Cost Analysis N287E Spring 2006 Professor: Joanne Spetz 10 May 2006
Dec 30, 2015
Ways to divide and analyze costs
Direct vs. Indirect Direct costs
Salaries, supplies, etc. Indirect costs
Benefits, depreciation, support departments
Variable vs. Fixed Variable costs Fixed costs Semi-fixed costs (step function) Semi-variable
What costs do you have control over?
All costs?Direct costs only?Variable costs only?
It’s very important to be clear about the control you have
Making decisions about the future requires information about
Avoidable costs Variable costs and some fixed costs
Sunk costs Fixed costs that cannot be undone
Incremental costs (marginal)Opportunity costs Other things you could have done
A note about opportunity cost
Other things you could have done have value Return on alternate investments Return from basic investment
This is why is discount future earnings and costs
Discounting future earnings
$100 received this year is more valuable than $100 received next year You could take the $100 this year and
invest it to get interest for next year
Thus, future earnings are discounted If “discount rate” is 5%, then next year
is worth 5% less than this year
Numerical example of discounting
$100 per year to be received for 5 yearsYear 1 - $100 no discountYear 2 - $100 discounted 5% = 100*.95 = $95Year 3 - $100 discounted 5% twice = (100)*(.95)*(.95) = $90.25Year 4 - $100 discounted 3 times = $85.74Year 5 - $100 discounted 4 times = $81.45
Measuring costs in a hospital
Units are categorized by Direct or indirect cost Revenue-producing or not
Nonrevenue units are usually indirect costs Indirect costs are allocated to
revenue-producing units to make pricing decisions
Ways to allocate indirect costs
Step-down method The department with the least service from
others allocated first Go in order form least to most Problem: results vary by order of allocation
Double-distribution method Go through the loop twice
Simultaneous equation method Create equations for allocation and solve the
math
The math problem
Fixed cost + (variable cost * quantity) = price * quantity
AFTER SOME ALGEBRA…
Quantity = (fixed cost)/(price-var cost)
ORPrice =((fixed cost)/quantity) + var cost
Creating a standard cost profile
A standard cost profile (SCP) is a cost breakdown for a single item/task
SCP for an IV…
Cost category
Quantity
Req’d fixed
Unit cost
Var cost
Av fixed cost
Av total cost
Direct labor .10 .05 $20 $2 $1 $3
Materials 1.00 0 $3 $3 $0 $3
Dept overhead
0 .25 $1 $0 $.25 $0.25
Allocated costs
0 .50 $4 $0 $2 $2
TOTALS $5 $3.25 $8.25
Average fixed cost = fixed units needed multiplied by unit cost
Assume that…
The nursing department was budgeted for 100 IV’sThe department did 90 IV’sTo do these IV’s, the hospital used 15 hours of labor and paid $22/hourWe can examine how this varied from our budget…
Efficiency variance
Efficiency variance = (actual Q – standard Q) * standard price
Standard Q = Var labor req * IV’s done + budget fixed=.10*90 + .05*100 = 9+5 = 14
Eff var = (15-14)*$20 = $20
Volume variance
Volume variance = (budget Q – actual Q) * av fixed cost
per unit
= (100-90) * $1 = $10
These add up…
Actual direct labor cost = $22*15=$330Standardized cost = 3*90=$270
Difference between these = $60Price variance = $30Efficiency variance = $20Volume variance = $10
These add to $60These tell us what share of overrun
came from price, efficiency, volume!
Standard treatment protocols
A cost sheet for a larger “product” E.g., an inpatient stay or diagnosis
It looks like a SCP, for the most partFor a STP, you can compute: Intensity variance =
(actual SU’s – std SU’s)*std cost per SU=(90-100) * $8.25 = -$82.5This is favorable because fewer IV’s were done than expected.
Variations in costs…
Average = mean =x = x/NVariance = ((xi -x)2)/(N-1)
Standard deviation = variance =
If “normally” distributed:68% will be within one std dev95% within two std devs99.7% within three std devs
More measures
Median Half of sample is above Half of sample is below
Percentiles 25th percentile = 25% are below
Use of these statistics
You want to investigate abnormally high or low costsYou want to do investigations only when the “payoff” is worth it Payoff defined by cost of investigation
and potential benefit of correction
You can use statistics to determine your “cutoff” for investigation
Payoff tables
Is the unit behaving properly?
Action In control Not in control
Investigate I I+C
Do not investigate
0 LI = cost to investigateC = cost to correctL = loss with no correction
Payoff tables and statistics
If P = probability of being in control… (1-P) = probability of not being in control
If we investigate: Cost = P(I)+(1-P)(I+C) = PI+I+C-PI-PC =
I+(1-P)C
It we do not investigate Cost = P(0)+(1-P)L = (1-P)L
Cost comparison
If the cost of investigating is greater than the cost of not investigating, we don’t investigate: If I+(1-P)C < (1-P)L investigate
I + C – CP < L – LP-CP < L-LP-I-CLP-CP < L-I-C(L-C)P < (L-C) – IP < ((L-C)-I)/(L-C) = 1-(I/(L-C))
How to determine P?
We can guess P based on distribution of data, or just make a best guessWe can focus on cases a certain number of standard deviations from mean to define P
When analyzing cost data…
One can examine: Prior period values
(variance over time) Departmental values
(variance within and across departments)
There are many numerical examples in Cleverly
As a nursing manager…
What can you do to control costs?
Identify sources of savings
Develop strategies for change
Identifying sources of savings
Reducing costs does not have to reduce quality
There is wide variation in nursing costs
Survey of 180+ acute care hospitals from ~1998
Total cost per patient day
Labor cost per pat. day
Case-mix index (CMI)
Average of top 25%
$323 $297 1.40
Median $235 $212 1.25
Average of lowest 25%
$188 $174 1.25
Even the best performers have variance
Hamel Hospital Total nursing cost per patient day =
$186 21% below $235 median
Within the hospital, cost variance per patient day (compared to Hamel) Critical care 7.4% better than median Med-surg 3.8% better than median Intermediate care 42.1% worse than median
Where do differences comes from?
Differences do not appear to come from Shifting tasks to “support” departments Reductions in skill mix
They do appear to come from Reduced overtime Reduced per-diem Fewer FTEs overall (is this good or bad?)
How do you compare your hospital’s costs?
Each hospital is unique
Start with national benchmarksOther approaches: Across-the-board reductions Bottom-up campaigns
Cost-saving strategiesBenchmarking Across-
board cutsBottom-up campaigns
Advantages “objective” “fair” “support from staff”
Effective at aggressively reducing cost
Good for morale
Disadvantages
No regard for individual needs
Penalizes top performers
Small cost savings
Staff resents this
Results are arbitrary
Usually focuses on less important causes of high costs
Creating a good report is important
Variance Report
Unit Actual Budget Variance Actual Budget Variance3N $195 $182 ($13) 51 11,070 8.90 8.10 (0.80)3W $217 $185 ($32) 88 2,739 10.10 9.40 (0.70)5N $145 $146 $1 53 11,316 6.80 7.20 0.40CCU $549 $464 ($85) 3 704 19.90 17.30 (2.60)ICU 1 $526 $486 ($40) 13 2,854 21.80 18.90 (2.90)ICU 2 $523 $489 ($34) 23 4,968 20.91 18.13 (2.78)Maternity $171 $180 $9 31 6,746 6.92 6.41 (0.51)6N $163 $149 ($14) 37 8,040 7.70 7.40 (0.30)9E $163 $147 ($16) 20 4,294 7.90 6.70 (1.20)Mother/Baby $322 $249 ($73) 19 4,040 13.10 10.70 (2.40)NICU $299 $309 $10 21 4,419 11.70 12.20 0.50
Total cost per patient day Worked hours per patient dayAverage daily census
Patient days
Problems:1. Comparable units not compared clearly2. Benchmarking by budget assumes budget was good3. No quality metrics4. No staff turnover metrics5. Patient days might miss stays under 24 hours6. No adjustment for turnover of patients7. No acuity adjustment
Creating a good report…Worked Hours per Patient Day, Med/Surg Unit 3N
0
1
2
3
4
5
6
7
8
9
10
Unit 3N Actual Unit 3N Budget Internal benchmark National benchmark
Problems:1. Budget might reflect historical
underperformance2. Why is the internal benchmark 6.8? This is
5N’s actual, but it is comparable?
A better report!Variance Report
Unit Actual BudgetNational
benchmarkVariance -
budgetVariance - benchmark
3N 51 11,070 8.90 8.10 7.77 -9.9% -14.5%3W 88 2,739 10.10 9.40 9.76 -7.4% -3.5%5N 53 11,316 6.80 7.20 7.77 5.6% 12.5%6N 37 8,040 7.70 7.40 8.71 -4.1% 11.6%9E 20 4,294 7.90 6.70 8.91 -17.9% 11.3%Maternity 31 6,746 6.92 6.41 11.18 -8.0% 38.1%Mother/Baby 19 4,040 13.10 10.70 6.80 -22.4% -92.6%CCU 3 704 19.90 17.30 18.87 -15.0% -5.5%ICU 1 13 2,854 21.80 18.90 18.70 -15.3% -16.6%ICU 2 23 4,968 20.91 18.13 18.84 -15.3% -11.0%NICU 21 4,419 11.70 12.20 10.57 4.1% -10.7%
Average daily census
Patient days
Worked hours per patient day
Units grouped by similarity
National benchmark
Unrealistic budget?
Beating the budget but not the benchmark
Another good reportVariance report
Unit Actual BudgetInternal
BenchmarkNational
BenchmarkVariance Budget
Variance Internal
Variance National
ICU 1 20.3 20.1 20.0 18.1 -1.0% -1.5% -12.2%ICU 2 20.8 20.9 20.0 18.1 0.5% -4.0% -14.9%ICU 3 20.0 19.8 20.0 15.8 -1.0% 0.0% -26.6%Tele 1 10.4 9.9 10.4 8.2 -5.1% 0.0% -26.8%Tele 2 10.7 10.2 10.4 9.1 -4.9% -2.9% -17.6%Surg 1 3W 8.5 8.2 8.1 6.5 -3.7% -4.9% -30.8%Surg 2 3E 10.3 8.6 8.1 6.5 -19.8% -27.2% -58.5%Med 1 5N 9.2 8.5 8.1 7.1 -8.2% -13.6% -29.6%Med 2 4N 8.1 7.9 8.1 6.8 -2.5% 0.0% -19.1%Med 3 4S 8.2 8.5 8.1 6.8 3.5% -1.2% -20.6%Med 4 7E 8.7 8.6 8.1 7.6 -1.2% -7.4% -14.5%
Worked hours per equivalent patient day Variance Analysis
Internal benchmarks are important
Compare to national benchmark – can be more aggressive?
Better range of comparisons
Still room to improve!
Some issues & ideas
Use the internal best performer to get ideas for improving other unitsMake units’ data comparable Use the same acuity system Make sure national benchmark has
same acuity system
The problem with midnight census
7am – 3pm 24 patients3pm – 11pm 29 patients11pm – 7am 20 patientspatient days
24.3 = average census
So… Actual HPPD Target HPPDPt days 6.26 6.01Blended ADC 5.15 6.01And adjust for admissions, discharges,
transfers
Be logical in figuring out where costs are
uncontrollable controllable
unit pat nurse regulations non-RN laborconfig mix comp. labor
overhead supplies too expensemany per FTE
cost too too FTEsper much many cost per richsupply orderedused direct indirect RN too skill
hours hours high mix
premium agepay mix