EP Research

Using Discrete Event Simulation to Model the Economic Value of Shorter Procedure Times on EP Lab Efficiency in the VALUE PVI Study

Marcin Kowalski, MD1;  J. Brian DeVille, MD2;  J. Thomas Svinarich, MD3;  Dan Dan, MD4;  Andrew Wickliffe, MD5; Charan Kantipudi, MD5;  Jason D. Foell, PhD6; Giovanni Filardo, PhD, MPH2;  Reece Holbrook, BSEE6;  James Baker, MD7;  Hassan Baydoun, MD8;  Mark Jenkins, MD9;  Peter Chang-Sing, MD10
1Staten Island University Hospital, Northwell Health, Staten Island, New York; 2The Heart Hospital, Baylor Plano, Plano, Texas; 3St. Anthony Hospital, Lakewood, Colorado; 4InterMountain Healthcare, Salt Lake City, Utah; 5Piedmont Healthcare, Atlanta, Georgia; 5Piedmont Healthcare, Atlanta, Georgia; 6Medtronic, Inc., Minneapolis, Minnesota; 7Saint Thomas Hospital, Nashville, Tennessee; 8Staten Island University Hospital, Staten Island, New York; 9Memorial University Medical Center, Savannah, Georgia; and 10Santa Rosa Memorial Hospital, Santa Rosa, California

Marcin Kowalski, MD1;  J. Brian DeVille, MD2;  J. Thomas Svinarich, MD3;  Dan Dan, MD4;  Andrew Wickliffe, MD5; Charan Kantipudi, MD5;  Jason D. Foell, PhD6; Giovanni Filardo, PhD, MPH2;  Reece Holbrook, BSEE6;  James Baker, MD7;  Hassan Baydoun, MD8;  Mark Jenkins, MD9;  Peter Chang-Sing, MD10
1Staten Island University Hospital, Northwell Health, Staten Island, New York; 2The Heart Hospital, Baylor Plano, Plano, Texas; 3St. Anthony Hospital, Lakewood, Colorado; 4InterMountain Healthcare, Salt Lake City, Utah; 5Piedmont Healthcare, Atlanta, Georgia; 5Piedmont Healthcare, Atlanta, Georgia; 6Medtronic, Inc., Minneapolis, Minnesota; 7Saint Thomas Hospital, Nashville, Tennessee; 8Staten Island University Hospital, Staten Island, New York; 9Memorial University Medical Center, Savannah, Georgia; and 10Santa Rosa Memorial Hospital, Santa Rosa, California

Atrial fibrillation (AF) is the most commonly diagnosed arrhythmia in the United States (US), with an annual incidence of 1.2 million in 2010, which is expected to increase to 2.6 million by 2030.1 Similarly, the prevalence of AF is expected to increase from 5.2 million in 2010 to 12.1 million by 2030.1 The increased prevalence of AF has led to an increase in economic burden, with an estimated annual incremental cost of $6.65 billion for the treatment of AF in the US.2 Catheter ablation is an established safe and effective treatment for AF, with pulmonary vein isolation (PVI) serving as a cornerstone of the ablation strategy.3,4 The number of AF ablations per year has steadily increased, from 3367 patients treated in 2000 to 12,006 patients treated in 2010.5

Catheter ablation procedures for AF are traditionally complex, long in duration, and require years of training before mastery is achieved. Conventionally, focal radiofrequency (RF) catheters have utilized a point-to-point ablation method, which can be associated with extensive ablation applications and increased procedure times.6 The cryoballoon is an anatomically designed catheter that has proven to be a safe and effective option for the treatment of patients with AF.7 Recent studies have demonstrated that the cryoballoon is associated with decreased procedure times compared with focal RF catheter ablation.6,8,9 However, the variability of the AF ablation procedure duration has not been previously investigated and few studies measure total electrophysiology laboratory (EP lab) occupancy. Consequently, the economic impact on hospitals from procedure duration, variability, and EP lab occupancy has not been evaluated. To understand the longer-term economic impact of procedural efficiencies, a discrete event simulation model is used in this current study. 

Methods

Patient population. The VALUE PVI study included 348 patients with paroxysmal AF who underwent ablation for AF with cryoballoon or focal RF catheters between January 1, 2011 and May 31, 2013 at seven high-volume ablation centers from geographically diverse sites within the US.8 Patient demographics, hospital resources, and procedural data were collected and reported in the VALUE PVI study.8 In the current study, the specific start and stop times of the ablation procedure for each VALUE PVI patient were evaluated using a discrete event simulation to assess the economic impact on hospitals associated with shorter AF ablation procedure times.

Discrete event simulation. A computer-based discrete event simulation was used to model the economic impact of changes in procedure/EP lab occupancy based on the variability of the ablation procedure start and stop times. Discrete event simulations are an established approach for the analysis of efficient use of resources in healthcare systems.10,11 The simulation was developed using SIMUL8 Professional version 21.0 (SIMUL8 Corporation). 

Model assumptions. The economic value model utilized assumptions of an AF ablation procedure based on the VALUE PVI study and feedback from hospital administrator surveys. Specifically, based on the average VALUE PVI procedure time and use of a block schedule, it was assumed that 2 PVI AF ablation procedures (either 2 RFs or 2 Cryoballoons) occur per day in an EP lab. The specifics of the block schedule information were derived from the raw VALUE PVI procedure time data and corroborated with the VALUE PVI investigators. In addition, other operational delays were assumed in the model, including: room set-up time prior to the procedure, availability of the electrophysiologist (EP), and the patient and EP co-availability prior to a procedure start. It is also assumed that the EP lab set-up for the second case cannot begin until the first procedure is complete and all other necessary resources will be available. Finally, there is an assumption that the EP lab resource can be used for another procedure when not being utilized for an AF ablation. If at least 60 minutes remain at end of the second ablation procedure on a given day, then it was assumed that another non-ablation procedure could be completed, including cardiac rhythm device implantation or replacement, for example. In the model, overtime was denoted as the amount of time that a procedure goes beyond the scheduled end of the regular shift.

Definition of model events. In the discrete event simulation, two separate simulation pathways were created: one to simulate AF ablation procedures performed with conventional RF catheters and the other for ablation procedures performed with cryoballoon catheters. In the simulation, the patient and EP paths into and out of the ablation procedure are represented by discrete events, which affected the start and stop times of the procedure (Figure 1). Specifically, individual patients entered the simulation at a time specified by the appropriate block schedule case start time for each line. Each patient was then subjected to a randomly assigned operational delay (most commonly 0 minutes, but it could range as high as 15 minutes) before being available to start the ablation case. Independent from the ablation case lines, the EPs in the simulation are kept busy with a series of tasks with randomly assigned lengths (most commonly 20 minutes, but ranging from 5-115 minutes). The ranges of delays were derived to best match the variability in the case start times observed in the VALUE PVI study. In the discrete event simulation model, if the patient was not available to start the case when the EP finished the preablation task, the EP would start a new task and not be available to start the case until the task was completed. Once the patient and EP are both available to start the case, the case time was randomly assigned based on the procedure time distributions from the VALUE PVI study, segregated by the technology that was used (RF or cryoballoon catheter ablation). 

Each “EP lab day” in the simulation represented 2 cases as described above. The block schedule for the second case for the cryoballoon line started 30 minutes earlier than the RF line based on the shorter average procedure time observed in VALUE PVI. If the first case of the day ended early, the second case still began at the time indicated by the block schedule. If the first case of the day ran so long as to overlap the beginning time of the second case (including set-up time), the second case was appropriately delayed. The simulation was programmed to track the following variables: (1) the number of overtime days, which were based on the days where cases resulted in the second case going past the normal shift end time; (2) cumulative overtime hours; and (3) number of days with additional EP lab usage, which occurred if the second case ended early (leaving enough time in a standard shift to do another EP lab procedure, including cardiac rhythm device management).

Model validation and sensitivity analysis. The economic model based on procedure time was validated by comparing the procedure start and stop times from the discrete event simulation to the data values recorded in the VALUE PVI study raw dataset. In addition, a sensitivity analysis was performed by varying the value of the mean procedure time by 15 minutes while keeping the procedure variability and block schedule times constant. This value was chosen as it represents approximately one-half of the mean procedure time difference between RF and cryoballoon procedures as reported in the VALUE PVI study.

Statistics. The discrete event simulation estimated the days in which the first case overlaps the second, days with overtime, cumulative overtime, and days with incremental procedures for patients undergoing ablation with cryoballoon or RF catheters. Gamma distributions were used to model RF and cryoballoon ablation procedure time variability. The simulation was run for 1000 days for each ablation technology, representing a total of 2000 RF cases and 2000 cryoballoon cases simulated (this represents an approximately 3-year workload for a large-volume AF ablation center). Also, the number of simulated cases was selected to ensure that the modeling output would not randomly depart from the measures recorded in the VALUE PVI study. With 2000 random samples from the gamma distributions, the sample mean is expected to be within ±1.7 minutes of the true mean for cryoballoon procedure times and within ±2.3 minutes for RF procedure times (with 95% confidence). Primary model outputs were reported as the number of days out of 1000 where conditions were met, as cumulative hours of overtime, and as average overtime on days where overtime occurred. Procedure time of day data were reported as mean and standard deviation. One VALUE PVI site had missing time of day data (because of hospital records release policy), so it was included for the procedure time analyses, but excluded from the time of day validation (due to the lack of actual time stamps in the data set). All descriptive raw procedure time data from the VALUE PVI study were reported as mean, 10th percentile, and 90th percentile. 

Results

Procedural parameters. The procedure times in the model are based on a deeper analysis of the raw procedure time data from the VALUE PVI study. Patients included were from the “Seven Cryo Sites” (n = 213) and the “RF” (n = 125) cohorts with validated procedure time data. The clinical characteristics, treatment details, and clinical outcomes of these patients were detailed in the original publication.8 EP lab occupancy used the original definition from VALUE PVI, which was the time denoted from the patient entrance into the procedure room until patient exit. While the VALUE PVI study reported an average reduction in lab occupancy time of 36 minutes with cryoballoon as compared with RF procedures, the detailed variability of the procedure times was not fully explored in the original manuscript. 

Cryoballoon lab occupancy times were shorter than RF across the distribution (cryoballoon mean = 246.8 ± 45.2 minutes; RF = 283.0 ± 52.7 minutes) (Figure 2A). The 10th percentile of lab occupancy times was 30 minutes shorter for cryoballoon vs RF procedures (cryoballoon = 198.0 minutes; RF = 228.0 minutes). The 90th percentile of lab occupancy times was 58 minutes shorter for cryoballoon vs RF procedures (cryoballoon = 304.0 minutes; RF = 361.6 minutes). The detailed procedure time distributions of the case times were found to best fit a one-sided gamma distribution (Figures 2B and 2C).

Block schedule times were determined from the VALUE PVI raw data procedure begin and end times, including: (1) scheduling a first case at 7:30 am for both catheters; (2) a second case at 12:30 pm for RF cases and 12:00 pm for cryoballoon cases; (3) the opportunity to begin an incremental procedure at 5 pm; and (4) the shift end/overtime begins at 6 pm (Table 1). The second case for cryoballoon was set 30 minutes earlier to reflect the shorter average procedure duration reported in the VALUE PVI study. A room switch/set-up time of 15 minutes was included in the simulation.

Impact on EP lab efficiency. The simulation using cryoballoon procedure times resulted in fewer days with overtime than RF procedures (422 days RF vs 60 days cryoballoon; absolute decrease 36.2%), and fewer cumulative hours of overtime over the full 1000-day simulation period (370 hours RF vs 27 hours cryoballoon; 92.7% reduction) (Figure 3). By dividing the cumulative hours of overtime by the number of days when overtime occurred, it was found that on days when overtime occurred, there was less average overtime with cryoballoon procedures as compared with RF (53 minutes RF vs 27 minutes cryoballoon). The simulation using cryoballoon procedure times also resulted in more days with at least 1 hour left in a normal shift (186 days RF vs 653 days cryoballoon; absolute increase 46.7%). During cryoballoon procedures, there were also fewer days found where the end time of the first ablation procedure interfered with the block scheduled begin time of the second case (554 days RF vs 461 days cryoballoon; absolute decrease 9.3%) (Table 2).

Model validation and sensitivity analysis. The procedure-time based economic model was validated by comparing the procedure start and stop times from the discrete event simulation to the values recorded in the VALUE PVI study raw dataset. The average actual vs simulated procedure start and stop times were comparable for “real-life” vs “simulated” cases (Table 3). There was close agreement between the actual and simulated lab occupancy times for both RF cases (283.1 ± 52.7 minutes VALUE PVI [real life] vs 278.4 ± 52.2 minutes [simulated]) and cryoballoon cases (246.8 ± 45.2 minutes VALUE PVI [real life] vs 240.7 ± 39.4 minutes [simulated]). The real “time of day” occurrences of patient entry to and exit from the lab were also in close agreement (Figure 4). The concurrence between the actual recorded lab occupancy times in the VALUE PVI study and the times modeled by the simulation were better than 97% in agreement. The sensitivity of the model was evaluated by varying the mean procedure time by 15 minutes while keeping the procedure variability and block schedule times constant. As shown in Table 4, varying model procedure times by 15 minutes resulted in consistent changes in the overlap of first and second cases, overtime, and incremental procedures.

Discussion

Main findings. The current study is the first to model the economic value resulting from procedural efficiencies associated with AF ablation performed with the cryoballoon catheter. This simulation demonstrates that the cryoballoon procedure is associated with some hospital cost benefits, including: (1) fewer days of overtime; (2) fewer cumulative overtime hours; (3) less variability of procedure duration; and (4) an increase in incremental procedures compared with focal RF procedures.

Procedure efficiency. Several single-center studies have shown that the procedure duration is shorter for cryoballoon compared with focal RF procedures.6 A recent meta-analysis of fourteen studies comparing cryoballoon and focal RF procedures showed that the cryoballoon catheter significantly reduced the total procedure time by a weighted mean of 29.65 minutes.6 It is remarkable that a small difference in procedure duration in the simulation translated into a significant burden on hospital resources over a longer duration of operation used in this discrete event simulation. An analysis of the VALUE PVI procedure times revealed the distributions do not follow a normal distribution curve. The procedure variability is demonstrated in Figures 2B and 2C, which shows right-sided skewness in the distribution of procedure duration. For more accurate descriptive purposes (when procedure times are reported in studies), they should be described in adequate detail (eg, reporting 10th/90th percentiles, which will decrease the masking of this variability when only mean data are reported). Importantly, this more rigorous description of the data shows that the longest cases of the cryoballoon procedure are nearly 60 minutes faster than the longest cases of the RF procedure. 

Recent studies have evaluated the impact of procedure efficiency on hospital resources.6,8,9 The VALUE PVI study was the first multicenter study to compare specific resources utilized in cryoballoon and focal RF procedures. Subsequently, the FAST PVI study showed that anatomically designed catheters (cryoballoon and pulmonary vein ablation catheter) are associated with faster procedure times and decreased hospital resources compared with focal RF procedures.9 Both studies suggest that the potential mechanism underlying the procedure time differences between focal RF and the cryoballoon procedure are due to the anatomical approach of the cryoballoon catheter, including less time creating three-dimensional electroanatomical maps, less time spent manipulating the catheter to new locations (no point-by-point movement), and less total ablation application time.

Ultimately, the decrease in overtime and increased predictability of procedure time/duration associated with the cryoballoon procedure may have positive benefits on overall hospital resources, including the ability to treat additional patients, the potential to reduce job stress on the EP staff (by lessening unpredicted overtime, which impacts not only one resource, but the full staff for the case going into overtime), and the potential to reduce support staff burnout and turnover.

Study limitations. The study limitations include the retrospective manner of data collection. The data were collected from only 7 centers, and the VALUE PVI data may not fully represent the true average and variability of procedure times across many users/hospital systems. Also, physicians examined in the VALUE PVI study are advanced users of the cryoballoon catheter system. In addition, the model makes assumptions, such as a fixed block schedule and operational delays, which may vary by institution. There are no studies available to do direct comparisons of the measures predicted by the model, although one benefit of modeling is to understand impacts that are not easily measured in the real world. 

The majority of the data (that were modeled in this study) were generated from the previous generation of catheters, including 94% from the Medtronic first-generation cryoballoon and 92% from the Biosense Webster non-contact force focal RF catheters. Evolving data from the next generation of cryoballoon and contact-force focal RF catheters suggest that these catheters will have improved procedural efficiencies compared with the preceding catheters. Nonetheless, it would be anticipated based on the initial published data on second-generation cryoballoon and contact-force focal RF catheters that the procedural efficiencies and positive economic impact identified in the current study will remain.12,13,14 However, further comparative studies similar to VALUE PVI with the next generation of cryoballoon and contact-force sensing focal RF catheters would be needed to confirm these findings.

Lastly, there is a potential cost difference in the materials (disposable and capital) used in cryoablation procedures compared with focal irrigated RF procedures, which impacts the total cost of the AF ablation procedure. The modeling of these material costs presents a significant challenge with the complexities of variable institutional pricing in addition to the degree of utilization of different adjunctive materials (eg, three-dimensional mapping, intracardiac echocardiography, contrast agent, etc). Therefore, we focused the objective of the study on assessing the economic value associated with a measure (procedure time) that is better understood and widely reported in AF ablation studies.

Conclusion

The current study developed a discrete event simulation to model the data from the VALUE PVI study. This model predicts meaningful economic benefits in terms that are understandable and actionable to those who have to make choices about how to best use EP lab resources and align overall hospital resource spending. The benefits of the cryoballoon procedure that were demonstrated in this study included more days where overtime is avoided, less cumulative overtime hours, and more days with time left for additional EP lab usage. This is the first study to translate the economic value of shorter procedure time and reduced variability of procedure duration associated with the cryoballoon procedure.

Disclosure: Dr. Kowalski is the lead investigator of this study and is a faculty member who receives compensation for physician teaching programs with Medtronic, Inc. Dr. DeVille served as a faculty member, receiving compensation for physician teaching programs with Medtronic, Inc., Stereotaxis, and Biosense Webster, and participates in research studies with each of these companies. Dr. Svinarich serves on an advisory board for Medtronic, Inc., and receives honoraria for teaching and speaking for Medtronic, Inc., Forest, and Sanofi-Aventis. Dr. Dan is a physician-training consultant for Medtronic, Inc., and is a speaker for Medtronic, Inc, St. Jude Medical, and Sorin. Dr. Wickliffe is a physician-training consultant for Medtronic, Inc. Dr. Foell and Reece Holbrook are employees of Medtronic, Inc. Dr. Baker is a faculty member who receives compensation for physician-teaching programs with Medtronic, Inc. Dr. Chang-Sing serves on the national EP advisory board for Medtronic, Inc., and he participated as PI and co-investigator for several Medtronic, Inc. clinical trials. Drs. Kantipudi, Filardo, Baydoun, and Jenkins report no disclosures regarding the content herein. 

Reprinted with permission from J Invasive Cardiol. 2016;28(5):176-182. 

References

  1. Colilla S, Crow A, Petkun W, et al. Estimates of current and future incidence and prevalence of atrial fibrillation in the US adult population. Am J Cardiol. 2013;112:1142-1147.
  2. Coyne KS, Paramore C, Grandy S, et al. Assessing the direct costs of treating nonvalvular atrial fibrillation in the United States. Value Health. 2006;9:348-356.
  3. Calkins H, Kuck KH, Cappato R, et al. 2012 HRS/EHRA/ECAS expert consensus statement on catheter and surgical ablation of AF: recommendations for patient selection, procedural techniques, patient management and follow-up, definitions, endpoints, and research trial design. Heart Rhythm. 2012;9:632-696.e621.
  4. Haissaguerre M, Jais P, Shah DC, et al. Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins. N Engl J Med. 1998;339:659-666.
  5. Deshmukh A, Patel NJ, Pant S, et al. In-hospital complications associated with catheter ablation of atrial fibrillation in the United States between 2000 and 2010: analysis of 93 801 procedures. Circulation. 2013;128:2104-2112.
  6. Xu J, Huang Y, Cai H, et al. Is cryoballoon ablation preferable to radiofrequency ablation for treatment of atrial fibrillation by pulmonary vein isolation? A meta-analysis. PLoS One. 2014;9:e90323.
  7. Packer DL, Kowal RC, Wheelan KR, et al. Balloon ablation of pulmonary veins for paroxysmal atrial fibrillation: first results of the North American Arctic Front (STOP AF) pivotal trial. J Am Coll Cardiol. 2013;61:1713-1723.
  8. DeVille JB, Svinarich JT, Dan D, et al. Comparison of resource utilization of pulmonary vein isolation: cryoablation versus RF ablation with three-dimensional mapping in the Value PVI study. J Invasive Cardiol. 2014;26:268-272.
  9. Klein G, Lickfett L, Schreieck J, et al. Comparison of ‘anatomically designed’ and ‘point-by-point’ catheter ablations for human atrial fibrillation in terms of procedure timing and costs in German hospitals. Europace. 2015;17:1030-1037. Epub 2015 Feb 5.
  10. Laker LF, Froehle CM, Lindshell CJ, et al. The Flex Track: flexible partitioning between low- and high-acuity areas of an emergency department. Ann Emerg Med. 2014;64:591-603.
  11. Caro JJ, Ward A, Deniz HB, et al. Cost-benefit analysis of preventing sudden cardiac deaths with an implantable cardioverter defibrillator versus amiodarone. Value Health. 2007;10:13-22.
  12. Di Giovanni G, Wauters K, Chierchia GB, et al. One-year follow-up after single procedure cryoballoon ablation: a comparison between the first and second generation balloon. J Cardiovasc Electrophysiol. 2014;25:834-839.
  13. Pandya B, Sheikh A, Spagnola J, et al. Safety and efficacy of second-generation versus first-generation cryoballoons for treatment of atrial fibrillation: a meta-analysis of current evidence. J Interv Card Electrophysiol. 2016;45:49-56. Epub 2015 Nov 18.
  14. Shurrab M, Di Biase L, Briceno DF, et al. Impact of contact force technology on atrial fibrillation ablation: a meta-analysis. J Am Heart Assoc. 2015;4:e002476.
/sites/eplabdigest.com/files/kowalski.pdf