Health System Perspective on Cost for Delivering a Decision Aid for Prostate Cancer Using Time-driven Activity-based Costing

Abstract

Background: 

Previsit decision aids (DAs) have promising outcomes in improving decisional quality, however, the cost to deploy a DA is not well defined, presenting a possible barrier to health system adoption.

Objectives: 

We aimed to define the cost from a health system perspective of delivery of a DA.

Research Design: 

Observational cohort.

Patients and Methods: 

We interviewed or observed relevant personnel at 3 institutions with implemented DA distribution programs targeting men with prostate cancer. We then created process maps for DA delivery based on interview data. Cost determination was performed utilizing time-driven activity-based costing. Clinic visit length was measured on a subset of patients. Decisional quality measures were collected after the clinic visit.

Results: 

Total process time (minutes) for DA delivery was 10.14 (UCLA), 68 (Olive View-UCLA), and 25 (Vanderbilt). Total average costs (USD) per patient were $38.32 (UCLA), $59.96 (Olive View-UCLA), and $42.38 (Vanderbilt), respectively. Labor costs were the largest contributors to the cost of DA delivery. Variance analyses confirmed the cost efficiency of electronic health record (EHR) integration. We noted a shortening of clinic visit length when the DA was used, with high levels of decision quality.

Conclusions: 

Time-driven activity-based costing is an effective approach to determining true inclusive costs of service delivery while also elucidating opportunities for cost containment. The absolute cost of delivering a DA to men with prostate cancer in various settings is much lower than the system costs of the treatments they consider. EHR integration streamlines DA delivery efficiency and results in substantial cost savings.

Many professional societies recommend shared decision-making (SDM) in evidence-based guidelines for quality care. Included in these is the American Urological Association, which cites “grade A” evidence for its use and adoption into routine clinical practice in counseling men with localized prostate cancer.¹ However, despite the demonstrated benefits of SDM, including enhanced patient knowledge regarding options and reduced decisional conflict, formal adoption of the approach has been incremental. Barriers to formal adoption of the approach include physician concerns regarding potential increased time in counseling among other items^1,2. One potential solution to this concern has been the deployment of previsit decision aids (DAs) to facilitate SDM. Previsit DAs have been implemented with similar outcomes in decisional quality but the incremental cost to deploy a SDM program is currently unknown and a potential barrier to widespread implementation.³

Many studies have assessed the value of DAs in aligning preferences with the use of invasive procedures. SDM represents an avenue towards reducing unnecessary health system utilization that translates to an economic benefit to the bearer of financial risk. A study completed in orthopedic surgery analyzed DA’s impact on the management of hip and knee osteoarthritis, ultimately finding that DA implementation led to a significant reduction in rates of elective surgery (hip replacement), lessening mean treatment-associated costs by 19% in the DA intervention group versus control.⁴ Furthermore, Loveman and colleagues suggested that increased focus on educational materials as part of diabetes care correlates with improved control of type 1 diabetes. The cost of these materials, £500−£600 per patient (as reported by sponsors), was found to be cost-effective with model analysis suggesting incremental cost savings ranging from £2579 to £3012 with added health benefits in the form of significantly improved HbA1c and life years saved over a 10-year time horizon.⁵ While these studies have assessed the reduction of clinical-related costs and financial strain on the health care system, cost accounting methodology applied to determine all-inclusive costs required to develop, implement, and sustain DA distribution and incorporation into the standard of care is scarce.

Time-driven activity-based costing (TDABC) is a costing methodology that can be used to identify inefficient business processes in health care settings characterized by multidisciplinary collaborative care delivery and complex billing and insurance policies^6,7. TDABC can accurately assign expenses of personnel compensation, equipment, space, material, and device/software technology to patient treatments and counseling.⁷ In a complex multipayer US system with varying reimbursement rates, TDABC enables users to estimate the cost of the delivery of health care services as opposed to those directly assessed to the payer or patient. As such, TDABC’s 7-step method is a promising tool for reliable, accurate cost accounting, and has been validated in several health care settings^8,9.

In this work, we aim to develop a process map for an SDM intervention based on a previsit DA and provide a foundation for understanding the return on investment in SDM as it relates to improved patient education regarding treatment options, decisional satisfaction, and implementation of a value-based health care model. TDABC can augment iterative quality improvement and cost optimization SDM efforts by demonstrating how changing personnel assignments and workflow practices can reduce the total cost of SDM delivery. In this paper, we use TDABC to determine the true cost of SDM delivery across 3 health care settings, with varying patient demographics and resource constraints, and outline infrastructure opportunities for cost containment. We hypothesize that compared with the cost of prostate cancer treatments that may not align with patient values, goals, or clinical indications, the measured cost of a pretreatment SDM intervention is low.

METHODS

Study Institutions

We performed an observational analysis of SDM programs targeting men with localized prostate cancer. The analysis was performed at 2 academic medical centers [UCLA Ronald Reagan Medical Center and Vanderbilt University Medical Center (VUMC)] and an academically affiliated safety net county hospital (Olive View-UCLA Medical Center).

Process Mapping and Data Acquisition

The SDM intervention was to issue a previsit DA to patients with newly diagnosed localized prostate cancer. At one institution (UCLA), the intervention was integrated into the electronic medical record (EMR). At Vanderbilt, the DA was issued by a staff member using a non-EMR software interface. At the County hospital, the DA was delivered electronically on an iPad in the waiting room but switched to telemedicine delivery mode. Details of the DA and decisional quality outcomes have been reported.¹⁰ We interviewed and observed a team of 3 urology faculty attending physicians and administrative staff including departmental chief financial officers, operations administrators, and clinic assistants to gain an understanding of current workflow practices, personnel, space, materials, and devices required for SDM module delivery. From this information, we created a process map of the step-by-step sequence of tasks, starting with screening patient eligibility for SDM module assignment and concluding with attending physician review of SDM individualized patient results.

Costing Methodology

We followed the TDABC methodology, as described by Kaplan and Porter⁷ of Harvard Business School. TDABC measures the costs of all resources—personnel time, space, devices, and materials—used to perform a specific process or service. TDABC implementation begins with the creation of process maps that portray the service delivery as a sequence of steps (eg, scheduling, procedure room set up, procedure, procedure room cleanup, and patient monitoring). For each step, costs from 4 primary resource expenditures (space, personnel, devices, and materials) are divided by the practical capacity of each resource. The resulting capacity cost rate for each resource is multiplied by its estimated time duration from the process maps. The additive costs of each step of the process map represent the overall procedure cost.

Cost Measurement Analysis

Capacity cost rates were derived for each personnel type, space, device, and materials used. Personnel costs included salary and fringe benefits costs as well as other expenditures, such as administrative support costs, malpractice insurance costs, and additional role-specific costs. The cost information was entered and accumulated in an Excel-based TDABC Model Template.

Physician and other personnel salaries were requested from human resources staff and incorporated into costing calculations as averages. Space-related costs consider new building construction costs, square footage, useful life, annual maintenance, and housekeeping costs. Spaces were assumed to have a 20-year useful lifespan before any renovation requirement. Device-specific (eg, computer workstations and reception printers) costs include acquisition cost, useful life, depreciation, and service contract costs. Material costs, where applicable, are calculated based on the purchase price of one-time consumable use and the quantity of each consumable required for DA delivery to the patient.

Each respective cost calculation was subsequently divided by its respective available capacity. This quantity (in minutes) was defined as the time that a specific resource was considered operational and involved in productive work. Duration of each step in the process map was manually timed and recorded and subsequently reported as an average length to completion (in minutes) over several observations of each workflow being completed. Space and device-related capacities are based on the number of hours the site or device is utilized as a part of regular working business and clinic operations. Personnel-specific capacities are calculated based on the entire calendar year with unavailable days excluded to account for weekends, vacations, allocated sick leave, and academic (continued medical education) days. (Tables 1,2, and 3) outlines the resources expended that contribute to the overall cost of each process map step.

TABLE 1 - Expended Resources Categorized by Process Map Step at UCLA Ronald Reagan Medical Center

Activity name	Space required	Device(s) required	Personnel involved
1. Screen weekly for new prostate cancer patients	Cubical office	PC Workstation Computer	Administrative assistant
2. SDM tool ordered per patient in EMR system	Cubical office	PC Workstation Computer	Administrative assistant
3. Send MyChart message notification	Cubical office	PC Workstation Computer	Administrative assistant
4. Patient completes SDM	—	—	—
5. Review of patient SDM results before the scheduled appointment	Examination room	PC Workstation Computer	Attending physician

EMR indicates electronic medical record; SDM, shared decision making.

TABLE 2 - Expended Resources Categorized by Process Map Step at Olive View Medical Center

Activity name	Space required	Device(s) required	Personnel involved
1. Identify eligible patients for SDM administration using screening criteria (weekly)	Physician office	PC Workstation Computer	Resident physician
2. Patient called and educated through phone on the utility of SDM as part of postbiopsy or e-consult follow-up	Cubical office	PC Workstation Computer	Administrative assistant
3. SDM patient account created	Cubical office	PC Workstation Computer	Administrative assistant
4. Patient called to schedule time to complete the SDM module	Cubical office	PC Workstation Computer	Administrative assistant
5. Patient completes SDM with assistance from LVN	Waiting room	iPad tablet	LVN
6. LVN prints SDM results in the clinic	Reception desk area	Receptionist office printer	LVN
7. Review of patient SDM results before the scheduled appointment	Examination room	PC Workstation Computer	Attending physician

LVN indicates licensed vocational nurse; SDM, shared decision making.

TABLE 3 - Expended Resources Categorized by Process Map Step at VUMC

Activity name	Space required	Device(s) required	Personnel involved
1. Identify eligible patients for SDM administration using screening criteria (weekly)	Cubical office	PC Workstation Computer	Clinic coordinator
2. SDM patient account created for eligible patients	Cubical office	PC Workstation Computer	Clinic coordinator
3. Patient completes SDM	—	—	—
4. SDM module results scanned into EMR, physician notified through appointment note	Cubical office	PC Workstation Computer	Clinic coordinator
5. Appointment addendum created to ensure MD aware of SDM results	Cubical office	PC Workstation Computer	Clinic coordinator
6. Review of patient SDM results before the scheduled appointment	Examination room	PC Workstation Computer	Attending physician

EMR indicates electronic medical record; MD, doctor of medicine; SDM, shared decision making; VUMC, Vanderbilt University Medical Center.

Cost Variance Analysis

Cost differences between sites can arise from multiple sources: differences in the compensation of personnel, variation in efficiency and productivity, and differences in the skill mix of personnel when delivering the same service. We used the accounting approach of variance analysis to measure the impact of each of these 3 sources of cost variation.¹¹ By separating out the effects of input prices and quantity of minutes used at 2 different institutions with similar personnel used but different workflows and assignments, users can determine the total potential amount of cost savings if efficient practices of a benchmark model at “site B” were replicated and instituted at “site A”.¹¹ For variance calculation purposes, we defined UCLA as the benchmark (site B) because it had the lowest overall process time and the average cost of SDM delivery. Vanderbilt University was chosen as site A due to its similar number and type of personnel in the workflow. Price variance equals (total process time at site B)*(average capacity cost rate at site A–average capacity cost rate at site B). Quantity variance equals (average capacity cost rate at site B)*(total process time at site A–total process time at site B).

Access Implications Analysis

Access to care is of paramount importance to the leaders of the Los Angeles County Department of Health Services, who oversee Olive View-UCLA Medical Center. We evaluated the implications for access improvement that accompany cost investment into the SDM program by measuring the time for consultation before and after implementation of the SDM program as well as the number of clinic visits needed for patients to arrive at their treatment decisions.

RESULTS

Shared Decision-making Module Delivery at UCLA Ronald Reagan Medical Center

The entire workflow at UCLA Medical Center included 5 total steps, each associated with a 100% activity probability (Fig. 1). Total workflow completion time from patient recruitment to attending physician review of patient individualized SDM results was 10.14 minutes excluding time for patient module completion, which was assumed to be completed asynchronously and noncontributory to overall costs. All steps are performed on a per-patient basis; therefore, time and cost expenditures are multiplied by the average new clinic patient volume to arrive at weekly and annual cost estimates.

FIGURE 1: Process map of DA delivery at Ronald Reagan UCLA Medical Center. Personnel category: “a” administrative assistant, “b” attending physician, steps 1–3 utilize cubical office space, step 5 utilizes examination room space. DA indicates decision aid; SDM, shared decision-making.

The entire process required contributions from 2 personnel roles: attending physicians and administrative assistants. Table 1 outlines space, material, and device resources expended in the workflow. The process map resulted in 3 overarching delivery steps: patient recruitment, outreach, and physician review of SDM results post-patient completion.

Cost Determination at UCLA Ronald Reagan Medical Center

Cost determination for the entire process at UCLA Ronald Reagan Medical Center is reported in Table 4. The total average cost per patient, in which DA is delivered is $38.32. A weekly clinic volume of 7 new and eligible prostate cancer patients was assumed based on a review of past scheduling. Total weekly costs for SDM for average clinic volume is $268.20, corresponding to an estimated annual cost of $12,230.10. Personnel costs represented the most significant cost expenditure, whereas costs related to purchase, service, and continued utilization of devices (computer workstation) was the lowest contributor to the total cost. The cost of acquisition and associated licensing fees of the SDM platform and software itself were excluded from calculations due to cost variability dependent on vendor and hospital institution negotiations.

TABLE 4 - Cost Determination of SDM Delivery at UCLA Ronald Reagan Medical Center by TDABC Expense Category

SDM delivery total costs (UCLA Ronald Reagan)
Expense category	Cost (USD);$	Percent total cost (%)
Personnel	37.76	98.54
Space/devices	0.56	1.46
Material	—	—
Total average cost per patient	38.32	100
Weekly cost: $268.20 Annual cost: $12,230.10 *Weekly and annual costs based on the projected clinical volume of 7 patients/week, for which SDM is delivered

SDM indicates shared decision-making; TDABC, time-driven activity-based costing.

Shared Decision-making Module Delivery at Olive View Medical Center

The entire workflow at Olive View Medical Center included 7 total steps, each with a 100% activity probability with the exception to step 5, which carries an 80% activity probability, respectively (Fig. 2). Total workflow completion time from patient recruitment to attending physician review of patient individualized SDM results was 68 minutes.

FIGURE 2: Process map of DA delivery at Olive View-UCLA Medical Center. Personnel category: “a” administrative assistant, “b” attending physician, “c” resident physician, and “d” licensed vocational nurse, step 1 utilizes physician office space, steps 2–4 utilize cubical office space, step 5 utilizes waiting room space, step 6 utilizes reception desk area, and step 7 utilizes examination room space. LVN indicates licensed vocational nurse; SDM, shared decision-making.

The entire process required contributions from 4 personnel roles: attending and resident physicians, administrative assistants, and licensed vocational nurses. Table 2 outlines space, material, and device resources expended in the workflow.

Cost Determination at Olive View Medical Center

Cost determination for the entire process completion at Olive View Medical Center is reported in Table 5. The total average cost per patient is $59.96. A weekly clinic volume of 3 new and eligible prostate cancer patients was assumed based on a review of past scheduling. Total weekly costs for SDM initiation and completion for average clinic volume is $179.87, corresponding to an estimated annual cost of $8,202.25.

TABLE 5 - Cost Determination of SDM Delivery at Olive View Medical Center by TDABC Expense Category

SDM delivery total costs (Olive View)
Expense category	Cost (USD); $	Percent total cost (%)
Personnel	58.92	98.26
Space/devices	1.04	1.74
Material	—	—
Total average cost per patient	59.96	100
Weekly cost: $179.87 Annual cost: $8,202.25 *Weekly and annual costs based on the projected clinical volume of 3 patients/week, for which SDM is delivered

SDM indicates shared decision-making; TDABC, time-driven activity-based costing.

We also estimated the improvement in clinic capacity and related access that accompanied the SDM intervention. We found that the median visit time with a physician decreased from 33 to 23.5 minutes (P = 0.0096) and that the proportion of patients who needed more than one visit to make a decision decreased from 46% to 13% (P < 0.0001)

Shared Decision-making Module Delivery at Vanderbilt University Medical Center

The entire workflow at VUMC included 6 total steps, each associated with a 100% activity probability (Fig. 3). Total workflow completion time from patient recruitment to attending physician review of patient individualized SDM results was 25 minutes.

FIGURE 3: Process map of DA delivery at Vanderbilt University Medical Center. personnel category: “b” attending physician, “e” clinic coordinator, steps 1–2, 4–5 utilize cubical office space, and step 6 utilizes examination room space. EMR indicates electronic medical record; MD, doctor of medicine; SDM, shared decision-making.

The entire process required contributions from 2 personnel roles: attending physicians and clinic coordinators. Table 3 outlines space, material, and device resources expended in the workflow.

Cost Determination at Vanderbilt University Medical Center

Cost determination for the entire process completion at VUMC is reported in Table 6. The total average cost per patient is $42.38. A weekly clinic volume of 6 new and eligible prostate cancer patients was assumed based on a review of past scheduling. Total weekly costs for SDM initiation and completion for average clinic volume is $254.26, corresponding to an estimated annual cost of $11,594.09.

TABLE 6 - Cost Determination of SDM Delivery at VUMC by TDABC Expense Category

SDM delivery total costs (VUMC)
Expense category	Cost (USD); $	Percent total cost (%)
Personnel	40.65	95.92
Space/devices	1.73	4.08
Material	—	—
Total average cost per patient	42.38	100
Weekly cost: $254.26 Annual cost: $11,594.09 *Weekly and annual costs based on the projected clinical volume of 6 patients/week, for which SDM is delivered

SDM indicates shared decision-making; TDABC, time-driven activity-based costing; VUMC, Vanderbilt University Medical Center.

Cost Variance Analysis

We defined UCLA Ronald Reagan as the most “mature” and efficient workflow of all our study institutions, as it has been using the DA the longest and has achieved EMR integration. UCLA’s total personnel only cost was $2.98 lower than Vanderbilt University's. Although this difference is small, the variance analysis revealed significant differences in the underlying cost drivers at the 2 institutions. The average cost capacity rates at UCLA and Vanderbilt are $3.72/min and $1.63/min, respectively, leading to a negative price variance of $21.21, a positive quantity variance of $55.25, and an efficiency variance of $24.19 between the two institutions.

DISCUSSION

In this work, we developed process maps outlining activities required for SDM delivery across 3 institutions. These process maps were then utilized to determine holistic average costs of SDM delivery as well as estimated weekly and annual costs based on typical patient clinical volumes. Our TDABC accounting assessment demonstrates a stepwise increase in the cost of delivery of similar services (irrespective of clinical volume) with Ronald Reagan UCLA being the least expensive, followed by VUMC and finally Olive View-UCLA Medical Center. Across all institutions, the largest cost contributor to SDM delivery is related to the compensation of involved personnel. Nevertheless, the costs associated with SDM are more than offset by other benefits, such as increased access and efficiency. For example, with our findings of 10 minutes of physician-involved consultation saved when SDM is utilized, using our calculated capacity cost rate of $3.55/min for attending-level physicians, this represents a $35.50 savings per patient from the health care system perspective. More broadly, our calculated per-patient SDM delivery cost is minimal compared with that estimated of the wide treatment options for prostate cancer. Laviana et al¹² demonstrated costs ranging from $7298 for active surveillance to $23,565 for external-beam radiation therapy at 5 years, with robotic prostatectomy notably costing $16,946.

SDM delivery can vary widely in time and cost expenditure, even if personnel types involved remain similar across sites. While all sites broadly utilized 2 personnel groups of physicians (residents and attendings) and administrative assistants, with the addition of a licensed vocational nurse at Olive View Medical Center, variation in costs are observed due to differing automation of steps in service delivery. For example, UCLA Ronald Reagan represents a more “mature” and streamlined process map compared with other studied institutions. A notable driver of cost savings at this institution is due to EMR integration of SDM screening and ordering. This integration allows personnel to screen with an EMR filter based on new prostate cancer patient appointment type, effectively circumventing manual personnel labor in screening medical records, pathology reports, and laboratories of relevant patients currently being performed at other sites. From an administrative perspective, we suggest that TDABC is a powerful tool that elucidates the cost implications of operational differences in SDM workflows, ultimately serving as a strong justification for SDM EMR integration to deliver maximum value to patients and health delivery systems.

The calculated efficiency variance of $24.19 per patient represents the total cost savings if efficiencies at UCLA were transferred to Vanderbilt. This savings figure is representative of a baseline average and is, therefore, amplified when typical clinic patient volumes are considered. This demonstrates that an overall negative price variance (higher personnel-related salary costs at UCLA compared with Vanderbilt) is overcome by a large disparity in efficiency and thus total process time (10.14 vs 25 min), emphasizing the importance of achieving a streamlined workflow that optimizes time expenditure and skill-level required of personnel to complete a given task. The efficiency variance serves as a valuable metric that administrators and departmental leadership can use to enact best practices (personnel involved and technology integration to reduce process time) from a model institution, with lower efficiency variances representing productive and cost-containing workflow modifications.

Costs to deliver the intervention ranged from $38.32 to $59.96 across our sites, and personnel costs were the largest component. Physician time to review the DA report was the largest component of personnel-related costs. We conceptualized and included this as part of the cost calculation of the intervention, but in a high-quality prostate cancer consultation, professional guidelines from the American Urological Association recommend assessing patient preferences for clinical outcomes. Thus, time spent before the visit for this purpose may shorten the clinical encounter, or could conceptually be included in the time for the encounter and not part of the intervention cost.

Although cost considerations are critical consideration for hospital administrators, SDM may help with access to care in resource-constrained environments. For example, our study included a county-level safety net hospital that serves a diverse patient population when compared with partner sites in terms of demographics, such as socioeconomic status, native language, technology availability, and usage competency. Due to the dynamic nature of providing a service that fits the needs of a specific patient population, the process maps created for the TDABC application are useful in determining how iterative and adaptive changes in personnel, materials, devices, and space affect the cumulative cost of SDM delivery. Our findings support that investment in these interventions yields positive outcomes that improve access, namely a significant reduction in the number and length of appointments required to reach a treatment decision for patients who received the SDM intervention.

Calculated costs are inevitably dynamic and subject to differences in health care structure and delivery, such as regional provider pay differences and availability of mid-level provider care support. However, through demonstrating a methodical process to elucidate health care system-facing costs of shared decision-making module delivery, we propose such approaches are helpful to iteratively assess cost trends and capture opportunities for health systems to efficiently lower service delivery costs. Our range of established per-patient costs of $38.32–$59.96 is within the range of existing literature regarding patient willingness to pay for decision support. Wilson et al¹³ report a median willingness to pay value ranging from $25 to $50, across men receiving usual care (websites and educational support materials) or being enrolled in a personal patient profile-prostate tailored web-based decision support system for men with newly diagnosed prostate cancer.

Our findings may interest stakeholders interested in providing support for SDM across several conditions. Though this DA is focused on prostate cancer, the workflow processes supporting it may be similar to those required for patients considering back surgery, coronary stent placement, or other high-cost procedures. As such, our cost estimates here may be helpful. In addition, as some payors, such as Medicare, begin to require evidence of SDM for services, such as percutaneous left atrial appendage closure, our data may provide a basis for payment for the costs incurred in delivering SDM programs.

Our study has several limitations. To fully and accurately capture an all-inclusive cost of SDM delivery, the acquisition cost of an SDM platform license is required. This cost was excluded from our calculations as quotes are highly variable and reliant on vendor-hospital negotiations. In addition, these costs will be readily apparent to health systems, whereas delivery costs are often harder to ascertain. In addition, the goal of sustainable, standard-of-care use of SDM requires seamless EMR integration including patient eligibility screening and automated module delivery. As health systems utilize varying EMRs, costs associated with information technology services and troubleshooting should be considered and are not captured in our calculated costs. Another limitation was the assumptions made regarding space and maintenance-related costs. If original build costs were unavailable, space costs were derived from new planned construction costs divided by total square footage, which may not always represent an accurate estimate. Lastly, when maintenance and utility records were not readily accessible, these figures were assumed to be comparable across geographical markets.

Our work fills a gap in the current understanding of the total resource and financial cost expenditure required to implement and sustain SDM delivery as a routine standard of care efficiently and practically. From a managerial perspective, we validate the utility of process mapping and TDABC accounting to improve operational efficiency and cost reduction. The overarching goal of such research is to facilitate the development of value-based health care models. As health care incorporates automated screening processes, big data and analytics, and technological innovations, our approach serves as a foundation to analyze whether proposed interventions truly result in cost savings and reductions in efficiency variance. Further multicenter cost-benefit analysis to juxtapose cost savings associated with a reduced number of patient visits and visit length against the incurred cost of delivery can be performed to augment the justification of SDM implementation.