0% found this document useful (0 votes)

108 views

Case Study 10 Hysterectomy: Hospital Costs and Outcomes Study For NSW Health

Uploaded by

kriesya riesya

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

108 views

Case Study 10 Hysterectomy: Hospital Costs and Outcomes Study For NSW Health

Uploaded by

kriesya riesya

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 68

Independent Pricing and Regulatory Tribunal

Case study 10 — Hysterectomy

Hospital costs and outcomes study for NSW Health

Other Industries
July 2010
Case study 10 - Hysterectomy
Hospital costs and outcomes study for NSW Health

Other Industries
July 2010
© Independent Pricing and Regulatory Tribunal of New South Wales 2010.
This work is copyright. The Copyright Act 1968 permits fair dealing for study, research,
news reporting, criticism and review. Selected passages, tables or diagrams may be
reproduced for such purposes provided acknowledgement of the source is included.

ISBN 978-921628-57-3 S9-56

The Tribunal members for this review are:

Mr James Cox, Acting Chairman and Chief Executive Officer
Ms Sibylle Krieger, Part Time Member

Inquiries regarding this document should be directed to a staff member:

Alison Milne (02) 9290 8443
Bee Thompson (02) 9290 8496

Independent Pricing and Regulatory Tribunal of New South Wales

PO Box Q290, QVB Post Office NSW 1230
Level 8, 1 Market Street, Sydney NSW 2000
T (02) 9290 8400 F (02) 9290 2061
www.ipart.nsw.gov.au

ii IPART Case study 10 - Hysterectomy

Contents

1 Introduction and executive summary 1

1.1 Why did we select hysterectomy as one of the case studies? 4
1.2 What was the scope of the hysterectomy case study? 4
1.3 What were the key findings of the hysterectomy case study? 5
1.4 What are the key implications of these findings? 9
1.5 List of recommendations 10
1.6 What does the rest of this report cover? 10

2 Main types of hysterectomy procedures and hysterectomy patients 11

2.1 Type of hysterectomy procedures and the conditions treated 11
2.2 Choice of procedure at the study hospitals 13
2.3 Hysterectomy is an area of declining activity 13

3 Number and mix of patients across study hospitals 15

3.1 Number of hysterectomy cases at study hospitals 15
3.2 Age of patients 17

4 Comparison of lengths of stay 18

4.1 Comparing length of stay for all hysterectomy cases 18
4.2 Comparing length of stay by surgical procedure 20
4.3 Variation between physicians in episode length of stay 21

5 Costs of providing inpatient care 24

5.1 Cost of nursing staff in wards 24
5.2 Imaging costs 32
5.3 Pathology costs 33
5.4 Blood use costs 34
5.5 Operating theatre time and costs 35

6 Configuration of care 38

7 Outcome, safety and quality indicators 40

7.1 How we developed a set of clinical indicators 40
7.2 Analysing indicators and risk-adjusting for patient characteristics 41
7.3 List of clinical indicators for hysterectomy and their availability 42

Case study 10 - Hysterectomy IPART iii

Contents

Appendices 45
A List of full recommendations from main report 47
B Risk-adjusted indicators provided by NSW Health 57

Glossary 58

iv IPART Case study 10 - Hysterectomy

1 Introduction and executive summary

NSW Health is currently coordinating a project that ultimately aims to improve

clinical practice and efficiency consistently across the NSW hospital system. The
project involves 6 components, and is designed to enable development of a
methodology that makes better use of available data to compare patient mix, costs,
clinical practice and outcomes and which can then be applied across other hospitals
to improve performance (see Box 1.1 for more information).

NSW Health asked the Independent Pricing and Regulatory Tribunal of NSW
(IPART) to conduct a costs and outcomes study that encompasses 3 components of
this larger project. The aim of the study was to provide information and analysis that
can be used by clinical experts to better understand the variation in clinical practice
in NSW hospitals, and the extent to which this variation can lead to differences in
hospital costs and clinical outcomes.

IPART’s study involved comparing costs, configurations of care and outcomes in 5

selected NSW hospitals:
 Royal Prince Alfred Hospital (RPAH)
 Royal North Shore Hospital (RNSH)
 John Hunter Hospital (JHH)
 Bankstown-Lidcombe Hospital (BLH), and
 Gosford Hospital (GH).

To do this, we analysed management practices at the hospital-wide level and did

detailed case studies of 11 specific clinical areas. As costs, configurations of care and
relevant indicators of outcomes vary significantly depending on the condition of the
patient and/or the procedure undertaken, these case studies allowed us to compare
the hospitals on a more like-with-like basis. This document discusses our findings in
one of these 11 clinical areas – hysterectomy. (See Box 1.2 for the full list of clinical
areas we examined, how they were selected, and how we conducted the case
studies.)

Case study 10 - Hysterectomy IPART 1

1 Introduction and executive summary

Box 1.1 NSW Health Project

IPART’s hospital costs and outcomes study is part of a larger, multi-stage project NSW Health is
coordinating with the assistance of other organisations. The terms of reference for this project
set out 6 components:
1. Audit the quality of current coding and costing data.
2. Analyse differences in costs between 3 principal tertiary referral hospitals and 2 other
principal referral hospitals.
3. Describe the different configurations of care that underpin different cost profiles.
4. Analyse available data on differences in adjusted admission rates and clinical outcomes for
the 5 selected hospitals.
5. Determine whether variations in configurations of care lead to different clinical outcomes.
6. Identify the extent to which clinical variation exists, with the aim of achieving clinical best
practice and maximum efficiency.

The first component is being completed by Health Outcomes International (audit of costing)
and Pavilion Health (audit of coding). The results will assist the NSW Department of Health in
further developing episode funding, in line with the national agreement by COAG to move to a
more nationally consistent approach to activity-based funding. IPART has completed the
second, third and fourth components through our hospital costs and outcomes study. The
results of this study will be used by clinical experts in completing the fifth and sixth
components.

The NSW Health project is part of its response to the findings and recommendations made in
the Report of the Special Commission of Inquiry into Acute Care Services by Commissioner
Garling.a

a Flowing from the NSW Government’s response to the Garling Inquiry (Caring Together - The Health Action Plan for
NSW (2009)), ‘four pillars’ of clinical improvement have been established – Clinical Excellence Commission (CEC),
Agency for Clinical Innovation (ACI), Bureau of Health Information (BHI) and Clinical Education and Training Institute
(CETI). IPART’s analysis on costs, clinical practice and outcomes is to be considered by the NSW Department of Health
and clinical experts in these agencies to assess whether variations in configurations of care lead to different clinical
outcomes and to identify the extent to which clinical variation exists, with the aim of achieving clinical best practice
and maximum efficiency.

2 IPART Case study 10 - Hysterectomy

1 Introduction and executive summary

Box 1.2 IPART’s case studies

To compare costs, configurations of care and outcomes in the 5 study hospitals, we focused on
11 specific conditions or procedures in detail (as well as undertaking a broad, hospital-wide
analysis). These conditions/procedures are:
 Hip joint replacement
 Major chest procedures
 Breast surgery
 Cholecystectomy
 Appendicectomy
 Stroke
 Cardiology – stents, pacemakers and defibrillators
 Tracheostomy, or ventilation for greater than 95 hours
 Cataract/lens procedures
 Hysterectomy, and
 Obstetric delivery.

In selecting these conditions/procedures, and the relevant indicators to compare for each, we
were advised by a clinical consultant (Dr Paul Tridgell) and a clinical reference group (Professor
Bruce Barraclough, Dr Anthony Burrell, Dr Patrick Cregan, Professor Phillip Harris, Professor
Clifford Hughes, Professor Brian McCaughan, Professor Peter McClusky, Dr Michael Nicholl,
Professor Ron Penny, Professor Carol Pollock and Dr Hunter Watt).

The case studies were selected to provide a range of surgical procedures and a range of
medical conditions that met one or more of the following criteria:
 high volumes
 high reported costs
 high variability in reported costs
 apparent differences in clinical practice, or
 a range of models of care.

To conduct the case studies, we visited each of the hospitals and spoke with a range of staff,
including clinical, nursing, management, finance, coding and administrative staff. We also
collected a range of clinical and financial data from NSW Health, relevant area health services
and hospitals. By analysing the data and speaking with clinical experts, we established the
most suitable data available for comparing hospitals on a like-with-like basis.

For further information on our methodology and broad findings on costs, outcomes and
configurations of care, see our report NSW Health costs and outcomes study by IPART for selected
NSW hospitals. Our detailed findings on the other case study areas can be found in our reports
on each area.

Case study 10 - Hysterectomy IPART 3

1 Introduction and executive summary

1.1 Why did we select hysterectomy as one of the case studies?

Hysterectomy was selected as one of the clinical areas for detailed study because it
involves:
 moderate volumes1
 large differences in reported costs between hospitals despite a relatively uniform
patient type and similar length of stay
 apparent differences in clinical practice (eg, surgical approach taken to perform
the hysterectomy).

1.2 What was the scope of the hysterectomy case study?

The hysterectomy case study compared the costs and configurations of care related
to this surgical procedure. We used diagnostic related groups (DRGs)2 to define the
procedures and identify the data included in the scope of the case study. This case
study involved a single DRG (see Table 1.1).

Table 1.1 DRGs included in the scope of the hysterectomy case study

DRG Description
N04Z Hysterectomy for non malignancy

For this case study, we found that it was necessary to go beyond DRGs and we
identified subgroups of patients based on the principal type of surgery performed
(‘principal procedure’) to meaningfully compare costs, configurations of care and
outcomes. See Chapter 2 for further details.

Unless specified otherwise in this case study, the data we analysed related to the
12-month period from 1 July 2008 to 30 June 2009.

1 In 2007/08, 26,073 hysterectomies were undertaken in Australian hospitals. See Australian

Institute of Health and Welfare, AR-DRG Data Cubes, Separation, patient day and average length of
stay statistics by Australian Refined Diagnosis Related Group (AR-DRG) Version 5.0/5.1, Australia,
1998-99 to 2007-08.
http://d01.aihw.gov.au/cognos/cgi-bin/ppdscgi.exe?DC=Q&E=/AHS/drgv5_9899-0708_v2)
2 Case allocation to a DRG is based on a combination of administrative patient data and 2 key
classification systems: the International Statistical Classification of Diseases and Related Health
Problems, 10th Revision, Australian Modification (ICD-10_AM) for diagnoses and the Australian
Classification of Health Interventions (ACHI) for procedures. The classification edition current
during the 2008/09 period was the 6th Edition. For the purpose of this study, the term DRG is
synonymous with AR-DRG 5.2.

4 IPART Case study 10 - Hysterectomy

1 Introduction and executive summary

1.3 What were the key findings of the hysterectomy case study?

To compare the costs and configurations of care for hysterectomies at the study
hospitals, we collected, analysed and compared data on:
 the type, number and mix of hysterectomy patients at each hospital
 the average length of stay for these patients at each hospital
 selected costs, or major clinical resources used to provide acute inpatient care for
these patients at each hospital
 the configurations of care used to provide and manage hysterectomy patient care
at each hospital
 indicators of outcome, safety and quality for hysterectomies for each hospital.

Our key findings are summarised below.

1.3.1 Types of hysterectomy procedures

As noted above, we found it was necessary to go beyond the DRG level to identify
groups of reasonably similar patients. Therefore, we broke the data into 4 subgroups
based on the principal type of hysterectomy procedure performed. These subgroups,
which have different implications for costs and length of stay, were abdominal,
laparoscopic, vaginal and laparoscopically-assisted vaginal hysterectomy.
Abdominal surgery is the oldest method and has the longest length of stay. Vaginal
hysterectomy is not feasible for patients with a very large uterus.

We found that about 50% of patients in the study hospitals had abdominal surgery in
2008/09, and about 34% had vaginal surgery. BLH had the highest rate of abdominal
surgery (62% of cases), and RNSH had the lowest (44% of cases).

We found that the choice of procedure depends partly on the condition to be treated,
but also on the skills and preferences of the clinical staff.

We also found that hysterectomy is an area of declining activity because

hysterectomies are increasingly being replaced by less invasive procedures such as
endometrial ablation3 and uterine artery embolisation4.

3 Endometrial ablation is a process that destroys the lining of the uterus, and is used to treat
uterine bleeding (WebMD, http://women.webmd.com/endometrial-ablation-16200).
4 Uterine artery embolisation is used to treat uterine fibroids by blocking both uterine arteries
with particles injected via the femoral and uterine arteries (National Institute for Health and
Clinical Excellence, http://www.nice.org.uk/IPG94).

Case study 10 - Hysterectomy IPART 5

1 Introduction and executive summary

1.3.2 Number and mix of patients

Our analysis shows that the average age of patients in all hospitals was 51 years old,
with only a small amount of variation between the hospitals. Almost all
hysterectomies were planned rather than emergency admissions.

We found that RNSH had a relatively low volume of hysterectomy cases relative to
its size, which probably reflects the volume of private cases undertaken in nearby
private facilities. GH had a high number of hysterectomies relative to its size.

1.3.3 Average length of stay

Our analysis indicates that the measure often used in National Hospital Cost Data
Collection (NHCDC) and DRG benchmarking – the average ‘acute episode length of
stay’, is appropriate for the hysterectomy DRG because there were very few transfers
in or out of the study hospitals, and very few patients that recorded more than one
acute episode. We have reported only the acute episode length of stay (LOS1).

We found that RNSH has the highest 23 hour surgery rate and the highest proportion
of 1 or 2 day stays. GH has a shorter average length of stay, but this is because it has
a significantly smaller share of stays of more than 4 days. The comparatively long
average length of stay at RPAH and JHH is largely due to a small number of cases at
these hospitals with a length of stay exceeding 7 days. We also found that the
average length of stay is higher for abdominal hysterectomies than for vaginal
hysterectomies.

Finally, we found that there is a great deal of variation between treating physicians in
length of stay of their patients and the proportion of patients staying for only 1 or 2
days.

1.3.4 Costs of inpatient care

To compare the costs related to the case study areas at the study hospitals, we
examined the management and use of a selection of clinical resources used directly
for patient care in that area. For hysterectomy, the main clinical resources we
examined were nursing staff in wards, imaging, pathology, blood use, and operating
theatre time. We had aimed to also estimate medical staff costs and pharmacy costs
for this case study, but were unable to obtain consistent comparisons within the
timeframe for this review.5

We separated costs for abdominal versus vaginal hysterectomies in our analysis of

direct nursing costs, blood use and operating theatre time, but did not break down
the hysterectomy cases into subgroups for our estimates of pathology or imaging
costs.

5 Medical staffing and pharmacy are discussed in our main report, NSW Health costs and outcomes
study by IPART for selected NSW hospitals, July 2010, Chapters 9 and 14.

6 IPART Case study 10 - Hysterectomy

1 Introduction and executive summary

Cost of nursing staff in wards

We found that the number of nursing hours per patient day (ie, the nursing staff-to-
patient ratio) is the main driver of the differences in the average cost per episode for
this DRG as a whole, partly because the length of stay is fairly uniform across the
hospitals (except JHH). However, we found that the average length of stay drives
the cost difference between different surgical procedures within each hospital.

We found that a more senior nursing staff mix – Clinical Nurse Specialists (CNSs)
and Registered Nurses (RNs) – was associated with fewer nursing hours per patient
day and a lower nursing cost per patient day. In particular, we found that JHH had
the lowest cost per patient day and the highest proportion of CNSs and RNs in its
staff mix, while BLH had lowest proportion of CNSs and RNs in its staff mix and the
highest cost per patient day.

The hospitals use inpatient fractions (IFRACs) to allocate staff time to acute care and
other staff responsibilities. We found that nursing costs are highly sensitive to the
IFRACs the hospitals apply to them. When all nursing costs are attributed to
inpatient care (ie, IFRAC=1), the cost variation between the hospitals is significantly
smaller than when the hospital’s IFRACs are used. Setting IFRAC equal to
1 increased RNSH costs the most, and changed its ranking from the lowest cost
hospital to the highest cost hospital (excluding RPAH).6

Imaging, pathology and blood use costs

We found that imaging costs for this DRG were low at all the hospitals – less than
$50 per episode. Blood use costs were also low and ranged from $24 per episode at
RNSH to $91 per episode at BLH. Blood use was much higher for abdominal
hysterectomies than for vaginal hysterectomies.

Pathology costs ranged between $221 and $276 at all the hospitals except GH, where
the cost was significantly lower, at $87 per episode. It appeared that the GH
pathology data excluded many item numbers for the higher cost tests, which
accounted for the difference. We believe that these tests were performed at another
lab and not captured in the hospital imaging data. All the study hospitals did a
significant proportion of these tests on the day of admission.

6 We excluded RPAH from this analysis because the gynaecology ward moved during the year,
making the costs unreliable.

Case study 10 - Hysterectomy IPART 7

1 Introduction and executive summary

Operating theatre times

We analysed operating theatre time for the various hysterectomy procedures. We

cannot directly compare our analysis of operating theatre times with the NHCDC
operating theatre costs, because the NHCDC costs include items that are not related
to time spent in surgery (such as goods and services and some anaesthetic
department costs).7 However, we found that there is a significantly wider variation
in the NHCDC costs than in operating theatre times.

National Hospital Cost Data Collection

The NHCDC reports estimates of average hospital costs based on data it collects
from hospitals around Australia.8 In this study, we had access to the study hospitals’
provisional de-identified patient-level data for 2008/09, as well as the overall
averages publicly reported by the NHCDC for different hospital groupings in
2007/08.9 The final NHCDC estimates for study hospitals for 2008/09 became
available towards the end of our study. We compared these to the provisional costs
and found that some of the costs had changed substantially. Give the limited time
available, we used the final costs only where these were substantially different from
the provisional costs.

We were able to compare study hospitals’ NHCDC estimates with our estimates of
nursing, imaging and pathology costs. For pathology we used the final NHCDC
costs. We found there is a much smaller variation in our nursing costs across the
study hospitals than is implied by the NHCDC costs. We also found a higher degree
of consistency in the hospitals’ use of imaging and pathology than is reflected in the
NHCDC. Finally, we found that the NHCDC operating theatre costs vary over a far
wider range than differences in operating theatre time would indicate.

1.3.5 Configurations of care

Due to time constraints, we did not have detailed discussions about configurations of
care for hysterectomies during our hospital visits.

However, we make the observation that JHH, RNSH and RPAH have specialist
gynaecology oncology units. These hospitals may have a more complex casemix
because some patients admitted through these units may be coded into this DRG
(N04Z - hysterectomy for non-malignancy). For example, we found a number of
cases with diagnoses like ‘carcinoma in situ of cervix, unspecified’ who were
included in this DRG.

7 NSW Health, Program and Product Data Collection Standards Product costing standards v1.0,
Reporting requirements v1.0, 2008/09, p 83.
8 In NSW, these cost estimates are often compiled by area health services, rather than individual
hospitals.
9 Commonwealth Department of Health and Ageing, National Hospital Cost Data Collection, Round
12, 2007-2008, September 2009.

8 IPART Case study 10 - Hysterectomy

1 Introduction and executive summary

1.3.6 Outcome indicators

NSW Health provided risk-adjusted 30-day mortality rates to compare hospital

outcomes for hysterectomy patients. Their analysis found that there were fewer than
5 deaths following hysterectomies in the 5 hospitals over the 3-year period 2005/06
to 2007/08. The number of deaths was too small to allow comparisons between the
hospitals.

1.4 What are the key implications of these findings?

DRG N04Z is not a uniform grouping but the differences are small compared with other
clinical conditions

Benchmarking studies of the performance of individual hospitals and the public and
private hospital sectors often use DRGs as the basis for comparing length of stay and
cost. This assumes that patients whose condition or procedure has been coded with
the same DRG are relatively similar. Models of casemix or episode-based funding
are based on similar assumptions.

The hysterectomy DRG does not recognise the complexity introduced by the range of
different conditions that hysterectomies are used to treat, and the different surgical
procedures that can be used. Ideally, comparisons between the hospitals should take
into account both of these factors. However, we note that the variations in length of
stay and cost between the different surgical procedures and hospitals are small
compared with DRGs for other clinical conditions, for example those for hip joint
replacement or major chest procedures.10

Nursing staff mix

A higher share of RNs and CNSs does not necessarily mean higher nursing costs. A
more senior staff mix can means that fewer nursing hours are required.

Further work on alternative procedures to hysterectomy

Hysterectomy is an area of declining activity because hysterectomies are increasingly

being replaced by more conservative procedures such as endometrial ablation and
uterine artery embolisation. Future comparisons by NSW Health or hospitals could
compare the costs and configurations of care for these other procedures, and the
extent to which they are used instead of hysterectomies.

10 See Case study 2 Major chest procedures Hospital costs and outcomes study for NSW Health and Case
study 1 - Hip joint replacement - Hospital costs and outcomes study for NSW Health.

Case study 10 - Hysterectomy IPART 9

1 Introduction and executive summary

1.5 List of recommendations

1 That any future studies of hysterectomy compare the costs and outcomes for
hysterectomies with the costs and outcomes of other procedures such as
endometrial ablation and uterine artery embolisation. 39

1.6 What does the rest of this report cover?

The rest of this report discusses the findings of the hysterectomy case study in more
detail:
 Chapters 2 and 3 discuss the main types of hysterectomy procedures and the
number and mix of patients at the study hospitals.
 Chapter 4 compares the length of stay at the study hospitals, and analyses the
length of stay by treating physician.
 Chapter 5 describes how we analysed the costs of hysterectomies by identifying
the main clinical resources used to provide inpatient care, then estimating and
comparing the level of each resource used at the study hospitals. It also compares
our cost estimates with estimates based on the provisional or final data reported
to NSW Health as part of the NHCDC.
 Chapter 6 makes some general observations about configurations of care. Due to
time constraints we did not have detailed discussions about configurations of care
for hysterectomies during our hospital visits.
 Chapter 7 discusses the indicators of outcome, safety and quality for hysterectomy
we identified as clinically meaningful. It then compares the available data on
these indicators across the study hospitals.
 The appendices contain the complete list of recommendations for our hospital
costs and outcomes study, more detailed information on the data sources for risk-
adjusted outcome indicators and the glossary.

10 IPART Case study 10 - Hysterectomy

2 Main types of hysterectomy procedures and
hysterectomy patients

2 Main types of hysterectomy procedures and

hysterectomy patients

To meaningfully compare data on the costs, configurations of care and outcomes for
a particular condition or procedure, the patients to which the data relate must be
reasonably similar – to allow ‘like-with-like’ comparisons. As Chapter 1 discussed,
the hysterectomy case study includes a single DRG to identify clinical and financial
data related to patients who had undergone hysterectomies at the study hospitals.
However, our analysis of the data and discussions with clinicians indicates that there
is a diverse mix of patients within this DRG and several types of hysterectomy
procedures are performed.

Patients undergo hysterectomies to address a range of medical conditions, and their

care requirements can be very different depending on their principal diagnosis. This
influences the configurations of care, associated costs and outcomes.

For this case study, we could not separate the data on hysterectomies into subgroups
by diagnosis because there were such a large number of diagnoses related to this
procedure. However we did separate the data into subgroups based on the type of
hysterectomy procedure performed, which enabled further analysis of the different
care requirements associated with different procedures. The sections below list the
main types of hysterectomy patients and the main hysterectomy procedures.

2.1 Type of hysterectomy procedures and the conditions treated

There are four main types of hysterectomy procedures, which have different
implications for costs and length of stay.11 These are:
 Abdominal surgery, which involves an incision through the abdominal wall. This
is the oldest method, and involves the longest hospital length of stay of about
5 days.
 Laparoscopic hysterectomy, which involves 3 to 4 small incisions through the
abdominal wall. This method reduces the length of stay to around 3 days, but is
capital intensive in terms of the equipment used for the procedure.
 Vaginal hysterectomy, which is performed entirely through the vaginal canal.
This method has the shortest length of stay of 1 to 2 days.

11 Comparisons of length of stay are discussed in Chapter 4. Length of stay estimates were also
provided by some study hospitals during the hospital visits, namely RNSH.

Case study 10 - Hysterectomy IPART 11

2 Main types of hysterectomy procedures and
hysterectomy patients

 Laparoscopically-assisted vaginal hysterectomy, which is also performed through

the vaginal canal but with the assistance of a laparoscope inserted through small
incisions in the abdominal wall.,12

Hysterectomy in this DRG is used to treat a number of different conditions, which

include:13
 Fibroids, which are non-cancerous, muscular tumours that grow in the wall of the
uterus.
 Endometriosis, when the tissue that lines the uterus grows outside the uterus.
 Prolapse of the uterus, the uterus slips from its usual place down into the vagina.
 Adenomyosis, when the tissue that lines the uterus is present inside the walls of
the uterus.
 Chronic pelvic pain, for which surgery may be a last resort.
 Abnormal vaginal bleeding.
 Very rarely, hysterectomy is needed to control bleeding during a caesarean
delivery14.

The choice of procedure depends partly on the clinical condition to be treated. For
example, vaginal procedures tend to deal with prolapse of the uterus while
abdominal surgery may be required when complications are expected or surgical
exploration is required. Vaginal hysterectomy is not feasible for patients with a very
large uterus (for example due to large fibroids). Laparoscopic procedures are more
appropriate to control bleeding.

But the choice of procedure may also depend on the skills and preferences of the
clinical staff. For example, we were advised that at RNSH there tend to be more
laparoscopic procedures due to the skill mix of the staff who often only do
gynaecological work. Clinicians who do vaginal and abdominal procedures tend to
cover obstetrics and gynaecology, as well as having more cases relating to
malignancy and pelvic floor problems.

12 Medline Plus, (http://www.nlm.nih.gov/medlineplus/ency/article/002915.htm).

13 Hysterectomy is also used to treat cancer of the uterus, ovary, cervix, or endometrium, but these
cases fall under a different DRG.
14 US Department of Health and Human Services,
(http://www.womenshealth.gov/faq/hysterectomy.cfm#b).

12 IPART Case study 10 - Hysterectomy

2 Main types of hysterectomy procedures and
hysterectomy patients

2.2 Choice of procedure at the study hospitals

In 2008/09, about 50% of patients at the study hospitals had an abdominal

hysterectomy and 34% had a vaginal hysterectomy (see Table 2.1). BLH had the
highest rate of abdominal surgery (62% of cases), and RNSH had the lowest (44% of
cases). RNSH had the highest rate of laparoscopically assisted vaginal surgery (21%
of cases compared with less than 10% of cases at the other hospitals).

Table 2.1 Proportion of hysterectomies by type of procedure

Hospital Abdominal Vaginal Laparoscopic Laparoscopically- Other Total

assisted vaginal
% % % % % %
RPAH 53 32 6 8 2 100
GH 46 46 2 5 2 100
RNSH 44 27 4 21 4 100
BLH 62 25 0 8 6 100
JHH 49 35 7 8 0 100
All study 50 34 5 9 2 100
hospitals
Note: ‘Other’ consists of those cases which did not have a procedure code for an abdominal, vaginal, laparoscopic or
laparoscopically-assisted vaginal hysterectomy. Totals may not add up due to rounding.
Source: HIE inpatient statistics, 2008/09 and IPART analysis.

2.3 Hysterectomy is an area of declining activity

Hysterectomy is an area of declining activity because hysterectomies are increasingly

being replaced by more conservative procedures such as endometrial ablation and
uterine artery embolisation. In NSW, there was a decline in activity of 15% in
Australian hospitals between 1998/99 and 2007/08.15

Future comparisons by NSW Health or hospitals could compare the costs and
configurations of care for these other procedures, but we have not done so in this
study. Table 2.2 shows the number of endometrial ablation and uterine artery
embolisation procedures carried out at the study hospitals in 2008/09, as well as the
number of hysterectomy cases for comparison. Endometrial ablation was a fairly
common procedure in our study hospitals, particularly at GH. Uterine artery
embolisation was less commonly used.

15 Australian Institute of Health and Welfare, AR-DRG Data Cubes, Separation, patient day and
average length of stay statistics by Australian Refined Diagnosis Related Group (AR-DRG)
Version 5.0/5.1, Australia, 1998-99 to 2007-08, (http://d01.aihw.gov.au/cognos/cgi-
bin/ppdscgi.exe?DC=Q&E=/AHS/drgv5_9899-0708_v2).

Case study 10 - Hysterectomy IPART 13

2 Main types of hysterectomy procedures and
hysterectomy patients

Table 2.2 Number of endometrial ablation and uterine artery embolisation

procedures

Hospital Hysterectomy cases Endometrial Uterine artery

ablation embolisation
procedures proceduresa
RPAH 200 7 6
GH 101 82 1
RNSH 52 14 11
BLH 52 5 0
JHH 204 26 2
All study hospitals 609 134 20
a The number of Transcatheter embolisation of blood vessels, pelvis, procedure code 35321-06, on patients in N07Z
(“Other uterine and adnexa procedures for non-malignancy”).
Source: HIE inpatient statistics, 2008/09 and IPART analysis.

14 IPART Case study 10 - Hysterectomy

3 Number and mix of patients across study hospitals

We identified the total number of hysterectomy cases at each hospital during the
study period and the proportion of these that were emergency admissions. We then
compared the age of patients at each hospital.

We found that the number of hysterectomy procedures performed at the study

hospitals varied significantly. Specifically, we found that RNSH performed
relatively few hysterectomies, possibly reflecting the volume of private cases
undertaken in nearby private facilities. We found little variation in the average age
of patients.

The sections below discuss our analysis of patient numbers and the age of patients in
more detail.

3.1 Number of hysterectomy cases at study hospitals

Table 3.1 provides a comparison of the total number of hysterectomies performed at

the study hospitals in the selected DRG, as well as the type of procedure performed.
Our data indicates that there were 609 patients who had hospital stays for
hysterectomies in the hospitals during 2008/09 and about half of these had
abdominal procedures.16

Table 3.1 Hysterectomy cases at study hospitals, by type of procedure

Hospital Abdominal Vaginal Laparoscopic Laparoscopically Other Total
assisted vaginal stays
RPAH 106 63 11 16 4 200
GH 46 46 2 5 2 101
RNSH 23 14 2 11 2 52
BLH 32 13 - 4 3 52
JHH 100 72 15 16 1 204
All study
hospitals 307 208 30 52 12 609
Note: See Box 3.1 for details on how we calculated the number of cases.
Source: HIE inpatient statistics, 2008/09 and IPART analysis.

16 We have only included cases if the first episode of the admission has a DRG of N01Z to avoid
counting any patients twice.

Case study 10 - Hysterectomy IPART 15

3 Number and mix of patients across study hospitals

JHH and RPAH had similar numbers of cases at 204 and 200, respectively while GH
had 101 cases. RNSH performed only 52 cases, as did BLH. This relatively low
volume of cases at RNSH probably reflects the volume of private cases undertaken in
nearby private facilities.

RPAH performed the most abdominal hysterectomies (106) while JHH performed
the most vaginal hysterectomies (72). GH and BLH performed very few laparoscopic
or laparoscopically-assisted vaginal hysterectomies.

Box 3.1 provides more detail on how we calculated the number of cases at each
hospital.

Box 3.1 How we calculated the number of hysterectomy cases

To calculate the number of hysterectomy cases at the study hospitals, we:

 used patient episode data for 2008/09
 counted adjoining episodes as part of the same stay (ie, adjoining episodes counted as one
case)
 only included patient data where the whole patient stay occurred within 2008/09 (ie, all
episodes and adjoining episodes had to start on or after 1 July 2008 and end on or before
30 June 2009 to be counted)
 only included patient data where the first episode in the year in the study hospital was
coded as a DRG for hysterectomy (ie, episode sequence number has to be 1).

The approach prevented double counting. It excluded patients that may be admitted for a
different condition and later be reclassified to a hysterectomy DRG.

Note that our approach means that the number of cases we identified will be less than the
number of separations in 2008/09.

3.1.1 Admissions and transfers

Almost all hysterectomies were planned admissions, and there were very few
transfers in or out of the study hospitals (see Table 3.2).

16 IPART Case study 10 - Hysterectomy

3 Number and mix of patients across study hospitals

Table 3.2 Transfers and emergency admissions for hysterectomy

Hospital Transfers in Transfers out Emergency admissions

% % %
RPAH 0 0 3
GH 1 0 1
RNSH 2 0 6
BLH 0 4 2
JHH 0 1 1
All study hospitals 0 1 2
Note: See Box 3.2 for details on how we identified emergency admissions and transfers.
Source: HIE inpatient statistics, 2008/09 and IPART analysis.

Box 3.2 How we identified emergency admissions and transfers

Emergency admissions were identified by linking emergency department attendance data with
admitted patient data where the time of arrival and departure in the emergency department
matched with the admission time.

Due to data quality issues with the transfer in and transfer out fields in the admitted patient
data, transfers were calculated using a linkage key developed by the Australian Institute of
Health and Welfare.

3.2 Age of patients

The average age of patients in all hospitals was 51 years old. Patients were on
average oldest at RPAH (54 years) and youngest at GH (47 years). The youngest
patient was 20 years old and the oldest was 90 years old (see Table 3.3). For the
study hospitals combined, only 9% of patients were more than 70 years old and 6%
were less than 35 years old.

Table 3.3 Average age of hysterectomy patients

Hospital Average age Age range

years years
RPAH 54 20-90
GH 47 29-77
RNSH 52 30-86
BLH 52 34-90
JHH 49 27-85
All study hospitals 51 20-90
Note: Age at date of admission.
Source: HIE inpatient statistics, 2008/09 and IPART analysis.

Case study 10 - Hysterectomy IPART 17

4 Comparison of lengths of stay

We examined the average length of stay of hysterectomy patients because it is one of

the factors that influence the cost of an individual’s hospital care. This is because a
large component of this cost is nursing care (and this cost increases with the length of
stay). In addition, differences in length of stay can point to differences in casemix or
clinical practice between hospitals.

For each of our clinical case studies we have calculated three different lengths of stay
measures. These are LOS1, which is the acute episode length of stay; LOS2, which is
the total length of stay in study hospital; and LOS3, which is LOS2 plus allowance for
time in other hospitals - one transfer in and one transfer out.

However in the case of hysterectomies, the majority of patients have planned rather
than emergency surgery and there are very few transfers in and out of the study
hospitals. Almost all patients are discharged after the acute episode. This means
that there is no observable difference in the acute episode length of stay (LOS1), the
hospital length of stay (LOS2) and the total number of days in hospital (LOS3). The
measure often used in NHCDC and DRG benchmarking – the average ‘acute episode
length of stay’, is therefore appropriate for the hysterectomy DRG.

We found that there is a great deal of variation between treating physicians in length
of stay of their patients and the proportion of patients staying for only 1 or 2 days.

4.1 Comparing length of stay for all hysterectomy cases

Table 4.1 compares the average length of stay for hysterectomy patients across the
study hospitals, using the acute episode measure (LOS1).

18 IPART Case study 10 - Hysterectomy

4 Comparison of lengths of stay

Table 4.1 Acute episode length of stay for all hysterectomy patients

Hospital Number of Average Number of cases Average length of

cases length of stay with LOS > 7 days stay excluding cases
with LOS >7 days

no. days no. days

RPAH 200 4.6 14 4.1
GH 101 3.6 2 3.5
RNSH 52 3.7 2 3.5
BLH 52 3.9 1 3.8
JHH 204 4.5 8 3.9
All study hospitals 609 4.3 27 3.8
Source: HIE inpatient statistics, 2008/09 and IPART analysis.

RPAH and JHH have the longest length of stay, as shown in Table 4.1. However, a
significant part of this is due to a few patients with a long length of stay, for example
14 of the 27 patients with a length of stay or more than 7 days are at RPAH and 8 are
at JHH. Excluding these patients significantly reduces the difference in the length of
stay between the study hospitals.

There are significant differences in the average length of stay for different conditions.
For example, the average length of stay ranged between 6 days for ‘benign neoplasm
of ovary’ to 1 day for ‘abnormal findings in specimens from female genital organs,
abnormal cytological findings’. This suggests that there is probably complexity that
is not adequately captured in the DRG system.

Figure 4.1 Proportion of hysterectomy stays by length of stay (days)

100% 7 days or
more
90% 9
7 6 days
80% 6
29
25 25 5 days
70% 10 12
60% 49 4 days
60
50%
36 13 3 days
40%
30% 24 59 2 days
53
20% 12
23 1 day
10% 33
22 5
0%
RPAH GH RNSH BH JHH

Data source: HIE inpatient statistics, 2008/09 and IPART analysis.

Case study 10 - Hysterectomy IPART 19

4 Comparison of lengths of stay

RNSH does 10% of cases on a 23 hour basis, and over 33% of cases have a length of
stay of less than 2 days, as shown in Figure 4.1. The other hospitals have much lower
23 hour surgery rates.

The shorter average length of stay at GH is due to a significantly smaller share of

stays of more than 4 days than at the other hospitals.

4.2 Comparing length of stay by surgical procedure

Table 4.2 compares the average length of stay for hysterectomy patients by surgical
procedure.

Table 4.2 Acute episode length of stay for abdominal and vaginal hysterectomies

Hospital Abdominal Vaginal

days days
RPAH 4.4 4.5
GH 4.1 3.1
RNSH 5.0 2.7
BLH 4.1 3.8
JHH 4.7 3.4
Note: The figures in this table exclude case outliers with a length of stay of 12 days or more. Laparoscopic or
laparoscopically-assisted vaginal hysterectomies are not included.
Source: HIE inpatient statistics, 2008/09 and IPART analysis.

The average length of stay varies depending on the surgical procedure adopted. The
average length of stay at the study hospitals varied from 4.1 days to 5.0 days for
abdominal hysterectomies and from 2.7 days to 4.5 days for vaginal hysterectomies
(excluding case outliers with a length of stay of 12 days or more). This is a further
complication in the casemix that the DRG system does not pick up.

Further analysis showed that the longer length of stay for vaginal hysterectomies at
RPAH is mainly due to the small number of patients that are admitted for only 1 or 2
days for this procedure (see Figure 4.2).

20 IPART Case study 10 - Hysterectomy

4 Comparison of lengths of stay

Figure 4.2 Length of stay for vaginal hysterectomies

100%
2 1
7 days or
90% 7
6 0 more
80% 2 11 6 days
70% 11 7
5 days
60% 19
50% 29 4 days
19 6
40%
3 days
30% 4
20% 16 2 days
14 18
10% 3
1 day
4
0%
RPAH GH RNSH BH JHH

Data source: HIE inpatient statistics, 2008/09 and IPART analysis.

4.3 Variation between physicians in episode length of stay

There is a great deal of variation between treating physicians in the proportion of

patients whose stay is either 1 or 2 days. Figure 4.3 shows this by presenting treating
doctors and the number of patients they treated compared with the proportion of
those patients whose length of stay was less than 1 or 2 day. We can see that for
2 physicians, all of their patients spend less than 2 days in hospital, and for
1 physician, half of their patients spend less than 1 day in hospital. However, a
greater majority of physicians treat patients who spend at least 2 days in hospital.

Case study 10 - Hysterectomy IPART 21

4 Comparison of lengths of stay

Figure 4.3 Proportion of patients with 1or 2 day stays per treating doctor

23HrRate
100%
90%
% length of stay <=23 hrs

80%
70%
60%
50%
40%
30%
20%
10%
0%
0 5 10 15 20 25 30 35 40 45 50

Number of hysterectomies per treating doctor

<=2Day
100%
90%
% length of stay <= 2 days

80%
70%
60%
50%
40%
30%
20%
10%
0%
0 5 10 15 20 25 30 35 40 45 50

Number of hysterectomies per treating doctor

Data source: HIE inpatient statistics, 2008/09 and IPART analysis.

There is also significant variation between treating physicians in the average length
of stay, as show in Figure 4.4. In this figure each physician is again identified by the
number of hysterectomies performed in 2008/09. The horizontal axis shows the
number of hysterectomies per physician, and the vertical axis shows the average
length of stay. Each physician is represented by a blue diamond.

For example, a physician who performed 35 hysterectomies kept his/her patients in

hospital for 3.3 days on average. There were 2 physicians who performed
22 surgeries each. One of these kept his/her patients in hospital for an average of
4.5 days, while the other kept his/her patients in hospital for an average of 3.7 days.

22 IPART Case study 10 - Hysterectomy

4 Comparison of lengths of stay

Figure 4.4 Average length of stay by number of hysterectomies per treating doctor

8
Average length of stay (days)

1
0 5 10 15 20 25 30 35 40 45 50

Number of hysterectomies per treating doctor

Note: An outlier of 18 days is excluded.

Data source: HIE inpatient statistics, 2008/09 and IPART analysis.

Case study 10 - Hysterectomy IPART 23

5 Costs of providing inpatient care

To compare the costs of caring for hysterectomy patients at the study hospitals, we
examined the major clinical resources used to provide inpatient care during their
‘acute episode’ (ie, using LOS1). The clinical resources we examined were:
 nursing staff in wards
 imaging
 pathology
 blood use
 operating theatres.

The sections below discuss our analysis of the estimated costs for each of these
resources across the study hospitals. For some resources, they also compare our
estimated costs with estimates based on provisional or final data reported to NSW
Health by area health services as part of the NHCDC.17

Note that given the scope of our task for this study, we were not able to undertake a
full bottom-up costing of all the above resource categories. Instead, we used a range
of approaches, based on the most reliable and consistent data we could obtain in the
time available. The methodologies we used are outlined in the sections below, and
discussed in more detail in our report, NSW Health costs and outcomes study by IPART
for selected NSW hospitals. Also note that while senior and junior medical staffing is a
major cost for hysterectomies, we were not able to compare the use and management
of these clinical resources due to the lack of consistent data.

5.1 Cost of nursing staff in wards

Nursing is one of the largest expenditure areas in hospital budgets, and is a

significant part of the cost of the acute care of patients having hysterectomies.

17 Due to time constraints we used final NHCDC data only where these were substantially
different from the provisional data. For this case study we used final NHCDC data only for
pathology costs.

24 IPART Case study 10 - Hysterectomy

5 Costs of providing inpatient care

5.1.1 Methodology

To analyse this cost and allow comparisons with the NHCDC, we focused on:
 the cost of nursing staff in wards (ie, excluding those of nursing staff in operating
theatres or senior nursing categories that provide area-wide or hospital-wide
functions)
 nursing costs for the acute episode only (ie, LOS1).

Our methodology for estimating nursing costs and its limitations are briefly
described in Box 5.1 and in more detail in Chapter 8 of the main report, NSW Health
costs and outcomes study by IPART for selected NSW hospitals.

We excluded RPAH from our analysis because the gynaecology ward at this hospital
moved during the year, making the costs unreliable.

Case study 10 - Hysterectomy IPART 25

5 Costs of providing inpatient care

Box 5.1 How we estimated nursing costs

IPART used a model for each hospital to allocate ward nursing costs to each DRG grouping and
compare nursing costs.

We calculated ‘nursing hours per patient day’, ‘nursing cost per patient hour’ and ‘nursing cost
per acute episode’ for 2008/09 by:
 Mapping the wards in each hospital to cost centres – so we could use these to extract
relevant payroll information for each ward.
 Extracting information from the payroll on nursing classification, nursing pay and nursing
hours worked for each ward.
 Applying inpatient fractions to our total ward nursing cost – so we only included nursing
costs for acute patient care. Note that some hospitals have a fraction of ‘1’ where other
hospitals may have fractions like ‘0.95’ for similar wards.
 Allocating ward nursing costs to all patients on the ward, based on their time on the ward
and the nursing service weights for their DRG.
 Allocating a cost of nursing care to each patient - for each step of the patient’s stay in acute
care. Note that from patient level episode information we attributed a cost to each ward
transfer during their ‘acute’ episode.

We then applied our estimate of nursing cost per hour to the average length of the acute
episode to obtain an estimate of the ‘nursing cost per episode’. We also calculated costs with
IFRAC = 1 for comparison.

Qualifications
 For our calculations, we included only direct costs of ordinary hours (excluding leave),
penalty rates and overtime, obtained from payroll data.
 The number of ‘nursing hours per patient day’ depends on the occupancy rates of the
wards. A higher occupancy rate reduces the hours per patient day but such a change can
cause other issues, like outliers or access block.
 The ‘nursing hours per patient day’ is the share of a patient’s use of the nursing staff based
on the nursing service weights. These service weights are not perfect and the mix of other
patients on the ward may impact on the nursing hours attributed to a patient and hence
their cost. The service weights do not take into account the generally higher cost of
patients at the start of their hospital stay.
 Some wards have a mix of more acute care with rehabilitation. Fewer nursing hours and
lower costs are attributed to the ‘acute’ episode in such wards compared with wards in
hospitals that have a greater separation of roles (acute wards separate from rehabilitation).
 Our nursing methodology excludes ‘wards’ like emergency departments where it is
particularly difficult to determine the inpatient fraction, but allocated a nursing cost for the
time spent in emergency.
 Hospitals with a shorter reported length of stay for the ‘acute’ episode may be expected to
have a higher number of nursing hours per day and higher daily nursing costs.

26 IPART Case study 10 - Hysterectomy

5 Costs of providing inpatient care

5.1.2 Use of IFRACs

The hospitals use inpatient fractions (IFRACs) to allocate staff time to acute care and
other staff responsibilities. We calculated two sets of inpatient nursing costs, the first
using the hospitals’ IFRACs and the second setting all IFRACs to 1.18 We did this
because we were concerned that IFRACs may not be consistently applied by the
hospitals.19

JHH has an IFRAC of 1, BLH and GH have IFRACs of 0.98 and 0.94 respectively,
while RNSH has an IFRAC of 0.79. These IFRACs may be valid, because ward
nursing staff (except at JHH) may spend time on non-inpatient activities (eg, staffing
outpatient clinics). However, the purpose of our analysis is to show how sensitive
the nursing cost estimates are to IFRACs. IFRACs also underlie the NHCDC
estimates of ward costs.

5.1.3 Ward nursing costs for all hysterectomies

Table 5.1 contains comparisons of our estimates of the average ward nursing costs for
the study hospitals (excluding RPAH). The last column shows the direct nursing
costs from the provisional data for the NHCDC, which we compare with our
estimates of cost.

Our analysis indicates that the cost per episode is affected by three main factors:
 IFRACs
 nursing hours per patient day (ie, nursing staff-to-patient ratio)
 nursing staff mix.

We found that these factors outweigh the impact of the length of stay on the average
cost per acute episode for the DRG as a whole, partly because the length of stay is
fairly uniform across the hospitals (except JHH). However, as discussed in section
5.1.4 below, we found that the length of stay accounts for most of the cost difference
between different surgical procedures within each hospital.

IFRACs affect both the nursing costs per acute episode and the nursing costs per
patient day (see Table 5.1 and Table 5.2). When all nursing costs are attributed to
inpatient care (ie, IFRACs=1), there is a smaller variation in nursing costs across the
study hospitals. Setting IFRACs to 1 has the biggest impact on RNSH, because the
IFRACs used at this hospital are comparatively low. When IFRACs are used, RNSH
has the lowest cost per episode ($807), but when IFRACs are set to 1, RNSH has the
highest cost per episode along with BLH (respectively $1,019 and $1,018).

18 Setting IFRACs to 1 means that we allocated 100% of the nursing time to inpatient care at the
hospitals for the purpose of comparison.
19 IFRACs may differ between the hospitals for valid reasons, because the ward nursing staff in
some hospitals may spend more time on outpatient activities than those in other hospitals. The
purpose of our analysis is to show how sensitive the nursing cost estimates are to IFRACs.

Case study 10 - Hysterectomy IPART 27

5 Costs of providing inpatient care

We found that the second factor, the number of nursing hours per patient day (which
reflects the nursing staff-to-patient ratio), outweighed the impact of length of stay on
the hospitals’ episode costs. For example, we estimated that the average cost of
nursing per acute episode is lower at JHH ($893) than at BLH ($1,002) when IFRACs
are used, even though JHH has a longer length of stay (4.5 days) than BLH (3.9 days).
The reason for this is that JHH has only 5.2 nursing hours per patient day compared
with 7.5 at BLH, which translates into a significantly lower cost per patient day at
JHH ($201) compared with BLH ($256). (See Table 5.2.)

The third factor is the nursing staff mix, which affects the average cost per nursing
hour as well as the number of hours per patient day (and hence the average cost per
patient day). A more senior staff means a higher cost per nursing hour, but fewer
hours per patient day. The net effect of this on the cost per patient day depends on
the extent to which the higher hourly cost of a more senior staff is offset by a lower
number of nursing hours.

Our analysis indicates that JHH has a far higher percentage of Registered Nurses
(RNs) and Clinical Nurse Specialists (CNSs) than any of the other hospitals (91%
combined), and therefore a higher average costs per nursing hour of $39 (Table 5.3).
But it still has the lowest cost per patient day ($201) because it has such a low
number of nursing hours per patient day (5.2 hours). Note however that the low
number of hours per patient at JHH may also be a consequence of other factors such
a high bed occupancy rate or staff shortages.

Conversely, BLH has the lowest percentage of RNs and CNSs (65%), and the lowest
costs per nursing hour ($34). But the high number of nursing hours per patient day
at BLH results in a significantly higher cost per patient day ($256).

The direct nursing costs from the provisional data for the NHCDC (Table 5.1, last
column and Figure 5.1) vary over a far wider range than our nursing cost estimates –
from $320 at JHH to $1,579 at GH. The costs for JHH are particularly low compared
with our cost estimates, while the costs for GH are significantly higher.

Table 5.1 Ward nursing costs per acute episode, with and without IFRACs

Hospital LOS1 Episode IFRAC Episode NHCDC

nursing cost nursing cost, direct ward
with IFRAC IFRAC =1 costs

days $ no. $ $
RPAH 4.6 na na na 410
GH 3.6 925 0.94 987 1,579
RNSH 3.7 807 0.79 1,019 1,066
BLH 3.9 1,002 0.98 1,018 1,266
JHH 4.5 893 1.00 895 320
Note: Episode nursing cost calculated using acute episode LOS measure (LOS1).
Source: IPART analysis from HIE inpatient statistics, 2008/09, payroll data and provisional cost data 2008/09, NHCDC.

28 IPART Case study 10 - Hysterectomy

5 Costs of providing inpatient care

Table 5.2 Nursing costs and hours per patient day, with and without IFRACs

Hospital Nursing costs per patient day Nursing hours per patient day
With IFRAC IFRAC=1 With IFRAC IFRAC=1
$ $ hours hours
RPAH na na na na
GH 254 271 7.0 7.5
RNSH 217 273 6.2 7.9
BLH 256 260 7.5 7.7
JHH 201 201 5.2 5.2
Source: HIE inpatient statistics, 2008/09, payroll data and IPART analysis.

Table 5.3 Nursing costs by award category, with IFRACs

All CNS RN EN AIN Students/

trainees

Case study 10 - Hysterectomy IPART 29

5 Costs of providing inpatient care

Figure 5.1 Comparison of direct ward nursing costs per acute episode ($)

1,800
1,600
1,400
$ per episode

1,200
1,000
800
600
400
200
-
RPAH GH RNSH BLH JHH

with IFRAC IFRAC=1 NHCDC

Source: IPART analysis from HIE inpatient statistics, 2008/09, payroll data and provisional cost data 2008/09, NHCDC.

5.1.4 Ward nursing costs for abdominal and vaginal hysterectomies

As previously discussed, this hysterectomy DRG covers a range of different

conditions and procedures. To compare the costs on a more like-with-like basis, we
calculated separate costs for abdominal and vaginal hysterectomies. To further
improve comparability, we excluded cases with a length of stay of 12 days or more.20

20 Most of these long-stay cases are at JHH and RPAH, which have specialised gynaecology
oncology units and have a higher proportion of complex cases in their casemix.

30 IPART Case study 10 - Hysterectomy

5 Costs of providing inpatient care

Table 5.4 Ward nursing costs for abdominal and vaginal hysterectomies (excluding
long staya), with and without IFRACs

Hospital LOS1 Episode nursing Episode nursing NHCDC direct

cost with IFRAC cost, IFRAC =1 ward costs

days $ $ $

Abdominal hysterectomies excluding long staya

RPAH 4.4 na 390
GH 4.1 1,065 1,133 1,802
RNSH 5.0 1,093 1,388 1,358
BLH 4.1 1,058 1,073 1,334
JHH 4.7 799 800 304
Vaginal hysterectomies excluding long staya
RPAH 4.5 na na 524
GH 3.1 800 843 1,386
RNSH 2.7 545 681 816
BLH 3.8 969 986 1,235
JHH 3.4 607 607 261
a Cases with a length of stay of 12 days or more excluded.
Note: Episode nursing cost calculated using acute episode LOS measure (LOS1).
Source: IPART analysis from HIE inpatient statistics, 2008/09, payroll data and provisional cost data 2008/09, NHCDC.

Table 5.5 Nursing costs and hours, with and without IFRACs

Hospital Nursing costs per patient day Nursing hours per patient day
With IFRAC IFRAC=1 With IFRAC IFRAC=1
$ $ hours hours
Abdominal hysterectomies excluding long staya
RPAH na na na na
GH 258 275 7.1 7.6
RNSH 218 277 6.2 7.9
BLH 257 261 7.5 7.6
JHH 171 171 4.4 4.4
Vaginal hysterectomies excluding long staya
RPAH na na na na
GH 255 269 7.1 7.5
RNSH 202 252 5.9 7.4
BLH 254 258 7.5 7.7
JHH 179 179 4.7 4.7
a Cases with a length of stay of 12 days or more excluded.
Source: HIE inpatient statistics, 2008/09, payroll data and IPART analysis.

Case study 10 - Hysterectomy IPART 31

5 Costs of providing inpatient care

At all the hospitals, the episode cost for abdominal hysterectomy is higher than for
vaginal hysterectomy (Table 5.4). The main reason for this difference is the longer
length of stay for abdominal hysterectomies. Abdominal hysterectomies also require
slightly more nursing hours per patient day at all the hospitals except BLH and JHH
(Table 5.5).

The NHCDC costs similarly show higher costs for abdominal hysterectomies than
vaginal hysterectomies, although the range is far wider than our analysis indicates.

5.2 Imaging costs

Table 5.6 sets out our calculations for the average cost of all imaging tests for
hysterectomy patients, during their acute episode and on the day of their admission.
It also sets out the average costs for selected high-cost tests (CT/MRIs, ultrasound
and fluoroscopy), as well as the direct and indirect imaging costs from the
provisional data for the NHCDC. Our methodology for calculating these costs is
explained in Box 5.2.

We found that the average cost of tests for hysterectomy patients is relatively low,
and varies from $47 at RPAH to $20 at RNSH. At RNSH most of the testing is done
on the day of admission, but this is not the case for the other hospitals. We also
found a higher degree of consistency in the hospitals’ use of imaging than is reflected
in the NHCDC (Table 5.6).

Table 5.6 Average value attributed to imaging during acute stay and on day of
admission ($)

All imaging CT/MRI Ultrasound Fluoroscopy NHCDC

tests direct and
indirect
costs

$ $ $ $ $

Acute episode
RPAH 47 21 2 3 193
GH 29 19 7 - 37
RNSH 20 9 2 - 13
BLH 35 31 - - 241
JHH 44 20 4 1 91
Day of admission
RPAH 6 5 0 0 na
GH 0 0 0 0 na
RNSH 16 9 2 0 na
BLH 1 0 0 0 na
JHH 2 0 0 1 na
Source: IPART analysis using data from hospital imaging services and provisional cost data 2008/09, NHCDC.

32 IPART Case study 10 - Hysterectomy

5 Costs of providing inpatient care

Box 5.2 Our approach to estimating imaging costs

We did not use a bottom-up costing approach to calculate imaging costs. Instead, we used
detailed information from imaging services on the number and type of tests performed, the
time and date. We use data for ‘acute’ episodes in 2008/09. As a proxy for cost, we attribute a
value based on the Medicare Benefits Schedule (MBS). Each test is valued at 100% of the MBS
fee.

More information is included in the main report, NSW Health costs and outcomes study by IPART
for selected NSW hospitals, Chapter 11 on imaging.

5.3 Pathology costs

Table 5.7 shows the average pathology costs for hysterectomy patients. Our
methodology is explained in Box 5.3.

Table 5.7 Average pathology costs

Acute episode Day of admission NHCDC direct and

indirect pathology
costs
$ $ $
RPAH 259 148 265
a a
GH 87 45 103
RNSH 221 135 186
BLH 251 183 241
JHH 276 162 547
a GH excludes many item numbers for high cost test.
Source: IPART analysis using data from hospital pathology services, 2008/09; and final cost data 2008/09, NHCDC.

We found that pathology costs are in the region of $250 per patient at all hospitals,
except GH. We also found that all the hospitals do a significant proportion of
pathology test on the day of admission. Pathology test for hysterectomy involve a
combination of low cost tests and a higher charge for examination of the removed
tissue.

Our analysis showed that the cost for pathology tests at GH was far lower than at the
other hospitals ($87). However, it appears that GH imaging data excludes many item
numbers for the higher costs tests. These tests are believed to be performed
elsewhere and not captured in the GH imaging data system. An examples is the item
number 72824:
Examination of complexity level 4 biopsy material with 1 or more tissue blocks, including
specimen dissection, all tissue processing, staining, light microscopy and professional
opinion or opinions - 2 to 4 separately identified specimens.

Case study 10 - Hysterectomy IPART 33

5 Costs of providing inpatient care

The MBS value of this item is currently $142. Adding this item to GHs costs would
increase its cost to $229, which is similar the other hospitals’ costs.

We found a far higher degree of consistency in the hospitals’ use of pathology for
diagnostic purposes than reflected in the NHCDC.

Box 5.3 Our approach to estimating pathology costs

We did not use a bottom up costing for pathology. Instead, we attributed a value for pathology
tests based on internal billing data between the hospitals and the pathology services. We also
used information on the time and date of tests.

Charging arrangements at each hospital are similar, but are not identical. All pathology
services base their charges on the Medicare Benefit Schedule.

More information is included in the main report, NSW Health costs and outcomes study by IPART
for selected NSW hospitals, Chapter 12 on pathology costs.

5.4 Blood use costs

We obtained detailed blood use data for each hospital, and calculated blood costs by
using an attributed cost of $250 per unit of blood. We found that the average cost of
blood for hysterectomy patients was relatively low. Blood use was much higher for
abdominal hysterectomies than for vaginal hysterectomies, and was particularly high
for abdominal surgery at BLH (Table 5.8).

Table 5.8 Average blood use costs for hysterectomy

Hospital All hysterectomies Abdominal Vaginal

hysterectomies hysterectomies
$ $ $
RPAH 33 57 4
GH 30 43 22
RNSH 24 54 0
BLH 91 148 0
JHH 45 43 63
Source: IPART analysis using data from hospital blood use data, 2008/09.

34 IPART Case study 10 - Hysterectomy

5 Costs of providing inpatient care

5.5 Operating theatre time and costs

We compared operating theatre time for the various hysterectomy procedures at the
study hospitals. We also compared operating theatre times with NHCDC operating
theatre costs.

We found that in general, operating theatre time (time from when surgery
commenced to when surgery was completed) was shorter for vaginal hysterectomies
than for abdominal hysterectomies, except at RNSH (Table 5.9). The average time of
surgery involving a laparoscope was longer than for either abdominal or vaginal
surgery. A laparoscopic hysterectomy took on average 62 minutes longer than an
open abdominal hysterectomy, while a laparoscopically-assisted vaginal
hysterectomy took 36 minutes longer than an open abdominal hysterectomy.

We cannot directly compare our analysis of operating theatre times with the NHCDC
operating theatre costs, because the NHCDC costs include items that are not related
to time spent in surgery (such as goods and services and some anaesthetic
department costs)21. However, we note that there is a significantly wider variation in
the NHCDC costs22 than in operating theatre times.23 For example, Figure 5.2 shows
that the NHCDC cost for:
 abdominal surgery at JHH is 2.7 times the cost at RNSH (which has the lowest
cost), while the surgery time at JHH is about 1.5 times the surgery time at RNSH
 vaginal surgery at JHH is 1.9 times the cost at RNSH, even though average theatre
time at JHH is about 8 minutes shorter than at RNSH.

21 NSW Health, Program and Product Data Collection Standards. Product costing standards v1.0,
Reporting requirements v1.0, 2008/09, p 83.
22 We used the provisional NHCDC costs for abdominal and vaginal surgery, scaled down in
proportion to the change in total operating theatre costs between the provisional and final
NHCDC data.
23 Operating theatre times may not be fully comparable between hospitals because they use
different systems to capture their theatre times.

Case study 10 - Hysterectomy IPART 35

5 Costs of providing inpatient care

Table 5.9 Operating theatre times and NHCDC costs for abdominal and vaginal
hysterectomies

Hospital Theatre time NHCDC costsa

Minutes $
Abdominal hysterectomy
RPAH 127.3 3,775
GH 93.7 3,377
RNSH 71.4 1,599
BLH 145.4 2,909
JHH 110.3 4,320
Vaginal hysterectomy
RPAH 107.0 3,345
GH 86.0 3,158
RNSH 84.6 1,894
BLH 111.8 2,909
JHH 77.0 3,519
a These are the provisional NHCDC costs for abdominal and vaginal surgery scaled down in proportion to the change
in total operating theatre costs between the provisional and final NHCDC data.
Note: Operating theatre times may not be fully comparable between hospitals because they use different systems to
capture their theatre times.
Source: IPART analysis using data from hospital operating theatres and NHCDC costs, 2008/09.

Figure 5.2 Ratio of time in surgery and NHCDC costs to RNSH time and costs for
abdominal and vaginal surgery (RNSH = 1)

Abdominal surgery Vaginal surgery

3.0 3.0
2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
RPAH GH RNSH BLH JHH RPAH GH RNSH BLH JHH

NHCDC costs Time in surgery NHCDC costs Time in surgery

Source: IPART analysis using data from hospital operating theatres and provisional NHCDC costs, 2008/09.

36 IPART Case study 10 - Hysterectomy

5 Costs of providing inpatient care

Overall, we believe that the NHCDC costs for 2008/09 could be misleading. They
could be improved by using standard clinical feeds (ie, actual patient data from
clinical systems) in the costing process, similar to the process we have used for our
analysis. The standard clinical data used in costing could also be used to provide
useful data for clinicians and others to compare resource use and performance.

We also believe that the Department needs to establish more ‘rules’ to ensure
consistent data for the NHCDC, and needs to audit the IFRACs and cost centres
regularly.

Case study 10 - Hysterectomy IPART 37

6 Configuration of care

The term ‘configurations of care’ refers to the way that hospitals choose to manage
and provide patient care, including their clinical practices. The particular
configurations of care within a hospital can be influenced by a complex array of
factors, including national or state-wide guidelines or protocols, the culture, practices
and controls of the individual hospital, the culture and practices of each clinical unit
and its leadership and the preferences of each clinician. Differences in the way
hospitals manage and provide patient care can also lead to differences in the costs
and outcomes of that care.

Due to time constraints we did not have detailed discussions about configurations of
care for hysterectomies during our hospital visits, and we can therefore make only
three very general observations.

The first observation relates to the regional role of RPAH, RNSH and JHH which,
along with 4 other NSW public hospitals, have specialist gynaecology oncology
units.24 These centres for gynaecological oncology were established by NSW Health
to improve clinical outcomes by concentrating treatment at a few hospitals. These
hospitals may have a more complex casemix, because some patients admitted
through these units may be coded into this DRG N04Z “hysterectomy for non-
malignancy”. For example, we found a number of cases with diagnoses like
‘carcinoma in situ of cervix, unspecified’ who were included in this DRG.

The second observation is that there are differences in the proportion of

hysterectomy cases performed on a 23 hour basis across the study hospitals. RNSH
performs hysterectomies on around 10% of patients in DRG N04Z “hysterectomy for
non-malignancy” on a 23 hour basis. The other hospitals had far lower 23 hour
surgery rates. See section 4.1.

24 NSW Health, (http://www.health.nsw.gov.au/gmct/gynaeonc/specialist_units.asp#para_3).

38 IPART Case study 10 - Hysterectomy

6 Configuration of care

The third observation concerns the use of alternative procedures to hysterectomy.

The Clinical Excellence Commission has raised concerns that hysterectomies may be
overused in some areas. They point to studies that show there are variations in rates
between regions that cannot be explained by underlying patterns of disease. They
found that hysterectomy rates in NSW for women under 35 years old are closely
correlated with socio-economic status. They also found that hysterectomy rates for
all age groups are significantly higher for area health services with large rural
populations.25

As previously mentioned, hysterectomy is an area of declining activity because

hysterectomies are increasingly being replaced by more conservative procedures
such as endometrial ablation and uterine artery embolisation. NSW Health could
investigate the extent to which these other procedures are being used instead of
hysterectomies in different hospitals and areas.

Recommendation
1. That any future studies of hysterectomy compare the costs and outcomes for
hysterectomies with the costs and outcomes of other procedures such as endometrial
ablation and uterine artery embolisation.

25 Clinical Excellence Commission, Quality and Safety of Health Care in NSW, version 10 (confidential
draft of Tuesday 5 July 2009), pp 130-133.

Case study 10 - Hysterectomy IPART 39

7 Outcome, safety and quality indicators

The terms of reference for this study required us to analyse available data on
differences in clinical outcomes across the 5 study hospitals. However, while there
are a number of safety and quality indicators being collected locally, at the state level
and through clinical registries, there are few clinically agreed outcome indicators. As
such, we found that data on only a few indicators of clinical outcomes are collected
consistently across hospitals, or on a state-wide (or national) basis. Therefore, we
worked with clinical experts to establish a set of outcome, safety and quality
indicators that are clinically relevant, and for which we could feasibly obtain data in
the timeframe for our study.

The sections below set out the clinical indicators we selected for hysterectomy.

7.1 How we developed a set of clinical indicators

To identify the indicators we should focus on for this study, we worked with a
number of eminent clinicians on our Clinical Reference Group26 to develop a set of
outcome indicators. We also consulted clinicians in study hospitals and sought
further advice from clinicians with specific expertise in the fields of interest, as well
as other relevant organisations.

Essentially, we aimed to establish a list of indicators that were:

 widely accepted as being clinically appropriate
 likely to be available from NSW hospitals, the NSW Department of Health or
other bodies, such as registries, and
 feasible for IPART to collect or calculate.

Following this clinical consultation, we established the set of indicators discussed in

section 7.3 below. Some of these indicators are not true outcome indicators, but are
safety and quality or process indicators. We were not able to obtain data for all of
these indicators.

26 In the early stages of the review when indicators were being selected, our Clinical Reference
Group comprised Professor Bruce Barraclough, Professor Clifford Hughes, Dr Michael Nicholl,
Professor Ron Penny and Dr Hunter Watt. A number of other clinicians later joined the Clinical
Reference Group and many other clinicians in study hospitals were consulted as part of this
process.

40 IPART Case study 10 - Hysterectomy

7 Outcome, safety and quality indicators

7.2 Analysing indicators and risk-adjusting for patient characteristics

It’s important to recognise that hospitals’ performance against many outcome

indicators is not simple to interpret and, when considered in isolation, can be
misleading. Therefore, this performance needs to be analysed within the appropriate
context.

In addition, hospitals treat patients with different mixes of illnesses, which can
influence the likelihood of adverse outcomes at the hospitals. To make meaningful
and fair comparisons of the performance of the study hospitals on some outcome
indicators, the analyses were risk-adjusted for factors outside the control of the
hospitals (ie, differences in patient characteristics – see Box 7.1).

Appendix B provides further details for the risk-adjusted indicator provided by NSW
Health, including the data sources used, the relevant time period for the data and the
adjustment factors applied.

Box 7.1 How data on indicators was risk-adjusted for differences in patient
characteristics

To meaningfully compare the performance of the study hospitals on some outcome indicators,
the hospitals’ data on these indicators needed to be risk-adjusted to account for differences in
patient characteristics that can influence the likelihood of adverse outcomes. In particular, NSW
Health adjusted data on mortality rates for the following patient characteristics:
 age
 sex
 comorbidity, and
 socio-economic status.a

To adjust for comorbidity, NSW Health used the Charlson index. This index simplifies the wide
range of comorbidities that may affect patients. It groups clinical conditions together (using
ICD 10), and assigns numerical weights (eg, 1, 2, 3) to them, based on the risk of dying
associated with the condition.b Adding together the numerical weights for a patient’s
comorbidities determines the patient’s combined Charlson index score, and therefore the
severity of their comorbidities

To make these adjustments, NSW Health used logistic regression in SAS 9.2. Where there were
sufficient numbers, it took repeated measures for the same person into account using multi-
level modelling. Where the number of events was too low to allow the above adjustment to be
carried out in full, the degree of adjustment was reduced and this was noted for each indicator.

a The ABS Index of Relative Socio-Economic Disadvantage (IRSD) was used to estimate socio-economic status. The
IRSD was assigned at Local Government Area level and grouped into quintiles from least disadvantaged to most
disadvantaged for analysis.
b Australian Institute of Health and Welfare, D Ben-Tovim, R Woodman, J Harrison, S Pointer, P Hakendorf & G Henley,
Measuring and reporting mortality in hospital patients, March 2009, p 18, (http://www.aihw.gov.au/publications/hse/hse-
69-10729/hse-69-10729.pdf).

Case study 10 - Hysterectomy IPART 41

7 Outcome, safety and quality indicators

7.3 List of clinical indicators for hysterectomy and their availability

Table 7.1 lists the 2 indicators we selected for hysterectomy. These indicators relate
to the outcomes of care.

Table 7.1 Clinical indicators for hysterectomy and data availability

No. Indicator Available?
1 30-day post-procedural infection rate No – data not available from NSW Health
2 30-day mortality rate (risk adjusted) Yes – data provided by NSW Health (see
Appendix B)

7.3.1 30-day mortality rate

NSW Health calculated and provided 30-day mortality rates to compare hospital
outcomes for hysterectomy patients. Box 7.2 explains the method used by NSW
Health to calculate mortality and survival rates.

However, the analysis found that there were fewer than 5 deaths following
hysterectomies in the 5 hospitals over the 3-year period 2005/06 to 2007/08. The
number of deaths was too small to allow comparisons between the hospitals.

42 IPART Case study 10 - Hysterectomy

7 Outcome, safety and quality indicators

Box 7.2 Calculating mortality rates

The NSW Department of Health’s Centre for Epidemiology and Research calculated risk-
adjusted odds ratios for mortality for patients treated in each study hospital in the
hysterectomy case study area, using the methodology outlined below. We note that the NSW
Department of Health does not usually undertake this type of analysis.

Data sources

The analysis for mortality and survival, apart from in-hospital mortality, was carried out using
linked records of the NSW Admitted Patient Data Collection (APDC) and NSW Registry of Births,
Deaths and Marriages death registration data. The analysis for in-hospital mortality was carried
out using linked records of the APDC. In-hospital deaths and deaths from all causes were
included for all relevant indicators.

Case-based analysis

As one person may have more than one admission for a specified condition, the analyses were
‘case-based’, where a case represents a hospital admission for a specified condition. This
means that, for example, if a person died after 2 hospital admissions for a specified condition
and the death occurred within the period specified by the indicator, then the case and
therefore the death would be counted twice.

Adjusting for risk and comparing hospitals

Indicators were adjusted for patient age, sex, comorbidity and socio-economic status as
described in Box 7.1.

Hospitals that were not significantly different in the adjustment model (at p<0.05) were
grouped.

Case study 10 - Hysterectomy IPART 43

7 Outcome, safety and quality indicators

44 IPART Case study 10 - Hysterectomy

7 Outcome, safety and quality indicators

Appendices

Case study 10 - Hysterectomy IPART 45

7 Outcome, safety and quality indicators

46 IPART Case study 10 - Hysterectomy

A List of full recommendations from main report

Consistency of DRG groupings

Our recommendations in this area are mainly aimed at making users of hospital data
aware of some of the limitations of using DRG groupings for hospital comparisons in
certain clinical areas.
1. That users of hospital cost and outcome data note that DRGs may contain a range of
patient types with varying clinical resource requirements, costs of care and expected
clinical outcomes. Therefore DRGs may not always provide the optimal basis for
comparing costs and outcomes among hospitals.
2. In light of Recommendation 1, that the NSW Department of Health, and other health
research bodies at both the state and national level, consider whether DRGs are a
suitable basis for determining funding and comparing performance among hospitals
(for various different types of hospital activity). Where they are not suitable, continue
research to develop better approaches for these areas.

Consistency of patient numbers

Our recommendations on patient numbers are aimed at making users of hospital

data aware of differences in patient counting practices and patient datasets between
hospitals that can affect hospital comparisons, to improve consistency of patient
counting practices between hospitals and lead to better integration of patient
datasets.
3. That users of hospital data note that there are differences in practices relating to
counting of patients that can affect hospital patient numbers and average cost
comparisons eg, counting differences relating to admission status, billing status,
location of care and collaborative care arrangements.
4. In light of Recommendation 3, that NSW Health clarifies and standardises
administrative procedures including guidelines for recording of non-inpatients of
various types, as well as ‘collaborative care’ patients.
5. That NSW Health considers ways of better integrating patient information held locally
by hospital clinical units (such as eye clinics and cardiac catheter labs) with the HIE
data set.

Case study 10 - Hysterectomy IPART 47

A List of full recommendations from main report

Consistency of lengths of stay

Our recommendations aim to improve consistency between hospitals on length of

stay measures, and to make users of hospital data aware of the limitations of
measures based on ‘acute episodes’.
6. That NSW Health monitors hospital practices relating to the classification of episodes
into care types and type-changing practices (eg, timing of type changes from acute to
rehabilitation care) and provide clear and consistent guidelines to hospitals, so
episode measures are more consistent among hospitals.
7. That users of hospital data note that 'acute episodes' often only represent a part of a
patient's hospital stay. Therefore, comparisons among hospitals using acute length of
stay measures or acute costs may produce misleading results. This is particularly
important for conditions that involve both acute and sub-acute care and/or transfers
between facilities.

Coding

We have made recommendations aimed at improving the quality of medical records

documentation and clinical coding in hospitals to both improve the quality of data
for clinical research as well as to more accurately reflect casemix complexity.
8. That NSW Health should continue to improve the quality of medical record
documentation and the accuracy and consistency of coding.
9. That hospitals should encourage consistent education on coding and facilitate
communication between clinical staff and coders regarding both the coding process
and the documentation required to code common clinical conditions, diagnoses or
complications, such as AMI, angina and chest pain.
10. Where pathology test information can be readily extracted (eg, Cerner sites), that
systems be developed so this information can be used to validate coding and support
work on variation in clinical practice and measuring clinical quality.
11. That NSW Health considers undertaking further analysis to identify pathology or
imaging tests that can be used to help target audits of coding and support work on
variation in clinical practice and measuring clinical quality – such as identifying types
of pathology tests that correspond closely with diagnosis coding.

48 IPART Case study 10 - Hysterectomy

A List of full recommendations from main report

Clinical costing

Our recommendations are aimed at improving the quality and consistency of clinical
costing data, and helping to ensure that quality costing data and clinical inputs to the
costing process (such as data from prosthesis, pathology and imaging systems) can
be used to inform hospital management about resource use, and clinicians about
clinical practice.
12. That the NSW Department of Health works with the area health services and hospitals
to apply a consistent set of rules for clinical costing covering cost centres and IFRACs
so that data are consistent and comparable between the hospitals.
13. That NSW Health regularly audits the accuracy of cost centres and IFRACs used for
clinical costing.
14. That NSW Health uses standard clinical data feeds (actual patient data) for clinical
costing where this is feasible and useful.
15. That the data used for clinical costing purposes be available to hospitals and clinicians
so they can undertake comparative analysis on clinical practices and performance.

Medical staff costs

Given our finding that there was a lack of consistency in the treatment of medical
staff costs and the difficulty this created in estimating medical staff costs for our
case study areas, we recommend:
16. That further work be undertaken to strengthen the quality and consistency of
available information on medical staff costs.

Prosthesis costs

Our recommendations on prosthesis costs are aimed at improving prosthesis

purchasing and making cost savings in this area. These should be considered in
conjunction with our recommendation that clinical experts should review the
appropriateness of clinical variation in prosthesis use and address this variation (see
Recommendation 31).
17. That NSW Health notes the variation in prostheses use among the study hospitals
including:
– drug-eluting stents versus bare metal stents
– single chamber pacemakers versus dual chamber pacemakers
– different types of components for hip replacement procedures.
18. That NSW Health notes the range of approaches to prosthesis controls and the
variation in prices currently paid for prostheses, including for exactly the same
models.

Case study 10 - Hysterectomy IPART 49

A List of full recommendations from main report

19. That NSW Health facilitates sharing of information on purchase prices for prostheses
to assist price negotiations with suppliers.
20. That NSW Health optimises prosthesis cost savings through tenders, supplier price
agreements and controlled approaches to prosthesis purchasing, noting that clinical
consultation and cooperation is essential as is retaining some flexibility to allow for
special orders when clinically indicated.

Imaging and pathology costs

Our recommendations are aimed at encouraging better use of imaging and

pathology data, and consideration of whether there should be standard treatment of
imaging and pathology within clinical costing and whether internal charges should
reflect actual costs. These recommendations should be considered in conjunction
with our clinical case studies, which include comparisons of imaging use, and
Recommendation 31, relating to clinical variation in imaging use for diagnosing
appendicitis.
21/25 That NSW Health notes that imaging and pathology data can be used to monitor
changes in imaging use and inform clinical practice, and that:
– All hospitals obtain detailed reports from pathology and imaging services on their
test ordering patterns, including the number of tests by major test type and the
cost of these tests.
– Hospitals routinely provide data to heads of clinical units to help inform them on
resource use and provision of care to improve patient outcomes and discuss
trends at management meetings – for example, summary reports that include
both the number of tests by test type, and the value (or preferably cost) of these
tests.
– NSW Health develops reports comparing the use of imaging and pathology tests
for clinical groupings and circulates these to area health services and hospitals.
22. That NSW Health considers whether, for clinical costing purposes, it is appropriate for
hospitals and area health services to base the value of imaging tests on the MBS rate
for these tests and, if so, what standard percentage of this rate is appropriate for use
by all hospitals given the actual costs of providing the test.
23. That NSW Health seeks to obtain detailed information from the pathology services on
the number and type of tests and the actual cost of undertaking a range of typical
tests for future comparisons of pathology costs.
24. That NSW Health addresses issues that prevent the actual costs associated with
specific pathology tests and ordering patterns being disclosed by pathology services.
26. That NSW Health considers whether the detailed cost estimates that pathology
services prepare as part of the benchmarking pathology project could be used for
more accurate pricing between pathology services and hospitals, to enable clinicians
to consider the actual cost of their clinical decisions.

50 IPART Case study 10 - Hysterectomy

A List of full recommendations from main report

Operating theatre costs

Our recommendations in relation to operating theatres aim to facilitate

improvements in theatre management arrangements, and the quality and consistency
of theatre data.
27. That NSW Health notes the differences in approaches to theatre management among
hospitals and consider if there is scope to share information about how the better
theatre arrangements are organised.
28. That NSW Health notes the issues regarding theatre data and work with the hospitals
to improve the completeness of datasheets and apply a consistent set of rules for
recording operating theatre times.
29. That NSW Health considers routine auditing of the quality of data on returns to
theatre and considers the best way for achieving accuracy and consistency in this
indicator.

Pharmacy costs

As we were not able to undertake a detailed comparison of pharmacy services and

costs, our recommendations focus on encouraging further analysis in this area.
30. That NSW Health:
– Notes the wide variation in the proportion of drugs dispensed versus held on
imprest across the study hospitals.
– Monitors the value of expired pharmacy stock and compares this among hospitals.
– Considers standardised guidelines for the return of unused medication, principally
to ensure patient safety but also to minimise wastage and reduce costs.
– Considers whether antimicrobial stewardship programs should be implemented at
the major hospitals where such programs are not currently in place. The purpose
of these programs would be to help prevent antimicrobial resistance and reduce
costs by preventing inappropriate use of antimicrobials.

Case study 10 - Hysterectomy IPART 51

A List of full recommendations from main report

Configurations of care – Review of clinical variations during Stages 5 and 6 of the

wider NSW Department of Health study

Our case studies identified a number of differences in the way care is provided
among study hospitals in specific clinical areas. We recommend that clinical experts
consider these clinical differences or clinical issues as part of Stages 5 and 6 of the
wider health study. This recommendation should be dealt with in conjunction with
Recommendation 36, relating to variation in indicators of safety, quality and
outcomes.
31. That NSW Health arranges for appropriate clinical expert groups to consider the
following clinical issues identified in our case studies; and that where appropriate,
NSW Health and the expert groups take steps to address clinical differences.
– Hip joint replacement:
o Note that separation of planned and emergency cases may reduce lengths of
stay for planned (arthritis) cases.
o Address the variation in the selection of hip prosthesis components (including
press fit, cementless hip stems versus cemented hip stems and ceramic femoral
heads versus metal femoral heads) among study hospitals.
– Major chest procedure:
o Note the different clinical pathways and high day of surgery admission rates for
thoracic surgery patients at RPAH compared with other study hospitals.
o Consider whether aspects of the model of care at RPAH are suitable to be used
in other hospitals.
– Breast surgery:
o Note the early discharge models at RNSH for breast surgery patients having
mastectomies and
o Consider whether such models should be followed more widely in NSW
hospitals and the types of patient cases they should be used for (eg, simpler,
unilateral cases or younger patients).
– Cholecystectomy:
o Note the variation in the proportion of patients with cholelithiasis or
cholecystitis who are operated on acutely as emergency admissions.
o Consider whether this variation has significant quality of care implications.
o Consider the relative costs and benefits of an emergency surgical services team
model for ensuring early diagnosis and treatment of conditions like
cholecystectomy and whether it should be more widely applied.
o Note that costing of cholecystectomy should take into account the costs of
prior related emergency department attendances. A similar approach should
be adopted for other clinical conditions that are likely to involve multiple prior
emergency department attendances.

52 IPART Case study 10 - Hysterectomy

A List of full recommendations from main report

o Consider the relative costs and benefits of cholecystectomies with and without
the use of fluoroscopy.
– Appendicectomy
o Note the variation in the use of imaging tests for diagnosing appendicitis.
o Consider establishing standard protocols for diagnosing appendicitis,
indicating when it is appropriate to use CT scans, MRIs and ultrasounds.
o As part of establishing standard protocols for diagnosing appendicitis, consider
whether CT scans, MRIs and ultrasounds should only be used for certain patient
groups (eg, older patients who are more likely to be suffering from other
conditions with symptoms similar to appendicitis).
o Consider the relative costs and benefits of laparoscopic versus open surgery for
appendicitis.
– Stroke
o Consider ways to reduce the proportion of stroke patients coded with a
principal diagnosis of 'stroke, not specified as haemorrhage or infarction'
(ICD10 code I64).
o Consider developing consistent guidelines for the administration of tPA.
o Consider including tPA administration as a procedure in coding standards.
o Consider ways to improve transfers of suspected stroke patients to stroke units
with minimum delay, including consultation with the Ambulance Service and
Emergency Departments.
o Investigate whether it is useful and possible to combine Ambulance Service
data on response time with hospital patient data to monitor time from call to
ambulance to arrival at an appropriate hospital.
o Consider the costs and benefits of providing more rehabilitation care in the
home.
o Pursue the collection of the data on outcome indicators from the National
Stroke Research Institute.
– Cardiology – Stents, Pacemakers and Defibrillators:
o Address the variation in the use of drug-eluting stents versus bare metal stents
among study hospitals.
o Address the variation in the types of pacemakers used among study hospitals.
o Investigate whether there are differences in treatment procedures, or waiting
times between presentation and procedure, for patients who present to
hospitals without a 24 hour cardiac catheter laboratory, compared to patients
who present to hospitals with a 24 hour cardiac catheter laboratory, and
whether any differences in procedure or waiting times have implications for
clinical outcomes.

Case study 10 - Hysterectomy IPART 53

A List of full recommendations from main report

o Consider ways of better integrating information held in cardiac catheter

laboratories with the HIE data set.
– Tracheostomy or ventilation greater than 95 hours:
o Note that at BLH, clinicians tend to perform surgical tracheostomies, whereas at
the other hospitals, these are usually performed percutaneously.
– Cataract/lens procedure:
o Assess the costs and benefits of toric lenses and develop guidelines for their
use in public hospitals.
– Hysterectomy:
o That any future studies of hysterectomy compare the costs and outcomes for
hysterectomies with the costs and outcomes of other procedures such as
endometrial ablation and uterine artery embolisation.

Improving outcome, safety and quality indicators

While current Commonwealth and State initiatives will improve outcomes data, we
have made recommendations that will assist this process.
32. That NSW Health enhances understanding and use of mortality, survival, unplanned
readmission and wound infection indicators and their risk adjustment by:
– continuing to contribute to the development of ACSQHC’s safety and quality
standards for these indicators
– refining the methodology used for standardising or risk-adjusting these indicators
– continuing to consult with clinicians regarding the agreed presentation of
mortality, survival unplanned readmission and wound infection information
– reporting this information on a more routine and regular basis consistent with
ACSQHC data sets.
33. That NSW Health encourages hospitals to put in place systems to facilitate accurate
coding of comorbidities and ensures that coding practices are consistent across
hospitals.
34. That NSW Health works with ACSQHC to negotiate more streamlined arrangements
for access to data held by third parties (such as clinical registries) for clinical analysis,
and makes these data available to hospitals and clinicians.
35. That NSW Health explores the possibility of providing outcomes information to
clinicians in a more systematic way as an aid to clinical improvement and a key
indicator of performance.

54 IPART Case study 10 - Hysterectomy

A List of full recommendations from main report

Indicators of safety, quality or outcomes, - review of clinical variations during stages

5 and 6 of the wider NSW Department of Health project

We have also made a number of findings relating to variations in indicators of safety,

quality or outcomes. Where we have observed apparent differences among
hospitals, these should be considered by clinical expert groups in completing stages 5
and 6 of the Department of Health’s wider project. These differences should be
considered in conjunction with differences in clinical practice (Recommendation 31).
36. That clinical expert groups consider the following clinical issues; and where
appropriate, NSW Health and clinical expert groups take steps to address clinical
variations as part of Stages 5 and 6 of the broader NSW Health review:
– Review the variations in outcome, safety and quality indicators among study
hospitals, including their:
o unplanned readmission rates
o wound infection rates for selected surgical procedures.
– Review the variation in mortality and survival rates for all major chest surgery
patients and consider whether to recommend changes to clinical practice or
conduct further investigation involving:
o a larger sample of hospitals, and
o more detailed analyses for ‘like patients’ (ie, lung cancer, infection-related
abscess/pyothorax and collapsed/punctured lung patients).
– Review the variation in the following clinical indicators for hip joint replacement
surgery at the study hospitals:
o wound infection rates
o unplanned readmission rates.
– Review the variation in wound infection rates for appendicectomy and
cholecystectomy surgery at the study hospitals.
– Note the variation in the following clinical indicators relating to obstetric delivery:
o caesarean section rates for ‘selected primipara’
o vaginal delivery rates following primary caesarean section
o caesarean section rates after induction of labour for ‘selected primipara’
o repeat caesarean section rates
o significant tear rates
and monitor changes arising from the implementation of the NSW Health policy
directive, Maternity – Towards Normal Birth in NSW, to determine whether this
policy effectively addresses the variation.

Case study 10 - Hysterectomy IPART 55

A List of full recommendations from main report

Additional outcome indicators

We made recommendations to consider the costs and benefits of collecting data for
the following areas where indicators are not commonly used.
37. That NSW Health considers the costs and benefits of collecting data and monitoring
performance against the following indicators:
– warfarin management
– visual outcomes for patients undergoing lens procedures.

We also made a recommendation to develop a set of standard indicators for

measuring care and/or outcomes in ICUs.
38. That NSW Health undertakes further work to develop a set of standard indicators for
measuring care and/or outcomes in ICUs.

Time Out audits

Finally, we made a recommendation to improve consistency in the number of cases

audited as part of the Time Out process relative to the number of separations.
39. That NSW Health specifies the number or proportion of patient cases that should be
audited as part of the Time Out process.

Next steps - wider application of this study

40. That NSW Health refines and develops useful aspects of this study for application
more widely to other hospitals, other health settings and other clinical conditions.

56 IPART Case study 10 - Hysterectomy

B Risk-adjusted indicators provided by NSW Health

Table B.1 indicates the data sources and risk adjustment factors used for the risk-
adjusted indicator provided by NSW Health.

Table B.1 Risk-adjusted indicator provided by NSW Health

No. Indicator Data Source Numerator & Risk-adjustment

denominator
1 30-day Linked records of the APDC and Numerator-Number of Age, sex,
mortality RBDM death registration data. deaths comorbidity
rate APDC records for 2005-06 to Denominator- Number of (Charlson index)
2007-08 and deaths to 30/9/2008 cases for DRG N04Z. and socio-
economic status.
Note: APDC - NSW Admitted Patient Data Collection. RBDM - Registry of Births, Deaths and Marriages. A case represents
a hospital admission for a specified condition. DRG - Diagnosis Related Group v 5.1. Charlson index (see Box 7.1).
Socioeconomic status (see Box 7.1).
Source: NSW Health.

Case study 10 - Hysterectomy IPART 57

Glossary

Term Abb. Definition

95% confidence A statistical term describing a range of values within which
interval we are 95% certain that the true population value lies.
Abdominal A procedure where the uterus is removed via direct open
hysterectomy surgery on the lower abdomen.
Activity-based ABF Funding that is based on the projected amount and type of
funding work of a facility, where standard prices are set for similar
work undertaken. This has also been referred to as casemix
or episode funding.
Acute care Clinical services provided to admitted or non-admitted
patients, including managing labour, curing illness or
treating injury, performing surgery, relieving symptoms
and/or reducing the severity of illness or injury, and
performing diagnostic and therapeutic procedures. Most
patients have acute or temporary ailments. The average
length of stay is relatively short.
Adenomyosis A condition where endometrial tissue that normally lines the
uterus is present in the muscular walls of the uterus.
Admission The process by which a person commences a period of
residential care in a health facility.
Admitted Patient APDC A database that covers all inpatient separations (discharges,
Data Collection transfers and deaths) from all Public (including Psychiatric),
Private, and Repatriation Hospitals, Private Day Procedures
Centres and Public Nursing Homes in NSW.
Agency for Clinical ACI A board-governed statutory health corporation that reports
Innovation to the NSW Minister for Health and the Director-General of
NSW Health.
Appendicectomy Surgical excision of the patient's appendix.
Assistant In Nursing AIN An employee that is not a registered nurse, enrolled nurse or
trainee nurse, who assists the Enrolled Nurses and Registered
Nurses by providing basic nursing care, working within a
plan of care under the supervision and direction of a
Registered Nurse.
Australian Council on An independent organisation dedicated to improving the
Healthcare Standards quality of health care through performance reviews,
assessment and accreditation.

58 IPART Case study 10 - Hysterectomy

Glossary

Term Abb. Definition

Average length of ALOS The average number of days each admitted patient stays in
stay hospital. This is calculated by dividing the total number of
occupied bed days for the period by the number of actual
separations in the period.
Bankstown- BLH One of the study hospitals included in the review.
Lidcombe Hospital
Bureau of Health BHI An independent, board-governed organisation established
Information by the NSW Government to be the leading source of
information on the performance of the public health system
in NSW.
Carcinoma in situ The presence of neoplastic cells or abnormal tissue that has
not invaded surrounding tissues.
Casemix The range and types of episodes of care of patients’ (the mix
of cases) treated by a hospital. This provides a way of
describing and comparing hospitals and other services for
planning and managing health care. Casemix classifications
put patients into DRG s with similar conditions that use
similar health-care resources, so that the activity and cost-
efficiency of different hospitals can be compared.
Casemix funding See Activity-based funding.
Cholecystectomy Excision of the gallbladder.
Clinical Excellence CEC A board-governed statutory health corporation with the CEO
Commission reporting directly to the NSW Minister for Health. A key role
of the Clinical Excellence Commission is building capacity for
quality and safety improvement in Health Services.
Clinical Nurse CNS A Registered Nurse/Midwife who applies a high level of
Specialist clinical nursing knowledge, experience and skills in
providing complex nursing/midwifery care directed towards
a specific area of practice, a defined population or defined
service area, with minimum direct supervision.
Comorbidity When a person has two or more health problems at the same
time.
Computed CT scan A non-invasive medical imaging method using X-rays and
tomography computer processing.
Diagnosis Related DRG A system used to classify hospital admissions into groups
Group with similar clinical conditions (related diagnoses) and
similar resource usage (hospital services). There are
approximately 500 coding classes. In Australian acute
hospitals, Australian refined DRGs are used (AR-DRGs). The
classification categorises episodes into groups with similar
conditions and similar usage of hospital resources, using
information in the hospital morbidity record such as the
diagnoses, procedures and demographic characteristics.
Endometrial ablation Removal of the endometrium (mucus membrane lining) of
the uterus.
Endometriosis A condition where the mucous membrane lining the uterus
(endometrium) occurs outside the uterus.

Case study 10 - Hysterectomy IPART 59

Glossary

Term Abb. Definition

Enrolled Nurse EN A person holding an Enrolled Nurse qualification who works
under the supervision of a Registered Nurse to provide
nursing care for patients in hospitals, nursing homes and a
variety of other health care organisations.
Episode funding See Activity-based funding.
Fibroids Non-cancerous, muscular tumours that grow in the wall of
the uterus.
Fluoroscopy An imaging technique that provides real-time moving
images of the internal structures of a patient through the use
of a fluoroscope.
Gosford Hospital GH One of the study hospitals included in the review.
Health Information HIE A database maintained by the NSW Department of Health
Exchange that contains a range of financial, patient and clinical
information from hospitals and area health services.
High dependency HDU An area or environment in a hospital that provides a higher
unit level of critical care and monitoring than is provided in a
general ward, but a lower level of care provided by an
intensive-care unit.
Hysterectomy Surgical removal of the uterus.
Independent Pricing IPART The independent economic regulator for NSW that is
and Regulatory undertaking this hospital study.
Tribunal of NSW
Inpatient fraction IFRAC A measure used in casemix costing. The proportion of total
(or operating) costs that are attributed to admitted patients.
Intensive care unit ICU An area or environment in a hospital that provides the
highest level of critical care and monitoring.
John Hunter Hospital JHH One of the study hospitals included in the review.
Laparoscopic A surgical procedure where ligamentous attachments of the
assisted vaginal uterus are severed using laparoscopic tool and the uterus is
hysterectomy removed through the vagina.
Laparoscopic Removal of the uterus via small incisions through the
hysterectomy abdomen.
Length of stay 1 LOS1 LOS1 is the episode length of stay in study hospital, ie, from
the start of the episode to the end of the episode of care.
Length of stay 2 LOS2 LOS2 is the total length of stay in study hospital, ie, from
admission to discharge at the study hospital.
Length of stay 3 LOS3 LOS3 is the total length of stay in study hospital plus up to 2
other hospitals - one transfer in and one transfer out.
Medical resonance MRI A medical imaging technique most commonly used in
imaging radiology to visualise detailed internal structures of the body
using a magnetic field.
Medicare Benefits MBS A listing of the Medicare services subsidised by the
Schedule Australian government.

60 IPART Case study 10 - Hysterectomy

Glossary

Term Abb. Definition

National Hospital NHCDC The NHCDC contains component costs per DRG based on
Cost Data Collection patient-costed and cost-modelled information. The NHCDC
enables DRG Cost Weights and average costs for DRGs for
acute in-patients to be produced.
NSW Health The broad term encompassing operational and other
structures including the NSW Department of Health, Area
Health Services, the Agency for Clinical Innovation, the
Clinical Excellence Commission and a range of clinical
taskforces.
Open Surgery An invasive medical procedure where an incision is required
for direct surgical access to the organs.
Principal referral Hospital within peer group (principal referral hospitals 1b)
hospital classified as an acute hospital, treating 25,000 or more acute
casemix weighted separations per annum, with an average
cost weight greater than 1 and 1 or fewer specialty services.
Principal tertiary Hospital within peer group (principal referral hospitals 1a)
referral hospital classified as an acute hospital, treating 25,000 or more acute
casemix weighted separations per annum, with an average
cost weight greater than 1 and having more than 1 specialty
service.
Prophylaxis Disease prevention, also called preventive treatment.
Registered nurse RN A qualified nurse who provides care for patients in a variety
of healthcare settings. These include public and private
hospitals, community and home-based services, nursing
homes and industry.
Royal North Shore RNSH One of the study hospitals included in the review.
Hospital
Royal Prince Alfred RPAH One of the study hospitals included in the review.
Hospital
Tracheostomy A surgical procedure to cut an opening into the trachea
(windpipe) so that a tube can be inserted into the opening to
assist breathing.
Uterine artery A non-invasive form of treating fibroids via blocking arteries
embolisation feeding the fibroids.
Vaginal A hysterectomy procedure performed through the vagina.
hysterectomy