Professional Documents
Culture Documents
Costs of Hospitalizations
Discussion Paper
June 2012
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
About CHSPR
The Centre for Health Services and Policy Research (CHSPR) is an independent research centre
based at the University of British Columbia. CHSPRs mission is to advance scientific enquiry
into issues of health in population groups, and ways in which health services can best be
organized, funded and delivered. Our researchers carry out a diverse program of applied health
services and population health research under this agenda. The Centres work is:
Independent
Population based
Policy relevant
Interdisciplinary
Privacy sensitive
CHSPR aims to contribute to the improvement of population health by ensuring our research is
relevant to contemporary health policy concerns and by working closely with decision makers to
actively translate research findings into policy options. Our researchers are active participants in
many policy-making forums and provide advice and assistance to both government and nongovernment organizations in British Columbia (B.C.), Canada and abroad.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Table of Contents
Executive Summary ........................................................................................................................ 3
Introduction ..................................................................................................................................... 5
Methods........................................................................................................................................... 6
Assumptions................................................................................................................................ 7
Day Surgery .......................................................................................................................................... 7
Inpatient Care ...................................................................................................................................... 12
Data ............................................................................................................................................... 12
Activity Based Costing ............................................................................................................. 15
Background and Principles ................................................................................................................. 15
Example: A Patients CT Scan............................................................................................................ 17
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Executive Summary
The primary objective of this project is the development of methodologies to estimate the
marginal costs of hospitalizations for different types of patients, both for day surgery and acute
inpatient. In this projects report, marginal cost refers to the change in total hospital
expenditures associated with treating one additional patient.
This analysis is motivated by the informational needs of provincial Ministries of health. The
analysis and interpretation assumes that the primary users of the findings will be provincial
Ministries of health and other significant purchasers of health care services, such as the B.C.
Health Services Purchasing Organization (HSPO).
Estimating marginal costs of hospitalized patients is complex; hospitals purchase a broad array
of equipment and services to provide care to patients. This includes nursing care and
sophisticated diagnostic and therapeutic equipment. Accordingly, a range of data sources is
required in order to derive estimates of patients marginal cost. These complementary datasets
include statistical and financial data from hospitals chart of accounts, activity-based costing
data, workforce data and lastly, clinical, demographic and administrative data abstracted from
the patients chart and submitted to the Canadian Institute for Health Information (CIHI). British
Columbia (B.C.) and Ontario have consented to have their anonymous data included in this
analysis. Ethics approval was provided by the University of British Columbia.
The report will first describe the methodology used to estimate marginal costs of patients
hospitalizations. Since the concept of marginal cost is based on the assumptions of the
availability of labour and equipment to treat additional patients, deriving estimates of marginal
costs are subject to important hospital-specific issues of constraints. Consequently, estimates of
marginal cost will be presented for an array of different assumptions which emulate operating
conditions within hospitals, each of which has critical operational and policy implications.
The first scenario assumes that there is currently excess capacity of most, if not all, hospital
resources (e.g., beds, OR time, nurses, imaging technologists, therapists) required to
accommodate one additional case. Is this assumption reasonable considering the wide-spread
existence of extensive hospital wait lists? In other words, is it logical to assume that hospitals
currently have already-paid-for resources that remain idle and are not being used? Without
passing judgment on the validity of this assumption, estimates of marginal costs are derived.
In the second scenario, estimates of marginal cost are based on differing assumptions regarding
the availability and costs of additional labour resources and equipment. While these assumptions
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
may be operationally realistic for a single incremental case, those same assumptions become
progressively more unrealistic as the number of incremental cases increases (e.g., to 10 or 100
additional cases).
Under the different scenarios, the analyses will present estimates of marginal cost as a
percentage of average cost. The rationale for presenting the results in this format is based on how
provincial Ministries of health examine hospitals costs. Currently, hospitals costs are compared
to the national average cost of patients with the same clinical profile (CIHI routinely provides the
average cost information to Ministries of health and hospitals).
The concept of marginal funding has considerable appeal to both funders and academics alike
since the method is seen as a way to motivate hospitals to improve efficiencies without actually
having to identify, or establish, that remediable inefficiencies or excess capacity actually exists.
However, if the assumptions about marginal costs change the determination of which scenarios
are more accurate, there is room for subjective opinions regarding which scenarios reflect the
hospitals ability to conduct additional cases. If this is the case, and if the assumptions about how
much excess capacity is readily available are wrong and funding proceeds on the basis of
marginal cost, serious service disruptions, substitutions between care types and/or budget deficits
may be an inevitable result.
In summary, while it has been demonstrated that it is possible to calculate hospital marginal costs
with some considerable degree of accuracy using available data and based on various scenarios
about which additional or excess resources are available, the application of marginal cost
information for hospital funding should be considered as a handle with care situation.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Introduction
The primary objective of this project is the development of methodologies to estimate the
marginal costs of hospitalizations for different types of patients. Marginal cost refers to the
change in total hospital expenditures associated with treating one additional patient. This
analysis is limited to hospital-based costs only; post-acute care costs, clinic visits, prescription
drug costs (other than those associated with a hospital stay) and physician claims are excluded.
Estimating the marginal costs of hospital care is complex; a hospitals revenue is used to
purchase a broad array of equipment and services to provide care to patients. The hospital
expenses associated with any given patient include direct patient care costs such as nursing and
therapies, but also unobserved (to the patient) costs, such as laboratory equipment and supplies,
as well as staff costs from support departments such as finance, human resources and hospital
senior management. Most, if not all, patient hospitalizations require a broad array of resources
from departments such as nursing, laboratory and diagnostic imaging services. Attributing all
relevant hospital services and equipment to individual episodes of patient care is not
straightforward, either conceptually or analytically; several data sources must be integrated to
estimate the proportion of a hospitals expenditures that can reasonable to allocated or attributed
to an individual patient.
To date, little work has been done with regard to estimating the marginal cost of hospitalizations
in Canadas hospitals. Canadian attempts to date to estimate the marginal cost of patients has
been limited to estimating cost per day in high volume case mix groups (CMG) in Alberta for
long stay patients. (1) The methods for deriving these estimates were based on clinical and
administrative data plus patient level cost data (described in more detail below) from Alberta
hospitals. The estimates of marginal cost were obtained by regressing hospitalization cost on
length of stay and were designed to estimate marginal costs per day after the acute portion of
hospital stays had transpired. International efforts to estimate marginal costs have been rough
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Methods
One important aspect of analyzing cost data is the challenge of how to present the findings in a
manner that potential users may be able to apply to their own settings. Consequently, a key
question underlying the development of this report is: who are the potential users of marginal
cost information and how will the information be used?
Since this analysis has been motivated by informational needs of provincial Ministries of health,
the analysis below assumes that the primary users of the findings will be provincial Ministries of
health and other significant purchasers of health care services, such as the B.C. Health Services
Purchasing Organization (HSPO).
This analysis assumes that the primary users of the methods and findings
regarding marginal costs of hospitalizations will be provincial Ministries of
health and other significant purchasers of health care services.
This analysis will present the estimates of marginal cost as a percentage of average cost. The
rationale for presenting the results in this format is based on how provincial Ministries of health
examine hospitals costs. Currently, hospitals costs are compared to the national average cost of
patients with the same clinical profile. CIHI provides this information to Ministries of health and
hospitals; the average costs are represented by Resource Intensity Weights (RIW).
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Due to sampling variability, patient heterogeneity and costing methodology variation within
Case Mix Groups (CMG) and Comprehensive Ambulatory Care Classification System (CACS)
groups, there can be considerable variability in patient costs within the same clinical group. To
reduce the likelihood of inaccurate estimates of marginal cost, only clinical groups with at least
100 patient hospitalizations are included in the analysis.
Only clinical groups (CMG+ and CACS) with at least 100 patient
hospitalizations are included in the analysis.
The analysis proceeds in three segments. In the first, day surgeries are analyzed. In the second,
patients admitted for surgical care are analyzed. Lastly, patients hospitalized for medical care are
analyzed. Separating hospital care into these three segments permits the underlying assumptions
and interpretation to vary according to the type of care. For instance, the assumptions regarding
hospitals capacity for day surgery are distinct from the assumptions regarding medical care.
Assumptions
Marginal costs are estimated for each CMG and CACS group, and reported as % of RIW,
since RIWs have established monetary values in the Canadian context and replicate (relative)
average costs.
The concept of marginal hospital cost is based on the assumption of the availability of latent
capacity (i.e., availability of labour and equipment) to treat additional patients. This assumption
raises many issues; for example, to conduct an additional (marginal) surgery, should nursing
costs be incorporated at regular hourly rates or at overtime rates?
In face of these important hospital constraint issues, estimates of marginal cost will be presented
for an array of different assumptions which simulate operating conditions within hospitals.
Day Surgery
Scenario 1
In scenario 1, it is assumed that there exists capacity in the hospital system to conduct a marginal
day surgery (one extra surgery). Under this assumption, additional purchases of major (or minor)
equipment are not required and all hospital labour costs are assumed fixed; the marginal cost of
providing nursing care, technologists, therapies, support and management staff is $0 and there
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
are no additional staff time or benefit expense costs. This scenario is based on the assumption
that these services are already funded under the existing global budget.
In this scenario, no overhead costs for support and management departments are attributed to the
marginal patients hospitalization. Marginal costs include only patient-specific supplies
(including food) and drugs. In this scenario, per unit supply and drug costs are not assumed to
change with the addition of one day surgery.
The rationale for this scenario is that all labour and equipment inputs are fixed. This is premised
on there being physical space and staff to accommodate a marginal day surgery. The specific
assumptions applied to the patient-level cost data from all departments are:
Labour costs: 0%
Patient-specific supplies: 100%
General supplies: 100%
Other supply costs: 100%
Overhead costs: 0%
Scenario 2A
In scenario 2A, it is assumed that there exists capacity in the hospital system to conduct a
marginal day surgery; however, in this scenario, all labour inputs are considered variable direct
costs; nursing care, technologists and other medical personnel are remunerated at their regular
rate. This scenario assumes that benefit hours and benefit expenses are $0. This scenario also
ignores standard minimum requirements for purchasing labour inputs (for example, a nurse is
required to have a minimum of a 4 hour shift. The same standard applies to some other unionized
hospital staff). Supply and drug costs are not assumed to change, nor are management costs
assumed to vary due to the marginal patient.
Similar to the preceding scenario, no overhead for support departments (management,
information services and equipment depreciation) is attributed to the marginal patients
hospitalizations since they are assumed to be remunerated under the current funding.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Scenario 2A: Labour costs from all clinical departments providing clinical
care are assumed variable. Additional purchases of major (or minor)
equipment are not required. Marginal costs will include supplies, drugs and
labour.
The rationale for this scenario is that one patients surgery can be accommodated by supplying
labour inputs (without overtime remuneration). The assumptions regarding the allocations of
patient costs for departments providing patient care are:
Scenario 2B
In this scenario, it is assumed that additional capacity in the hospital system for a marginal day
surgery can only be managed by remunerating nursing labour inputs at 1.5 times their regular
rate. All other labour inputs, such as technologists and therapies are remunerated at the regular
rate. Supply and drug costs are not assumed to change.
Under this scenario, benefit hours and benefit expenses increase, but reflect that some portion of
the benefit expenses are already fully paid by the regular hours. Due to the additional labour
expenses under this scenario, some portion of benefits expenses are attributable to the marginal
patient (since only some portion of benefit expenses increase with additional worked hours; for
example, holiday hours). Other benefit expenses do not increase with additional worked hours,
such as health insurance premiums. While the exact percentage is debatable, this scenario
assumes that benefit expenses of a marginal hospitalization is 50% of benefit expenses.
Sensitivity analysis of this assumption (not shown) finds that the results are not sensitive to
moderate deviations from 50%.
This scenario assumes that additional equipment is not required and overhead departments
(management and support departments and equipment depreciation) are not attributed to the
additional patients hospitalization cost.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
The rationale for this scenario is that one patients surgery can only be accommodated with
existing capacity and equipment, though only by supplying additional nursing labour inputs at
overtime rates.
Scenario 2B: Additional purchases of major (or minor) equipment are not
required, but nursing labour costs are remunerated at initial overtime rates.
The assumptions regarding the allocations of patient costs for departments providing patient care
are:
Labour costs:
o Nursing: 150%
o Non-Nursing: 100%
Scenario 2C
In this scenario, it is assumed that additional capacity in the hospital system for a marginal day
surgery can only be managed by remunerating labour inputs at 1.5 times their regular rate, the
initial overtime rate. This assumption is applied to all labour inputs, irrespective of the clinical
department. Supply and drug costs are not assumed to change.
By increasing the cost of direct labour inputs, the benefit hours and benefit expenses also
increase. As above, this assumption reflects the fact that some portion of the benefit expenses are
already fully paid by the regular hours.
This scenario also assumes that additional major (or minor) equipment is not required and
overhead departments (management and support departments and equipment depreciation) costs
are not attributed to patients hospitalizations.
The rationale for this scenario is that one patients surgery can only be accommodated with
existing capacity and equipment by supplying additional labour inputs at (initial) overtime rates.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
10
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Scenario 2C: Additional purchases of major (or minor) equipment are not
required, but labour costs from all clinical departments providing clinical
care are assumed to be remunerated at overtime rates.
The assumptions regarding the allocations of patient costs for departments providing patient care
are:
Scenario 2D
In scenario 2D, it is assumed that a marginal day surgery can be accomplished only by
remunerating nursing labour inputs at 2.0 times their regular rate. Other labour inputs are
remunerated at their regular rate (such as therapists, pharmacists and technologists). Supply and
drug costs are not assumed to change.
In this scenario, it is assumed that benefit hours and benefit expenses are 30% of completed
hospitalizations to reflect that a significant portion (70%) of benefit expenses are already paid
within regular working hours. This scenario assumes that additional major or minor equipment is
not required.
The rationale for this scenario is that one patients surgery can only be accommodated by
supplying additional nursing labour inputs at premium overtime levels.
Scenario 2D: Additional purchases of major (or minor) equipment are not
required, but nursing labour costs are assumed to be remunerated at
premium overtime rates.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
11
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
The assumptions regarding the allocations of patient costs for departments providing patient care
are:
Compensation costs:
o Nursing: 200%
o Non-Nursing: 100%
Category
Compensation
Nursing
Non-nursing
Patient-specific supplies
General supplies
Other supply costs
Overhead fixed costs
Overhead variable costs
1
0%
0%
100%
100%
100%
0%
0%
Scenario
2B
2A
100%
100%
100%
100%
100%
0%
0%
2C
150%
100%
100%
100%
100%
0%
50%
2D
150%
150%
100%
100%
100%
0%
50%
200%
100%
100%
100%
100%
0%
30%
Table 1. Summary of scenarios (and assumptions) under which marginal costs are estimated.
Inpatient Care
The scenarios described above for day surgery care are similarly applied to inpatient care.
Inpatient care is further stratified into surgical care and medical care, and the results are
presented separately.
Data
All data used in this analysis is anonymous and cannot be linked to other data sources to identify
individual patients. The data are used with permission of the British Columbia Ministry of
Health Services (MoHS) and the Ontario Ministry of Health and Long-Term Care (MOHLTC).
Ethics approval for analysis of anonymized secondary data was obtained from the Behavioural
Research Ethics Board at the University of British Columbia.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
12
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
In order to calculate an estimate of marginal cost, very detailed hospital financial and statistical
data are required. Such level of detail is provided by CIHIs Standards for Management
Information Systems in Canadian Health Service Organizations (MIS Standards). (2)
All Canadian hospitals are required to report their financial and statistical data to their respective
health Ministries using standardized MIS Functional Centre (i.e., departmental) categories and a
standardized set of financial chart of accounts. Both the Functional Center and account
categories are highly discriminate (i.e., specific).
Definition: Functional Centre
A Functional Centre (FC) is an operational subdivision of a hospital, such as a clinical
department, used to record the budget, revenue and expense statistics which pertain to the
clinical activity being carried out in the department. (2) FCs are hierarchical and financially
(even if not clinically) independent by definition and design, such that aggregating across FCs
provides hospital-level detail.
For hospitalization costing purposes, Functional Centers (departments) are classified into two
(unique) separate categories.
Direct Department
Direct departments (or Functional Centers) are those that provide patient-specific services and
include inpatient (e.g., Orthopedic Ward, ICU) and outpatient nursing departments (e.g., Day
Surgery Unit, Endoscopy), diagnostic departments (e.g., Labs, Diagnostic Imaging), and
therapeutic departments (e.g., physiotherapy, pharmacy).
Indirect (Overhead) Department
Indirect (overhead) departments provide corporate-level (or support), rather than patient-level,
services to other departments (both Direct and Indirect). Examples are Human Resources,
Materials Management, Information Technology, Housekeeping, Finance/Accounting, Senior
Management, etc. Generally, indirect departments are synonymous with the term overhead.
The costs of nurse managers, in the nursing units, that predominantly perform administrative
functions are allocated to patients (in the unit) as direct costs.
The costs of both direct and indirect MIS Functional Centers are reported at the account level.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
13
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Account Code
MIS account codes can specify a very precise resource cost. For example, account codes permit
the distinction, within a Functional Center, between the labour of hands-on patient care staff
versus the cost of non-hands-on staff like supervisors and clerks. Account codes can further
distinguish between salary costs (MIS uses the term compensation) versus benefit (e.g.,
employment insurance, pension, etc.) labour costs.
We propose to employ hospital financial, workforce/workload and clinical data from fiscal years
2008/2009 and 2009/2010. There are a variety of complementary data sources used for this
project, including B.C.s Health Authority Management Information System (HAMIS), B.C.s
and Ontarios version of CIHIs Discharge Abstract Database (DAD), B.C.s Health Sector
Compensation Information System (HSCIS) and summaries of Ontarios Ontario Cost
Distribution Methodology database.
Health Authority Management Information System (HAMIS)
HAMIS is the B.C. MoHs database of financial and statistical (service recipient workload)
reports of provincial health organizations (and health service providers, or hospitals). The data in
HAMIS, which are based on hospitals general ledger data, include detailed revenue, expenditure
and statistical totals across fiscal years from submitting organizations. HAMIS includes financial
and statistical amounts by department or functional centre. Data at this level of specificity are
transmitted regularly to the MoH through the Health Authority Reporting Program (HARP).
HAMIS data are not patient-identifiable data; rather these summary level data provide a financial
and statistical summary of the expenditures and activities of the hospitals. Key information from
HAMIS used in this study includes statistical reports of earned hours, a hospitals reporting of
the number of labour input hours for: 1) management and operational support personnel (MOSP)
and 2) unit producing personnel (UPP). For unit producing personnel, earned hours are stratified
by employee worked hours and purchased service personnel hours by department, a proportional
breakdown of permanent staff and contract staff expenditures.
Discharge Abstract Database (DAD)
The DAD is a comprehensive dataset whose coverage is the population of hospitalizations in
B.C. and Ontario. Maintained by CIHI, the DAD includes patient-level demographic,
administrative and clinical data for all acute inpatient episodes and day surgeries in (B.C.)
hospitals. Ontario hospitalization data is included in the analysis to compensate for small
samples of patient level cost data (discussed in more detail later; Ontarios day surgery data are
maintained in CIHIs National Ambulatory Care Reporting System, or NACRS).
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
14
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
For inpatient hospitalizations, the DAD contains an exhaustive list of interventions that occurred
during the hospitalization, demographic details such as age and sex of the patient, and
administrative details such as province of residence and discharge location. While the episodes
of hospital care within the DAD contain potentially identifying information, the subset of DAD
data provided to support this study from B.C. and Ontario had all potentially identifying
information removed.
Health Sector Compensation Information System (HSCIS)
Wage rate data from (HSCIS) will be helpful to inform the calculation of marginal costs. For
instance, labour contracts will be instrumental in determining whether overtime rates for nursing
care are to be remunerated at time-and-a-half or double-time. This information will also guide
whether departmental level staff inputs should be categorized as Fixed or Variable costs.
Activity-based Costing Dataset (ABC)
ABC is the detailed departmental level expenditure data associated with a patients hospital
activity. This source of data is described in detail in the next section.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
15
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Only hospitals with activity-based cost information are capable of deriving costs for all aspects
of patients activity. Activity-based cost data help derive marginal cost estimates since it values
each activity and contribution of overhead.
Unlike the financial and statistical reporting requirements mandated for hospitals in each
province, hospitals are not currently required to implement activity-based costing by provincial
Ministries of health. As a result, as of 2011, only a small subset of Canadian hospitals has
activity-based costing. In Ontario, over 40 hospitals derive activity-based costs in acute inpatient,
rehabilitation, chronic care and mental health inpatient hospitals. In Alberta, eight hospitals have
the ability to derive activity-based costing, though Alberta opted not to participate in this study.
In B.C., only three hospitals derive activity-based costs.
The principles of activity-based costing are based on accurately attributing intermediate product
costs to patients. While there are published standards in Canada for activity-based costing
methods (by CIHI), these standards are not enforced within hospitals. Ontario is alone among
provinces providing financial and technical support to these hospitals. In Alberta and B.C., while
each hospital tries to adhere to general principles of attributing hospital costs to patients,
implementations are not centrally supported, nor are methods standardized between hospitals.
How does activity-based costing relate to financial and statistical data reported into B.C.s
Health Authority Management Information System (HAMIS) and Ontarios Cost Distribution
Methodology (OCDM) database? The financial and statistical data reported into HAMIS and the
OCDM represent a summary of department-level expenditures and activities, such as counts of
activities and total expenditures. While valuable, these data sources are limited in their ability to
understand patients costs since they do not link department activity to individual patients. For
example, HAMIS and the OCDM include departmental counts of activity, but they do not
differentiate across patients in terms of intensity of departmental resource consumption.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
16
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
In contrast, activity-based costing data are more detailed, providing patient- and departmentlevel summaries of expenditures. While the two sources are complementary, the aggregation of
expenditures in each source will not be precisely the same; differences will occur due to
differences in cost allocation methods and to patient stays that span multiple fiscal reporting
periods. Gross discrepancies in aggregate expenditures between the two sources are used as a
measure to assess the completeness and comprehensiveness of reported activity-based costing
data.
The objective of activity-based costing methodology is to be revenue neutral; that is, the sum of
the costs of all individual hospitalizations must be exactly equal to the same-period global budget
expenditures. Or, in other words, if hospitals were instead billing the Ministry separately for each
individual patient case (i.e., fee-for-service billing or activity-based-funding ABC) using their
activity-based costing figures, the totals would precisely match current globally funded
expenditures.
While there are 3 components of costs associated with the CT scan, activity-based costing
proceeds by analyzing the sub-categorizations within each component.
For the imaging technologist who conducts the scan, there are earned hours (to conduct
the scan), benefit hours (such as holiday days - hours of absence that are funded benefits
associated with the number of worked hours) and benefit expenses (such as pension and
disability insurance). Since the imaging technologist is generally a salaried employee of
the hospital, the imaging technologist expense is most often considered a fixed cost.
For the imaging equipment, there are fixed costs associated with the amortization of the
original purchase price of (or lease costs associated with) the imaging equipment,
insurance on the equipment and expenses associated with the maintenance of the
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
17
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
equipment. Depreciation is amortized over the life of the equipment. There are variable
costs associated with the storage media for each CT scan.
Costs associated with indirect or overhead departments such as Information Systems,
Human Resources and Materials Management can be considered fixed, as an additional
patient hospitalization will not necessitate additional expenditures. However, for other
indirect departments, such as Food Services, Admitting and Medical Records, some
additional expenditures are required and a portion of their costs may thus be considered
variable.
In addition to the costs of the CT scan described above, this imaging department functional
centre (FC) consumes services from other departments not directly related to the production of
CT scans. For example, imaging also consumes some portion of hospital overhead costs for
departments such as finance, information services and administration. As well, fixed
housekeeping expenses may be attributed to this department based on the square footage of the
imaging department (relative to all other departments). Other departments, such as finance, may
have algorithms for allocating their costs (fixed and variable) to the diagnostic imaging
department as overhead costs. Since overhead can be a significant component of patient costs,
the methods (and guidelines) for attributing hospital overhead costs to the clinical departments
are described in the next section.
Labour (Compensation)
For the purposes of reporting staffing costs, hospital personnel are categorized into one of 3
unique occupational groups:
Unit producing personnel. Unit producing personnel (UPP) are those hospital staff whose
primary function is the execution of hands on or patient care activities of the
department. Examples of unit producing personnel include: registered nurses, laboratory
technologists, pharmacists (712**, 713** and 714**). Personnel in departments
considered as overhead, such as housekeeping or food services are not included.
Management and support personnel. Management and support personnel are those
hospital staff whose primary function is the management or support of clinical
departments. Examples of management and support personnel are managers, clerks, and
medical personnel assuming management responsibilities (711**).
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
18
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Medical personnel. Medical personnel are those physicians compensated by the hospital.
Examples may include pathologists, psychiatrists, and radiologists.
For all hospital employees, earned hours (worked and benefit hours) and benefit expenses are
attributed to the department that provides services. For example, the earned hours and benefit
expenses of all RNs who provide nursing care to patients in medical beds are attributed to the
medical department of the same patients.
Hospital staff that work in multiple departments have their earned hours and benefit hours split
among their departments according to the proportional distribution of hours.
Earned hours and benefit expenses of hospital managers and support staff are attributed to the
management department where the work was performed.
Allocating Unit Producing Personnel (UPP) Compensation Costs
Compensation costs are allocated to patients on the basis of actual workload (time) or relative
workload, or relative value unit (RVU). For example, nursing costs are attributed to patients
based on nursing workload statistics. In contrast, RVUs distribute compensation costs for UPP
and management and support personnel to intermediate products which are subsequently
attributed to patients.
A simple example of workload-based distribution of compensation costs is highlighted in Table
2. Consider an imaging technician who has completed 40 CT scans and 40 MRIs. Since the
workload of CT scans is 60 RVU (1.5 RVU times 40), and the workload of MRIs is 40 (1.0 RVU
times 40), 60% of the imaging technicians compensation costs are attributed to the CT scans.
This portion of the imaging technicians compensation costs are then evenly attributed to the 40
patients receiving CT scans.
Function
CT Scan
MRI
Relative
Workload (RVU)
1.5
1.0
Number
Completed
Workload
40
40
Relative
Cost
60%
40%
60
40
Table 2. Example of workload units used to allocate an imaging technicians compensation costs to diagnostic
imaging patients.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
19
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
In contrast to hospital management personnel (whose costs are allocated as overhead), salaries of
unit nurse managers (whose roles are primarily administrative within the unit) are distributed to
patients as direct costs. Nurse managers salary costs are attributed to patients as direct costs on
the basis of the nursing workload of patients (in the unit). For example, if each of the patients on
the unit attracts equal amounts of (nursing) workload, then the nurse managers salary costs are
distributed equally across the patients. On the other hand, if a single patient attracts one-half of
the UPP workload of nurses in the unit, then one-half of the nurse managers costs are attributed
to this patient.
Allocating Medical Personnel Compensation Costs
Some physicians, such as staff pathologists, are employees of the hospital. The costs of salaried
physicians are attributed to the clinical department that provides the services and to patients on
the basis of relative workload.
Supplies
Patient Specific (Traceable) Supplies
The costs of some supplies or expenses are tracked and directly attributed to the patient. These
tend to be high cost supplies or expenses where amounts tend to vary according to the clinical
needs of the patient. Examples of traceable supplies include: drugs, artificial prosthetics,
pacemakers and catheters for cardiac catheterizations. In some hospitals, traceable supplies also
include meals, in others, meal costs are distributed to patients. These costs are allocated to the
departments responsible for the patient, and then attributed to the patient.
General Supplies
The costs of some supplies are attributed to the department but not tracked. These supplies tend
to be less expensive and broadly used for patients and are attributed to the departments that
consume the supplies. For example, disposable linens are attributed to the department that uses
the linens. Clinical department supply items would include sutures, staples, dressings, needles,
syringes and gloves. Supply items for pharmacy departments would include antineoplastics, antiinfectives and parenteral nutrition products. These costs are distributed to patients in that
department.
Sundry
The costs of some expenses are not easily categorized as compensation, supplies or equipment.
(2) Examples include postage, courier expenses, workshop expenses, staff conference travel,
membership expenses and insurance. These costs are attributed to the department that consumes
the supplies and distributed to the patient.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
20
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Equipment
This category is used to describe the cost associated with the purchase (or lease) and use of
major equipment, software, and minor equipment. The cost of use, maintenance, amortization,
and insurance of major equipment is allocated to departments that consume the equipment
and whose costs are then allocated to patients that use the equipment. Building and new
construction (and amortization) is not included in patients cost calculations.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
21
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
22
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
considered to have some variable costs). For example, pharmacy (pharmacist) costs are
considered fixed while the drug costs are variable.
It should also be noted that building depreciation costs, since they are not funded globally, are
not included in any case cost calculations. These costs are allocated to a separate capital account,
not an operations account.
Variable Costs
Variable costs are those costs that are deemed to vary directly with the volume of patients, and
the volume of services provided to them. For example, prostheses costs for surgical patients will
vary directly with the volume of joint replacement cases.
Marginal Costs
Marginal cost represents the total cost of producing one additional unit; in the present context,
one additional unit represents one additional case of a specific type of day surgery or inpatient
hospitalization.
Marginal cost is the sum of select (depending on specific assumptions/scenarios) variable costs
of all direct departments (and indirect departments, if they are deemed to have any variable
costs). This includes variable costs in indirect departments allocated to direct departments.
An example of a hypothetical cancer patients cost data is shown in Table 3. In this example, 3
absorbing departments provided services to this patient: nursing, laboratory and imaging. For
each department, direct compensation, supply and equipment costs are shown. The sixth column
shows the overhead costs from management and support departments.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
23
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Department
Labour Costs
Patient-Specific
Supplies
General
Supplies
Equip
-ment
$12,000
$300
$50
$200
$300
$200
$100
$12,500
$600
Medical Oncology
Nursing
Laboratory
Imaging
Overhead
Total
$50
$3,000
$15,400
$50
$100
$200
$400
$200
$500
$850
$1,400
$200
$650
$3,700
$18,250
Table 3. Example of patient cost data, including direct compensation, supply and equipment costs plus overhead
costs.
In this example, the marginal costs are the sum of labour ($12,500), patient specific supplies
($600) and general supplies ($200) used in the treatment of this patient. In scenario 1, labour
costs are consider fixed and are not included in the calculation of marginal costs. The fixed costs
are represented by the sum of equipment and overhead ($650 and $3,700).
Types of Patients
Provided that there are thousands of possible diagnostic conditions and surgical interventions,
figuratively, there are an infinite number of combinations of conditions and procedures that can
occur during episodes of hospital care. Fortunately, methods have been developed over the
course of several decades to use diagnostic and procedure information abstracted from the
patient chart to summarize the types of patients that hospitals treat. (3) (4) (5)
The systems that use clinical information to derive descriptors of hospital care are known as case
mix systems. These systems consist of two components. The first component is the deterministic
algorithm that uses patient diagnostic and procedure information to assign patients into unique
case mix groups and the second is the estimation of patients (average) cost within the same
group.
The case mix groups are characterized by clinical homogeneity of patients in the same group and
homogeneity of patients costs. For example, one group may consist of all hospitalized patients
diagnosed with pneumonia, while another may consist of all patients undergoing spinal fusion. In
Canada, CIHI applies the CMG case mix system to all hospital discharge summaries (abstracts)
submitted to its DAD.
In this analysis, the patients in the same CMG will be considered sufficiently homogeneous to
not warrant disaggregation for the purposes of estimating marginal costs. This assumption is also
supported by examining the count of hospitalizations for the same diagnosis or procedure code;
the count may be very small relative to the total number of patients in the same CMG. This
approach assumes that the marginal costs associated with all patients in the same CMG are
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
24
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
equivalent; this assumption may limit the applicability of the findings if the figures are applied to
specific procedures or diagnostic codes that are a subset within a CMG group.
To simplify interpretation and application of the findings, and to maintain a larger number of
patients in each case mix group, the methods are usually applied only to base CMG groups
(referred to as typical hospitalizations). Patients whose hospitalizations included effects which
CIHI determines to modify cost weights, namely flagged interventions, intervention events and
out-of-hospital indicators, are removed from the dataset. Long stay, death and transfer patients
are also not included in the analysis. Therefore, it may not be appropriate to apply the findings to
such cases. These exclusion criteria remove approximately 12% of patients from the dataset. All
age groups and comorbidity levels are included in the analysis.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
25
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Clinical activity abstracted from the patient chart is linked to patient cost
data. The average hospitalization cost in each CMG is calculated. The
average cost in each CMG, relative to the average cost of all cases, is the
resource intensity weight (RIW) for the CMG.
While the above description is applicable to acute inpatient hospital activity, the principles are
generalizable to day surgery (same day surgery). In this setting, there is an analogous case mix
system that is based on the patients procedure (the reason for the day surgery). Day surgery
patients are defined by the CACS case mix system.
The methods for estimating marginal costs are extended to day surgeries
defined by the comprehensive ambulatory classification (CACS) case mix
system.
In this analysis, the CMG classification is the basis for clinical grouping of acute inpatient
activity, and the CACS classification is the basis for clinical grouping of same day surgery
activity.
The same methodology described in this report could be reapplied each year to CIHIs updated
clinical and cost data (and updated methodology year). The CMG and CACS methodology
changes from year to year to reflect the addition of new diagnostic or procedure codes or to
reflect changes in the heterogeneity of patients in the same group. In this analysis, the CMG and
CACS methodology from 2010 are used, but the methodology year can similarly be updated.
The methodology developed in this report could also be applied to finer subgroups of patients.
For instance, patients hospitalizations could be grouped according to age categories. Then, the
methodology could be applied separately to each age category to derive estimates of marginal
cost for each age category. The principle is also true for long stay patients or in-hospital deaths
excluded from this analysis.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
26
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Results
Results Day Surgery CACS
Scenario 1
In scenario 1, it is assumed that there exists capacity in the hospital system to conduct a marginal
day surgery. Under this assumption, additional purchases of major (or minor) equipment are not
required and all hospital labour costs are fixed. The findings under this assumption are shown in
Table 4 for a sample of CACS groups.
In Table 4, the 5 CACS groups with the highest percentage of marginal costs and the 5 CACS
groups with the lowest percentage of marginal costs are shown. A complete list of findings under
this scenario is shown in Appendix 1.
CACS
CACS Title
2705
1525
1536
1100
1033
Peritoneal Dialysis
Pacemaker
Oth Card Interv-Perc Translum
Cataract Extractions
Crpl Tnl Rel Open-Loc/No Anae
CACS
RIW
0.0830
1.3822
1.1043
0.1974
0.0397
1045
2300
1020
1500
1973
0.1523
0.1336
0.1378
0.1188
0.0726
Marginal Cost
(% of Average)
50.9%
47.7%
40.9%
37.5%
35.6%
14.0%
12.0%
10.2%
10.0%
8.6%
Table 4. Partial results for estimating the marginal costs of patients assigned to CACS groups under scenario 1.
The results from Table 4 demonstrate significant variability in marginal costs between CACS
groups under scenario 1. Peritoneal dialysis and pacemaker have the highest marginal costs, at
approximately 50% of the average cost. Under this scenario, labour costs are fixed and variable
costs are driven by consumable costs. Thus, for each of CACS 2705 (Peritoneal Dialysis) and
CACS 1525 (Pacemaker), there are high consumable costs.
At the other end of the spectrum are day surgeries whose marginal costs are low relative to the
average cost. For example, CACS 1500 (Cardioversion), the marginal cost is estimated to be
10% of the average cost. In this CACS group, consumables are a small portion of the
hospitalization cost.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
27
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Scenario 2A
In scenario 2A, it is assumed that there exists capacity in the hospital system to conduct a
marginal day surgery, though labour inputs are considered variable costs remunerated at their
regular rate. This scenario assumes that benefit hours and benefit expenses are $0. The findings
under this assumption are shown in Table 5 for a sample of CACS groups (a complete list of
findings is shown in Appendix 1).
CACS
CACS Title
1033
1120
1525
1536
1900
CACS
RIW
0.0397
0.2220
1.3822
1.1043
0.3685
2610
1625
1810
2300
1045
0.1624
0.0836
0.2044
0.1336
0.1523
Marginal Cost
(% of average)
77.4%
71.9%
69.8%
69.8%
69.8%
53.8%
53.1%
49.9%
49.3%
44.2%
Table 5. Partial results for estimating the marginal costs of patients assigned to CACS groups under scenario 2A.
Scenario 2B
In this scenario, it is assumed that additional capacity in the hospital system for a marginal day
surgery can only be managed by remunerating nursing labour inputs at 1.5 times their regular
rate. All other labour inputs, such as technologists, pharmacists and therapists are remunerated at
the regular rate. The findings under this assumption are shown in Table 6 for a sample of CACS
groups (a complete list of findings is shown in Appendix 1).
CACS
CACS Title
1033
1120
1900
2706
1706
CACS
RIW
0.0397
0.2220
0.3685
0.0802
0.2840
2300
2800
2610
1810
1045
Lithotripsy
Angiography
Percutaneous Needle Lymph Int
Other Hepatobiliary Intrvn
Other Nerv Syst Intrv-Grade 1
0.1336
0.3063
0.1624
0.2044
0.1523
Marginal Cost
(% of avg)
98.4%
95.4%
87.9%
85.8%
84.9%
65.0%
60.8%
60.0%
59.8%
56.4%
Table 6. Partial results for estimating the marginal costs of patients assigned to CACS groups under scenario 2B.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
28
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Scenario 2C
In scenario 2C, it is assumed that additional capacity in the hospital exists for a marginal day
surgery and can be managed without new technology or equipment, though only by remunerating
labour inputs at 1.5 times their regular rate. The findings under this assumption are shown in
Table 7 for a sample of CACS groups (a complete list of findings is shown in Appendix 1).
CACS
CACS Title
CACS
RIW
1033
1120
2706
1900
1706
0.0397
0.2220
0.0802
0.3685
0.2840
Marginal Cost
(% of
Average)
98.5%
95.7%
88.6%
88.4%
85.5%
2325
2610
2300
1810
1045
0.0933
0.1624
0.1336
0.2044
0.1523
71.5%
71.0%
67.9%
64.6%
59.3%
Table 7. Partial results for estimating the marginal costs of patients assigned to CACS groups under scenario 2C.
Scenario 2D
In scenario 2D, it is assumed that a marginal day surgery can only be accomplished by
remunerating nursing labour inputs at 2.0 times their regular rate. Other labour inputs are
remunerated at their regular rate. This scenario assumes that additional major or minor
equipment is not required. The findings under this assumption are shown in Table 8 for a sample
of CACS groups (a complete list of findings is shown in Appendix 1).
CACS
CACS Title
CACS
RIW
1033
1120
2706
1900
1230
0.0397
0.2220
0.0802
0.3685
0.2078
Marginal Cost
(% of
Average)
119.3%
118.8%
106.3%
104.4%
103.6%
1510
1045
1810
2800
2610
Cardiac Study
Other Nerv Syst Intrv-Grade 1
Other Hepatobiliary Intrvn
Angiography
Percutaneous Needle Lymph Int
0.4333
0.1523
0.2044
0.3063
0.1624
71.4%
68.6%
68.2%
66.9%
65.6%
Table 8. Partial results for estimating the marginal costs of patients assigned to CACS groups under scenario 2C.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
29
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
CMG Title
161
70
176
177
315
Implantation of Cardioverter/Defibrillator
Cochlear Implant
PCI without MI/Shock/Arrest/Heart Failure
Management of Pacemaker/Epicardial Lead
Bilateral Hip/Knee Replacement
302
180
380
500
504
Volume
Weighted CMG
RIW
5.4142
4.7330
1.3024
1.6631
2.4851
Marginal Cost
(% of Average)
2.1341
2.4704
1.8992
1.5351
1.3089
12.7%
12.2%
12.1%
11.4%
10.9%
56.4%
44.4%
42.4%
40.4%
38.6%
Table 9. Partial results for estimating the marginal costs of patients assigned to CMG groups under scenario 1.
Scenario 2A
In scenario 2A, it is assumed that there exists capacity in the hospital system to conduct a
marginal inpatient surgery, though labour inputs are considered variable costs remunerated at
their regular rate. The findings under this assumption are shown in Table 10 for a sample of
CMG groups (a complete list of findings is shown in Appendix 2).
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
30
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
CMG
CMG Title
161
315
319
320
326
Implantation of Cardioverter/Defibrillator
Bilateral Hip/Knee Replacement
Revised Knee Replacement without Infection
Unilateral Hip Replacement
Shoulder Replacement
53
54
500
387
179
Volume
Weighted CMG
RIW
5.4142
2.4851
1.9949
1.7888
1.6145
Marginal Cost
(% of Average)
0.5204
0.3918
1.5351
0.8038
0.8207
56.8%
56.2%
55.9%
55.0%
54.0%
76.2%
71.4%
70.4%
70.2%
69.7%
Table 10. Partial results for estimating the marginal costs of patients assigned to CMG groups under scenario 2A.
Scenario 2B
In this scenario, it is assumed that additional capacity in the hospital system for a marginal
inpatient surgery can only be managed by remunerating nursing labour inputs at 1.5 times their
regular rate. All other labour inputs, such as technologists, pharmacists and therapists are
remunerated at the regular rate. The findings under this assumption are shown in Table 11 for a
sample of CMG groups (a complete list of findings is shown in Appendix 2).
CMG
CMG Title
224
113
315
54
387
304
179
178
167
169
Volume
Weighted CMG
RIW
3.1050
3.0153
Marginal Cost
(% of Average)
3.0966
1.6326
2.4851
90.2%
89.0%
87.7%
0.3918
0.8038
1.2628
0.8207
1.3729
75.8%
73.1%
72.2%
68.8%
68.4%
91.7%
90.3%
Table 11. Partial results for estimating the marginal costs of patients assigned to CMG groups under scenario 2B.
Scenario 2C.
In scenario 2C, it is assumed that additional capacity in the hospital exists for a marginal
inpatient surgery and can be managed without new technology or equipment, though only by
remunerating all labour inputs at 1.5 times their regular rate. The findings under this assumption
are shown in Table 12 for a sample of CMG groups (a complete list of findings is shown in
Appendix 2).
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
31
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
CMG
CMG Title
167
224
169
113
650
423
54
52
178
179
Volume
Weighted CMG
RIW
3.1050
3.0966
3.0153
1.6326
1.4850
Marginal Cost
(% of Average)
0.8466
0.3918
0.4508
1.3729
0.8207
78.1%
77.6%
77.2%
75.3%
74.2%
95.0%
94.9%
94.1%
93.7%
92.8%
Table 12. Partial results for estimating the marginal costs of patients assigned to CMG groups under scenario 2C.
Scenario 2D
In scenario 2D, it is assumed that a marginal inpatient surgery can be accomplished by
remunerating nursing labour inputs at 2.0 times their regular rate. Other labour inputs are
remunerated at their regular rate. This scenario assumes that additional major or minor
equipment is not required. The findings under this assumption are shown in Table 13 for a
sample of CMG groups (a complete list of findings is shown in Appendix 2).
CMG
CMG Title
113
81
176
304
387
179
178
167
224
169
Volume
Weighted CMG
RIW
3.1050
3.0966
3.0153
Marginal Cost
(% of Average)
1.6326
0.8374
107.9%
106.6%
1.3024
1.2628
0.8038
0.8207
1.3729
87.5%
87.4%
85.7%
78.9%
75.5%
110.9%
109.2%
108.7%
Table 13. Partial results for estimating the marginal costs of patients assigned to CMG groups under scenario 2D.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
32
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
are not required and all hospital labour costs are fixed. The findings under this assumption are
shown in Table 14 for a sample of medical CMG groups with the highest and lowest percentage
of marginal costs. A complete list of findings is shown in Appendix 3.
CMG
CMG Title
440
546
203
521
64
594
577
549
585
584
Volume
Weighted CMG
RIW
0.3781
0.5917
0.9505
0.4883
0.7728
Marginal Cost
(% of Average)
0.3264
0.2350
0.6479
1.4081
2.7540
6.8%
6.6%
6.3%
6.3%
5.6%
24.5%
23.0%
22.8%
18.5%
18.3%
Table 14. Partial results for estimating the marginal costs of medical patients assigned to CMG groups under
scenario 1.
Scenario 2A
In scenario 2A, it is assumed that there exists capacity in the hospital system to admit a marginal
medical patient, though labour inputs are considered variable costs remunerated at their regular
rate. The findings under this assumption are shown in Table 15 for a sample of CMG groups (a
complete list of findings is shown in Appendix 3).
CMG
CMG Title
130
440
141
39
147
Respiratory Failure
Disease/Disorder of Thyroid/Parathyroid Gland
Upper/Lower Respiratory Infection
Status Epilepticus
Asthma
814
199
203
521
198
Observation/Evaluation
Cardiac Valve Disease
Unstable Angina/Atherosclerotic Heart Dis w CC
Fibroid/Prolapse/Fistula/Other Disorder
Congenital Cardiac Disorder
Volume
Weighted CMG
RIW
0.8268
0.3781
0.5771
0.5146
0.3904
Marginal Cost
(% of Average)
0.3310
1.0503
0.9505
0.4883
0.7612
59.2%
59.0%
59.0%
58.5%
55.4%
68.7%
67.9%
67.4%
67.0%
66.8%
Table 15. Partial results for estimating the marginal costs of medical patients assigned to CMG groups under
scenario 2A.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
33
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Scenario 2B
In this scenario, it is assumed that additional capacity in the hospital system to admit a marginal
medical patient. Under this scenario, this can only be managed by remunerating nursing labour
inputs at 1.5 times their regular rate. All other labour inputs, such as technologists, pharmacists
and therapists are remunerated at the regular rate. The findings under this assumption are shown
in Table 16 for a sample of CMG groups (a complete list of findings is shown in Appendix 3).
CMG
CMG Title
584
141
130
592
587
549
199
521
203
198
Volume
Weighted CMG
RIW
2.7540
0.5771
0.8268
0.7748
0.4403
Marginal Cost
(% of
Average)
95.7%
94.0%
93.5%
93.2%
92.9%
0.6479
1.0503
0.4883
0.9505
0.7612
76.2%
74.9%
74.1%
73.7%
67.1%
Table 16. Partial results for estimating the marginal costs of medical patients assigned to CMG groups under
scenario 2B.
Scenario 2C
In scenario 2C, it is assumed that additional capacity in the hospital exists to admit a marginal
medical patient. In this scenario, this can only be managed without new technology or
equipment, though only by remunerating all labour inputs at 1.5 times their regular rate. The
findings under this assumption are shown in Table 17 for a sample of CMG groups (a complete
list of findings is shown in Appendix 3).
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
34
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
CMG
CMG Title
130
141
810
809
584
Respiratory Failure
Upper/Lower Respiratory Infection
Palliative Care
Awaiting Placement
Newborn/Neonate 1500-1999 gr, Gest Age 32-34 Wks
814
199
203
521
198
Observation/Evaluation
Cardiac Valve Disease
Unstable Angina/Atherosclerotic Heart Disease w CC
Fibroid/Prolapse/Fistula/Other Disorder
Congenital Cardiac Disorder
Volume
Weighted CMG
RIW
0.8268
0.5771
1.3790
1.3035
2.7540
Marginal Cost
(% of
Average)
99.9%
99.1%
97.8%
97.7%
97.5%
0.3310
1.0503
0.9505
0.4883
0.7612
84.3%
81.3%
79.6%
79.5%
75.6%
Table 17. Partial results for estimating the marginal costs of medical patients assigned to CMG groups under
scenario 2C.
Scenario 2D
In scenario 2D, it is assumed that a marginal medical patient can be admitted only by
remunerating nursing labour inputs at 2.0 times their regular rate. Other labour inputs are
remunerated at their regular rate. This scenario assumes that additional major or minor
equipment is not required. The findings under this assumption are shown in Table 18 for a
sample of CMG groups (a complete list of findings is shown in Appendix 3).
CMG
CMG Title
584
585
587
592
589
762
199
521
203
198
Volume
Weighted CMG
RIW
2.7540
1.4081
0.4403
0.7748
0.4813
Marginal Cost
(% of
Average)
122.7%
118.5%
118.4%
117.7%
117.3%
1.0059
1.0503
0.4883
0.9505
0.7612
91.9%
88.5%
88.3%
85.1%
76.4%
Table 18. Partial results for estimating the marginal costs of medical patients assigned to CMG groups under
scenario 2D.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
35
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Summary of Findings
Results Day Surgery CACS
The marginal costs of day surgery cases vary between CACS groups (Table 19). Under scenario
1 (fixed labour costs), variations in marginal costs between CACS groups are driven by the
costliness of consumables in different CACS groups. Under scenario 2D (overtime labour rates),
the situation is reversed, as the role of the intensity of labour costs becomes more important, and
marginal costs are predominantly affected by the amount of labour inputs.
2705
1525
1536
1100
1033
1900
1706
1705
1510
Peritoneal Dialysis
Pacemaker
Oth Card Interv-Perc Translum
Cataract Extractions
Crpl Tnl Rel Open-Loc/No Ana.
Shoulder Intervention
Complex Hernia Intervention
Non-Complex Hernia Interv
Cardiac Study
CACS
RIW
0.0830
1.3822
1.1043
0.1974
0.0397
0.3685
0.2840
0.2840
0.4333
1245
2730
1410
1985
1045
2300
1020
1500
1973
0.1356
0.1069
0.1764
0.1117
0.1523
0.1336
0.1378
0.1188
0.0726
CACS
CACS Title
Scenario
1
50.9%
47.7%
40.9%
37.5%
35.6%
34.9%
30.8%
30.6%
29.7%
16.0%
16.0%
15.3%
15.2%
14.0%
12.0%
10.2%
10.0%
8.6%
75.4%
74.3%
66.2%
75.1%
56.4%
65.0%
68.1%
71.2%
73.7%
Scenario
2D
84.2%
89.6%
82.1%
97.5%
119.3%
104.4%
101.2%
99.2%
71.4%
77.5%
84.1%
74.9%
77.4%
59.3%
67.9%
78.2%
76.9%
82.5%
93.4%
85.1%
76.7%
92.8%
68.6%
80.7%
80.1%
87.5%
88.6%
Table 19. Partial results for estimating the marginal costs of patients assigned to CACS groups, sorted from highest
proportional marginal cost to lowest proportional marginal cost, scenario 1.
36
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
CMG
CMG Title
161
70
176
177
315
326
2
327
319
Implantation of Cardio./Defib.
Cochlear Implant
Perc. Coronary Interv.
Management of Pacemaker
Bilateral Hip/Knee Repl.
Shoulder Replacement
Intracranial Vessel Intv., Perc.
Other Joint Replacement
Revised Knee Repl. wo Inf,
75
511
501
81
302
180
380
500
504
Weighted
CMG RIW
5.4142
4.7330
1.3024
1.6631
2.4851
1.6145
2.5063
1.6101
1.9949
Scenario
1
56.4%
44.4%
42.4%
40.4%
38.6%
38.3%
38.2%
37.1%
35.1%
0.6128
0.3947
1.2670
0.8374
2.1341
2.4704
1.8992
1.5351
1.3089
13.3%
13.3%
13.0%
12.9%
12.7%
12.2%
12.1%
11.4%
10.9%
81.2%
84.5%
78.6%
84.1%
78.2%
85.4%
84.8%
76.2%
78.4%
Scenario
2D
93.1%
93.2%
87.5%
90.9%
99.7%
98.4%
89.3%
96.7%
99.8%
83.7%
88.0%
83.7%
85.6%
83.9%
91.3%
90.8%
81.6%
84.2%
101.2%
103.2%
94.7%
106.6%
96.1%
103.7%
103.0%
91.8%
94.6%
Table 20. Partial results for estimating the marginal costs of patients assigned to surgical CMG groups, sorted from
highest proportional marginal cost to lowest proportional marginal cost, scenario 1.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
37
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Weighted
CMG
440
546
203
521
64
198
762
629
199
588
593
589
587
594
577
549
585
584
CMG Title
Disease/Disorder of Thyroid Gl
Ectopic Pregnancy Treated
Surgically/Non-Major Interv
Unstable Angina Heart
Disease with Cardiac Cath
Fibroid Disorder
Major Ophthalmology Disorder
Congenital Cardiac Disorder
Complication of Transpl Organ
Aplastic Anemia
Cardiac Valve Disease
Newborn/Neonate 2000-2499
gr, Gestational Age 37+ Wks
Newborn/Neonate 2500+ gr
Short Gestation/Low BW
Newborn/Neonate 2500+ gr
Maj Respiratory Compl
Newborn/Neonate 2000-2499
gr, Gestational Age 35-36 Wks
Newborn/Neonate 2500+
grams, Jaundice
Normal Newborn Mult/Caes
Abortion for Fetal Anomaly
Treated Medically
Newborn/Neonate 1500-1999
gr Gestational Age 35+ Wks
Newborn/Neonate 1500-1999
gr Gestational Age 32-34 Wks
CMG
RIW
0.3781
0.5917
Scenario
1
24.5%
23.0%
Scenario
2D
101.0%
99.9%
0.9505
22.8%
59.0%
73.7%
79.6%
85.1%
0.4883
0.7728
0.7612
1.0059
0.9257
1.0503
18.5%
18.3%
18.3%
18.2%
17.9%
17.7%
58.5%
61.2%
55.4%
61.5%
64.2%
59.0%
74.1%
82.9%
67.1%
78.6%
81.5%
74.9%
79.5%
87.2%
75.6%
85.9%
89.0%
81.3%
88.3%
98.1%
76.4%
91.9%
96.4%
88.5%
0.2342
7.3%
64.2%
91.5%
94.6%
116.2%
0.2464
7.2%
63.5%
90.3%
93.3%
114.7%
0.4813
7.1%
66.1%
92.8%
97.0%
117.3%
0.4403
6.8%
64.6%
92.9%
95.5%
118.4%
0.3264
6.8%
62.9%
88.7%
92.2%
112.8%
0.2350
0.6479
6.6%
6.3%
63.4%
59.6%
90.0%
76.2%
93.4%
86.3%
114.4%
92.6%
1.4081
6.3%
64.4%
92.7%
95.3%
118.5%
2.7540
5.6%
65.2%
95.7%
97.5%
122.7%
Table 21. Partial results for estimating the marginal costs of patients assigned to medical CMG groups, sorted from
highest proportional marginal cost to lowest proportional marginal cost, scenario 1.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
38
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
39
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Scenario 2 estimates marginal cost based on several assumptions regarding the availability and
costs for additional labour resources required to accommodate one additional case. There are
inherent assumptions about the unit of incremental labour that can be purchased by the hospital
(i.e., a single additional hour of nurse time, versus a minimum call-in time of four hours, versus a
full-time-equivalent) as well as the appropriate wage rate for incremental labour resources (i.e.,
average hourly rate versus premium/overtime rate). While these assumptions may be
operationally realistic for a single incremental case, those same assumptions become
progressively more unrealistic as the number of incremental cases increases (e.g., to 10 or 100
additional cases). It should also be noted that some of these assumptions may be antithetical to
current Ministry policies. For example, the Ontario MOHLTC has established standards for fulltime versus part-time nurse hours. So, marginal funding based on part-time staffing costs would
seem to be inconsistent with a policy that requires hospitals to increase the ratio of full-time
nurses.
There is some evidence that the concept of marginal cost, and its corollary concept of
economies-of-scale, have limited, if any, practical relevance to actual hospital operations. In
practice, larger hospitals are generally not less costly than smaller hospitals for identical cases
(e.g., typical hernia repair or normal vaginal delivery). In other words, the cost of a hernia repair
in a 400-bed hospital is generally more, rather than less, than in a 300-bed hospital. In light of
this, it seems that the assumption that one can fund a 300-bed hospital for additional cases at the
marginal cost rate (until it can reach, for example, a 400-bed capacity) could be challenging to
defend.
Do these findings necessarily suggest that hospitals will decline accepting incremental volumes
even when they are fully aware that incremental funding, which is based on a scenario-driven
marginal cost, may be grossly inadequate and operationally unsustainable? The answer would
seem to be, not necessarily, since a) hospitals are highly motivated to increase their revenues and
b) they also have considerable operational latitude on how to accommodate these new cases (in
the absence of actually having excess resources) without incurring any additional costs.
In other words, hospitals are able to reallocate resources from the types of cases that are not
funded based on volume (i.e., that portion of a hospitals activity that is globally funded) to those
cases that are. Such reallocation of resources inevitably results in increases in activity of certain
types, and reductions in patient services in other areas. These reductions are commonly
manifested in increased wait times for elective surgery (not a priority list) and MRIs. The overcommitment of existing bed capacity to accommodate additional cases also frequently leads to
ER backups and cancellation of elective surgeries.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
40
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Consequently, from the perspective of Ministries of health, it would appear risky to decide to
fund only a portion of hospital services on an activity-volume basis, while a substantial portion
of services are funded globally (and the associated volumes not defined or monitored).
Alternatively, in lieu of reallocating resources, some hospitals might opt to purchase whatever
additional resources are deemed necessary, exceeding the marginal funding rate, and thus incur a
year-end deficit. Historically, in these situations, Ministries have, ultimately (and with few
exceptions) covered the hospitals deficit. On the other hand, should a Ministry decline to
provide the additional funding and insist that the hospital manage within the existing budget,
hospitals will likewise insist this can only be accomplished by reducing service volumes. In the
end, the Ministry will be put in the unenviable position of being seen to sanction hospital service
reductions.
The risks associated with marginal-cost-based funding have been tacitly acknowledged in the
Ontario Wait Time Strategy funding rates which are based on the assumption that most hospital
costs are variable. These rates include 100% of all direct department (e.g., nursing, labs,
imaging, therapy, etc.) costs plus 30% of indirect department (e.g., housekeeping, food services,
laundry and linen, medical records, etc.) costs. (Still, there is no way to guarantee that no
substitution is taking place).
The concept of marginal funding has considerable appeal to both funders and academics because
it is seen as a way to motivate providers to improve efficiencies (without actually having to
identify or establish that remediable inefficacies and significant excess capacity actually exist).
However, if the assumptions about how much excess capacity is really available are wrong,
serious service disruptions and/or budget deficits will inevitably result.
In summary, then, while it has been demonstrated that it is possible to calculate hospital marginal
costs with some considerable degree of accuracy based on various scenarios about which excess
resources are available, the application of marginal cost information for hospital funding should
be considered as a handle with care situation.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
41
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
For Policymakers
1. If purchasing additional hospital-based care, what range of scenarios presented above
credibly reflects operational characteristics of hospitals?
a. Should the ranges of scenarios reflect the availability of labor (and equipment) in
different regions? Should they?
2. Do hospitals know their costs well enough to be able to respond to financial incentives?
a. If hospitals respond, and change the mix of services they provide, can equity of
access be assured?
b. If hospitals respond, but do not know their costs, will marginal pricing create
financial risk for hospitals?
c. Can policy makers effectively monitor changes in hospitals clinical care and
financial management in a timely manner?
3. If purchasing significant volumes of hospital care (e.g., 400 knee replacements), can
hospitals achieve economies of scale that would result in unintended surpluses?
For Hospitals
1. What information is needed in order to determine whether marginal funding is a
legitimate opportunity to increase activity and revenue?
2. What range of scenarios reflects hospitals operational constraints and opportunities?
3. If providing additional hospital-based care based on marginal funding, are hospitals
costs of different types of care known to the hospital (say, CMGs or CACS)?
a. Is there confidence around hospitals clinical and financial data, and around
knowing patients costs?
b. Do hospitals marginal costs change with increasing volume?
4. Are the findings sensitive to teaching status or rural location differences?
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
42
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Acknowledgements
Financial support for this project was provided by the Canadian Institute for Health Information
(CIHI). The funders had no role in developing the methods or interpreting the results.
Access to anonymous data was provided by the B.C. Ministry of Health and the Ontario Ministry
of Health and Long-Term Care.
I thank the following people for providing their critical input into improving this report: Drs.
Chas. Botz, M. Barer, B. Evans, and N. Repin, P. Weeks, D. Mooney and several CIHI staff. Dr.
Jason M. Sutherland is a Scholar of the Michael Smith Foundation for Health Research.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
43
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
References
1. Dean, S. Personal communication. 2011.
2. Canadian Institute for Health Information. Standards for Management Information
Systems in Canadian Health Services Organizations. Ottawa: Canada : Canadian Institute for
Health Information, 2009.
3. Fetter, R.B., Shin, Y., Freeman, J.L., Averill, R.F. and Thompson, J.D. Case mix
definition by diagnosis-related groups. Medical Care. 1980, 18(2)S:1-53.
4. Busse, R., Schreygg, J. and Smith, P.C. Editorial: hospital case payment systems in
Europe. Health Care Management Science. 2006, pp. 9(3), 211213.
5. Sutherland, Jason M. Hospital payment mechanisms: An overview and options for Canada.
Cost Drivers and Health System Efficiencies Series, Paper 4. Ottawa : Canadian Health Services
Research Foundation, 2011. ISBN: 978-1-927024-00-3.
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
44
ESTIMATINGTHEMARGINALCOSTSOFHOSPITALIZATIONS
Appendix 1.
See file: CACS Results: Marginal Costs
Appendix 2.
See file: Surgical CMG Results: Marginal Costs
Appendix 3.
See file: Medical CMG Results: Marginal Costs
UBCCENTREFORHEALTHSERVICESANDPOLICYRESEARCH
45