A new pathway for elective surgery to reduce cancellation rates

The project involved all of the surgical departments at Førde Hospital, including in- and outpatient ophthalmology, general surgery, gynecology, orthopedics, and ear, nose, and throat. Additionally, the hospital has a small odontology unit that accounted for less than 5% of the total operations.

Four different project groups were established to improve different aspects of the elective surgery pathway. Redesign of this pathway was the first project that was run in accordance with the new improvement strategy of the local health authority. Table 1 displays the main activities during the project period. A strong motivator for the leaders who initiated the project was to improve patient satisfaction. Patient advocacy groups were invited to take part in the project groups but declined to participate. Instead, actual patient cases were used to focus improvement efforts, and patient-centered interventions were a core idea in the redesign process [14, 15].

The entire pathway for elective surgery was redesigned, focusing on earlier patient assessment, improved communication between staff, improved management, improved planning, and patient participation in the planning of their elective operations. Table 2 shows a description of the clinical pathway and a detailed description of the different intervention elements, including the intended improvements. Ideas for improvements stemmed from discussions in the project groups, recommendations in the literature, and a site visit to a hospital with low cancellation rates.

As part of the interventions, a new day-surgery center was designed within the existing premises. All patients met at this center before their elective operation, and all day-surgery patients were discharged from this center without admission to a surgical ward. A computer application was introduced during the project. It provided an overview of referrals, waiting lists, and surgery schedules in all departments. A new position, a capacity coordinator, was created with the mandate to plan and coordinate the surgery program across different departments 6 months ahead. The implementation of the new pathway began when the new day-surgery center was opened in March 2008.

Another project during the study period reduced turnover time between operations by improving logistics and coordination between the facilities for preparation, surgery, and recovery. In addition, the durations of surgical procedures were continuously monitored to get a more realistic picture of the actual time used, thereby improving scheduling of surgery. Furthermore, one operating theatre was designated for emergency cases.

We collected qualitative and quantitative data from the hospital between April 2010 and February 2012. We obtained the number of planned and performed operations and cancellations per month from the hospital’s patient administrative system, and then calculated the monthly CRs. A cancellation was defined as a planned operation that was cancelled within 24 hours of the scheduled time. We considered a decrease in monthly CRs and an increase in the number of performed operations to be an improvement.

To compare CRs before and after the interventions we used the statistical method recommended by Dexter et al., since this method has been shown to be the most robust for this purpose [16]. We calculated the CRs for each month, and transformed the CRs using the Freeman-Tukey double arcsine transformation, and then applied a Student’s t-test on the transformed rates using SPSS 18.0 [16].

The cancellation rate is nonconformance per unit, and we therefore used a u-chart to analyze whether changes in CRs coincided with the interventions and whether improvements were sustained [17].

Cancellations vary by specialty [18]. The number of performed operations increased after the interventions. Thus, an increase in the number of performed operations at a department with low CRs could disproportionately affect the total number of cancellations at the hospital. To assess this effect, we calculated the expected number of cancellations for each department for the time period after the interventions (i.e., the product of pre-intervention CRs and the number of scheduled operations after the interventions). We then calculated the expected cancellation rate at the hospital for the time after the interventions (i.e., the sum of the expected number of cancellations for each department divided by the total number of scheduled operations).

We recorded the volume of elective surgery, emergency surgery, and consultations at the outpatient clinics, because an increase in the volume of one of these activities could affect the volume of the others. The Department of Ophthalmology, which accounts for less than 1% of the total number of emergency cases, was excluded from the count of emergency cases because of incomplete data. CRs from the odontology unit were not available for the entire period, and it was excluded from our study. Moreover, we recorded the number of cancellations per month due to overriding emergency cases.

Changes in the ratio of capacity and demand could influence the cancellation rate [19]. We measured demand as scheduled operations per month and capacity as hours available for surgery per week and the number of full-time equivalents per year for the involved departments. Due to changes in the data system of the hospital, information about full-time equivalents was only available after January 2008.

Utilization of the list of scheduled operations, and in particular list over-runs, can influence CRs [19]. We recorded the proportion of list over-runs for each department. In line with Pandit et al. [19] a list finishing >10% after the scheduled end time was classified as over-running. Data for these calculations were only available for the time after the interventions. Finally, we recorded the number of cancellations per month caused by the hospital not being able to finish the scheduled list.

We interviewed a strategic sample of employees at the hospital (n = 20) to understand how the pathway for elective surgery was redesigned and to identify factors that had contributed to sustained improvements. Interviewees had different professional backgrounds (i.e., physicians, nurses, secretaries, and leaders) and worked in different departments involved in the clinical pathway. The degree of interviewees’ participation in the improvement projects varied. Some had participated in the design of interventions in their project groups and others were not directly engaged in the interventions. The interviews were semi-structured and based on an interview guide that covered the following topics: local problem, setting, context, intended improvement, planning of interventions, implementation of interventions, outcomes, and efforts to sustain outcomes. The interviews were taped, transcribed, and transferred to HyperRESEARCH 2.8.3 software (ResearchWare, 2009) for coding. We developed an initial coding scheme from the themes in the interview guide, and codes were added as the data were analyzed [20]. We interpreted the relationship between the codes to identify distinctive elements of the interventions and factors that influenced success of the improvement process [21].

The protocol of the study was presented to the regional ethical review board, and a formal ethical review was not deemed necessary. The study was approved by the Norwegian Social Science Data Services. All the interviewees participated voluntarily based on informed consent, and could withdraw from the study at any time.

The entire pathway for elective surgery, from referral to discharge, was redesigned. Changes were implemented across departmental borders, and frontline staff were broadly involved [22]. Improved management and surgery planning, redesign of work processes, training of staff, and early clinical evaluation have been suggested as strategies to reduce CRs [6]. At our hospital, all of these strategies were applied. The interventions included various elements that were linked to the local context. The distinct effects of the separate elements can therefore not be disentangled.

Sanjay et al. [4] suggest allowing patients to select the time for surgery to give them earlier notice of their operating day, and to send them a reminder. Involving patients in these ways can increase satisfaction with treatment decisions during the initial consultation, which is a strong predictor of whether a patient will attend the surgery [23]. In the new pathway, patients participated in planning the date of their surgery and received the actual date of the operation before they left the outpatient clinic. It is therefore likely that these measures contributed to reducing cancellations.

Early preoperative assessment has previously been shown to reduce CRs [7‐10]. Van Klei et al. [8] showed that cancellations attributable to medical reasons decreased from 2.0% to 0.9% for patients who had attended a pre-assessment clinic. Ferschl el al. [7] found a CR of 5.3% among patients who visited an anesthesia preoperative medicine clinic, in contrast to 13.0% for those who did not. Rai and Pandit [9] found that nurse-led pre-assessments in an elective surgical center reduced cancellations. O’Regan et al. [10] demonstrated that a process-oriented multidisciplinary approach for patients who undergo bypass surgery led to improved patient outcomes and lower CRs. Our findings are consistent with these studies.

Cancellations of elective surgery due to emergency cases were practically eliminated after the interventions. The number of emergency cases was the same before and after the interventions and can thus not explain this finding. It is most likely that the designated day-time theatre for emergency cases contributed to reducing CRs, which is supported by previous studies [24, 25].

The increase in performed operations exceeded what can be explained by the reduction in CRs alone. More patients were referred to the hospital for elective surgery mainly because of the closure of a local hospital in the region. The number of performed operations increased after the interventions, while the capacity was reduced 3 months after the interventions. This finding indicates that the hospital managed to increase the efficiency of their operating lists [19].

Since the CRs varied across specialties, the disproportional increase in scheduled operations among the surgical specialties could have contributed to reduced CRs after the interventions [18]. The calculated expected CR after the interventions for the hospital was 8.2%, while the observed rate before the interventions was 8.5%. The difference between the calculated and the observed CR indicates that a disproportional increase in scheduled operations may have contributed to the decrease in CRs after the interventions. The effect is, however, small, and can only explain a small portion of the observed reduction in CRs.

A reduction in emergency cases could have affected the CRs. However, the number of emergency cases per month was the same before and after the interventions, making this an unlikely explanation for the reduction in CRs. The increase in the number of performed operations was not caused by a subsequent fall in consultations at the outpatient clinic, as this number increased after the interventions.

Reduced pressure on the service could contribute to reduced CRs [26]. In our case the number of scheduled and performed operations increased. This increase indicates that there was a corresponding increase in demand. The capacity, measured as hours per week available for surgery, increased until 3 months after the interventions and then decreased, while the number of full-time equivalents remained unchanged after the interventions. Thus, the pressure on the service increased after the interventions, making it unlikely that change in the ratio between demand and capacity contributed to reduced cancellations.

An increased tendency to over-run operating lists after the interventions could have contributed to the observed reduction in CRs. Data about this were not available for the pre-intervention period. The proportion of over-running lists for the ophthalmology- and ear, nose, and throat departments was low after the interventions, at 1.2% and 2.8%, respectively. It is therefore unlikely that list over-runs at these departments increased after the interventions, thereby contributing to reduced CRs. For the three remaining departments, we cannot exclude the possibility that list over-runs increased after the interventions. Before the interventions, 4.2 cancellations per month were due to the fact that the hospital was not able to finish the scheduled surgery list as planned. These cancellations could have been avoided by over-running the lists [19]. However, there was no significant reduction of cancellations caused by the hospital not being able to finish the scheduled program after the interventions. This finding indicates that even if list over-runs may have increased after the interventions it cannot be a strong factor for explaining the observed reduction in CRs.

The redesign of the surgical pathway was embedded in the new strategies to improve the performance of the clinical system. This integration secured a solid foundation in the top management without compromising professional entrepreneurship of middle managers and frontline professionals. Consistent with earlier studies [27, 28], we found that this strategy created a basis for improvement by providing guidance about tools and techniques that were important for the success of the project.

The improvement strategy was also important for securing a wide representation of clinical staff in the project groups and for setting the context for the project. In the strategy, the top management emphasized system improvement by equally addressing professional patient and management quality, as well as staff satisfaction. The inclusion of all these dimensions contributed to acceptance by clinicians and other staff, consistent with other studies [12, 29].

Improvement groups interacted in an informal network across departments. This network continued after the project period. Frontline employees were engaged in suggesting adoptions and modifications of the interventions. The presence of the middle managers in the actual work processes allowed them to follow the implementation daily and adapt and re-implement changes when needed. The hospital increased the effectiveness of the interventions by adapting them to contextual changes, as indicated by Fixsen et al. [30]. This flexibility seems to be a key factor for sustaining the outcomes.

The optimal use of information technology contributes to the success of high-performing institutions [31]. In our case, the new software for planning surgery integrated the schedules of all the departments thereby improving the scheduling of operations for the whole hospital.

Moreover, the surgery coordinator supported scheduling of operations up to 6 months in advance by matching available slots for surgery and the expected duration of procedures based on previous experiences. The planning processes became more dynamic because waiting lists were taken into account when assigning slots for surgery. Altogether, these measures ensured a better utilization of the total capacity of the operating theatres.

CRs of approximately 5% in our study are still high compared to van Klei et al., [8] who reported a rate below 1%. Further improvements can probably be achieved by fully implementing the aforementioned changes. The process of scheduling surgery could likely be further improved by applying the approach described by Pandit and Tavare [32] using a formula to predict the likely duration of an operating list.

An observational and retrospective study design has the limitation of information bias and confounding, and we cannot prove causality between interventions and the observed outcomes. The improvement project in our case is an example of complex, context-dependent interventions that evolved over time. Such projects are less suitable for strictly controlled, prospective, experimental study designs that could avoid these limitations [33]. Nonetheless, our design makes long-term follow-up feasible and allows us to learn from a successful case by combining qualitative and quantitative data and different analytical methods [34]. Therefore we found it appropriate to use the chosen design.

Our calculation of the number of full-time equivalents before the interventions was based upon data from January 2008, because prior data were not available. These data might not be representative of the pre-intervention period. However, the mean number of full-time equivalents after the interventions was 280 compared with 279 before the interventions, indicating no increase during the post-intervention period. This finding is also supported by data from the interviewees stating that there was no substantial change in the number of full-time equivalents during the study period.

We used the number of scheduled operations per month as a measure for demand. A more precise measure for demand would have been minutes of operating capacity needed per week, but data about this were not available [26]. However, we argue that the number of scheduled patients per month also reflects service demand. Since the number of scheduled patients increased after the interventions, while the capacity was reduced, it is likely that the pressure on the service increased.

In summary, the strength of our study is the long observation time with sustained improvements. Moreover, through our use of quantitative and qualitative methods, we were able to identify factors that contributed to the changes. By using statistical process control we could demonstrate a clear association between interventions and improvements. Data from interviews were consistent across departments and professional borders and did not reveal other organizational changes or quality improvement projects that could have influenced CRs. Finally, our findings are consistent with previous studies. The hospital in our study resembles other district hospitals and the interventions implemented here can likely be adapted to other hospitals of similar size and complexity.