Advances in Bioscience and Biotechnology, 2013, 4, 118-125 ABB
http://dx.doi.org/10.4236/abb.2013.41A017 Published Online January 2013 (http://www.scirp.org/journal/abb/)
Reducing hospital inpatient complications: A four-year
experience
Ronald Lagoe1*, Joseph Bick2
1Hospital Executive Council, Syracuse, USA
2Associate for Clinical Intelligence, St. Joseph’s Hospital Health Center, Syracuse, USA
Received 8 November 2012; revised 14 December 2012; accepted 20 January 2013
ABSTRACT
This study described the use of administrative data
and a computer software algorithm, Potentially Pre-
ventable Complications, to support reduction of inpa-
tient hospital complications. The study was carried
out between 2008 and 2012 in St. Joseph’s Hospital
Health Center in Syracuse, New York. The hospital
generates approximately 23,000 inpatient discharges
annually. The study employed summary tables for
individual inpatient complications and patient spe-
cific spreadsheets to evaluate and follow adverse out-
comes. The spreadsheets were employed by hospital
staff to determine whether patient medical records
confirm each complication identified by the software.
This process resulted in improvement of the accuracy
of administrative data describing inpatient complica-
tions. The administrative data and the software were
also used in conjunction with medical records to
identify patients who received program interventions
and still experienced inpatient complications. This
process enabled hospital staff to ensure that interven-
tions were being provided and evaluate their effec-
tiveness. The study demonstrated that, at the aggre-
gate level, the inpatient complication rate per 1000
discharges declined by 33.4 percent, from 56.11 to
37.37 between 2008 and 2011. The principal drivers
of this decline were high volume complications such
as pneumonia, where the rate declined by 45.7 per-
cent and urinary tract infection where the rate de-
clined by 23.7 percent. The project provided a means
of communicating and managing outcomes data that
could be implemented and understood by a wide
range of health care providers.
Keywords: Hospital Outcomes; Hospital Complications;
Quality of Care
1. INTRODUCTION
In recent years, increased attention has developed con-
cerning improvement of hospital and health care outcomes
in the United S tates. Th is d eve lo p ment has re su lted fro m a
combination of factors related to inpatient complications,
hospital readmissions, and other indicat ors.
A major cause of this development has been research
that demonstrates the relationship between adverse health
care outcomes and higher costs to providers and payers.
Studies have demonstrated that patients with inpatient
complications such as pneumonia, urinary tract infection,
and clostridium difficile colitis have much longer hos-
pital stays and related labor and pharmaceutical costs
than those who do not [1-3].
Related to these costs is a new urgency to reduce
health care expenses in society. In the United States, all
major health care payers, including Medicare, Medicaid,
and private insurance, are under a large amount of pres-
sure to reduce spending. This situation strongly suggests
that current increases in these expenditure can no longer
be sustained [4,5]. The potential to control these costs
through financial penalties for adverse outcomes holds
the potential to reduce health care spending while im-
provin g patient care [6] .
In addition to these factors, efforts to improve health
care outcomes are benefiting from the development of
electronic software for analysis of patient specific data.
Tools such as the Potentially Preventable Complications
system developed by 3M Health Information Systems
can analyze large amounts of administrative data at the
aggregate and patient specific levels. The information
produced by these tools can guide clinical management
initiatives [7,8].
This study described the use of administrative data and
one of these algorithms to support reduction of inpatient
complications in a large urb an hospital in Syracuse, New
York during a four year period. It demonstrated the use
of this software to identify and address adverse outcomes
at the aggregate and specific levels.
*Corresponding a uthor.

R. Lagoe, J. Bick / Advances in Bioscience and Biotechnology 4 (2013) 118-125 119
2. POPULATION AND METHODS
This study involved the use of administrative data to
identify and manage inpatient complications in St. Jo-
seph’s Hospital Health Center, in Syracuse, New York
during a four year period. This hospital has the largest
inpatient volume (23,832 inpatient discharges excluding
well newborns in 2011) of the acute care facilities in
Syracuse. Other hospitals in Syracuse that participated in
the program included Crouse Hospital (20,540 dis-
charges) and Upstate University Hospital at Community
General (6959 discharges) (Hospital Executive Council,
Unpublished data, 20 12).
The program involved the staffs of St. Joseph’s Hos-
pital Health Center and the Hospital Executive Council,
the cooperative planning organization for the Syracuse
Hospitals. Historically, the Syracuse Hospitals have
worked through the Council to improve health care effi-
ciency and outcomes in Central New York [9].
3. DATA DEVELOPMENT
The study was developed as part of a demonstration pro-
gram including the Hospital Executive Council, St. Jo-
seph’s Hospital Health Center, and 3M Health Informa-
tion Systems. This program involved the use of the 3M
Potentially Preventable Complications software and hos-
pital administrative da ta to identify and manage inpatien t
complications.
The Potentially Preventable Complications software
includes extensive log ic for identifying inpatient hospital
complications in administrative data. At the summary
level, it identifies complications at the patient specific
level based on specific secondary diagnoses. These di-
agnoses, which are identified by the reporting indicator
as Not Present on Admission, are assumed to be candi-
dates for inpatient complications. The software then
screens these diagnoses with specific exclusion criteria
such as logical sequelae of principal diagnoses and ser-
vices such as trauma and neonatal care for which inpa-
tient complications have not been clearly identified.
In the demonstration program with 3M Health Infor-
mation Systems, the Hospital Executive Council gener-
ated Potentially Preventable Complications data for St.
Joseph’s Hospital Health Center each month beginning
in 2009. These data included patient specific spread-
sheets and summary tables for PPCs that were the subject
of clinical interventions at the Hospital.
In response to this information, the Hospital staff
identified whether patient records confirmed that each
patient experienced the complication(s) identified by the
software and whether they received the clinical interve-
netions designed to avoid these adverse outcomes. This
information was used to monitor development of the
program by the Hospital Executive Council and for in-
ternal follow up by the hospital.
4. HOSPITAL INTERVENTIONS
The charts of patients that were identified as having a
particular potentially preventable complication were re-
viewed in detail to identify commonalities. The first in-
tervention was to determine whether the documenttation
supported the diagnosis that resulted in the PPC. Coding
errors had to be eliminated before clinical interventions
could be put in place. This gave the hospital a more ac-
curate picture of the scope of the problem. This was ac-
complished by reviewing each clinical record and by
comparing them to coding summaries and coding defini-
tions [10].
The charts fell into two distinct categories: The first
were records where the documentation did not support
the coding. The other category included charts where the
clinical information did not support the diagnosis. By
educating coders about errors and common mistakes,
coding errors were systematically eliminated. By edu-
cating providers about how diagnosis of pneumonia and
urinary tract infections were coded, their documentation
regarding these complications became clearer.
With administrative “slack” taken out of the system
the true clinical situations could be identified and miti-
gated. Over 100 chart reviews were conducted to define
what specific outcome issues there was at the hospital.
The staff wanted to be as specific as possible for the pa-
tient populations so that they could effectively deploy
resources to reduce the complications of interest. The
hospital focused on four complications: Pneumonia, Uri-
nary Tract Infection, Clostridium Difficile Colitis, and
Decubitus Ulcers.
5. PNEUMONIA
Charts were examined to identify if patients were re-
ceiving the basic care documented in the literature to
prevent hospital acquired pneumonia [11]. Specifically,
nursing charting was reviewed to determine if the pa-
tients with the complication received mouth care, had the
use of incentive spirometry documented, were out of bed
ambulating, and if they had the head of their bed elevated
[12,13]. It was clear from chart reviews that incentive
spirometry and mouth care were opportunities to im-
prove care. Units and services with the highest numbers
of cases were selected for education regarding incentive
spirometry use and documentation. Daily rounding was
instituted to ensure that nurses were reminding patients
to use their incentive spirometers and that they docu-
mented its use.
6. URINARY TRACT INFECTION
The data from each source were imported into a Micro-

R. Lagoe, J. Bick / Advances in Bioscience and Biotechnology 4 (2013) 118-125 120
soft Access database and compared using medical record
number, account number and admission date. Records
for review were then sorted by comparing them with
valid culture results to arrive at the study group. Chart
reviews were conducted to determine what care was
documented. These reviews included the collection of
the infection indicators of white blood cell count and
temperature associated with culture collection. The re-
views determined for each case if there was a urinary
catheter placed during the patients admission, how many
days it was in place, and the number of times care was
charted in association with it. For those infections where
a urinary catheter was in place, the number of catheter
days was compared with the number of times catheter
care was charted to approximate a rate of care episodes
per catheter day.
The key intervention in reducing hospital acquired
urinary tract infection was to reduce the overall usag e of
urinary catheters. To do this, the problem was addressed
by a number of approaches. First, using the CDC’s
guidelines for reducing catheter associated urinary tract
infections, standard indications for catheter insertion
were implemented [14]. Three interventions were effec-
tive in reducing the number of catheter days per patient
(the in process measure). Firstly, an RN driven foley
removal protocol for inpatients was instituted. Essen-
tially, if the indications no longer existed, the catheter
was removed automatically by the nurse. Secondly, for
surgical patients, an education program for physician
assistants, residents and nurse practitioners about the
correct indications for indwelling catheters was instituted.
Finally, it was determined that the majority of urinary
catheters were inserted in the emergency department. As
a result, an education program for the both the nurse and
providers in the emergency department about the indica-
tions, and effects of indwelling catheters for patients
during the entire hospitalization was carried out. All of
these interventions resulted in approximately a 20 per-
cent reduction of catheter days per patients and a pre-
cipitous decline in hospital acquired urinary tract infec-
tions.
7. CLOSTRIDIUM DIFFICILE COLITIS
Cases identified by the 3M Poten tially Preventable Com-
plication (PPC) software were compared with those re-
ported to the National Health Safety Network (NHSN) at
the Centers for Disease Control and Prevention (CDC)
[15-17]. They were also cross referenced with clostrid-
ium difficile colitis cases identified by Care fusion/Med
mined Virtual Surveillance Indicators. After cases were
identified and cross referenced, a total of 124 unique
records identified as hospital onset were reviewed in
detail using data found in the clinical documentation
system, the orders system and the pharmacy system. Fi-
nally, patient room assignments were reviewed.
The initial review was carried out using the pharmacy
system, Horizon Meds Management (HMM). The re-
viewer reviewed notations made by pharmacist regarding
antibiotic prescribing and indications. This documenta-
tion is not part of the patient’s clinical record. The re-
viewer attempted to identify the patient’s original infec-
tion and the antibiotics used to treat it. When this was not
documented in HMM, the clinical record was reviewed.
The findings of infection requiring antibiotic treatment
are listed in Table 1.
Next the antibiotics used to treat these infections were
review and counted. Vancomycin and Flagyl were ex-
cluded from this count. The results are listed in Table 2.
It is worthwhile to know that 84 percent of patients
received at least one of the antibiotics listed in Table 2
and 45 percent of patients received at least two. These
findings are consistent with the findings noted by Pepin
et al. in their study “Emergence of Fluotoquinolones as
the predominant risk factor for Clostridium Difficile-
Associated Diarrhea: A cohort study during an epidemic
in Quebec” [18].
8. DECUBITUS ULCERS
It was clear from an extensive review of patient records
and careful comparisons of other data sets that a previ-
ously identified quality improvement opportunity in the
care of the decubitus ulcer PPC was far more theoretical
Table 1. St. Joseph’s Hospital Health Center, infections treated
with antibiotics in patients with C. Diff.
Infection Count
Pneumonia or rule out pneumonia 42
Urinary tract infections or urosep sis 22
Other infections including sepsis, preop, prophylaxis, cellulitis,
and wound infections 60
Table 2. St. Joseph’s Hospital Health Center, antibiotics com-
monly prescribed to patients who later developed C. Diff infec-
tions.
Antibiotic Number of C. Diff cases presecribed
Zosyn 47
Ciprofloxacin 46
Cephalosporins 34
Ampicillin/Amoxicillin10
Moxifloxacin 8
Clindamycin 4
Zyvox 3
Azactam 2

R. Lagoe, J. Bick / Advances in Bioscience and Biotechnology 4 (2013) 118-125 121
than practical. Coding errors, careful nursing documenta-
tion, poor initial patien t condition and co -morbidities and
additional PPCs have conspired to negate any real finan-
cial gain to be had from further improvements to decu-
bitus care alone.
9. DATA ANALYSIS
Analysis of the data involved the impact of interven tions
for individual inpatient complications, as well as the de-
velopment of summary tables including numbers of
complications, at risk populations, and rates per 1000
discharges for St. Joseph’s Hospital Health Center. In
order to define the full and most recent impacts of the
program, one analysis included data for January-Decem-
ber 2008-2011, while the other included data for January-
March 2008-2012. In both analyses, data for individual
and aggregate complications were identified.
10. RESULTS
The first component of the results involved the individ-
ual complications addressed by the project. Data con-
cerning all individual complications are summarized in
Tables 3 and 4 which follow.
With respect to pneumonia, increased incentive spiro-
metry use and documentation resulted in the direct re-
duction of hospital acquired pneumonia. Changes in com-
plication rates for pneumonia are identified in Tables 3
and 4 which follow.
With respect to urinary tract infection, the association
between urinary catheters and urinary tract infections
was examined. The review looked at all of these infec-
tions, as well as their relationship to indwelling urinary
catheters. It found that 78% (n-76) of St. Joseph’s hospi-
tal acquired urinary tract infections were in patients who
had been catheterized during their admission. On average,
a patient who developed a urinary tract infection had an
indwelling catheter for 12 days. St. Joseph’s standard for
catheter care is that it be performed and documented
once daily. The rev iew found that only 25 percent (n-19)
of patients with positive urine cultures received this care
correctly. In 34 percent of the cases the patient received
less than what was required (n-23) and curiously, 45
percent (n-34) of the patients received too much care,
that is to say they had more episodes of care then rec-
ommended. Changes in complication rates for urinary
tract infection are identified in Tables 3 and 4 which
follow.
With respect to clostridium difficile colitis, another
factor examined was the timeliness of isolation precau-
tions associated with the onset of symptoms. To deter-
mine this all of the nursing orders for “en teric” and “con-
tact precautions C. difficile” were compared with orders
for “Stool for C. difficile”. The basic presumption of this
comparison was that for a “Stool for C. difficile” to be
ordered there was reason to believe that the patient is in
some way symptomatic of the disease. As such, if there
was a reasonable expectation that a patient is symptom-
matic, they were placed isolation in accordance with the
hospitals isolation manual and the CDC’s current best
practice guidelines. The examination included all inpa-
tients who were placed on isolation and all of those who
had an order for stool for C. diff. This produced 607
unique order combinations for inpatients that were ad-
mitted for at least 48 hours. On average, patients who
were placed on isolation in greater than one hour waited
five hours for the is ol at ion order to be entered.
Finally, an examination of positive clostridium diffi-
cile colitis results was made, irrespective of the onset to
determine if any other environmental factors could be
identified. This review specifically looked at the rooms
patients with positive results stayed in. It found that 15
patient rooms (7 percent of rooms) accounted for 54 (20
percent) of positive results. This was significant in that
environmental contamination with C. Diff spores is a
major contributing factor in developing the disease in pa-
tients with other risk factors. Changes in complication
rates for clostridium difficile colitis are identified in Ta-
bles 3 and 4 which follow.
With respect to decubitus ulcers, prolonged length of
stay seemed to contribute to the development of the
complication. In other words, the ulcer resulted from the
extended stay, rather than the long stay resulting from
the ulcer. Changes in complication rates for decubitus
ulcer are identified in Tables 3 and 4 which follow.
The second component of the results involved Poten-
tially Preventable Complications at the aggregate and
diagnosis specific levels for January-December 200 8- 20 1 1.
Relevant data are summarized in Table 3.
This information demonstrates that, at the aggregate
level, the PPC rate per 1000 discharges at the hospital
declined by 33.4 percent, from 56.11 to 37.37 during this
period. This occurred as the number of PPCs declined by
13.4 percent from 1035 to 896, while the at risk popula-
tion increased by 30.0 percent, from 18,446 to 23,975. It
was notable that the hospital was able to reduce compli-
cations at a time when its inpatient population was in-
creasing substantially.
At the PPC specific level, the principal drivers of the
decline were high volume diagnoses that were addressed
by specific interventions. For pneumonia (PPC 04) the
rate declined by 45.7 percent from 14.07 to 7.64 per
1000 discharges between 2008 and 2011. Most of this
reduction occurred between 2009 and 2011. For urinary
tract infection (PPC 16), the rate declined by 23.7 per-
cent from 8.37 to 6.39 between 2008 and 2011. As a re-
sult of an increase between 2008 and 2009, all of the
eline occurred during the last three years. d
c

R. Lagoe, J. Bick / Advances in Bioscience and Biotechnology 4 (2013) 118-125 122
Table 3. Potentially preventable complications, St. Joseph’s Hospital Health Center, January-December 2008-2011.
Patients w PPC At Risk Population Major PPC Rate/1000 Discharges
Major PPC Category 2008 20092010201120082009201020112008 2009 20102011
01 Stroke & Intracranial Hemorrhage 37 49 49 47 17,86617,12119,78321,189 2.07 2.86 2.482.22
02 Extreme CNS Complications 7 7 16 18 16,89816,16618,83120,079 0.41 0.43 0.850.90
03 Pulm Ed/Rsp Fail w/Vent 116 94 49 32 16,55815,49517,99719,235 7.01 6.07 2.721.66
04 Pneum & Other Lung Infect 200 175 16812814,21913,29515,48416,743 14.07 13.16 10.857.64
05 Aspiration Pneumonia 44 76 64 49 16,91516,14718,63920,271 2.60 4.71 3.432.42
06 Pulmonary Embolism 19 20 22 25 17,92517,10119,73621,160 1.06 1.17 1.111.18
07 Shock 81 80 10179 17,69116,85519,59320,982 4.58 4.75 5.153.77
08 Congestive Heart Failure 75 84 72 82 15,19514,60917,51718,122 4.94 5.75 4.114.52
09 Acute Myocardial Infarct 74 55 67 74 17,15416,33618,90320,165 4.31 3.37 3.543.67
10 Ventricular Fibrillation/Cardiac Arrest 80 73 80 65 18,11317,31119,96121,382 4.42 4.22 4.013.04
11 Peripheral Vascular Comp Except
Venous Thrombosis 8 15 20 13 18,02617,20919,86521,286 0.44 0.87 1.010.61
12 Venous Thrombosis 38 50 41 49 17,91017,09519,74521,143 2.12 2.92 2.082.32
13 Major Gastrointestinal Comp w
Transfus or Sign Bleed 8 12 0 0 17,21116,52218,99820,428 0.46 0.73 0.000.00
14 Major Liver Complications 10 21 17 12 17,93817,12519,74321,084 0.56 1.23 0.860.57
15 Clostridium Difficile Colitis 29 38 74 85 18,11317,31119,96121,382 1.60 2.20 3.71 3.98
16 Urinary Tract Infe ctio n 141 145 16212716,85416,05518,45619,884 8.37 9 .03 8.786.39
17 Renal Failure with Dialysis 15 16 0 0 16,43315,33217,98118,992 0.91 1.04 0.000.00
18 Post-Hemorrh & Oth Acute Anemia w
Transfusion 91 27 0 0 14,44813,92916,20817,494 6.30 1.94 0.000.00
19 Decubitus Ulcer 53 42 49 43 17,53816,58119,23221,033 3.02 2.53 2.552.04
20 Septicemia & Severe Infections 121 102 10373 17,61416,76819,47920,874 6.87 6.08 5.293.50
21 Post-Op Wound Inf & Deep Wound
Disruption w Proc 4 3 0 0 7,7797,0553,7994,125 0.51 0.43 0.000.00
22 Reopening Surgical S ite 13 18 0 0 7,7247,0413,7654,102 1.68 2.56 0.000.00
23 Post-Op Hemorrhage & Hematoma w Hem Cntrl Procedure or I&D
Procedure 26 22 5 5 7,9687,2373,9194,257 3.26 3.04 1.281.17
24 Accidental Puncture/Laceration During
Invasive Proc 95 78 88 58 9,2528,5515,0685,590 10.27 9.12 17.3610.38
25 Post-Procedure Foreign Bodies 1 1 1 2 7,9767,2433,92111,292 0.13 0.14 0.260.18
26 Encephalopathy 28 26 37 28 17,09316,32118,95319,052 1.64 1.59 1.951.47
27 Iatrogenic Pneumothrax 14 22 35 25 14,99414,39417,03418,344 0.93 1.53 2.051.36
28 Mechanical Com plication of Device,
Implant & Graft 26 24 19 23 17,64516,82219,35920,796 1.47 1.43 0.981.11
29 Inflamm & Other Complication of
Devices, Implants or Grafts Except
Vascular Infection 42 36 96 43 17,64516,82219,41620,796 2.38 2.14 4.942.07
30 Infections Due to Central Venous
Catheters 20 14 17 7 18,01717,17519,76421,404
1.11 0.82 0.860.33
31 Obstetrical Hemorrhage w Tr an sf us ion 7 7 6 5 1,9251,8921,8881,898 3.64 3.70 3.182.63
32 Obstetrical Lacerations & Oth Trauma
w/o Instrumentation 19 31 38 23 1,9241,8881,8941,902 9.88 16.42 20.0612.09
33 Obstetrical Lacerations & Oth Trauma
with Instrumentation 4 3 1 0 963 27 11 18 4.15 111.11 90.910.00
34 Major Puerperal Infection and Other
Major Obstetrical Complications 4 5 2 3 1,9461,9161,9061,920 2.06 2.61 1.051.56
35 Post-Op Resp Failure w Trache os to my 18 10 0 0 6,0935,4591,9622,256 2.95 1.83 0.000.00
Discharges w/One or More PPCs 1,035 993 1,02689618,44617,63521,70023,975 56.11 56.31 47.2837.37
Hospital Executive Council—3M Health Informati on S ystems PP C demonstration program.

R. Lagoe, J. Bick / Advances in Bioscience and Biotechnology 4 (2013) 118-125 123
Table 4. Potentially preventable complications, St. Joseph’s Hospital Health Center, January-March 2008-2012.
Major PPC Category Patients w PPC At Risk Population Major PPC Rate/1000 Dchgs
2008 2009 2010 20112012200820092010201120122008 2009 2010 20112012
01 Stroke & Intracranial
Hemorrhage 13 11 9 139 4,6554,3844,6425,1715,4622.79 2.51 1.94 2.511.65
02 Extreme CNS Complications 2 1 4 7 5 4,4184, 13 54,4134,9005,1610.45 0. 24 0.91 1.430.97
03 Pulm Ed/Rsp Fail w/Vent 40 31 12 109 4,3593,9624,1944,7304, 8829.18 7.82 2.86 2.111.84
04 Pneum & Other Lung Infect 52 42 50 27303,5973,4043,5774,0654,29214.46 1 2.34 13.98 6.646.99
05 Aspiration Pneumonia 13 13 21 16144,4154,1304,3844,9005,3602.94 3.15 4.79 3.272.61
06 Pulmonary Embolism 10 3 9 4 3 4,6684,3844,6315,1525,4792.14 0.68 1.94 0.780.55
07 Shock 24 20 25 22154,6304,3214,5875,1315,4215.18 4.63 5.45 4.292.77
08 Congestive Heart Failure 16 28 16 28153,9153,7554,4984,4044,6194.09 7.46 3.56 6.363.25
09 Acute Myocardial Infarct 24 8 18 27 114,4864,1944,4524,9755,0845.35 1.91 4.04
5.43 2.16
10 Ventricular
Fibrillation/Cardiac Arrest 24 9 12 17194,7204,4324,6905,2235,5215.08 2.03 2.56 3.253.44
11 Peripheral Vascular Comp
Except Venous Thrombosis 3 4 8 2 1 4,6944,4104,6665,1925,4890.64 0.91 1.71 0.390.18
12 Venous Thrombosis 13 10 5 13124,6694,3804,6345,1535,4602.78 2.28 1.08 2.522.20
13 Major Gastrointestinal Comp
w Transfusion or Sign Bleed 2 6 0 0 0 4,4864,2394,4544,9615,2610.45 1.42 0.00 0.000.00
14 Major Liver Complications 1 4 5 4 3 4,6834,3834,6475,1595,4370.21 0.91 1.08 0.780.55
15 Clostridium Difficile Colitis 8 9 13 19184,720 4,432 4,690 5,2235,5211.69 2.03 2.77 3.643.26
16 Urinary Tract Infection 34 37 41 28344,4054,1174,3294,8515,1407.72 8.99 9.47 5.776.61
17 Renal Failure with Dialysis 0 3 0 0 0 4,2813,8864,1734,7324,4950.00 0.77 0.00 0.000.00
18 Post-Hemorrh & Oth Acute
Anemia w Transfusion 22 11 0 0 0 3,7603,6193,8104,2854,5895.85 3.04 0.00 0.000.00
19 Decubitus Ulcer 12 9 14 118 4,5744,2554,4905,0345,6032.62 2.12 3.12 2.191.43
20 Septicemia & Severe
Infections 34 25 22 22114,6194,3034,5565,1095,3827.36 5.81 4.83 4.312.04
21 Post-Op Wound Inf & Deep Wound Disruption w
Procedure 0 1 0 0 0 2,0471,7519079791,1580.00 0.57 0.00 0.000.00
22 Reopening Surgical Site 5 3 0 0 0 2,0341,7498969661,1412.46 1.72 0.00 0.000.00
23 Post-Op Hemorrhage & Hematoma w Hem Cntrl Proc
or I&D Proc 9 7 1 4 1 2,0961,8049251,0181,1994.29 3.88 1.08 3.930.83
24 Accidental Puncture/Laceration During
Invasive Procedure 24 19 12 24132,4182,1351,2001,3761,4859.93 8.90 10.00 17.448.75
25 Post-Procedure Foreign Bodies 1 1 0 0 0 2,0971,8059261,0146,6490.48 0.55 0.00 0.000.00
26 Encephalopathy 9 3 9 11 4 4,4664,1724,4484,9434,1712.02 0.72 2.02 2.230.96
27 Iatrogenic Pneumothrax 5 4 8 6 6 3,8643,7313,9954,4924,7161.29 1.07 2.00 1.341.27
28 Mechanical Complication of
Device, Implant & Graft 6 9 4 3 5 4,5994,3224,5485,0775,3811.30 2.08 0.88 0.590.93
29
Inflamm & Other
Complication of Devices,
Implants or Grafts Ex Vascular
Infection
13 10 10 104 4,5994,3224,5485,0775,3812.83 2.31 2.20 1.970.74
30 Infections Due to Central
Venous Catheters 7 6 5 0 3 4,7124,4124,6465,1625,6571.49 1.36 1.08 0.000.53
31 Obstetrical Hemorrhage w
Transfusion 2 3 1 1 2 4474684404364314.47 6.41 2.27 2.294.64

R. Lagoe, J. Bick / Advances in Bioscience and Biotechnology 4 (2013) 118-125 124
Continued
32 Obstetrical Lacerations & Oth
Trauma w/o Instrumentation 4 8 11 3 4 4574654414354278.75 17.20 24.94 6.909.37
33 Obstetrical Lacerations & Oth
Trauma with Instrumentation 1 1 0 0 1 4578 3 3 4 2.19 125.00 0.00 0.00250.00
34 Major Puerperal Infection and Other Major Obstetrical
Complications 1 1 0 2 1 4574744454384312.19 2.11 0.00 4.572.32
35 Post-Op Resp Failure w
Tracheostomy 4 2 0 0 0 1,6201,3654535536472.47 1.47 0.00 0.000.00
Discharges w/One or More
PPCs 282 239 245 2372024,8204,5105,3895,5276,64958.51 52.99 45.46 42.8830.38
Hospital Executive Council—3M Health Informati on S ystems PP C demonstration program.
The lower volume complications that were addressed by
specific interventions contributed less to the o verall PPC
decline. The rate for clostridium difficile colitis in-
creased from 1.60 to 3.98 between 2008 and 2011. The
rate for decubitus ulcer declined from 3.02 to 2.04.
In addition to pneumonia and urinary tract infection,
PPCs that were not addressed by specific interventions
also contributed to declines in the overall rate. These
included pulmonary embolism (PPC 03), ventricular fib-
rillation (PPC 10), and septicemia (PPC 20).
In order to provide more updated information con-
cerning Potentially Preventable Complications at St. Jo-
seph’s Hospital Health Center, the third component of
the study included information for January-March 2008-
2012. These data are summarized in Table 4.
This information demonstrated that between January-
March 2008 and 2012, the aggregate PPC rate at the hos-
pital declined by 48.1 percent, from 58.51 to 30.38 per
1000 discharges. This decline progressed throughout the
five year period. Between the two most recent periods,
January-March 2011 and 2012, the reduction was 29.2
percent, from 42.88 to 30.38 per 100 0 di scharges.
As in the annual data, high volume PPCs that were
addressed by interventions were major drivers of the
aggregate reduction. These included pneumonia (PPC
04), where the rate declined by 51.7 percent, from 14.46
to 6.99 per 1000 discharges and urinary tract infection
(PPC 16) where the rate declined by 14.4 percent, from
7.72 to 6.61 per 1000 discharges, during the five year
period. It was notable that the rates for both of these
complications increased between January-March 2011
and 2012.
As in the annual data, smaller volume PPCs did not
contribute greatly to the overall decline for this time pe-
riod. The rate for clostridium difficile colitis (PPC 15)
increased, while the rate for decubitus ulcer declined.
As in the annual data, additional PPCs that were not
addressed by specific interventions also contributed to
declines in the overall rates. These included pulmonary
edema and respiratory failure (PPC 03), an 80.0 percent
reduction; septicemia (PPC 20), a 72.3 percent red uction ,
and postoperative hemorrhage (PPC 23), an 80.7 percent
reduction.
11. DISCUSSION
This study described the use of administrative data and
computer software to support the reduction of inpatient
complications in a large urban hospital d uring a four year
period. It demonstrated how these resources could be
employed to identify and improve these outcomes for a
wide range of diagno s es.
The interventions to reduce complications imple-
mented by St. Joseph’s Hospital Health Center were de-
rived from research literature and local experience.
Based on recommendations from published research,
they wer e adapted to the needs an d resources of th e hos-
pital. The results of the study demonstrated that the id en-
tification and use of these interventions were largely
successful.
The experience of St. Joseph’s Hospital Health Center
demonstrated how aggregate complications data could be
used to iden tify address specific diagnoses for interv ene-
tions. Using the Potentially Preventable Complications
software, the hospital staff was able to select complica-
tions with relatively high volumes, such as pneumonia
and urinary tract infections, that would have the largest
impact on aggregate outcomes and related costs. The
staff was also able to identify complications with lower
volumes, such as clostridium difficile colitis and decubi-
tus ulcer that were of interest.
The administrative data and computer software were
useful in identifying patient specific issues with respect
to documentation. The spreadsheets that were developed
from these resources contributed to improvements in the
coding of administrative data that clarified the actual
numbers of complications that occurred. This process
improved evaluatio n of hospital quality assurance efforts
to address these outcomes, as well as the accuracy of
administrative data being used by reporting agencies in
the public area.

R. Lagoe, J. Bick / Advances in Bioscience and Biotechnology 4 (2013) 118-125 125
The administrative data and computer software also
made it possible to identify patients who received the
program inventions, but also experienced the complica-
tions. From this perspective, it provided information for
evaluation of the impact of interventions over time on a
patient specific basis.
Through these applications, the staff of St. Joseph’s
Hospital Health Center was able to use these resources at
both the patient specific and aggregate levels to improve
care and reduce related costs. The aggregate data pro-
vided perspectives concerning this information across a
wide range of individual complications and through total
frequencies and rates.
In summary, this approach to improving patient out-
comes was simple and direct. It provided a means of
communicating and managing outcomes data that could
be understood and used by a wide variety of health care
providers.
REFERENCES
[1] Fuller, R.L., McCullough, E.C., Bao, M.Z. and Averill,
R.F. (2009) Estimating the costs of potentially prevent-
able hospital acquired complications. Health Care Finan-
cing Review, 30, 17-32.
[2] Lagoe, R.J., Johnson, P.E. and Murphy, M.P. (2011) In-
patient hospital complications and lengths of stay: A short
report. BMC Research Notes, 4, 135.
doi:10.1186/1756-0500-4-135
[3] Hoonhout, L.H., de Bruijne, M.C., Wagner, C., Zegers,
M., Waaijman, R., Spreeuwenberg, P., Asscherman, H.,
van der Wal, G. and van Tulder, M.W. (2009) Direct me-
dical costs of adverse events in Dutch hospitals. BMC
Health Services Research, 9, 27.
doi:10.1186/1472-6963-9-27
[4] Dentzer, S. (2011) Urgent measures for an old problem.
Health Affairs, 30, 1626. doi:10.1377/hlthaff.2011.0961
[5] Marcus, A. (2009) Bending the curve: The twists and
turns. Health Affairs, 28, 1256-1258.
doi:10.1377/hlthaff.28.5.1256
[6] Skinner, J., Chandra, A., Goodman, D. and Fisher, E.S.
(2009) The elusive connection between health care spen-
ding and quality. Health Affairs, 28, w119-w123.
doi:10.1377/hlthaff.28.1.w119
[7] Hughes, J.S., Averill, R.F. and Goldfield, N.J. (2006)
Identifying potentially preventable complications using a
present on admission indicator. Health Care Financing
Review, 27, 63-82.
[8] Lagoe, R.J. and Westert, G.P. (2010) Evaluation of hos-
pital inpatient complications: A planning approach. BMC
Health Services Research, 10, 200.
doi:10.1186/1472-6963-10-200
[9] Lagoe, R., Pasinski, T., Kronenberg, P., Quinn, T. and
Schaengold, P. (2006) Linking health services at the com-
munity level. Canada Healthcare Quarterly, 9, 60-65.
[10] National Center for Health Statistics (2009) International
classification of diseases, tenth revision, clinical modifi-
cation (ICD-10-CM).
http://www.cdc.gov/nchs/about/otheract/icd9/icd10cm.ht
ml
[11] Centers for Disease Control and Prevention (1997) Gui-
delines for prevention of nosocomial pneumonia. Mor-
bidity and Mortality Weekly Report, 46, 1-79.
[12] Niederman, M.S., Craven, D.E., Bonten M.J., et al. (2005)
Guidelines for the management of adults with hospital ac-
quired, ventilator-associated, and healthcare associated—
Pneumonia. American Journal of Respiratory Critical
Care Medicine, 171, 388-416.
doi:10.1164/rccm.200405-644ST
[13] McEachern, R. and Campbell Jr., G.D. (1998) Hospital-
acquired pneumonia: Epidemiology, etiology, and treat-
ment. Infectious Disease Clinics of North America, 12,
761-779. doi:10.1016/S0891-5520(05)70209-9
[14] Cohen, S.H., Gerding, D.N., Johnson, S., Kelly, C.P., Loo,
V.G., McDonald, L.C. and Wilcox, M.H. (2010) Clinical
practice guidelines for clostridium difficile infection in
adults: 2010 update by the society for healthcare epide-
miology of America and the infectious diseases society of
America. Infection Control & Hospital Epidemiology, 31,
431-455. doi:10.1086/651706
[15] Gould, C.V., Umscheid, C.A., Agarwal, R.K., Kuntz, G.
and Pegues, D.A. (2009) Guideline for prevention of ca-
theter-associated urinary tract infection. Healthcare Infec-
tion Control Practices Advisory Committee, Atlanta.
[16] Grossman, S. and Mager, D. (2010) Clostridium difficile:
Implications for nursing. MEDSURG Nursing, 19, 155-
158.
[17] Novell, M.J. and Morreale, C.A. (2010) The relationship
between inpatient fluoroquinolone use and clostridium dif-
ficile-associated diarrhea. Annals of Pharmacotherapy, 44,
826-831. doi:10.1345/aph.1M696
[18] Pépin, J., Saheb, N., Coulombe, M.A., Alary, M.E., Cor-
riveau, M.P., Authier, S. and Lanthier, L. (2005) Emer-
gence of fluoroquinolones as the predominant risk factor
for clostridium difficile-associated diarrhea: A cohort stu-
dy during an epidemic in Quebec. Clinical Infectious Dis-
ease, 41, 1254-1260.