Problem

Cervical cancer is the third most common gynaecologic cancer in developed countries and the second most common cancer in females in developing countries.1 While early-stage cervical cancer may be cured with surgery alone, many women present with locoregionally advanced disease. The standard of care for higher-stage cervical cancer consists of external beam radiation therapy (EBRT) with concurrent chemotherapy followed by a brachytherapy (BT) boost to the cervix.2 It has been well established that prolonged treatment times from the start to finish of definitive chemoradiation (CRT) are associated with increased cancer recurrence and decreased survival.3–8 Despite the accepted importance of treatment duration, completing treatment in a timely fashion still faces practical challenges.

As a National Cancer Institute (NCI)-designated comprehensive cancer centre and the only university-based academic medical centre in the state, our institution has become the main referral site for women diagnosed with advanced cervical cancer, serving not only the metropolitan area, but also patients from rural Minnesota and bordering states. Although a departmental policy was in place to keep the total treatment time (TTT) to 56 days or less, as per Radiation Therapy Oncology Group protocols, an initial informal audit showed a disappointingly low compliance rate.9 Our aims were, thus, to identify reasons of treatment prolongation, to implement quality improvement (QI) measures and to compare TTT before and after these implementations among cervical cancer patients treated with definitive CRT.

Background

In the USA, approximately 13 000 new cases of cervical cancer are diagnosed annually with 4100 cancer-related deaths.10 CRT has been the standard of care for locoregionally advanced disease since 1999 when several randomised trials showed a survival advantage with the addition of concurrent cisplatin-based chemotherapy to radiation therapy (RT).11 12 The treatment regimen consists of EBRT with concurrent weekly cisplatin followed by a BT boost.2 13 EBRT is typically delivered in 25–28 fractions, 5 days per week, over 5–5.5 weeks. BT is administered over 3–5 fractions, at least 48 hours apart, either interdigitated with or after the EBRT. Patients with parametria or sidewall disease may also require an additional EBRT boost, given on days without BT treatment.

TTT, defined as the time from first radiation fraction to treatment completion date, has been shown to significantly impact patient outcomes following definitive radiotherapy. In the era when radiation was used alone, treatment times beyond 7–9 weeks were shown to reduce locoregional control (LRC) and overall survival.3–8 Similar findings have been reported for patients treated with CRT.14–16 In one series, treatment times over 56 days were associated with a 17% increase in pelvic failures compared with shorter treatment times.14 Each extra day of treatment beyond 6–8 weeks can result in a loss of 0.6% per day in pelvic control.16 While the mechanism is not fully understood, it is hypothesised that an increased proliferation of tumour cells in response to treatment-induced cell killing through a phenomenon known as clonogenic accelerated repopulation is responsible for the decreased tumour control with prolonged treatment.17 In addition to worse clinical outcomes, prolonged treatment times have also been shown to significantly reduce overall patient satisfaction.18

Despite the recognised importance of TTT, completing treatment under 56 days can be challenging. We set out to review TTT for patients treated with definitive CRT at our institution, to identify sources of delays in treatment and to implement QI measures to enhance the delivery of care.

Baseline measurement

In the initial phase of the study, we sought to categorise patient and treatment characteristics and assess the prevalence of prolonged treatment time at our institution. After obtaining institutional review board approval, we retrospectively reviewed 24 consecutive patients treated with definitive CRT using a combination of EBRT and high-dose-rate (HDR) BT from July 2011 to August 2013. Only patients treated with curative intent were included.

Radiation treatment characteristics of these patients are presented in table 1. All patients were treated with EBRT followed by image-guided HDR tandem and ring (T&R) (n=17, 70.8%) or interstitial BT (n=7, 29.2%) applications. Before starting BT, patients were usually required to undergo a diagnostic MRI to assess treatment response to EBRT and to allow selection of appropriate boost technique. A Smit sleeve was sewn into the endocervical canal in all patient receiving T&R BT to facilitate applicator insertion. The median number of BT treatments was five with T&R administered at least 48 hours between fractions, whereas interstitial BT was delivered in a two times per day fashion over 3 days. A total of 21 patients (87.5%) received a parametrial boost (PB) via anterior–posterior split-pelvic (SP) fields that occurred before, during or after BT. The majority of patients received EBRT at our university main campus (n=20, 83.3%); however, patients were also treated at university affiliate sites (n=2, 8.3%) and outside institutions (n=2, 8.3%), and subsequently came to the university for BT.

Concurrent chemotherapy with weekly cisplatin was given in all patients staged FIGO IB2 or higher. For patients treated with chemotherapy, 15 (65.2%) received all five planned weekly doses. Chemotherapy was withheld in the remaining patients primarily due to cytopenia, and occasionally for severe nausea, dehydration and diarrhoea, at the discretion of the treating gynaecologic oncologist.

The median TTT among the 24 pre-intervention patients was 54 days (range 40–69 days). Of these, 14 (58.3%) had TTT≤56 days, whereas 10 (41.7%) exceeded 56 days. With an approximately 40% baseline rate of failing to achieve target TTT, QI efforts were initiated with the goal of reducing this rate by at least half.

Design

A QI team was established, comprising of physicians who specialise in gynaecologic malignancies, a resident physician as well as an oncology nurse.

A root cause analysis (RCA) was performed by the attending and resident physicians to determine the cause of treatment delay in the retrospective cohort. Manual chart review of the electronic health record (EHR) was performed to identify the primary cause of delay. For some patients, it was difficult to ascertain a primary cause for delay from the EHR alone. In these cases, members of the patient’s treatment team were contacted for insight into the cause of the prolonged treatment. Of the 10 patients who had TTT greater than 56 days, the primary causes of prolonged treatment could be assigned to three general categories, which is represented in a Pareto plot (figure 1). Four patients experienced treatment breaks due to toxicity, including cytopenia requiring transfusions and severe nausea, vomiting and dehydration necessitating inpatient fluid resuscitation. One of these patients was admitted to an outside community hospital, which resulted in further delays due to logistic challenges. Three patients had treatment prolongation due to gaps between the end of EBRT at an outside institution and the start of BT at our institution. Another three patients had prolonged treatment times because of difficulties scheduling one or more phases of their RT treatment. This group included one patient who had difficulty making treatment appointments due to a lack of reliable transportation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Pareto plot illustrating the primary causes of treatment prolongation among the 10 patients with TTT greater than 56 days. Patients treated prior to June 2013 represent the pre-intervention cohort. Quality improvement interventions were implemented from July 2013 to July 2017. Patients treated after July 2017 represent the post-intervention cohort. b/w, between; TTT, total treatment time; w/, with.

We used a process map to visually display the workflow of how a patient is scheduled for treatment. Process mapping of our workflow revealed several potential sources for delays that were not apparent prior to creating a formal diagram of our workflow. The established protocol for the first fraction of BT included Smit sleeve placement and applicator insertion in the operating room (OR) followed by a planning MRI, a treatment planning CT, creation of the first BT plan and treatment delivery, all occurring in the same day. Such a long process necessitated the first OR timeslot of the morning, which was in high demand and required booking weeks in advance. Process mapping also revealed that the PB via split pelvis field was often added on after BT was complete, instead of intercalated into the BT schedule, thus prolonging TTT. For patients receiving EBRT at an outside facility who were referred to the university for BT, a stalling point was found in the timely scheduling of consultation appointment and BT procedures with a university radiation oncologist. Finally, this exercise showed that there was a knowledge gap among scheduling staff. Many of the medical assistants and schedulers were not aware that reducing treatment times was a priority in the treatment of cervical cancer, and therefore were not making a conscious effort to expedite the coordination of care.

To reduce treatment prolongation, we implemented improvement initiatives using Plan-Do-Study-Act (PDSA) cycles over the course of 4 years. To evaluate the efficacy of our new action plan, a post-intervention analysis of treatment times was prospectively conducted on 81 subsequent cervical cancer patients treated with definitive CRT and compared with that of the pre-intervention cohort using the χ² test. The sustainability of our intervention was tested by evaluating an additional 47 patients after all PDSA cycles had been implemented and the formal QI project had ended.

Strategy

PDSA cycle 2

Plan: We planned to examine and redesign our workflow in an effort to reduce/eliminate treatment delays. We also sought to establish collaborative relationships with the outside providers to encourage referral of patients early in the course of EBRT so that elapsed time between EBRT and BT could be minimised.

Do: We reached out to OR scheduling team to leverage more first-case slots for our patients. We lobbied outside physicians to place referral orders early in course of patients’ EBRT. Furthermore, we asked the physicians to declare an intention for a PB at the time of initiation of EBRT, rather than towards the end.

Study: We found that while OR was supportive of our efforts, reserving more first-case start time was not always achievable, and thus separating the Smit sleeve placement from first insertion of T&R may be necessary. In some instances, OR or BT slots could not be secured due to lack of timely orders from physicians. For patients requiring PB, it could be a source of delay when it is given after all five BT were complete.

Act: In order to make more OR times workable, Smit sleeve placement was scheduled on a separate day prior to first BT. On the actual days of BT, T&R applicators could be inserted in the clinic or sedation suite, significantly reducing the demand for OR. Additionally, for patients requiring PB, the delivery of SP fields was shifted from following BT to interdigitating with BT on days BT was not given. All physicians at our institution were also asked to place BT-related orders within the first 5 days of initiating EBRT to allow more flexibility in scheduling with OR for Smit sleeve placement, sedation suite for applicator insertion and radiology for treatment planning MRI. The improved workflow is shown in figure 2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Re-designed workflow and timeline of a cervical cancer treatment course incorporating external beam radiation therapy with concurrent weekly cisplatin, five fractions of brachytherapy (BT) and up to five fractions of a parametrial boost that finishes with a TTT of less than 56 days. HDR, high-dose rate; OR, operating room.