Average versus incremental costs

Practical Costing
Henry A. Glick, Ph.D.
Pharmacoeconomics
March 19, 2012
www.uphs.upenn.edu/dgimhsr
“Economic” Measures of Cost
• Standard economic assumption
– Purchase price = marginal cost (= average cost)
• Health care (particularly U.S.)
– Purchase price ≠ cost or there is no price to observe
• Difference due to
– Health care consumers not having adequate
information
– High levels of insurance
– Regulation
– Hospital internal pricing policies; free care
– Economics of scale / fixed costs
Marginal Costs
• Always want to use marginal cost when evaluating a
therapy, not average cost (total cost / total units of
service), no matter which perspective is adopted
– Marginal costs: Costs incurred by adding an extra unit
of service or saved by reducing a unit of service
(excludes “truly” fixed costs)
• Perspective can affect how we define marginal costs
– Payments are unlikely to represent marginal costs
from the social perspective, but from the payer’s
perspective they often do represent marginal costs
• From the social perspective (and others?) we don’t
usually use marginal costs
– Unknown or unappreciated
1

Marginal Costs (2)
• In health care, marginal costs generally differ from
average costs
• What is counted as a marginal cost may vary depending
on how we define the program we are evaluating
– Creation of a new program versus expansion of an
existing program
– Temporary program or permanent
– Small program or large
Variable vs Fixed Costs
• Economics textbook assumption: In the short run, labor
is a variable input while capital is a fixed input
• Increasing output from its current level in the “short run”
implies the addition of labor but no addition of capital
(which is assumed to be fixed in the short run)
• Given enough time, capital can be varied
• Implication: short run marginal cost will usually be higher
than long run marginal cost because there is more
freedom to adjust to the ideal level of labor and capital in
the long run
“Discretionary” Fixed Costs
• The fact that some costs aren’t observed to change
doesn’t mean they are fixed
• Technically, fixed costs are those that are impossible or
foolish to change
– e.g., we might elect to maintain a constant staffing
level independent of number of patients on a ward,
but that’s a choice we make
– We could rent buildings and add or subtract rented
space as volume changes
• Cost of discretionary “fixed” costs equals their net benefit
in their next best use
– Often 0, but not always
2

Transactions ≠ Social Costs
• Not all transactions represents social opportunity costs
– Pure profits (monopoly or fortuitous)
– Fixed costs
– External benefits
– Transfers
• Social security
• Disability payments
• Unemployment benefits
Translating Service Use to Cost
• If we have collected information about medical service
use, need to translate it into a cost that can be compared
with the measure of study effectiveness
• Common to
– Identify medical service use
– Identify price weights (unit costs) and multiply service
use times price weights
– Sum the products
• Alternatively, administrative data -- such as billing
records -- may provide measures of both medical service
use and cost (or expenditure)
Costing Specific Services
3

Outline, Costing Specific Services
• Readily available U.S. data
–Hospital costs
• Charges and cost-to-charge ratios
• DRGs
• National Inpatient File (HCUP)
– Physician fees
• Medicare Fee schedule
• Hospital outpatient bills
– Other federal fee schedules
– Drug prices
Hospital Costs
• Readily available U.S. hospital cost data
– Charges and Cost to Charge Ratios (Medicare Cost
Report)
– Diagnosis Related Groups (DRGs)
– National Inpatient File
• Hospital accounting data
Cost-To-Charge Ratios
(Medicare Cost Report)
• In US, total charges represent amounts hospitals "bill"
for services, but not how much hospital services actually
cost nor the specific amounts that hospitals receive in
payment
• CMS cost-to-charge ratios represent the quotient of CMS
recognized costs divided by hospital charges
(operating+capital CCR)
• Web resource: FY 2011 Impact File
http://www.cms.gov/AcuteInpatientPPS/FFD/itemdetail.asp?filterType=
none&filterByDID=0&sortByDID=1&sortOrder=descending&itemID=
CMS1255464&intNumPerPage=10
4

Cost-To-Charge Ratios
(Medicare Cost Report) (Cont.)
• Advantages
– Readily available from all hospitals
– Uniform methodology
• Disadvantages
– Works on averages (how bad is that?)
– Includes fixed costs
– Not easy to estimate costs of intermediate products
(i.e., bundles of services such as a procedure)
New
England
Mid
Atlantic
South
Atlantic
E.N.
Central
E.S.
Central
2011 Operating Cost-To-Charge Ratios *
Large
Urban
.468
.433
.300
.276
.352
.335
.363
.328
.328
.286
Urban Classification
Other
Urban
.479
.437
.347
.315
.334
.327
.376
.357
.300
.286
Rural
.524
.481
.486
.451
.392
.380
.452
.440
.376
.345
* Row 1, Simple average; Row 2, Weighted by # of beds
W.N.
Central
W.S.
Central
Mountain
Pacific
Puerto
Rico
Total
2011 Operating Cost-To-Charge Ratios
.349
.321
.294
.263
.243
.229
.250
.239
.475
.467
.321
.297
.381
.349
.328
.278
.375
.348
.317
.295
.561
.549
.354
.329
Source: FY_2011_FR_Impact_File.zip
.438
.405
.394
.344
.475
.440
.412
.401
-
-
.419
.391
Total
.482
.439
.344
.297
.355
.337
.388
.352
.345
.304
.394
.349
.343
.283
.373
.310
.284
.259
.506
.492
.359
.320
5

Diagnosis Related Groups (DRGs)
• Are used by Medicare and a large number of insurers to
reimburse hospital services
• DRGs are a set of relative weights for hospitalizations
(identified by ICD-9-CM code and level of concomitant
conditions)
– i.e., the cost of other hospitalizations is reported in
relative terms compared to the cost of an average
hospitalization
– e.g., a hospitalization with a DRG weight of 2.0 is
considered to be twice as costly as a hospitalization
with a DRG weight of 1.0
Initial Development
• DRG weights initially based on relative average charge
for different types of hospitalization
– Hospitalizations assigned to DRGs
– Average hospital charge per DRG estimated
– Relative weight derived by dividing DRG-specific
average hospital charge by average charge for all
hospitalizations
• Charges for hospitalizations in individual DRGs were not
very homogeneous
– Use of the DRG weights explained only about 10% of
the variation in charges
Changes, October 2007
• Changed relative weights to reflect costs rather than
charges
– Unweighted mean of weights relatively unchanged,
although medical weights on average increased ~1%
while surgical weights on average decreased ~1%
• Expanded concomitant condition coding
– Without; With; With major
• Eliminated age coding
• Changed outlier threshold
• Renumbered DRGs
6

Advantages / Disadvantages
• Advantages
– Readily available
– Uniform methodology
• Disadvantages
– Different conditions in a DRG may have widely
varying costs
– Cannot (without adjustment) distinguish different
levels of intensity of care within a DRG
– Based originally on charges; did not explain a large
amount of variation
– Do not include physician fees
DRG Web Resource
http://www.cms.gov/acuteinpatientpps/downloads/FY_2011
_FR_Table_5.zip
Surg/
DRG
MDC Med
001
002
263
264
280
281
282
283
284
285
PRE
PRE
05
05
05
05
05
05
05
05
SELECTED DRG DATA (Table 5, 2011)
Surg
Surg
Surg
Surg
Med
Med
Med
Med
Med
Med
Description
Heart transplan or implant w/MCC
Heart transplant or implanc w/o MCC
Vein ligation & stripping
Other circulatory system O.R. proc
Acute MI, discharged alive w/MCC
Acute MI, discharged alive w/CC
Acute MI, discharged alive w/o
CC/MCC
Acute MI, Expired w/MCC
Acute MI, Expired w/CC
Acute MI, expired w/o CC/MCC
Rel
Wght
26.3441 31.6 31.9
13.6127 17.6 22.6
1.7565
2.5305
1.8503
1.1912
0.8064
1.7151
0.8888
0.5712
Geo
LOS
3.5
5.5
5.2
3.6
2.2
3.2
2.1
1.4
Arith
LOS
5.6
8.2
6.6
4.4
2.8
5.1
2.0
1.8
7

DRG Reimbursement
• Relative weights aren't payments
– Payment rates set by Congress annually
• Payment includes base payment, divided into laborrelated
and nonlabor
– Labor-related adjusted by wage index
– Nonlabor adjusted by cost of living adjustment
• Additional percentage add-ons are made if hospital:
– Treats a high-percentage of low income patients
(disproportionate share)
– Is a teaching hospital (indirect medical education)
–Has outlier cases
National Inpatient sample (NIS / HCUP)
• Largest all-payer inpatient care database in the US
• Contains clinical and resource use information included
in a typical discharge abstract
– Only national hospital database containing charge
(cost) information on all patients, regardless of payer
• Has data from approximately 8 million hospital stays
each year
– In 2008, represents 20% stratified sample of U.S.
community hospitals
National Inpatient sample (NIS / HCUP)
• Allows for hospital costs by subgroup
– Diagnosis Related Group (DRG) in use on discharge
date
– Diagnoses and procedures
• Hospital, county, and state identifiers for links to other
datasets
• Severity and patient level controls
– Age, gender, race, expected primary and secondary
payer, Median household income for patient's ZIP
Code
8

Other HCUP Databases
• National Kids Inpatient Database (KID)
• Nationwide Emergency Department Sample (NEDS) &
Statewide Emergency Department Databases (SEDD)
• State Ambulatory Surgery Databases (SASD)
HCUP Calculator
http://hcupnet.ahrq.gov/HCUPnet.jsp?Id=BCA3534C7A0C1
C28&Form=SelLAY&JS=Y&Action=%3E%3ENext%3E%
3E&_LAY=Researcher
“National Statistics on All Stays” Create your own
statistics…
All Hosp, 2009
LOS
4.6
(0.0)
Charge
$30,655
(758)
Cost
$9173
(126)
U.S. Physician Data: The Medicare Fee Schedule
• Used by Medicare to reimburse physician services
• Fee schedule is a set of relative value units (RVUs) for
medical services (identified by CPT code)
• RVUs report relative cost of providing a service
– i.e., some service designated as the baseline level of
work (RVU = 1.0); the work for all other services are
reported in relative terms compared to this baseline
– e.g., a service with an RVU of 2.0 represents twice
the work of a service with an RVU of 1.0
9

U.S. Physician Data: The Medicare Fee Schedule (2)
• Separate RVUs were estimated for the work involved in
a procedure (physician work RVUs), malpractice, and for
practice expenses (facility fees)
• Payment = RVU * $/RVU (set annually by Congress)
Initial Development
• RVUs were developed in an extensive study conducted
by clinicians, economists, statisticians, measurement
psychologists, and experts in survey research
• Expert panels of physicians within each specialty
enumerated their different tasks and assigned RVUs to
these tasks within their specialty
• Expert panels of physicians from multiple specialties
developed cross walks between RVUs in each of the
specialties (e.g., an RVU of 1.0 in specialty A might have
been determined to be equivalent to an RVU of 0.95 in
speciality B)
Advantages and Disadvantages
• Advantages
– Readily available
•Files
www.cms.hhs.gov/PhysicianFeeSched/PFSRVF/list.asp
• Look up calculator
https://www.cms.gov/apps/physician-feeschedule/overview.aspx
– Uniform methodology
• Disadvantages
– May be difficult to identify services provided in
individual physician / patient encounters
10

CPT/
HCPCS
99201
99203
99205
Selected 2011 Medicare Fee Schedule RVUs
Description
Office/op visit, new: problem
focused hist & exam; straight
forward med decis making; ~10 min
Office/op visit, new: detailed hist &
exam; med decis making of low
complexity; ~30 min face-to-face
Office/op visit, new: comprehensive
hist & exam; med decis making of
high complexity, ~60 min
* $33.9764 / RVU
Work Tot NF Tot Fac
RVU RVU RVU
0.48
1.42
3.17
1.25
3.13
5.95
0.78
2.28
4.93
“Bundled” Office Visit Costs
• Medical Expenditure Panel Survey Stat Brief #166
Expenses for Office-Based Physician Visits by Specialty,
2004*
– Dentistry: 155
–GP: 101
– Pediatrics: 98
– Ob/Gyn: 153
– Opthalmology 206
– Orthopedics 210
– Psychiatry: 95
– Dermatology: 142
– Other: 235
* Also see #157, Regional Differences in Total and Out-of-Pocket
Expenditures for Selected Types of Office-Based Visits, 2004
Medical Expenditure Panel Survey Statistical Briefs
• #111: Expenses for a Hospital Emergency Room Visit,
2003
– Average expenditure*: $560
– Average if surgical procedure performed: $904
– Average if 1+ nonsurgical services † provided: $637
– Average if no special service provided: $302
* Direct payments from individuals, private insurance, Medicare, Medicaid,
Workers’ Compensation, and miscellaneous other sources. Expenditures
include those for both facility and separately billed physicians’ services
† Laboratory test, sonogram or ultrasound, X-ray, mammogram, MRI, CAT
scan EKG, ECG, EEG, vaccination, anesthesia, and other miscellaneous
diagnostic tests
11

2011 Clinical Diagnostic Laboratory Fee Schedule
HCPC
85520
85525
85530
85536
85540
85547
85576
85597
85610
85612
Heparin neutralization
Heparine-protamine tolerance
Iron stain peripheral blood
WBC alkaline phosphatase
RBC mechanical fragility
Blood platelet aggregation
Phospholipid platelet neutralization
Prothrombin time
Viper venom prothrombin time
16.67
19.95
9.11
12.11
12.11
30.23
25.30
5.53
13.47
http://www.cms.hhs.gov/ClinicalLabFeeSched/02_clinlab.asp#TopOfPage
HCPC
A5500
A5501
A5503
A6010
A6023
V2475
V2750
Short Description
Heparin assay
Nat Limit
18.43
2011 Medicare Durable Good Fee Schedule
Short Description
Diab shoe for density insert
Diabetic custom molded shoe
Diabetic shoe w/roller rockr
Collagen based wound filler
Collagen dressing>48 sq in
Tint, any color/solid/grad
Anti-reflective coating
Ceiling
80.03
240.06
39.12
32.48
199.62
12.34
22.93
Mid point
24.90
22.53
26.96
12.31
16.36
16.36
40.85
34.19
7.47
18.2
Floor
60.02
180.05
29.34
27.61
169.68
9.25
17.20
http://www.cms.hhs.gov/DMEPOSFeeSched/LSDMEPOSFEE/
HOME HEALTH VISIT COSTS *
N, Mean
Visit Type Agen. Cost S.D. Min Max
Skilled nurse 4609 87.68 35.10 12.65 756.89
Physical therapist 4287 104.40 43.37 12.29 663.63
Occupat therapist 3492 107.78 52.31 10.20 773.86
Speech therapist 3292 110.50 58.60 10.84 724
Med/Soc services 3661 139.77 82.01 10.20 883.8
Home health aid 4603 40.30 20.97 5.67 848.7
* Cost data used in establishing the Home Health Agency(HHA) Per-Visit Cost Limits for
fiscal periods beginning after October 1, 1999. The cost and statistical data were
obtained from Free Standing Medicare HHA cost reports (Form 1728-94) for full cost
reporting periods beginning on or after October 1, 1994 and settled by March 1998.
Data may no longer be maintained by CMS
12

Medicare Fee Schedules Home Pages
• http://www.cms.hhs.gov/FeeScheduleGenInfo/
• http://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/
Drug Prices
• Average wholesale price (AWP) has been commonly
used in the literature as an estimate of drug prices
• AWP is an unregulated number that has been found by
courts to be a fairly artificial construct that has been
gamed by the pharmaceutical industry
• Historically, AWP was the average price charged by
wholesalers to providers, like doctors and pharmacies,
derived from an industry-wide formulaic 20%-%25
markup between actual acquisition cost and AWP
• Over time, actual margins have fallen to 2%-3%, and the
20%-25% markups no longer reflect actual wholesaler
margins
ASP
• For non-Part D drugs, Medicare has moved to
reimbursing drugs by use of a small mark-up of average
sales price (ASP)
– ASP represents the sales price including volume
discounts, prompt pay discounts, cash discounts, free
goods that are contingent on any purchase
requirement, chargebacks, and rebates
– Applies to physician-administered drugs (e.g.,
chemotherapy infused in physicians’ offices)
13

AWP / ASP Comparison
• The US Inspector General's office* has reported:
– The median difference between ASP and AWP is
49%
– For generic drugs, ASP is 68% less than AWP at the
median
– For single source brand drugs, ASP is 26% less than
AWP
– For multisource branded drugs, ASP is 30% less
* Medicaid Drug Price Comparison: Average Sales Price To Average
Wholesale Price, Office of the Inspector General, DHHS, June 2005
HCPCS
90658
90716
90647
J0130
J9260
J1438
J3370
J2020
J9182
J3480
J7330
Description
Selected 2009 ASP Prices
Flu vaccine, 3+ years, im
Chicken pox vaccine, sc
HIB, PRP-OMP, im
Abciximab injection
Methotrexate sodium inj
Etanercept injection
Vancomycin hcl injection
Linezolid injection
Etoposide injection
Inj potassium chloride
Cultured chondrocytes, impnt
0.5 ML
10 MG
50 MG
25 MG
500 MG
200 MG
10 MG
2 MEQ
https://www.cms.gov/McrPartBDrugAvgSalesPrice/01_overview.asp#TopOfP
age
AMP
Dosage
0.5 ML
0.5 ML
1
Payment limit
$13.218
$77.651
$23.34
$438.872
$2.201
$180.782
$3.017
$30.23
$0.463
$0.015
$23,102
• Alternatively, Medicaid upper limits are based on
average manufacturers price (AMP) plus a dispensing
fee for pharmaceuticals
– AMP represents the average price paid to the
manufacturer by wholesalers for drugs distributed to
the retail pharmacy class of trade including discounts
and other price concessions provided by the
manufacturer, but without regard to customary prompt
pay discounts extended to wholesalers
14

AWP / AMP Comparison
• The US Inspector General's office* has reported:
– For generic drugs, AMP on average is between 65%
(simple average) and 74% (sales-weighted average)
less than AWP
– For single-branded drugs AMP on average is 25%
less than AWP
– For multisource branded drugs, AMP on average is
between 35% and 40% less than AWP
* Medicaid Drug Price Comparisons: Average Manufacturer Price to
Published Prices, Office of the Inspector General, DHHS, June 2005
Thomson Reuters Redbook
• Provides
–AWP
– Wholesale Acquisition Cost (WAC)
– Direct Price (DP)
– Suggested Retail Price (SRP)
– Federal Upper Limit (FUL)
VA Federal Supply Schedule (FSS)
• Some of the most complete / accessable pharmaceutical
price lists
– FSS Big 4 prices (for VA, DoD, PHS, and Coast
Guard) based on vendor’s most favored commercial
customer pricing or statutorily-required pricing
calculations
– FSS Dual prices -- for Other Government Agencies
authorized to use the FSS program -- based on most
favored commercial customer pricing negotiations
• Web address:
http://www.pbm.va.gov/DrugPharmaceuticalPrices.aspx
15