A hospital’s risk A 30-day time period is used for the six ...€¦ · Hip fracture surgery 100...
32
Performance Profile: bhi.nsw.gov.au July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15 Acute myocardial infarction Ischaemic stroke 55 2.60 Congestive heart failure 106 0.89 Pneumonia 162 1.27 Chronic obstructive pulmonary disease 249 1.08 Hip fracture surgery 100 1.85 Total hip replacement 82 0.33 Total knee replacement 125 1.04 Murwillumbah District Hospital A hospital’s risk-standardised readmission ratio (RSRR) is the ‘observed’ number of readmissions that occurred among its patient cohort divided by the ‘expected’ number of readmissions among its patients 1 . For this report, readmission is defined as a return to acute care 2 . Funnel plots with 95% and 99.8% control limits around the NSW ratio are used to interpret the ratios and identify outlier hospitals – those with ‘special cause’ variation that may warrant further investigation. The RSRR does not enable direct comparisons between hospitals. It assesses each hospital’s results given its particular case mix. Slightly different approaches are used for the conditions. A 30-day time period is used for the six acute conditions while a 60-day period is used for the elective surgeries. The analyses focused on acute conditions only consider readmission episodes that are classed as acute emergencies while analyses for the elective surgeries also include some ‘planned’ readmissions, such as planned returns to theatre for wound wash-outs. RSRRs do not distinguish readmissions that are avoidable from those that are a reflection of the natural course of illness. 1 Statistically significant result No significant difference <50 cases 95% control limits Higher than expected No different than expected Readmission this period: Lower than expected * Index cases exclude those with <30 days follow up information. Note: More than 5% of total patients underwent hip fracture surgery at another hospital. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 < 50 index hospitalisations, results not shown
Transcript of A hospital’s risk A 30-day time period is used for the six ...€¦ · Hip fracture surgery 100...
Performance Profile: bhi.nsw.gov.au
July 00 –
June 03
July 03 –
June 06
July 06 –
June 09
July 09 –
June 12
July 12 –
June 15
Acute myocardial infarction
Ischaemic stroke 55 2.60
Congestive heart failure 106 0.89
Pneumonia 162 1.27
Chronic obstructive
pulmonary disease249 1.08
Hip fracture surgery 100 1.85
Total hip replacement 82 0.33
Total knee replacement 125 1.04
Murwillumbah District Hospital
A hospital’s risk-standardised readmission ratio (RSRR) is the
‘observed’ number of readmissions that occurred among its
patient cohort divided by the ‘expected’ number of readmissions
among its patients1. For this report, readmission is defined as a
return to acute care2.
Funnel plots with 95% and 99.8% control limits around the NSW
ratio are used to interpret the ratios and identify outlier hospitals –
those with ‘special cause’ variation that may warrant further
investigation. The RSRR does not enable direct comparisons
between hospitals. It assesses each hospital’s results given its
particular case mix.
Slightly different approaches are used for the conditions.
A 30-day time period is used for the six acute conditions while
a 60-day period is used for the elective surgeries. The analyses
focused on acute conditions only consider readmission episodes
that are classed as acute emergencies while analyses for the
elective surgeries also include some ‘planned’ readmissions,
such as planned returns to theatre for wound wash-outs.
RSRRs do not distinguish readmissions that are avoidable from
those that are a reflection of the natural course of illness.
1
Statistically significant result
No significant difference <50 cases
95% control limits
Higher than expected
No different than expected
Readmission this period: Lower than expected
* Index cases exclude those with <30 days follow up information.
Note: More than 5% of total patients underwent hip fracture surgery at another hospital.
0.0 0.5 1.0 1.5 2.0 2.5 3.0
< 50 index hospitalisations,
results not shown
Performance Profile: bhi.nsw.gov.aubhi.nsw.gov.au 2
If a hospital’s RSRR lies on the grey bar, its
readmissions are within the range of values
expected for an in control NSW hospital of
similar size
The length of the bar for each condition reflects
the tolerance for variation. It is wider for hospitals
admitting a small number of patients
Readmissions are
lower than expected
Readmissions are
higher than expected
0
1
2
3
0 20 40 60 80 100 120 140 160 180 200 220
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Expected number of returns to acute care (readmissions) within 30 days
Hospital with higher
readmissions
Hospital with lower
readmissions
Hospital within the range of values expected for
an in control NSW hospital (inside the funnel)
Hospital with higher readmissions
(between 95% and 99.8% control limits)
Greater tolerance for variation
for hospitals with fewer
expected readmissions
Reflects patient volume and
case mix at the hospital
99.8% limits95% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Acute myocardial infarction
3
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital 4
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total index cases for ischaemic stroke 60 14,471
Average length of stay (days) 5.5 8.3
Patients transferred in from acute care in another hospital 37 1,943
Discharge destination:
Home 12 7,760
Other 48 6,711
5
*Age was not a significant factor in the final model of 30-day readmission following hospitalisation for ischaemic stroke.
4.0
13.3
19.9
31.7
22.9
25.0
30.8
30.0
22.4
% index cases
15–44 45–64 65–74 75–84 85+
This hospital
NSW
6.4
4.3
1.9
1.3
-1.0
-1.2
-4.8
-5.9
-6.5
-10.3
-20 -15 -10 -5 0 5 10 15 20
Fluid and electrolyte disorders
Chronic pulmonary disease
Female
Congestive heart failure
Deficiency anaemia
Liver disease
Weight loss
Renal failure
Diabetes, complicated
Cardiac arrhythmia
% difference from NSW (index cases with factor recorded)
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total readmissions following index hospitalisation for ischaemic stroke 16 1,539
Readmitted to the hospital where acute care was completed 16 1,188
Readmitted to a different hospital 0 351
Of these:
To an urban public hospital
To a regional or rural public hospital
To a private hospital
6
Distribution of reasons for returns to acute care
Number of, and reasons for, returns to acute care following hospitalisation for ischaemic stroke, by days post discharge
Same principal diagnosis Condition related to principal diagnosis Potentially related to hospital care
(relevant at any time)
Potentially related to hospital care
(time sensitive, ≤ 7 days post discharge)
Potentially related to hospital care
(time sensitive, 8–30 days post discharge)
Other conditions
17.2 10.0
25.0
22.4
25.0
6.5
6.3
14.4
43.8
29.5
0 10 20 30 40 50 60 70 80 90 100
% returns to acute care
This hospital
NSW
2 2
4
1
4
2
10
2
4
6
8
10
12
1–7 days 8–14 days 15–21 days 22–30 days
Num
ber
of re
turn
s t
o a
cute
care
Days post discharge
Performance Profile: bhi.nsw.gov.au
19.610.2
16.7
0
1
2
3
4
0 50 100 150 200 250 300 350 400 450
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio
(Ob
serv
ed
/Exp
ecte
d)
Expected number of returns to acute care (readmissions) within 30 days
Murwillumbah District Hospital
Hospital-specific RSRRs report the ratio of actual or ‘observed’
number of returns to acute care to the ‘expected’ number of
returns. A competing risk regression model draws on the NSW
patient population’s characteristics and outcomes to estimate the
expected number of returns for each hospital, given the
characteristics of its patients.
An RSRR less than 1.0 indicates lower-than-expected returns to
acute care, and a ratio higher than 1.0 indicates higher-than-
expected returns. Small deviations from 1.0 are not considered to
be meaningful. Funnel plots with 95% and 99.8% control limits
around the NSW ratio are used to identify outliers.
7
99.8% limits95.0% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
In order to make fair comparisons, a number of risk adjustments
are made to readmission data. These take into account patient
factors that influence the likelihood of returning to acute care
within 30 days. The table below illustrates the effect of statistical
adjustments on this hospital’s results.
Unadjusted ratio
Age and sex
standardised ratio
Risk-standardised
readmission ratio
The RSRR is calculated on the basis of three years of data.
It takes account of differences in patient characteristics so that
assessments of hospital performance are fair. To give an
indication of results within the three-year period, the figure below
shows the RSRR result for July 2012 – June 2015 alongside
differences between this hospital and the NSW result for annual
unadjusted readmission rates.
Lower than
expected
No different
than expected
Higher than
expected
RSRR:
Unadjusted
readmission
rate percentage
point difference
from NSW result
The extent to which comorbidities are coded in the patient record
may affect risk adjustment. Therefore the ‘depth of coding’10 has
been assessed across NSW hospitals. In July 2009 – June 2012,
the average depth of coding was 7.4 diagnoses in this hospital
and 6.3 in NSW public hospitals; and in July 2012 – June 2015,
there were 7.9 diagnoses in this hospital and 7.0 in NSW public
hospitals.
RSRR
July 2012 –
June 2015
July 12 –
June 13
July 13 –
June 14
July 14 –
June 15
Lower than
expected
No different
than expected
Higher than
expectedRatio:
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
1. Data refer to patients who were discharged from this hospital, between July 2012 and June 2015, following an acute hospitalisation with ischaemic stroke as principal diagnosis (ICD-10-AM code I63).
2. Returns to acute care are to any NSW hospital in the 30 days (for acute conditions) or 60 days (for elective surgeries) following discharge, and are attributed to the last discharging hospital. For patients
whose acute hospitalisation ended in discharge home, a return to acute care involves readmission to hospital; while for patients whose acute hospitalisation ended with a 'discharge' to non-acute care, a
return involved a move back into an acute care setting regardless of whether they physically left the hospital.
3. For calculation of average length of stay, index admissions that were transferred in from, or transferred out to, another acute care hospital were excluded. Unreasonably long episodes are trimmed on the
basis of the Diagnosis Related Group (DRG) of the episode. The trim point is the third quartile plus 1.5 x the interquartile range of all in-scope episodes in each DRG.
4. Discharge destinations are based on the mode of separation of the index case. For episodes coded as 'Discharged by hospital' or 'Discharged on leave', patients are considered to be destined for their
place of usual residence. All other modes of separation are deemed to indicate a discharge destination other than a patient’s place of usual residence.
5. Age at admission date.
6. Comorbidities are identified from the hospital discharge records using the Elixhauser comorbidity set (plus dementia) with a one year look-back from the admission date of the index case. Only those
conditions that were shown to have a significant impact on readmission (P<0.05) are shown.
7. Hospitals are classified as urban and regional/rural using the geocoded address of the hospital assigned to ABS statistical areas (SA2) and the Australian remoteness index for areas.
8. Reasons for return to acute care are classified according to a draft specification made available to BHI by the Australian Institute of Health and Welfare. Principal diagnoses for the return to acute care
episode, are stratified as: the same as the index hospitalisation; related to that of the index hospitalisation (same ICD-10-AM chapter); potentially related to hospital care (i.e. complications and adverse
events) using various time horizons; and, other reasons. Percentages may not add to 100% due to rounding.
9. Results for hospitals with <1 expected readmission are not shown. Peer hospitals are identified according to the NSW Ministry of Health’s peer grouping as of April 2012.
10. The depth of coding has been defined as the average number of secondary diagnosis coded for the index cases. The one year look back method which is used for risk adjustment, to some extent
accounts for possible lower depth of coding in some hospitals.
Details of analyses are available in Spotlight on Measurement: Measuring return to acute care following discharge from hospital, 2nd edition.
Data source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health.
8
Ob
serv
ed
(unad
juste
d) ra
teE
xp
ecte
d r
ate
Ris
k–sta
nd
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Statistically significant result
0.00.20.40.60.81.01.21.41.61.82.02.22.42.62.83.03.23.43.63.84.04.24.44.64.85.05.25.4
0
4
8
12
16
20
24
28
32
36
July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total index cases for congestive heart failure 107 33,450
Average length of stay (days) 4.9 6.1
Patients transferred in from acute care in another hospital 12 3,216
Discharge destination:
Home 68 28,883
Other 39 4,567
9
*Age was a significant factor in the final model of 30-day readmission following hospitalisation for congestive heart failure.
10.9
15.0
18.3
38.3
35.3
43.0
34.1
% index cases
15-44 45–64 65–74 75–84 85+
This hospital
NSW
6.7
3.8
3.4
0.6
0.6
0.1
-0.5
-1.1
-2.0
-2.3
-4.7
-6.7
-8.9
-8.9
-9.9
-15.1
-15.4
-30 -20 -10 0 10 20 30
Female
Chronic pulmonary disease
Other neurological disorders
Metastatic cancer
Abuse drug/alcohol/psychoses
Dementia
Peptic ulcer disease, excluding bleeding
Rheumatoid arthritis/collagen
Hypothyroidism
Liver disease
Cardiac arrhythmia
Previous congestive heart failure admission
Coagulopathy
Deficiency anaemia
Renal failure
Diabetes, complicated
Fluid and electrolyte disorders
% difference from NSW (index cases with factor recorded)
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total readmissions following index hospitalisation for congestive heart failure 20 7,602
Readmitted to the hospital where acute care was completed 19 6,256
Readmitted to a different hospital 1 1,346
Of these:
To an urban public hospital 0
To a regional or rural public hospital 1
To a private hospital 0
10
Distribution of reasons for returns to acute care
Number of, and reasons for, returns to acute care following hospitalisation for congestive heart failure, by days post discharge
Same principal diagnosis Condition related to principal diagnosis Potentially related to hospital care
(relevant at any time)
Potentially related to hospital care
(time sensitive, ≤ 7 days post discharge)
Potentially related to hospital care
(time sensitive, 8–30 days post discharge)
Other conditions
35.0
37.1
5.0
8.0 6.8
25.0
8.5
15.0
18.8
20.0
20.8
0 10 20 30 40 50 60 70 80 90 100
% returns to acute care
This hospital
NSW
5
21
5
3
1
2
1
0
2
4
6
8
10
12
1–7 days 8–14 days 15–21 days 22–30 days
Num
ber
of re
turn
s t
o a
cute
care
Days post discharge
Performance Profile: bhi.nsw.gov.au
8.1
-11.2 -13.8
0
1
2
3
4
0 50 100 150 200 250 300 350 400 450
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio
(Ob
serv
ed
/Exp
ecte
d)
Expected number of returns to acute care (readmissions) within 30 days
Murwillumbah District Hospital
Hospital-specific RSRRs report the ratio of actual or ‘observed’
number of returns to acute care to the ‘expected’ number of
returns. A competing risk regression model draws on the NSW
patient population’s characteristics and outcomes to estimate the
expected number of returns for each hospital, given the
characteristics of its patients.
An RSRR less than 1.0 indicates lower-than-expected returns to
acute care, and a ratio higher than 1.0 indicates higher-than-
expected returns. Small deviations from 1.0 are not considered to
be meaningful. Funnel plots with 95% and 99.8% control limits
around the NSW ratio are used to identify outliers.
11
99.8% limits95.0% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
In order to make fair comparisons, a number of risk adjustments
are made to readmission data. These take into account patient
factors that influence the likelihood of returning to acute care
within 30 days. The table below illustrates the effect of statistical
adjustments on this hospital’s results.
Unadjusted ratio
Age and sex
standardised ratio
Risk-standardised
readmission ratio
The RSRR is calculated on the basis of three years of data.
It takes account of differences in patient characteristics so that
assessments of hospital performance are fair. To give an
indication of results within the three-year period, the figure below
shows the RSRR result for July 2012 – June 2015 alongside
differences between this hospital and the NSW result for annual
unadjusted readmission rates.
Lower than
expected
No different
than expected
Higher than
expected
RSRR:
Unadjusted
readmission
rate percentage
point difference
from NSW result
The extent to which comorbidities are coded in the patient record
may affect risk adjustment. Therefore the ‘depth of coding’10 has
been assessed across NSW hospitals. In July 2009 – June 2012,
the average depth of coding was 6.3 diagnoses in this hospital
and 4.8 in NSW public hospitals; and in July 2012 – June 2015,
there were 5.4 diagnoses in this hospital and 5.9 in NSW public
hospitals.
RSRR
July 2012 –
June 2015
July 12 –
June 13
July 13 –
June 14
July 14 –
June 15
Lower than
expected
No different
than expected
Higher than
expectedRatio:
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
1. Data refer to patients who were discharged from this hospital, between July 2012 and June 2015, following an acute hospitalisation with congestive heart failure as principal diagnosis (ICD-10-AM codes
I11.0, I13.0, I13.2, I50.0, I50.1, I50.9).
2. Returns to acute care are to any NSW hospital in the 30 days (for acute conditions) or 60 days (for elective surgeries) following discharge, and are attributed to the last discharging hospital. For patients
whose acute hospitalisation ended in discharge home, a return to acute care involves readmission to hospital; while for patients whose acute hospitalisation ended with a 'discharge' to non-acute care, a
return involved a move back into an acute care setting regardless of whether they physically left the hospital.
3. For calculation of average length of stay, index admissions that were transferred in from, or transferred out to, another acute care hospital were excluded. Unreasonably long episodes are trimmed on the
basis of the Diagnosis Related Group (DRG) of the episode. The trim point is the third quartile plus 1.5 x the interquartile range of all in-scope episodes in each DRG.
4. Discharge destinations are based on the mode of separation of the index case. For episodes coded as 'Discharged by hospital' or 'Discharged on leave', patients are considered to be destined for their
place of usual residence. All other modes of separation are deemed to indicate a discharge destination other than a patient’s place of usual residence.
5. Age at admission date.
6. Comorbidities are identified from the hospital discharge records using the Elixhauser comorbidity set (plus dementia) with a one year look-back from the admission date of the index case. Only those
conditions that were shown to have a significant impact on readmission (P<0.05) are shown.
7. Hospitals are classified as urban and regional/rural using the geocoded address of the hospital assigned to ABS statistical areas (SA2) and the Australian remoteness index for areas.
8. Reasons for return to acute care are classified according to a draft specification made available to BHI by the Australian Institute of Health and Welfare. Principal diagnoses for the return to acute care
episode, are stratified as: the same as the index hospitalisation; related to that of the index hospitalisation (same ICD-10-AM chapter); potentially related to hospital care (i.e. complications and adverse
events) using various time horizons; and, other reasons. Percentages may not add to 100% due to rounding.
9. Results for hospitals with <1 expected readmission are not shown. Peer hospitals are identified according to the NSW Ministry of Health’s peer grouping as of April 2012.
10. The depth of coding has been defined as the average number of secondary diagnosis coded for the index cases. The one year look back method which is used for risk adjustment, to some extent
accounts for possible lower depth of coding in some hospitals.
Details of analyses are available in Spotlight on Measurement: Measuring return to acute care following discharge from hospital, 2nd edition.
Data source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health.
12
Ob
serv
ed
(unad
juste
d) ra
teE
xp
ecte
d r
ate
Ris
k–sta
nd
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Statistically significant result
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
0
5
10
15
20
25
30
35
40
July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total index cases for pneumonia 166 46,422
Average length of stay (days) 4.9 5.6
Patients transferred in from acute care in another hospital 11 4,505
Discharge destination:
Home 138 40,374
Other 28 6,048
13
*Age was a significant factor in the final model of 30-day readmission following hospitalisation for pneumonia.
13.3
11.5
21.1
20.2
19.9
19.7
21.7
26.2
24.1
22.4
% index cases
18–44 45–64 65–74 75–84 85+
This hospital
NSW
8.5
1.5
0.8
-0.3
-0.9
-0.9
-1.0
-1.7
-2.2
-2.7
-2.8
-3.2
-3.4
-3.6
-4.0
-4.1
-4.3
-4.6
-5.9
-10.3
-20 -15 -10 -5 0 5 10 15 20
Female
Abuse drug/alcohol/psychoses
Rheumatoid arthritis/collagen
Depression
Metastatic cancer
Congestive heart failure
Lymphoma
Liver disease
Other neurological disorders
Diabetes, complicated
Deficiency anaemia
Pulmonary circulation disorders
Solid tumour without metastasis
Chronic pulmonary disease
Renal failure
Weight loss
Hypertension
Fluid and electrolyte disorders
Diabetes, uncomplicated
Cardiac arrhythmia
% difference from NSW (index cases with factor recorded)
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total readmissions following index hospitalisation for pneumonia 26 6,543
Readmitted to the hospital where acute care was completed 19 5,304
Readmitted to a different hospital 7 1,239
Of these:
To an urban public hospital 6
To a regional or rural public hospital 1
To a private hospital 0
14
Distribution of reasons for returns to acute care
Number of, and reasons for, returns to acute care following hospitalisation for pneumonia, by days post discharge
Same principal diagnosis Condition related to principal diagnosis Potentially related to hospital care
(relevant at any time)
Potentially related to hospital care
(time sensitive, ≤ 7 days post discharge)
Potentially related to hospital care
(time sensitive, 8–30 days post discharge)
Other conditions
19.2
19.8
15.4
19.6 6.8
11.5
8.3
23.1
16.1
26.9
29.4
0 10 20 30 40 50 60 70 80 90 100
% returns to acute care
This hospital
NSW
3
1 1
1
1 1
1
1
3
1 14
2 2
3
0
1
2
3
4
5
6
7
8
9
1–7 days 8–14 days 15–21 days 22–30 days
Num
ber
of re
turn
s t
o a
cute
care
Days post discharge
Performance Profile: bhi.nsw.gov.au
-3.1
8.1
2.0
0
1
2
3
4
0 50 100 150 200 250 300 350 400 450
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio
(Ob
serv
ed
/Exp
ecte
d)
Expected number of returns to acute care (readmissions) within 30 days
Murwillumbah District Hospital
Hospital-specific RSRRs report the ratio of actual or ‘observed’
number of returns to acute care to the ‘expected’ number of
returns. A competing risk regression model draws on the NSW
patient population’s characteristics and outcomes to estimate the
expected number of returns for each hospital, given the
characteristics of its patients.
An RSRR less than 1.0 indicates lower-than-expected returns to
acute care, and a ratio higher than 1.0 indicates higher-than-
expected returns. Small deviations from 1.0 are not considered to
be meaningful. Funnel plots with 95% and 99.8% control limits
around the NSW ratio are used to identify outliers.
15
99.8% limits95.0% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
In order to make fair comparisons, a number of risk adjustments
are made to readmission data. These take into account patient
factors that influence the likelihood of returning to acute care
within 30 days. The table below illustrates the effect of statistical
adjustments on this hospital’s results.
Unadjusted ratio
Age and sex
standardised ratio
Risk-standardised
readmission ratio
The RSRR is calculated on the basis of three years of data.
It takes account of differences in patient characteristics so that
assessments of hospital performance are fair. To give an
indication of results within the three-year period, the figure below
shows the RSRR result for July 2012 – June 2015 alongside
differences between this hospital and the NSW result for annual
unadjusted readmission rates.
Lower than
expected
No different
than expected
Higher than
expected
RSRR:
Unadjusted
readmission
rate percentage
point difference
from NSW result
The extent to which comorbidities are coded in the patient record
may affect risk adjustment. Therefore the ‘depth of coding’10 has
been assessed across NSW hospitals. In July 2009 – June 2012,
the average depth of coding was 3.7 diagnoses in this hospital
and 3.7 in NSW public hospitals; and in July 2012 – June 2015,
there were 3.8 diagnoses in this hospital and 4.8 in NSW public
hospitals.
RSRR
July 2012 –
June 2015
July 12 –
June 13
July 13 –
June 14
July 14 –
June 15
Lower than
expected
No different
than expected
Higher than
expectedRatio:
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
1. Data refer to patients who were discharged from this hospital, between July 2012 and June 2015, following an acute hospitalisation with pneumonia as principal diagnosis (ICD-10-AM codes J13, J14,
J15, J16, J18).
2. Returns to acute care are to any NSW hospital in the 30 days (for acute conditions) or 60 days (for elective surgeries) following discharge, and are attributed to the last discharging hospital. For patients
whose acute hospitalisation ended in discharge home, a return to acute care involves readmission to hospital; while for patients whose acute hospitalisation ended with a 'discharge' to non-acute care, a
return involved a move back into an acute care setting regardless of whether they physically left the hospital.
3. For calculation of average length of stay, index admissions that were transferred in from, or transferred out to, another acute care hospital were excluded. Unreasonably long episodes are trimmed on the
basis of the Diagnosis Related Group (DRG) of the episode. The trim point is the third quartile plus 1.5 x the interquartile range of all in-scope episodes in each DRG.
4. Discharge destinations are based on the mode of separation of the index case. For episodes coded as 'Discharged by hospital' or 'Discharged on leave', patients are considered to be destined for their
place of usual residence. All other modes of separation are deemed to indicate a discharge destination other than a patient’s place of usual residence.
5. Age at admission date.
6. Comorbidities are identified from the hospital discharge records using the Elixhauser comorbidity set (plus dementia) with a one year look-back from the admission date of the index case. Only those
conditions that were shown to have a significant impact on readmission (P<0.05) are shown.
7. Hospitals are classified as urban and regional/rural using the geocoded address of the hospital assigned to ABS statistical areas (SA2) and the Australian remoteness index for areas.
8. Reasons for return to acute care are classified according to a draft specification made available to BHI by the Australian Institute of Health and Welfare. Principal diagnoses for the return to acute care
episode, are stratified as: the same as the index hospitalisation; related to that of the index hospitalisation (same ICD-10-AM chapter); potentially related to hospital care (i.e. complications and adverse
events) using various time horizons; and, other reasons. Percentages may not add to 100% due to rounding.
9. Results for hospitals with <1 expected readmission are not shown. Peer hospitals are identified according to the NSW Ministry of Health’s peer grouping as of April 2012.
10. The depth of coding has been defined as the average number of secondary diagnosis coded for the index cases. The one year look back method which is used for risk adjustment, to some extent
accounts for possible lower depth of coding in some hospitals.
Details of analyses are available in Spotlight on Measurement: Measuring return to acute care following discharge from hospital, 2nd edition.
Data source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health.
16
Ob
serv
ed
(unad
juste
d) ra
teE
xp
ecte
d r
ate
Ris
k–sta
nd
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Statistically significant result
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
0
6
12
18
24
July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total index cases for chronic obstructive pulmonary disease 258 47,359
Average length of stay (days) 5.1 5.3
Patients transferred in from acute care in another hospital 11 3,367
Discharge destination:
Home 230 42,937
Other 28 4,422
17
*Age was a significant factor in the final model of 30-day readmission following hospitalisation for chronic obstructive pulmonary disease.
23.3
21.8
33.3
31.4
28.3
32.9
15.1
14.0
% index cases
45–64 65–74 75–84 85+
This hospital
NSW
8.0
4.7
2.2
-0.5
-0.5
-0.5
-1.5
-2.6
-3.0
-4.0
-4.1
-5.0
-6.2
-7.3
-7.5
-20 -15 -10 -5 0 5 10 15 20
Abuse drug/alcohol/psychoses
Depression
Previous COPD admission
Lymphoma
Cardiac arrhythmia
Congestive heart failure
Metastatic cancer
Solid tumour without metastasis
Weight loss
Coagulopathy
Deficiency anaemia
Renal failure
Female
Diabetes, complicated
Fluid and electrolyte disorders
% difference from NSW (index cases with factor recorded)
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total readmissions following index hospitalisation for chronic obstructive pulmonary disease 59 10,293
Readmitted to the hospital where acute care was completed 53 8,696
Readmitted to a different hospital 6 1,597
Of these:
To an urban public hospital 4
To a regional or rural public hospital 2
To a private hospital 0
18
Distribution of reasons for returns to acute care
Number of, and reasons for, returns to acute care following hospitalisation for chronic obstructive pulmonary disease, by days post
discharge
Same principal diagnosis Condition related to principal diagnosis Potentially related to hospital care
(relevant at any time)
Potentially related to hospital care
(time sensitive, ≤ 7 days post discharge)
Potentially related to hospital care
(time sensitive, 8–30 days post discharge)
Other conditions
45.8
54.9 9.9
11.9
4.7
10.2
11.0
27.1
16.0
0 10 20 30 40 50 60 70 80 90 100
% returns to acute care
This hospital
NSW
11
6 5 5
2
7
2
2
2
3
7 4
2
0
5
10
15
20
25
1–7 days 8–14 days 15–21 days 22–30 days
Num
ber
of re
turn
s t
o a
cute
care
Days post discharge
Performance Profile: bhi.nsw.gov.au
5.6
0.6
-2.0
0
1
2
3
4
0 50 100 150 200 250 300 350 400 450
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio
(Ob
serv
ed
/Exp
ecte
d)
Expected number of returns to acute care (readmissions) within 30 days
Murwillumbah District Hospital
Hospital-specific RSRRs report the ratio of actual or ‘observed’
number of returns to acute care to the ‘expected’ number of
returns. A competing risk regression model draws on the NSW
patient population’s characteristics and outcomes to estimate the
expected number of returns for each hospital, given the
characteristics of its patients.
An RSRR less than 1.0 indicates lower-than-expected returns to
acute care, and a ratio higher than 1.0 indicates higher-than-
expected returns. Small deviations from 1.0 are not considered to
be meaningful. Funnel plots with 95% and 99.8% control limits
around the NSW ratio are used to identify outliers.
19
99.8% limits95.0% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
In order to make fair comparisons, a number of risk adjustments
are made to readmission data. These take into account patient
factors that influence the likelihood of returning to acute care
within 30 days. The table below illustrates the effect of statistical
adjustments on this hospital’s results.
Unadjusted ratio
Age and sex
standardised ratio
Risk-standardised
readmission ratio
The RSRR is calculated on the basis of three years of data.
It takes account of differences in patient characteristics so that
assessments of hospital performance are fair. To give an
indication of results within the three-year period, the figure below
shows the RSRR result for July 2012 – June 2015 alongside
differences between this hospital and the NSW result for annual
unadjusted readmission rates.
Lower than
expected
No different
than expected
Higher than
expected
RSRR:
Unadjusted
readmission
rate percentage
point difference
from NSW result
The extent to which comorbidities are coded in the patient record
may affect risk adjustment. Therefore the ‘depth of coding’10 has
been assessed across NSW hospitals. In July 2009 – June 2012,
the average depth of coding was 3.6 diagnoses in this hospital
and 3.2 in NSW public hospitals; and in July 2012 – June 2015,
there were 3.6 diagnoses in this hospital and 4.1 in NSW public
hospitals.
RSRR
July 2012 –
June 2015
July 12 –
June 13
July 13 –
June 14
July 14 –
June 15
Lower than
expected
No different
than expected
Higher than
expectedRatio:
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
1. Data refer to patients who were discharged from this hospital, between July 2012 and June 2015, following an acute hospitalisation with COPD as principal diagnosis (ICD-10-AM code J41, J42, J43,
J44, J47, and J20 and J40 if accompanied by J41, J42, J43, J44 and J47 in any secondary diagnoses).
2. Returns to acute care are to any NSW hospital in the 30 days (for acute conditions) or 60 days (for elective surgeries) following discharge, and are attributed to the last discharging hospital. For patients
whose acute hospitalisation ended in discharge home, a return to acute care involves readmission to hospital; while for patients whose acute hospitalisation ended with a 'discharge' to non-acute care, a
return involved a move back into an acute care setting regardless of whether they physically left the hospital.
3. For calculation of average length of stay, index admissions that were transferred in from, or transferred out to, another acute care hospital were excluded. Unreasonably long episodes are trimmed on the
basis of the Diagnosis Related Group (DRG) of the episode. The trim point is the third quartile plus 1.5 x the interquartile range of all in-scope episodes in each DRG.
4. Discharge destinations are based on the mode of separation of the index case. For episodes coded as 'Discharged by hospital' or 'Discharged on leave', patients are considered to be destined for their
place of usual residence. All other modes of separation are deemed to indicate a discharge destination other than a patient’s place of usual residence.
5. Age at admission date.
6. Comorbidities are identified from the hospital discharge records using the Elixhauser comorbidity set (plus dementia) with a one year look-back from the admission date of the index case. Only those
conditions that were shown to have a significant impact on readmission (P<0.05) are shown.
7. Hospitals are classified as urban and regional/rural using the geocoded address of the hospital assigned to ABS statistical areas (SA2) and the Australian remoteness index for areas.
8. Reasons for return to acute care are classified according to a draft specification made available to BHI by the Australian Institute of Health and Welfare. Principal diagnoses for the return to acute care
episode, are stratified as: the same as the index hospitalisation; related to that of the index hospitalisation (same ICD-10-AM chapter); potentially related to hospital care (i.e. complications and adverse
events) using various time horizons; and, other reasons. Percentages may not add to 100% due to rounding.
9. Results for hospitals with <1 expected readmission are not shown. Peer hospitals are identified according to the NSW Ministry of Health’s peer grouping as of April 2012.
10. The depth of coding has been defined as the average number of secondary diagnosis coded for the index cases. The one year look back method which is used for risk adjustment, to some extent
accounts for possible lower depth of coding in some hospitals.
Details of analyses are available in Spotlight on Measurement: Measuring return to acute care following discharge from hospital, 2nd edition.
Data source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health.
20
Ob
serv
ed
(unad
juste
d) ra
teE
xp
ecte
d r
ate
Ris
k–sta
nd
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Statistically significant result
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0
6
12
18
24
30
36
42
July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total index cases for hip fracture surgery 103 14,581
Average length of stay (days)
Patients transferred in from acute care in another hospital 103 2,728
Discharge destination:
Home 37 4,873
Other 66 9,708
21
*Age was not a significant factor in the final model of 30-day readmission following hospitalisation for hip fracture surgery.
Note: More than 5% of total patients underwent hip fracture surgery at another hospital.
5.8
18.5
12.7
40.8
33.4
35.9
48.2
% index cases
50–64 65–74 75–84 85+
This hospital
NSW
6.4
5.3
2.5
1.5
0.6
0.0
-0.4
-1.5
-1.6
-5.8
-8.6
-20 -15 -10 -5 0 5 10 15 20
Chronic pulmonary disease
Fluid and electrolyte disorders
Paralysis
Other neurological disorders
Coagulopathy
Deficiency anaemia
Renal failure
Cardiac arrhythmia
Diabetes, complicated
Dementia
Female
% difference from NSW (index cases with factor recorded)
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total readmissions following index hospitalisation for hip fracture surgery 20 1,485
Readmitted to the hospital where acute care was completed 14 1,135
Readmitted to a different hospital 6 350
Of these:
To an urban public hospital 5
To a regional or rural public hospital 1
To a private hospital 0
22
Distribution of reasons for returns to acute care
Number of, and reasons for, returns to acute care following hospitalisation for hip fracture surgery, by days post discharge
Same principal diagnosis Orthopaedic complications Potentially related to hospital care
(relevant at any time)
Potentially related to hospital care
(time sensitive, ≤ 7 days post discharge)
Potentially related to hospital care
(time sensitive, 8–30 days post discharge)
Other conditions
5.0
7.1
10.0
17.0
10.0
8.0
15.0
11.6
15.0
22.4
45.0
33.8
0 10 20 30 40 50 60 70 80 90 100
% returns to acute care
This hospital
NSW
1
1
1
1
13
2
1
4
2
3
0
1
2
3
4
5
6
7
8
1–7 days 8–14 days 15–21 days 22–30 days
Num
ber
of re
turn
s t
o a
cute
care
Days post discharge
Performance Profile: bhi.nsw.gov.au
14.8 12.5
1.4
0
1
2
3
4
0 50 100 150 200 250 300 350 400 450
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio
(Ob
serv
ed
/Exp
ecte
d)
Expected number of returns to acute care (readmissions) within 30 days
Murwillumbah District Hospital
Hospital-specific RSRRs report the ratio of actual or ‘observed’
number of returns to acute care to the ‘expected’ number of
returns. A competing risk regression model draws on the NSW
patient population’s characteristics and outcomes to estimate the
expected number of returns for each hospital, given the
characteristics of its patients.
An RSRR less than 1.0 indicates lower-than-expected returns to
acute care, and a ratio higher than 1.0 indicates higher-than-
expected returns. Small deviations from 1.0 are not considered to
be meaningful. Funnel plots with 95% and 99.8% control limits
around the NSW ratio are used to identify outliers.
23
99.8% limits95.0% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
In order to make fair comparisons, a number of risk adjustments
are made to readmission data. These take into account patient
factors that influence the likelihood of returning to acute care
within 30 days. The table below illustrates the effect of statistical
adjustments on this hospital’s results.
Unadjusted ratio
Age and sex
standardised ratio
Risk-standardised
readmission ratio
The RSRR is calculated on the basis of three years of data.
It takes account of differences in patient characteristics so that
assessments of hospital performance are fair. To give an
indication of results within the three-year period, the figure below
shows the RSRR result for July 2012 – June 2015 alongside
differences between this hospital and the NSW result for annual
unadjusted readmission rates.
Lower than
expected
No different
than expected
Higher than
expected
RSRR:
Unadjusted
readmission
rate percentage
point difference
from NSW result
The extent to which comorbidities are coded in the patient record
may affect risk adjustment. Therefore the ‘depth of coding’10 has
been assessed across NSW hospitals. In July 2009 – June 2012,
the average depth of coding was 10.3 diagnoses in this hospital
and 8.3 in NSW public hospitals; and in July 2012 – June 2015,
there were 9.6 diagnoses in this hospital and 9.2 in NSW public
hospitals.
RSRR
July 2012 –
June 2015
July 12 –
June 13
July 13 –
June 14
July 14 –
June 15
Lower than
expected
No different
than expected
Higher than
expectedRatio:
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
1. Data refer to patients who were discharged from this hospital, between July 2012 and June 2015, following an acute hospitalisation with hip fracture as principal diagnosis and treated with surgery (ICD-
10-AM codes for hip fracture S72.0, S72.1, S72.2 accompanied with a fall codes W00-W19 and R29.6 and treated with a surgical procedure).
2. Returns to acute care are to any NSW hospital in the 30 days (for acute conditions) or 60 days (for elective surgeries) following discharge, and are attributed to the last discharging hospital. For patients
whose acute hospitalisation ended in discharge home, a return to acute care involves readmission to hospital; while for patients whose acute hospitalisation ended with a 'discharge' to non-acute care, a
return involved a move back into an acute care setting regardless of whether they physically left the hospital.
3. For calculation of average length of stay, index admissions that were transferred in from, or transferred out to, another acute care hospital were excluded. Unreasonably long episodes are trimmed on the
basis of the Diagnosis Related Group (DRG) of the episode. The trim point is the third quartile plus 1.5 x the interquartile range of all in-scope episodes in each DRG.
4. Discharge destinations are based on the mode of separation of the index case. For episodes coded as 'Discharged by hospital' or 'Discharged on leave', patients are considered to be destined for their
place of usual residence. All other modes of separation are deemed to indicate a discharge destination other than a patient’s place of usual residence.
5. Age at admission date.
6. Comorbidities are identified from the hospital discharge records using the Elixhauser comorbidity set (plus dementia) with a one year look-back from the admission date of the index case. Only those
conditions that were shown to have a significant impact on readmission (P<0.05) are shown.
7. Hospitals are classified as urban and regional/rural using the geocoded address of the hospital assigned to ABS statistical areas (SA2) and the Australian remoteness index for areas.
8. Reasons for return to acute care are classified according to a draft specification made available to BHI by the Australian Institute of Health and Welfare. Principal diagnoses for the return to acute care
episode, are stratified as: the same as the index hospitalisation; orthopaedic complications; potentially related to hospital care (i.e. complications and adverse events) using various time horizons; and,
other reasons. Percentages may not add to 100% due to rounding.
9. Results for hospitals with <1 expected readmission are not shown. Peer hospitals are identified according to the NSW Ministry of Health’s peer grouping as of April 2012.
10. The depth of coding has been defined as the average number of secondary diagnosis coded for the index cases. The one year look back method which is used for risk adjustment, to some extent
accounts for possible lower depth of coding in some hospitals.
Details of analyses are available in Spotlight on Measurement: Measuring return to acute care following discharge from hospital, 2nd edition.
Data source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health.
24
Ob
serv
ed
(unad
juste
d) ra
teE
xp
ecte
d r
ate
Ris
k–sta
nd
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Statistically significant result
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
0
2
4
6
8
10
12
14
16
18
20
22
July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total index cases for total hip replacement 93 8,312
Average length of stay (days) 4.4 5.4
Discharge destination:
Home 92 7,084
Other 1 1,228
25
*Age was a significant factor in the final model of 60-day readmission following hospitalisation for total hip replacement.
5.4 54.8
33.3
24.7
33.9
15.1
24.0
% index cases
18–44 45–64 65–74 75–84 85+
This hospital
NSW
-0.2
-0.3
-0.7
-0.8
-1.3
-2.1
-2.4
-4.8
-6.7
-8.0
-9.4
-20 -15 -10 -5 0 5 10 15 20
Lymphoma
Solid tumour without metastasis
Pulmonary circulation disorders
Depression
Liver disease
Chronic pulmonary disease
Renal failure
Diabetes, complicated
Diabetes, uncomplicated
Cardiac arrhythmia
Female
% difference from NSW (index cases with factor recorded)
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total readmissions following index hospitalisation for total hip replacement 2 764
Readmitted to the hospital where acute care was completed
Readmitted to a different hospital
Of these:
To an urban public hospital
To a regional or rural public hospital
To a private hospital
26
Distribution of reasons for returns to acute care
Number of, and reasons for, returns to acute care following hospitalisation for total hip replacement, by days post discharge
Orthopaedic complications
(within time specified)
Orthopaedic complications
(outside time specified)
Potentially related to hospital care
(within time specified)
Potentially related to hospital care
(outside time specified)
Other conditions
<10 readmissions
Detailed results not shown
<10 readmissions
Detailed results not shown
<10 readmissions
Detailed results not shown
Performance Profile: bhi.nsw.gov.au
-2.9-9.7 -10.0
0
1
2
3
4
0 20 40 60 80 100 120 140
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio
(Ob
serv
ed
/Exp
ecte
d)
Expected number of returns to acute care (readmissions) within 60 days
Murwillumbah District Hospital
Hospital-specific RSRRs report the ratio of actual or ‘observed’
number of returns to acute care to the ‘expected’ number of
returns. A competing risk regression model draws on the NSW
patient population’s characteristics and outcomes to estimate the
expected number of returns for each hospital, given the
characteristics of its patients.
An RSRR less than 1.0 indicates lower-than-expected returns to
acute care, and a ratio higher than 1.0 indicates higher-than-
expected returns. Small deviations from 1.0 are not considered to
be meaningful. Funnel plots with 95% and 99.8% control limits
around the NSW ratio are used to identify outliers.
27
99.8% limits95.0% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
In order to make fair comparisons, a number of risk adjustments
are made to readmission data. These take into account patient
factors that influence the likelihood of returning to acute care
within 60 days. The table below illustrates the effect of statistical
adjustments on this hospital’s results.
Unadjusted ratio
Age and sex
standardised ratio
Risk-standardised
readmission ratio
The RSRR is calculated on the basis of three years of data.
It takes account of differences in patient characteristics so that
assessments of hospital performance are fair. To give an
indication of results within the three-year period, the figure below
shows the RSRR result for July 2012 – June 2015 alongside
differences between this hospital and the NSW result for annual
unadjusted readmission rates.
Lower than
expected
No different
than expected
Higher than
expected
RSRR:
Unadjusted
readmission
rate percentage
point difference
from NSW result
The extent to which comorbidities are coded in the patient record
may affect risk adjustment. Therefore the ‘depth of coding’10 has
been assessed across NSW hospitals. In July 2009 – June 2012,
the average depth of coding was 1.6 diagnoses in this hospital
and 2.5 in NSW public hospitals; and in July 2012 – June 2015,
there were 1.3 diagnoses in this hospital and 2.6 in NSW public
hospitals.
RSRR
July 2012 –
June 2015
July 12 –
June 13
July 13 –
June 14
July 14 –
June 15
Lower than
expected
No different
than expected
Higher than
expectedRatio:
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
1. Data refer to patients who were discharged from this hospital, between July 2012 and June 2015, following an acute hospitalisation for an elective total hip replacement (ACHI codes 49318-00, 49319-
00).
2. Returns to acute care are to any NSW hospital in the 30 days (for acute conditions) or 60 days (for elective surgeries) following discharge, and are attributed to the last discharging hospital. For patients
whose acute hospitalisation ended in discharge home, a return to acute care involves readmission to hospital; while for patients whose acute hospitalisation ended with a 'discharge' to non-acute care, a
return involved a move back into an acute care setting regardless of whether they physically left the hospital.
3. For calculation of average length of stay, index admissions that were transferred in from, or transferred out to, another acute care hospital were excluded. Unreasonably long episodes are trimmed on the
basis of the Diagnosis Related Group (DRG) of the episode. The trim point is the third quartile plus 1.5 x the interquartile range of all in-scope episodes in each DRG.
4. Discharge destinations are based on the mode of separation of the index case. For episodes coded as 'Discharged by hospital' or 'Discharged on leave', patients are considered to be destined for their
place of usual residence. All other modes of separation are deemed to indicate a discharge destination other than a patient’s place of usual residence.
5. Age at admission date.
6. Comorbidities are identified from the hospital discharge records using the Elixhauser comorbidity set (plus dementia) with a one year look-back from the admission date of the index case. Only those
conditions that were shown to have a significant impact on readmission (P<0.05) are shown.
7. Hospitals are classified as urban and regional/rural using the geocoded address of the hospital assigned to ABS statistical areas (SA2) and the Australian remoteness index for areas.
8. Reasons for return to acute care are classified according to a draft specification made available to BHI by the Australian Institute of Health and Welfare. Principal diagnoses for the return to acute care
episode, are stratified as: orthopaedic complications using various time horizons; potentially related to hospital care (i.e. complications and adverse events) using various time horizons; and, other reasons.
Percentages may not add to 100% due to rounding.
9. Results for hospitals with <1 expected readmission are not shown. Peer hospitals are identified according to the NSW Ministry of Health’s peer grouping as of April 2012.
10. The depth of coding has been defined as the average number of secondary diagnosis coded for the index cases. The one year look back method which is used for risk adjustment, to some extent
accounts for possible lower depth of coding in some hospitals.
Details of analyses are available in Spotlight on Measurement: Measuring return to acute care following discharge from hospital, 2nd edition.
Data source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health.
28
Ob
serv
ed
(unad
juste
d) ra
teE
xp
ecte
d r
ate
Ris
k–sta
nd
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Statistically significant result
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0
5
10
July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total index cases for total knee replacement 133 14,961
Average length of stay (days) 5.1 5.6
Discharge destination:
Home 130 12,362
Other 3 2,599
29
*Age was a significant factor in the final model of 60-day readmission following hospitalisation for total knee replacement.
33.8
29.8
45.9
40.1
18.8
26.4
% index cases
18–44 45–64 65–74 75–84 85+
This hospital
NSW
4.8
-0.2
-0.6
-0.6
-1.0
-4.8
-7.4
-7.7
-13.3
-30 -20 -10 0 10 20 30
Abuse drug/alcohol/psychoses
Liver disease
Blood loss anaemia
Depression
Pulmonary circulation disorders
Cardiac arrhythmia
Fluid and electrolyte disorders
Diabetes, complicated
Female
% difference from NSW (index cases with factor recorded)
Performance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
Total readmissions following index hospitalisation for total knee replacement 15 1,727
Readmitted to the hospital where acute care was completed 4 1,011
Readmitted to a different hospital 11 716
Of these:
To an urban public hospital 11
To a regional or rural public hospital 0
To a private hospital 0
30
Distribution of reasons for returns to acute care
Number of, and reasons for, returns to acute care following hospitalisation for total knee replacement, by days post discharge
Orthopaedic complications
(within time specified)
Orthopaedic complications
(outside time specified)
Potentially related to hospital care
(within time specified)
Potentially related to hospital care
(outside time specified)
Other conditions
33.3
29.5
20.0
13.4
26.7
10.5
13.3
14.2
6.7
32.3
0 10 20 30 40 50 60 70 80 90 100
% returns to acute care
This hospital
NSW
4
1
2
1
3
1
1 1
1
0
1
2
3
4
5
6
7
8
1–14 days 15–28 days 29–42 days 43–60 days
Num
ber
of re
turn
s t
o a
cute
care
Days post discharge
Performance Profile: bhi.nsw.gov.au
0.6 1.4
-2.4
0
1
2
3
4
0 20 40 60 80 100 120 140
Ris
k-s
tand
ard
ised
read
mis
sio
n r
atio
(Ob
serv
ed
/Exp
ecte
d)
Expected number of returns to acute care (readmissions) within 60 days
Murwillumbah District Hospital
Hospital-specific RSRRs report the ratio of actual or ‘observed’
number of returns to acute care to the ‘expected’ number of
returns. A competing risk regression model draws on the NSW
patient population’s characteristics and outcomes to estimate the
expected number of returns for each hospital, given the
characteristics of its patients.
An RSRR less than 1.0 indicates lower-than-expected returns to
acute care, and a ratio higher than 1.0 indicates higher-than-
expected returns. Small deviations from 1.0 are not considered to
be meaningful. Funnel plots with 95% and 99.8% control limits
around the NSW ratio are used to identify outliers.
31
99.8% limits95.0% limits
Higher than expected:
No different than expected:
Lower than expected:
This
hospital
Peer
hospitals
Other
hospitals
In order to make fair comparisons, a number of risk adjustments
are made to readmission data. These take into account patient
factors that influence the likelihood of returning to acute care
within 60 days. The table below illustrates the effect of statistical
adjustments on this hospital’s results.
Unadjusted ratio
Age and sex
standardised ratio
Risk-standardised
readmission ratio
The RSRR is calculated on the basis of three years of data.
It takes account of differences in patient characteristics so that
assessments of hospital performance are fair. To give an
indication of results within the three-year period, the figure below
shows the RSRR result for July 2012 – June 2015 alongside
differences between this hospital and the NSW result for annual
unadjusted readmission rates.
Lower than
expected
No different
than expected
Higher than
expected
RSRR:
Unadjusted
readmission
rate percentage
point difference
from NSW result
The extent to which comorbidities are coded in the patient record
may affect risk adjustment. Therefore the ‘depth of coding’10 has
been assessed across NSW hospitals. In July 2009 – June 2012,
the average depth of coding was 1.6 diagnoses in this hospital
and 2.1 in NSW public hospitals; and in July 2012 – June 2015,
there were 1.6 diagnoses in this hospital and 2.4 in NSW public
hospitals.
RSRR
July 2012 –
June 2015
July 12 –
June 13
July 13 –
June 14
July 14 –
June 15
Lower than
expected
No different
than expected
Higher than
expectedRatio:
Performance Profile: bhi.nsw.gov.auPerformance Profile: bhi.nsw.gov.auMurwillumbah District Hospital
1. Data refer to patients who were discharged from this hospital, between July 2012 and June 2015, following an acute hospitalisation for an elective total knee replacement (ACHI codes 49518-00,
49519-00, 49521-00, 49521-01, 49521-02, 49521-03, 49524-00, 49524-01).
2. Returns to acute care are to any NSW hospital in the 30 days (for acute conditions) or 60 days (for elective surgeries) following discharge, and are attributed to the last discharging hospital. For patients
whose acute hospitalisation ended in discharge home, a return to acute care involves readmission to hospital; while for patients whose acute hospitalisation ended with a 'discharge' to non-acute care, a
return involved a move back into an acute care setting regardless of whether they physically left the hospital.
3. For calculation of average length of stay, index admissions that were transferred in from, or transferred out to, another acute care hospital were excluded. Unreasonably long episodes are trimmed on the
basis of the Diagnosis Related Group (DRG) of the episode. The trim point is the third quartile plus 1.5 x the interquartile range of all in-scope episodes in each DRG.
4. Discharge destinations are based on the mode of separation of the index case. For episodes coded as 'Discharged by hospital' or 'Discharged on leave', patients are considered to be destined for their
place of usual residence. All other modes of separation are deemed to indicate a discharge destination other than a patient’s place of usual residence.
5. Age at admission date.
6. Comorbidities are identified from the hospital discharge records using the Elixhauser comorbidity set (plus dementia) with a one year look-back from the admission date of the index case. Only those
conditions that were shown to have a significant impact on readmission (P<0.05) are shown.
7. Hospitals are classified as urban and regional/rural using the geocoded address of the hospital assigned to ABS statistical areas (SA2) and the Australian remoteness index for areas.
8. Reasons for return to acute care are classified according to a draft specification made available to BHI by the Australian Institute of Health and Welfare. Principal diagnoses for the return to acute care
episode, are stratified as: orthopaedic complications using various time horizons; potentially related to hospital care (i.e. complications and adverse events) using various time horizons; and, other reasons.
Percentages may not add to 100% due to rounding.
9. Results for hospitals with <1 expected readmission are not shown. Peer hospitals are identified according to the NSW Ministry of Health’s peer grouping as of April 2012.
10. The depth of coding has been defined as the average number of secondary diagnosis coded for the index cases. The one year look back method which is used for risk adjustment, to some extent
accounts for possible lower depth of coding in some hospitals.
Details of analyses are available in Spotlight on Measurement: Measuring return to acute care following discharge from hospital, 2nd edition.
Data source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health.
32
Ob
serv
ed
(unad
juste
d) ra
teE
xp
ecte
d r
ate
Ris
k–sta
nd
ard
ised
read
mis
sio
n r
atio (R
SR
R)
Statistically significant result
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
0
4
8
12
16
20
July 00 – June 03 July 03 – June 06 July 06 – June 09 July 09 – June 12 July 12 – June 15
