

























Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
A report by the Office of Inspector General (OIG) on errors in the assignment of Diagnosis Related Groups (DRGs) for pulmonary edema and respiratory failure (DRG 87) under the Prospective Payment System (PPS). The report identifies overpayments to hospitals due to incorrect coding and recommends actions for recovery of overpayments.
Typology: Lecture notes
1 / 33
This page cannot be seen from the preview
Don't miss anything!
.. stRV/CE
to"
:flf...ua
AUGUST 1989
10
DRG 87: PULMONARY EDEMA AND
RESPIRATORY FAILURE
OAI. 12-88-
AUGUST 1989
Diagnosis related group (DRG) 87 pays for
five, very specifc, ICD- CM codes group to DRG 87 as principal diagnoses. Disorders such
disease, asthma, bronchitis, and
pneumonia also involve
lung pathology, but should group to other DRGs with lower relative
weights.
assigned to a different DRG.. This error rate significantly exceeds the 18.
percent for all DRGs from the National DRG Valdation Study.
Hospitals overpaid themselves in 96.4 percent of the incorrect DRG
assignments. This ratio significantly exceeds the National DRG Validation
Study s 59.7 percent of overpayments.
The codng errors principally derived from physician mis-specification of the
pricipal diagnosis or hospital resequencing of the diagnoses to substitute
corrct principal diagnosis.
diect the peer
review organizations (PROs) to review prospectively DRG 87
The HCFA should diect the PROs to educate physicians and hospitals about the
The HCFA disagrees with the fist recommendation and believes that it has implemented the
second recommendation. The Offce of Inspector Genera has altered this report to accom
modte HCFA's comments and continues to believe that adopting these recommendations
could recover a projected $35.5 milion annually. The HCFA has alady implemented a third
recommendation made in the drt of this inspection.
.................................. ... .. ... .... .... ......... .......... ..... .................................... ... .. .. ... ........ ... ...... ...................................... ... ...... ........................... ............................ ....................... ........................... .................................. ....................... ............................. .......................... .......................... ................................... ........................... ................................... ................................... ................................... ............................. ............................................................................ .............................. ............... ............... .............. .......... .............. .. .. ......... ...... .. .. .. .. ... ...
Background
PPS vulnerabilities
Claims processing.........
Assignment errrs.
Soure of errors...
Reasons for assignment errors
Financial effects
Corrct DRG assignments..................................
Clinical review results..................
Appendix A- I: DRG 87 discharges from all PPS hospitals......................... A-
Appendix A-2: DRG 87 sampling fre.... A-
Appendix A-3: DRG 87 hospital demography........................
........... A-
Appendix A-4: DRG 87 hospital demography comparson...
...................... A ..........
Appendix A- S: DRG 87 patient demogrphy
Appendix A-6: DRG 87 patient demogrphy comparson...
....................... A-
Appendix B- 1: DRG 87 assignment errors
Appendix B- 2: DRG 87 assignment errors comparson
........................... B-
Appendix B- 3: DRG 87 assignment errors by patient demography
.................. B-
errr..
. C-
Appendix C-
error comparson....
........................ C
On October 1, 1983, the Health
Car Financing
(HCFA) began implementing a
new system of payment for inpatient hospital services under the
penses under Medicare.
Under PPS, hospitals received a pre-established payment for each discharge, based upon the
diagnosis related group (DRG) to which the discharge is assigned.
PPS classifed discharges
into clinicaly coherent groups which used simlar amounts of hospital resources, based on
charge status. Each of the 473 DRGs had an associated
relative. weight, which represented the
average cost for hospital care provided to patients with diagnoses
that DRG as a
hospital retaned any
ces, and suffered losses on those patients consuming more.
The shift from cost-based, retrospective reimbursement to prospective payment constituted
A fixed payment per discharge induced hospitals to implement economies and reduce unneces
sar servces. The total payments to the hospitas provided the same financial resources for
Where the cost-
reimbursement system rewarded longer hospita stays and more costly
rewarded earlier discharges and less costly procedures. One
of the first consequences of the
new payment system was a drop in average length of hospita stay for Medicare patients.
PPS vulnerabilties
The advent of PPS created new opportnities for manipulation or "
tal revenues from
maitain quality of
care Congress established the peer review organizations (PROs) to monitor hospital activities.
The Office of the Inspector General (OIG) conducted The National DRG Validation Study
(NDRGVS) to
qualty of care performed
700
medical records established that assignment
errors resulted in $300 milion in overpayments to hospitals, and that
the majority of overpay
ments could be trced to assignment errors afecting a small number of DRGs. This report is
one in a series examning assignment accuracy of one of the DRGs
identied as having the
highest impact on overpayments under PPS and the
gratest potential for cost
recovery.
PPS gamg takes two principal forms: optimization and creep. "
" strategies ad
maximi hospital reimburements by selecting the most expensive
among viable alternative principal diagnoses or adding more
mits optization, which flows from
the basic incentive strctu of the PPS system.
DRG creep
" results from codg practices which do not conform to codng rules.
Sources of
DRG crp include:
Misspecijcation:
The attending physician wrtes an incorrct principal diagnosis
(defied by the Uniform Hospital Discharge Data Set (UDS)
as "that condition
established after study to be chiefly responsible for occasioning the
procedurs on the attestation
sheet.
Miscoding: The hospita assigns incorrct numeric codes to diseases or procedures
Resequencing: The hospital substitutes a seconda diagnosis for the corrct principal
diagnosis.
Auditig and review practices seek to curail ilegal creep by identiying discharges in which
codng rules are misapplied or ignored.
Under PPS, the hospita fies a clai for Medicare reimbursement upon discharging the
beneficiar. At the time of discharge, the attending physician attests to the principal diagnosis
which caused the patient s adssion to the hospita, seconda diagnoses, and procedures
(diagnostic and therapeutic) provided. The
physician s attestation statement into numeric codes based on the International Classification
of Diseases, Ninth Revision
(P
organizations, working under contract with HCFA , enter the hospital' s codes
into the GROUPER computer progr which assigns the appropriate DRG for reimbursement.
Hospita reimbursement is calculated by multiplying the " relative weight " of each DRG
category by a standadized amount, as
tive weight of each DRG vares above or below 1.
00 accordig to the average amount of
patients in that diagnostic group. The higher the relative weight,
Mis-assignment of the ICD- CM categories, or erroneous as
signment or sequencing of patient diagnoses, can thus have significant financial implications.
This study examnes erroneous assignment in a single DRG: 87, pulmonar edema and
respirtory failur. At the time covered by this inspection , only five ICD- CM codes group
to DRG 87: thee types of pulmonar edema and
This study used a strtified two-stage sampling design based on hospitals to select medical
records for review. The fist stage used simple random sampling without replacement to
select up to 80 hospitas in each of thee bed size
strata: Less than 100 beds (small), 100 to
299 beds (medium), and 300 or more beds (large)
The second stage of the design employed
systematic random sampling to select at least 25 DRG 87
discharges between October 1, 1984 and March 31 , 1985.
100 beds
10029 beds
1:.:.:.:.:.:.:.:.
30+ beds
DRG 87 All DRGs
Figure 2: Samplil)g frame
The OIG contrcted with the Health Data Institute
(HI) of Lexington
, Massachusetts to
contrctor "blinded" the ICD- CM codes
by covering them , and assigned an identification number to each record. An Accredited
Record Technician or Registered Record Admnistrator proficient in ICD- CM codng
reviewed the enti record to substantiate the principal diagnosis, other diagnoses, and proce
dures indicated by the attending physician in the
which did not support the assigned DRG classification were referred to physician reviewers.
The physician reviewers designated the corrct Uniform Hospita Discharge Data
diagnosis, and additional diagnoses and/or proedures which were substantiated by the patient
records. The GROUPER computer program processed the reabstracted ICD- CM codes to
determne corrct DRGs. A full discussion of the methodology and fmdings of the contractor
record review is available in the fmal report of the National DRG Validation Study (available
from OIG Public Affais).
DRG 87 was chosen for this inspection because of its high relative weight (1.
high rate of errors. The OIG contrcted with BOTEC Analysis of Cambridge, MA to examne
data for DRG 87 in greater detal
assignment errors, and to make recom
mendations for recovery of overpayments.
..::_ ::'...--....-... .-.-.:.;._" ':..-----..=-"-' ' , ..--- - -.-.-- -.- --.---..-- -.-..-
Sample characteristics
In FY 1985, 90, 917 of the 8.3 milion PPS discharges (1.1 percent) grouped to DRG 87. The
National DRG Validation Study estiated that approximately equal proportons came from
each bed size strta. Small hospitals therefore biled for DRG 87 at a higher rate than large
25). In the fist half of FY 1985, the 239 hospitals selected in stage-one of the sample
design (the sampling
fre) biled for 222 396 discharges of which 2,417 came from DRG 87
(1. 1 percent). The sampling frame reflects their higher frequency of discharges from larger
hospitals, while the sampling
(Chi-square 3.
1). (Appendix A , df 1, P
Additionally, the two-stage sample design
' of
an event at a parcular hospita). The appendices, tables, and char therefore report in
dividual totas weighted by both discharges and hospitas.
Yes
DRG 87: Urban
All DRGs: Urban
DRG 87: Teaching
All DRGs: Teaching
DRG 87: Profi
All DRGs: Profi
rt "" - ..
100
Percent of discharges
Figure 3: Hospital demography
Like all Medicar discharges as measured by the National DRG Validation Study, the majority
Haenszel chi-square 0. 16, df 1 , P 75),
nonteaching (Mantel-
59, df I , P 5), and nonprofit (Mantel-Haenszel
3J In comparson to all PPS discharges, a slightly larger proportion of
DRG 87 discharges came from
-------- ..------------------------- --- - --. '"..... ..,. .- ------------------- .... ------------------- .-------------------------------------- ---------------------------------------------- ----------------- - -- - - -- -- - - - - - -- -- - -- - - -- - - - - - - - -. ....... ------------------- --- - -- - - -- - - - - -- - -- - -. .............. ---------------------------------------------------- --- - --
Patient demographics differ between the correct and incorrct subsamples of discharges. In
discharges assigned incorrctly, patients were on average slightly older, more frequently
reimbursements lower, in discharges incorrectly assigned to DRG 87. (Appendix B-
The overwhelmng majority of errors in assignment for DRG 87 resulted in overpayments to
hospitals. Weighted by discharges, hospitals overpaid themselves in 96.4 percent of
mis-assignments. In other words, hospitals biled discharges with lower weighted diagnoses
as DRG 87 which, in 1985, had a relative weight of 1.5368. This rate signifcantly exceeds
the 59. 6 percent overpayments rate for all discharges (Mantel-
P c: 0. 5). Combining the 56. 0 percent error rate with this 96.4 percent overpayment rate gives
an effective overpayment rate of 54. 0 percent. This rate is nearly five times the 11. 1 percent
for the National DRG Validation Study. (Appendix C-
Percent
DRG 87: Overpaid
All DRGs: Overpaid
CI
DRG 87: Underpaid
EJ All DRGs: Underpaid
..100 10029 30+
Bed size
Figure 5: Direction of errors
Hospitals of al
tyes overpaid themselves on
DRG 87 errors. However , except for nonteach
ing status (Mantel-
, df 1 , P c: 0. 05), these demographic trnds did not
signficantly exceeded those in the National DRG
Validation Study when
some differences in patient demographics among
over- and under- paid discharges, the number of underpayments was too small to allow mean
ingful analysis of these dierences. (Appendix C-
Almost all the errors in this sample can be traced to the hospitas
' medical records practices.
Fifty- thee cases were miscoded as DRG 87 and
biled accordingly. Only one case
was coded
biled incorrectly as DRG 87 by the hospita.
!:::::::::::::::::::::::::::::::
Reasons for assignment errors
Physicians made the majority of errors by incorrctly specifying the diagnoses or procedures
on patient attestation sheets. This mis-specification caused 48. 2 percent of the mis-assign-
physician mis-specified a procedure. 29. 6 percent of errors occurd when the hospita rese
quenced the narative to substitute a seconda diagnosis for the corrct principal diagnosis. A
smaller number of errors resulted when the medical records deparent selected incorrect
numeric codes for correct nartive diagnoses. The distrbution of errors in this sample be
tween physicians and other hospital deparents approximated the trnds in the entie Nation
al DRG Valdation Study. (Appendix E-
t.:.:.:.:.:.:.:.:
Misspeclfcation
Miscing
Resuenclng
DRG 87 All DRGs
Oter
Figure 6: Reasons for errors
The reasons for errrs in biling discharges to DRG 87 vared slightly by hospital
demogrphics. (Appendix E-2) Narative (physician) errors occured parcularly frequent
discharges from small hospitals and mid-sized facilties.
Large hospitals, in
resequencing errors. Teaching hospitals miscoded records at
over twice the rate of non-teaching hospitas, but non-teaching
facilties resequenced diag
noses nearly twice as often as their non-teaching counterpars.
In discharges from for- profit
hospitals, narative errors appeared
frequently, but no resequencing errors. Patient age, sex,
length of stay, and mortity were roughly similar across error
tyes. (Appendix E-
Financial effects
After reabstrction , the average relative weight for DRG 87 discharges in this sample dropped
from 1.5368 to 1.2277. Discharge weighted, the
decrease amounted to a net overpayment of
3 percent and applies to FY 1984- 87 without modfication.
(Appendix F
The majority of cases corrctly assigned to DRG 87 also involved patients with respiratory
faiure. The balance of corrctly biled discharges have acute lung edema -- not otherwise
specified, as their principal diagnosis. (Appendix G-
Percent
DRG 87
All DRGs
Unneeed admissions Poor quality care Premature discharges
Figure 8: Clinical incidents
Reviewers concluded that only 1.3 percent of discharges in this sample constituted unneces
sar admssions ("an admssion in which the care received by the patient was either not
needed or did not require the use of the inpatient setting. "
) (Appendix H- l) However, this rate
did not dier signifcantly from the 10.
2, df 2, P .. 0. 25).
Two discharges were judged to have been
frm the National DRG Valdation Study (Chi-square 1.66, df 2, P.. 0. 50). (Appendix H-
8 percent of discharges in this sample (discharge-weighted) evidenced "
quality of
in discharges from small hospitals, which had over twice the rate of poor
qualty compared to
other hospitals in this sample, and substantialy more problems than discharges from small
hospitals in the full National DRG Validation Study.
The Health Care Financing Admnistration should dict the peer review organizations
The HCF A should diect the PROs to educate physicians and hospitals about the proper
codng of DRG 87.
second recommendation. The OIG has considered the HCFA' s comments on the drt of this
report and contiues to believe that adopting these recommendations could recover a projected
the draft of this inspection.
Appendix A-4: DRG 87 hospital demography comparison
Percent Bed size Weighted percentage
distribution c:100 100- 299 300+ Sample Discharge Hospital
Urban
DRG 87 10. 71.4 100 53. 65. 44.
NDRGVS 19.
Rural DRG 87
NDRGVS
Teaching DRG 87 14. 52. 18.
NDRGVS
Non- DRG 87 100 85. 48.
teaching
NDRGVS 97.4 81. 44. 74.
Profit DRG 87
NDRGVS 17.
Non- DRG 87
profit NDRGVS 90.
Appendix A-5:
demography
Bed size Weighted average
c: 100- 299 300+ Sample Discharge Hospital
Age (years)
Sex (% male)
LOS (days)
Payment ($) 2868 4217 5476 3987 4328 3720
Mortality (%)
(%)($)
Appendix A-
Bed size Weighted average
c:100 100- 299 300+ Sample Discharge Hospital
Age DRG 87 74. 72. 67. 72. 71.
(years) NDRGVS
Sex DRG 87 69. 50.
(% male) NDRGVS 43. 45. 48. 45.
LOS DRG 87 11.
(days) NDRGVS
Payment DRG 87 2868 4217 5476 3987 4328 3720
NDRGVS 1849 2923 3807 2860 3074 2508
Mortality DRG 87
NDRGVS