments for a limited selection of minor acute illnesses. In addition, patients can be seen at their primary care office; however, before scheduling an appointment for an acute concern, they are requested to initially speak with an RN for phone triage. An appointment is scheduled if the concern cannot be addressed over the phone. Most patients seen in the primary care provider's office for an acute concern therefore represent a specific population, and have already been evaluated by nurse triage and were recommended to be seen for an F2F visit, while those seen at MCEC have not typically had any previsit triaging. Thus, in our system, patients evaluated at the MCEC retail clinic likely are more representative of a typical population seeking F2F care for UTI than those who seek care at their primary care provider's office.

A retrospective review was completed of initial encounters for Mayo Clinic Rochester primary care empaneled female patients, ages 18–65 years, who had an F2F visit at MCEC, an RN phone protocol encounter, or an eVisit for urinary symptoms between August 1, 2016, and May 1, 2017. F2F visits and eVisits were retrieved by searching for diagnostic codes, while RN phone protocol calls were identified by using Mayo Clinic's Advanced Cohort Explorer (ACE) search tool. ACE uses text searching of electronic medical records (EMRs) to identify specific words and/or phrases. We used ACE to search EMRs for the specific protocol (by reference number) and triaging tools nurse's use for documentation of RN phone protocols for UTIs or for triaging for urinary symptoms.

Exclusion criteria for the study included age less than 18 years and over 65 years. Patients who had been seen for urinary complaints or treated for a UTI in the prior 30 days before the selected encounter were excluded. Other exclusion criteria included hospitalization, residence in a long-term care facility, or urinary procedure or catheterization in the previous 2 weeks. Women who were pregnant at the time of the encounter were also excluded. In addition to the above exclusion criteria, patients not reporting dysuria as a presenting symptom are unable to complete the eVisit questionnaire and are referred in for in-person evaluation.

We randomized 150 visits from each encounter type for manual chart review. We reviewed encounters for age at the time of the visit and the following outcomes: treatment with an oral antibiotic (yes vs. no), follow-up within 30 days of the initial encounter for urinary symptoms (yes vs. no), whether urine testing (urinalysis, urine dip, or urine culture) was done at initial encounter (yes vs. no), whether patient was advised to have an F2F follow-up visit by the encounter provider (yes vs. no), time to first follow-up (if follow-up occurred), and type of follow-up encounter (F2F office visit, portal message, eVisit, phone call, emergency department visit, hospitalization). Of note, patients completing RN phone protocol encounters for whom urine testing was ordered completed testing at a laboratory without having an F2F visit with a health care provider.

In addition to the follow-up rates, clinical outcome data were collected, including antibiotic extension or changes in antibiotic regimen at the follow-up encounter, which might suggest a failure of initial treatment. Rates of urinary sepsis or progression to kidney infection were also collected in the charts reviewed. Kidney infection was defined in our chart review as a provider diagnosis of kidney infection, suspected kidney infection, or change in antibiotic therapy appropriate to cover pyelonephritis at follow-up visit.

Results

During the study period, there were 1,673 RN phone protocol encounters, 779 F2F visits, and 538 eVisits for urinary symptoms. The average age of patients seeking care via RN phone protocol was significantly older than patients seeking care via eVisit or F2F care (Table 1). Patients evaluated in F2F encounters were significantly more likely to have urine dipstick testing and urine cultures performed at their initial encounter than those evaluated initially by RN phone protocol. No urine testing was done for patients who were evaluated by eVisit; however, despite this significant difference in urine testing between the different encounter types, there was no difference in antibiotic treatment rates between encounter types. Patients who had an initial eVisit or RN phone protocol encounter were significantly more likely to be advised by the evaluating health care provider to have follow-up compared with patients evaluated initially by an F2F visit. Despite this advice, there were no significant differences in follow-up rates between the different initial encounter types.

Of the patients who had follow-up within 30 days, there were no differences in initial treatment rates between the three initial encounter types (Table 2). There were also no differences in average time to first follow-up or number of follow-up encounters between initial encounter types. Same-day follow-up was more likely to occur for patients who initially had non-F2F care. Follow-up was more likely to occur at the primary care team's office for patients who initially had an F2F or RN phone protocol encounter, while those initially evaluated by eVisit were more likely to have follow-up at MCEC (Table 3).

There was no difference in apparent treatment failures between the three initial encounter types, as measured by antibiotic retreatment with an extension of the initially prescribed antibiotic, or an antibiotic change in therapy within the subsequent 30 days between initial encounter types. There were no statistically different clinical outcomes between the three initial encounter types as evaluated by diagnosis of pyelonephritis within the 30-day follow-up period (Table 4). There were no hospitalizations or insistences of sepsis within 30 days of initial encounter in any patients.

Discussion

Contrary to the findings of the only previous study comparing treatment rates in patients utilizing eVisits versus F2F visits for UTI symptoms, ⁹ our study did not identify a difference in antibiotic prescribing rates between eVisits and F2F care. We also did not find any difference in antibiotic prescribing rates between patients seeking care by eVisit compared with RN phone protocol care. This lack of difference has important implications to address the concerns of eVisit overprescribing in a time of antimicrobial stewardship. The eVisit antibiotic prescribing rate in our study was lower than in the study by Mehrotra et al., 81% versus 99%. Antibiotic prescribing rates for F2F encounters differed also with our study reporting higher antibiotic prescribing rates than the study by Mehrotra et al., 83% versus 49%. ⁹ However, the F2F encounters were very different in the two studies. Mehrotra et al. studied care provided in a primary care setting, where relationships with providers may be established, communication easier, and consistent follow-up provided, while our F2F encounters were provided in a retail care setting, which has a different set of expectations of care (no prior relationship with the patient, serving acute management of conditions, not allowing for follow-up or consistent care). In addition, treatment rates for all encounter types in our study (81–83%) appear compatible with what is in the literature for treatment rates for women seeking care for urinary symptoms/UTI, with Schauberger et al. reporting treatment rates of 75%. ¹⁷

Another important factor when comparing the two studies is that average age and age range are not reported by Mehrotra et al. ⁹ The average age for patients in our study was 39.6 years. It is accepted that women of reproductive age are more likely to have an uncomplicated UTI. Those at an extreme ages have the potential for more comorbidities or causative factors for urinary complaints, ^19,20 impacting the amount of testing ordered and provider comfort with empiric prescribing. Generally, it is recommended that children, adolescents, and elderly patients should have urine testing completed before initiating antibiotic therapy. ^19,20

Providers were significantly more likely to order urine testing in the setting of an F2F visit, similar to the findings of Mehrotra et al. ⁹ It should be noted once again that the providers at our institution do not have the ability to order urine testing for eVisit patients. If urine testing is felt to be warranted in an eVisit, the patients are directed to seek F2F follow-up, often being referred to MCEC. The RN phone protocol patients may have urine testing ordered by a health care provider if they fail to meet criteria to be treated by protocol. This testing is at the discretion of the covering health care provider and is not universally offered to all RN phone protocol patients who do not meet criteria to be treated. However, in our study it was more common that RN phone protocol patients who did not meet all inclusion criteria for treatment were recommended to follow-up (19%) rather than just completing urinary testing (5%).

The follow-up rates of patients in our study were not statistically different between encounter types, but were higher than those in the study by Mehrotra et al. ⁹ The reason for this difference is likely multifactorial and includes that we used a longer follow-up interval (30 days) compared with the study by Mehrotra et al. In addition, our study included all patients with urinary symptoms, not just those ultimately thought to have a UTI, and/or given a UTI diagnosis code. As previously mentioned, the inability for the eVisit providers to order urine testing is also likely a significant factor for higher follow-up rates. Nearly half of the eVisit patients who sought follow-up did so on the same day as the initial encounter indicating a need for further evaluation rather than failed treatment or worsening symptoms. Furthermore, the F2F visits in our study occurred in a retail clinic setting, MCEC. This setting has a limited menu of care offerings with the usual practice of referring patients to the primary care clinic if further care is felt to be warranted such as sexually transmitted infection testing, pelvic examination, prophylaxis for frequent UTIs, intervention for overactive bladder, or referral to urology. Primary care F2F appointments as an initial/primary encounter type were not included in our study, as most patients attempting to schedule an appointment for urinary symptoms initially are evaluated by nurse protocol over the phone. Those who fail the phone protocol are then given the option for primary care appointments or referral to MCEC.

While the study by Mehrotra et al. used follow-up rates as an extrapolation/inference of misdiagnosis or treatment failure to comment on clinical outcomes, our study went beyond the follow-up rate analysis and looked at treatment failure as defined by changes to the treatment plan at initial encounter (extension of treatment duration or change in prescribed antibiotic) as well as identifying those who went on to be treated in the emergency department or developed kidney infection or urosepsis. Follow-up prescribing rates were the same across all three initial encounter types with ∼15% of those visits resulting in treatment modifications. There were no cases of urosepsis, and very few emergency department visits or episodes of pyelonephritis at follow-up, again without any difference between encounter types. No patients with kidney infection had signs and symptoms serious enough to require hospitalization.

Conclusions

eVisits for UTI and urinary symptoms offer patients a convenient option for care without an increased use of antimicrobials, follow-up, or adverse clinical outcomes when compared with F2F visits or RN-administered phone protocols.