Download Efficiency and Effectiveness

St. Jude Children’s Research Hospital
International Outreach Program
Seventh Meeting
Calvo MacKenna BMT Cost Study
December 19, 2007
Agenda
– Discuss Chapter 12 and 15
– End of Meeting Administration
Chapter 12 Outline
Chapter 12: Measuring Productivity
– Productivity
• A measurement of inputs required to produce an
output
• An area where the use of cost accounting ratios
provide information for improving management of a
health care organization
• Cost accounting has begun to focus on the
problems related to productivity measurement
Productivity Measurement
– The measurement of productivity
– Hard to determine in industries, including
health care
– The difficulties arise from the problems related
to quality and outcomes measurement
– Outputs are difficult to define in health care
– The actions of employees determines how
productive the health care organization will be
Productivity Defined
– The ratio of any given measure of output to
any given measure of input over a specified
time period
– Most common productivity measure is output
per labor hour
• Measures only part of the organization’s
productivity
• Does not account for the amount of capital
equipment used in producing the output
– Either state the results in dollar terms or in
physical units
Total Productivity
– The ratio of total outputs to total inputs
– The amount of output per unit of input
– Usually expressed in dollar terms
– Shows the financial benefits of improved
productivity
– Considers the increase in wages or the price
of supplies, as well as the quantity of inputs
consumed
– Must exercise caution to keep the unit of
evaluation constant over time
Total Productivity
Total Outputs
Total Productivity =
Total Inputs
- From a finance standpoint we measure productivity
as follows,
P=
Operating revenue
Supplies + labor + capital + overhead
Total Productivity Problems
– When using dollar amounts, inflation can
cause problems
• Have to keep the calculations in constant dollars
– Potential changes in quality
• Productivity measures assume that quality is held
constant
– The total productivity ratio is inadequate to
segregate the impact of case-mix changes
from the impact of productivity changes
Partial Productivity
Total Outputs
Partial Productivity =
Partial Inputs
– A measure of output compared with a partial measure
of input
– More frequently measured in physical units
– Useful for labor negotiations, monitoring continued
efficiency of operations, and identifying places to
improve the operations of health care organizations
Partial Productivity Example
– We will consider the technician labor hours necessary
for a computed tomography (CT) scan
– We are assuming that CT scans come in only two
types, head and body
Scan
Labor
Hours
Number
Total
Actual Number
Actual
Current Volume ×
Type
per Scan
of Scans
Hours
of Scans
Labor Hours
Base Period Hours
Head
1
100
100
120
N/C
120 × 1.0 = 120
Body
1.5
80
120
70
N/C
70 × 1.5 = 105
180
220
190
215
225
Totals
*N/C = Data not collected
Statistics from
Partial Productivity
Actual output hours
– Index of output = Baseline output hours × 100% = 102.3%
• Our output increased by 2.3% over the base period
Actual labor hours
– Index of labor hours = Baseline labor hours ×100% = 97.7%
• We actually worked only 97.7% as many hours in the
current period as in the baseline case.
– Index of output per labor hour =
Index of output
Index of labor hours
× 100% = 104.7%
• Our output in the current period per labor hour is 104.7%
of that in the baseline period.
Productivity and Indirect Costs
– An alternative to the traditional comparison is
to develop a productivity measure based on
the comparison of direct costs and indirect
costs
– Many departments in a health care
organization use some resources in direct
patient contact and other resources indirectly
– If there is a relationship between hours of
direct patient care time and indirect time then
we can monitor this on a monthly basis to see
if productivity is being improved or maintained
Managing Discretionary Costs
– A major challenge for management
– Costs incurred in departments that are essential to
the organization but that do not have simple inputoutput relationships.
• Examples: personnel, marketing, legal, finance,
administration, housekeeping and security
– Productivity is difficult to assess in these
circumstances.
– Health care costs can be classified into three broad
categories
• Engineered costs
• Committed costs
• Discretionary costs
Engineered Costs
– Costs for which there is a specific input-output
relationship
– These relationships can be readily observed
– Normally include the direct materials and direct labor
cost
– Can be controlled by using flexible budget variance
reports
– Examples
• More patient days require more meals from the dietary
department
• More X-rays require more sheets of X-ray film
Committed Costs
– Costs that cannot be changed in the short run, such
as during the coming year
– They are generally reviewed as part of the capital
budget process
– Once committed to they usually do not vary from year
to year
– Changes in volume of services provided often have
no effect on the committed costs
– Examples
•
•
•
•
The depreciation cost on a nursing home building.
Long-term leases
Depreciation on equipment
Insurance
Discretionary Costs
– Are costs that are incurred, typically each
year, in an amount that is approved as part of
the normal budget process
– Sometimes referred to as managed costs
– The budget for discretionary costs is generally
based on negotiation.
Zero-Base Review
– Zero-base Budgeting
•
•
•
•
•
Requires that all costs be examined and justified
Requires examination of alternatives
Expensive and time-consuming
Usually done once every five years
Most effective in terms of discretionary costs
because they do not have a clear relationship
between inputs and outputs
• Helps you understand what types of objectives are
being accomplished by discretionary cost centers
and what resources are being devoted to
accomplishing the various objectives
Work Measurement
– Another approach to dealing with discretionary costs
is to perform a review of activities with the hope of
converting a discretionary cost center into a
engineered cost center
– Work measurement is a technique that evaluates
what a group of workers needed to accomplish the
task efficiently
– Example: the mail room
• It would be hard to relate mail room costs with patient days.
• But you could relate pieces of mail sent out and pieces of
mail received and sorted
– With the proper measurement of the type of work
being performed, it might be possible to reclassify
many costs as engineered.
Efficiency and Effectiveness
– Effectiveness
• A measure of the degree to which the organization
accomplishes its desired goal
• Measuring effectiveness
– A set of goals should be established for the discretionary cost
center
– Must evaluate to see if the goals have been met
– If the goals are met then the department is effective
– Efficiency
• A measure of how close an organization comes to minimizing
the amount of resources used to accomplish a result
• For a given result, the organization should attempt to
minimize the cost of the resources required
Effectiveness & Efficiency
– By examining effectiveness and efficiency
one can determine
1. Whether or not the goals were achieved
2. The cost of what was actually achieved
– We must look at both effectiveness and
efficiency
– There can be trade offs between the two
•
•
Sometimes to be effective we have to be less
efficient
Healthcare tends to sacrifice some efficiency for
effectiveness
Effectiveness & Efficiency
–Goal is to have effectiveness and
efficiency in balance
Effectiveness & Efficiency
– Approaches to ensure the efficiency and
effectiveness of a discretionary cost
department
1.Thorough review of all budget elements
2.Development of monitoring tools for assessing
efficiency and effectiveness
3.Introduction of competitive market forces
4.Leadership
5.Formalized spending approval mechanism
6.Promotion of an organizational culture
Chapter 14 Outline
Chapter 14: Dealing with Uncertainty
– Health care managers make many decisions after first
making financial estimates.
– Potential revenues and costs are predicted based on a
number of calculations, and decisions are based off of
these.
– A degree of uncertainty is inherent in such estimates.
– We will look at four common approaches that help
improve the predictions made and used for decisions.
The Expected Value Technique
– Expected value analysis
• Estimates the costs or revenues based on the
likelihood of each possible outcome
• Key focus is on the possible events that might
occur
• Management can take actions to affect the
outcomes, but events are occurrences that are
beyond the control of managers
• A weighted average of the outcomes with the
probabilities of each outcome serving as the
weights
• Outcomes are measured in monetary terms
Using Expected Value
– To calculate, first establish a probability distribution
for the possible states of nature
– Example:
• Suppose that there is a 25% chance that the competitor will
be a lithotriptor and a 75% chance that it will not, if we buy
one.
• Suppose that the profits from 1,000 patients are expected to
be $50,000, whereas the loss related to 700 patients will be
$150,000
Lithotriptor Project
Probability
Payoff
We buy/they buy
0.25
(150000)
We buy/ they do not buy
0.75
50000
1.00
Using Expected Value
– To determine the expected value of this project:
We buy/they buy
.25 × ($150,000) = ($37,500)
We buy/ they do not buy
.75 ×
Expected value
50,000 =
=
37,500
$
0
– The large potential loss is so great that it just offsets the
benefit from the probable gain, and the expected financial
result is zero.
– The hospital may still buy the machine to keep physicians
happy and to improve the hospital’s reputation and service to
the community
Expected Value
Example (pg. 309)
Full Range of Possible Events
– In making decisions based on expected value,
we must consider all possible states of nature
that might occur.
– We have only considered 2 options:
• We buy and they do not
• We buy and they buy
– But there are two more options that need to
be considered:
• We do not buy and they do not buy
• We do not buy and they buy
Expected Value Continued
Example (pg. 310)
Subjective Estimates
– Subjective probability
• A probability that is based on a manager’s estimate rather
than known odds
• All our estimates are based on it
– Managers must use their knowledge, experience, and
judgment to make a best guess about the likelihood of
each event.
– There is always a risk that unknown negative factors
not taken into account will determine the probabilities.
– Therefore, many organizations demand a substantial
positive expected value to move forward with a
project.
Subjective Estimates
Example (pg. 311)
– The probability changes
Competitor
buys
Competitor
does not buy
We buy
10%
90%
We do not buy
80%
20%
Another Expected Value Example
– Suppose that we are planning the staff level for the
emergency department (ED).
– The ED managers know that it is most economical to
staff the ED at just the level needed to provide care to
all patients who arrive.
Cost/Patient ($)
Low Medium High
Staff
Staff Staff
ER Cases
Probability
20,000-25,000
.10
50
60
70
25,001-30,000
.40
55
50
60
30,001-35,000
.35
60
50
55
35,001-40,000
.15
70
60
50
– The cost per patient is minimized when the staffing
pattern is correct for the volume of patients.
Simulation Analysis
– A tool that can take the various errors in all the
estimates into account and provide managers with the
likelihood of actual results being substantially different
from the budgeted projection.
– Avoids ignoring the possibility of unfavorable
outcomes compounding other unfavorable outcomes
– Helps improve the accuracy of the decisions
managers make
– Frequently used in health care
Sensitivity Analysis
– Allows managers to consider various possible actual
results and see their impact on the projection.
– Tells a manager how sensitive the profits of the
venture are to any given change in one or more
variables.
– Done using a spreadsheet program, such as Excel
– Requires management judgment in determining a
reasonable set of “what-if” assumptions.
Probabilistic Example
– Based on a manager’s subjective estimate of
what is most likely to occur.
– The most likely outcome is generally the one
the manager has used to make the original
estimates for the projection.
– When making a projection, the probabilistic
nature of the estimate is often ignored.
– Does not give the manager a sense of how
likely or unlikely the result is to occur.
Simulation Technique
– Similar to a sensitivity analysis, except it goes further
in generating a value for each variable in the model.
– The key to simulation analysis is that the value
chosen for each variable is based on a probability
distribution.
– Managers must indicate how likely they believe a
variable is to take on a variety of different values.
– Will select each of the values on a random basis,
adjusted for probability.
Implementing Simulation Results
– The manager must use the simulation results
in making a decision.
– Relatively profitable and unprofitable
organizations can afford to take risks, but
those that fall in the middle are in a difficult
position.
– Most small to medium size health care
organizations fall in the middle.
Limitations of Simulation
Analysis
– Primary concerns
• May be prohibitively costly
• The model may be built incorrectly
• The probability estimates may be incorrect
Network Cost Budgeting
– A popular management tool used for planning and
controlling nonroutine, complex projects
– May be beneficial for some ongoing activities
– Uses arrow diagrams, that finds a critical path that
determines the least amount of time for project
completion
– Refers to combining the techniques of network
analysis with cost accounting to generate the most
cost-effective approach to the project.
– Often assumed that the network is laid out in an
optimal fashion, but it can indicate other possible
timelines.
• Use overtime or extra staff reduces the amount of time, but
the cost usually increases
Network Analysis
– Uses a diagram to show the relationships in a
complex project
– Arrows represent specific activities that must
be done
– Crucial aspect: identifying the activities that
must be completed before some other activity
or activities can commence
Network Analysis Example
– Assume that a hospital has determined midway
through the year that revenues are below
expectations and a financial crisis is anticipated.
– The Decision: Select five departments that have the
potential for budget cuts that would help offset the
crisis
– A zero-based review is conducted in each one.
• Each element of the budget is reviewed and its role
determined
– Based on the information from the review, a decision
would be made concerning how much to cut from the
operating budget of each department.
Network Analysis
Example (pg. 320)
– Each activity starts at a numbered point
– Each activity has a number at its start and at
its completion
– The activity can be identified by these two
numbers
• Examples: Activity 1-2 or Activity 2-4
Network Analysis
Example (pg. 321)
– An alternative presentation is to place each
activity into a “node”
– You identify each activity by referring to it by
the name or description in the node
Critical Path
– Always found as an element of network
analysis
– To find: it is necessary to determine the
expected length of time for each activity
– Critical path method (CPM)
• A program of techniques that indicates the cost
and time for each element of a complex project
and indicates cost/time trade-offs where applicable
Critical Path Example
– Assumed that an initial plan was developed as
follows:
•
•
•
•
•
•
•
•
Develop an audit plan, assign staff, schedule audits: 31 days
Audit Department A: 22 days
Audit Department B: 15 days
Audit Department C: 13 days
Audit Department D: 9 days
Audit Department E: 19 days
Review audit results: 31 days
Make budget modification decisions (top management), hear
appeals, make final decision: 31 days
– All days are shown, including weekend days.
– In this plan, it was not assumed that any work would
take place on weekends.
Critical Path
– Decided there would be two teams of auditors
– Fastest approach
• One team: Audits A and then E
• Other team: Audits B, C, and D
Critical Path
Example (pg. 322)
– One route or path through the network is
1-2-3-5-6-7-8
• The total days: 130
– The other route or path through the network
is 1-2-4-6-7-8
• The total days: 134
Critical Path
– Determined by totaling the days required for
each possible route through the network and
finding the longest one.
– All tasks must be undertaken.
– The project can be achieved in no less time
than the time it takes for the longest path.
Benefits of Planning
– Planning takes time
– Conducting a zero-base budget review requires a careful audit
plan.
– Meetings must be held to determine the goals of the audit and
the specific audit activities.
– Auditors and departments have to schedule a mutually
convenient time to do the audit.
• Tasks make up Activity 1-2
– A review of the plan by the top management can result in
determining how important each activity is
– If the activity is important enough, then the individuals involved
will be told to set all else aside temporarily.
– Audits should begin as soon as the auditors are ready.
Revised Network
Example (pg. 323)
– Revised network has a new critical path that is only
65 days long.
• More than half if the project time has been eliminated by
indicating priority.
• Reflects the significant benefits that can be obtained by
creating and reviewing a plan before starting work on a
project.
Working Time vs. Slack Time
– Much of the time in any project is waiting time, also
known as slack time
– Waiting time between activities is identified in
network analysis
– Waiting or slack time can be found within an activity
– Solution to the problem of slack time within an
activity is to break the activity into a number of
smaller activities.
– Each waiting period needs to be indicated in the
plan.
Work Time vs. Slack Time
Example (pg. 324)
Consideration of Costs
– Assumption of network analysis
• Finding the critical path and minimizing it also
minimizes costs
– Knowing which paths have slack and allowing
that slack also keeps costs down
– Cost minimization is not necessarily optimal
• It may be beneficial to push a project through, even
if at a higher cost
– Network cost budgeting helps determine the
optimal time and cost budget for the project
Network Cost Calculation
– The accountants need to review each activity and the
various individuals involved
– Example: Activity 1-2 requires meetings for planning
activities.
• Which members of the meetings are paid an annual salary
with no overtime, which will be paid overtime?
• What is the extra cost per overtime?
• How much overtime would have to be worked to reduce the
activity required time by one day?
– This process needs to be done for each activity.
– Once done, we can determine which activity is the
least costly to shorten by one day.
Network Cost Example
Activity 1-2
– Example: Suppose the
following information was
calculated for Activities 1-2
and 2-4
Original days:
10
Extra cost for 9 days:
$800
Extra cost for 8 days:
1,100
Extra cost for 7 days:
1,100
Extra cost for 6 days:
NF*
Activity 2-4
– The next step is to
determine the benefits of a
shortened project and
compare the costs of the
least costly way to reduce
one day with the benefits of
a reduction
Original days:
22
Extra cost for 21 days:
$400
Extra cost for 20 days:
400
Extra cost for 19 days:
400
Extra cost for 18 days:
400
Extra cost for 17 days:
600
Extra cost for 16 days:
600
Extra cost for 15 days:
NF*
NF*= Not Feasible
Linear Programming
– A mathematical technique that allows the user to
maximize or minimize the value of a specific objective
in a constrained environment
– Unconstrained maximization or minimization is rarely a
realistic goal.
– Constraints generally exist and must be considered as
we attempt to maximize or minimize the objective.
– The more complex an organization is, the more
difficult it becomes to determine exactly how to
maximize objectives.
Linear Programming
– It develops a series of equations based on the
stated objective and subject to its various
constraints
– Equations are solved at the same time
– The more accurate the equations are that
state the existing relationship, the more
accurate the resulting model and its solution
will be.
Developing Equations
– First equation to be developed: objective
function
• It states the relationship between the objective and
the other variables in the process
• Regardless of the objective chosen, it can be
modified to meet the mission of the organization by
the various constraints that will be added to the
model
Example
– If the hospital wanted to maximize revenue, the
equation would set revenue equal to the sum of the
price of each product multiplied by the units of each
PAD product
– If we offered only 3 products: PAD1, PAD2, PAD3
then revenue is defined as follows:
Revenue = PAD1 rate × # of PAD1 patients
+ PAD2 rate × # of PAD2 patients
+ PAD3 rate × # of PAD3 patients
– Increasing the number of patients in any of the three
PADs would increase the total revenue and would
make the hospital more profitable.
Example
– We want to know the average length of stay
(LOS) of each type of patient, and the
maximum number of patient days that can be
attained given the hospital’s number of beds.
– Example
•
•
•
•
•
PAD1 patients: LOS= 5 day
PAD2 patients: LOS= 10 days
PAD3 patients: LOS= 7 days
Hospital has 100 beds for 365 per year
Full capacity= 36,500 patients
Equation for a Constraint
Revenue = PAD1 rate × # of PAD1 patients
+ PAD2 rate × # of PAD2 patients
+ PAD3 rate × # of PAD3 patients
Revenue = 5 × # of PAD1 patients
+ 10 × # of PAD2 patients
+ 7 × # of PAD3 patients ≤ 36,500
Linear Programming Example
– If the hospital believes that PAD3 is essential
to the community that it must be offered even
if it doesn’t earn a profit, then a constraint can
be established that sets the output level of
PAD3 greater than a certain level, such as 0.
# of PAD3 patients > 0
– Setting the level at zero will make the
organization fairly aggressive because of the
probable fixed costs.
Linear Programming Example
– If the hospital wants to guarantee some level
of service for that PAD, such as 1,000 patients
or more per year, the constraint can be set at
that level:
# of PAD3 patients ≥ 1,000
Example Continued
– We can create a
spreadsheet that
calculates total
revenue, if we know
the expected payment
rate for each PAD.
PAD
Expected
Payment
(A)
Volume
Length
of Stay
Revenue
Bed
Days
(B)
(C)
(A×B)
(B×C)
PAD1
$1,000
0
5
$0
0
PAD2
$15,000
0
10
$0
0
PAD3
$5,000
0
7
$0
0
$0
0
Total
Constraints
– Also need to establish additional constraints,
many typically involve costs.
– If we are maximizing revenues, the objective
is constrained by the fact that we must at least
break even.
• A constraint would be that revenue must exceed
the sum of all costs.
Example Continued
– Cost information is added to example:
DRG
Variable Costs
per Patient
Volume
Total Variable
Costs
PAD1
$500
0
$0
PAD2
$14,750
0
$0
PAD3
$4,950
0
$0
Fixed Costs
$1,000,000
Total
Cost
$1,000,000
Example Continued
– Sheet 2 from the Excel workbook in the
documents section of C4K website
• Shows the results of the linear program with a
breakeven constraint added to the earlier model.
• Adding the breakeven constraint had changed the
solution.
– Total revenue has fallen to $45,566,667
– We are now serving all 3 PADs
Equations
– Generic format of equations:
Maximize: R= p1x1 + p2x2 + …+ pnxn
Subject to:
a11x1 + a12x1 + … + a1nxn ≤b1
a21x1 + a22x2 + … + a2nxn ≤b2
am1x1 + am2x2 + … + amnxn ≤bm
n
Maximize: R = Σ pjxj
j=1
m
Subject to: Σ aijxj ≤ bi
j=1
p1 = the price of the first product
x1 = the output level for the first product
n = the number of variables or products
m = the number of constraints
j=1
Linear Programming
– Once the equations have been specified and
the data input into a computer, a solution that
tells the optimum mix of patients is generated.
– May seem unrealistic but even if it seems
unreasonable, it can help point you in the
proper direction.
– Linear programming can significantly improve
decision making for the health care
organization.
End of Meeting
– Minutes
– Final Meeting
• Chapter 18
– Open Discussion