the Case Mix Programme Quality Improvement

Document Sample
the Case Mix Programme Quality Improvement Powered By Docstoc
					Experience from ICU audits:
  the Case Mix Programme
        Kathy Rowan
          Director




                        ICNARC 2001
Case Mix Programme:
• National, comparative audit of patient outcome
• 160 recruited/trained units -
  143 (60.9%) adult general units in England
 [(50%)Wales (50%) N.Ireland]
• Units in
     28 + 24 ) (32.5%) university or affiliated
     108       (67.5%) non-university
• Confidential
• Case mix (risk) adjusted

• Consecutive admissions                 ICNARC 2001
Case Mix Programme - history:

•   1987-   ICS UK APACHE II study
•   1991    Proposal to Government
•   1994    Establishing ICNARC
•   1995    Establishing the Case Mix Programme
•   1996-   CMP “roll-out”
•   1997-   CMP development



                                       ICNARC 2001
Case Mix Programme recruitment:

•   Visit
•   Data collection training (2 days)
•   Pilot data collection period (6 weeks)
•   Full data collection




                                             ICNARC 2001
Case Mix Programme dataset:

• Alphanumeric identifiers (unit/admission)
• Sociodemographic (age, sex, postcode)
• Organisational (source, date/time in/out etc.)
• First 24 hour physiology forAPACHE II/III,
  SAPS II, MPM0/24
• Outcome (unit/hospital, immediate/ultimate)




                                         ICNARC 2001
Case Mix Programme data entry/export:

• Dataset specification (rules/definitions)
• Raw data
• Software/software developers




                                         ICNARC 2001
Case Mix Programme data validation:

•   Local/central
•   Completeness
•   Illogicalities
•   Inconsistencies
•   Reliability (optional)




                                  ICNARC 2001
                                                ICNARC / units
                                               visit
                          units                                     1
                                               data collection training              units
                     local review/use                                             data collection
                     of report                 Annual Meeting of the              and entry
comparative case mix,                          Case Mix Programme
outcome and activity report             2                                    2                         data
(within six weeks of                                                                                   (via disk
receipt of complete,                                                                                   or modem)
accurate data)

               ICNARC
                                               six month                                   ICNARC
        produce data analysis report              cycle                                   data validation
                                                                                          produce data
                                                                                          validation report
                                               data validation
                                               report                      units              data validation report
                complete and                                                                  (within four weeks of
                accurate data     ICNARC                            correct data              receipt of data) and
                                  data                              reliability study         reliabilty study sample
                                  validation                        (re-collect data on
                                                corrected and       random sample of 20
                                                reliability study
                                                data                patients)

                                                                                             ICNARC 2001
Case Mix Programme feedback:

•   Comparative data analysis reports
•   Ad hoc specific analyses
•   Annual Meeting
•   Publications




                                        ICNARC 2001
 Data completeness - admission variables (n=103) [01/07/99 to 29/12/99]


                                      Variable
                                    Postcode
                                Date of birth
                                          Sex
           Date of admission to your hospital
               Date of admission to your unit
               T ime of admission to your unit
Managed by your unit team prior to admission                                   Complete
*Date/T ime managed by your unit team (n=2)
               Planned admission to your unit                                  Missing
      Admission for pre-surgical preparation
             Source of admission to your unit
*Location immediately prior to source (n=86)
               Cardiopulmonary resuscitation
  *Date of original admission to a unit (n=8)
*Date of original admission to hospital (n=18)
                                            0 10 20 30 40 50 60 70 80 90 100
                                                     Proportion (%)



                                                                               ICNARC 2001
Distribution of Acute Physiology Score for all eligible admissions to
your unit in this cycle from 01/07/99 to 29/12/99 (n=99) compared with
the last cycle (n=131) and the CMPD (n=19,421)
                             Proportion (%)
        35.0
                                                                             Your unit, last cycle (%)
        30.0                                                                 Your unit, this cycle (%)
        25.0                                                                 CMPD (%)

        20.0
        15.0
        10.0
           5.0
           0.0
                                 0-

                                        5-

                                               10

                                                       15

                                                               20

                                                                       25

                                                                                 30

                                                                                         35

                                                                                                 40

                                                                                                         45

                                                                                                                 50

                                                                                                                         55
                                    4

                                           9

                                                 -1

                                                         -1

                                                                 -2

                                                                         -2

                                                                                   -3

                                                                                           -3

                                                                                                   -4

                                                                                                           -4

                                                                                                                   -5

                                                                                                                           -6
                                                   4

                                                           9

                                                                   4

                                                                           9

                                                                                     4

                                                                                             9

                                                                                                     4

                                                                                                             9

                                                                                                                     4

                                                                                                                             0
   Your unit, this c yc le (n)   5.0    22.0   24.0    19.0    20.0    5.0        2.0    2.0     0.0     0.0     0.0     0.0
   Your unit, this c yc le (%)   5.1    22.2   24.2    19.2    20.2    5.1        2.0    2.0     0.0     0.0     0.0     0.0
                C M PD (%)       8.2    29.8   30.6    18.1     8.2    3.3        1.3    0.3     0.1     0.0     0.0     0.0




                                                               Acute Physiology Score



                                                                                                                 ICNARC 2001
Distribution of APACHE II probability of death before discharge from
your hospital following care in your unit for all eligible admissions to
your unit in this cycle from 01/07/99 to 29/12/99 (n=95), with the last cycle
(n=124) and the CMPD (n=17,552)
                                 Proportion (%)
             35.0
                                                                      Your unit, last cycle (%)
             30.0
                                                                      Your unit, this cycle (%)
             25.0                                                     CMPD (%)
             20.0
             15.0
             10.0
                 5.0
                 0.0
                                     0-

                                            10

                                                     20

                                                               30

                                                                          40

                                                                                   50

                                                                                            60

                                                                                                        70

                                                                                                                    80

                                                                                                                                90
                                       9

                                               -1

                                                        -2

                                                                  -3

                                                                             -4

                                                                                      -5

                                                                                                  -6

                                                                                                              -7

                                                                                                                          -8

                                                                                                                                      -1
                                                 9

                                                          9

                                                                    9

                                                                               9

                                                                                        9

                                                                                                    9

                                                                                                                9

                                                                                                                            9

                                                                                                                                        00
       Your unit, this c yc le (n)   25.0   15.0     10.0      10.0       10.0     10.0     6.0         4.0         2.0         3.0
       Your unit, this c yc le (%)   26.3   15.8     10.5      10.5       10.5     10.5     6.3         4.2         2.1         3.2
                    C M PD (%)       31.9   22.0     14.6       9.7        7.4      5.2     3.6         2.6         2.2         0.7



                                                            APACHE II probability (%)



                                                                                                                                             ICNARC 2001
     Proportion of admissions dying in hospital (%)

50

40

30

20

10

 0
                Intensive Care Units (ICUs) n=91

                                                   ICNARC 2001
      Ratio of observed to expected hospital deaths

2.5

 2

1.5

 1

0.5

 0
                 Intensive Care Units (ICUs) n=91


                                                    ICNARC 2001
Observed number of deaths versus expected number of deaths estimated
using the UK APACHE II equation for all eligible admissions to your unit
in this cycle from 01/07/99 to 29/12/99 (n=91) compared with the CMPD

        Observed number of deaths
  250                                         the x axis represents the expected number
            Your unit                         deaths estimated using the UK APACHE II
  200                                         equation

                                              the y axis represents the observed number of
  150
                                              deaths before discharge from hospital
                                              following care in your unit
  100
                                              the relationship between the observed
   50                                         number of deaths and the expected number
                                              of deaths + 95% confidence intervals for
    0                                         your unit is compared with those for the
        0    50    100 150 200          250   units in the CMPD
            Expected number of deaths




                                                                           ICNARC 2001
Case mix     Observed   Expected   Mortality ratio (95%
adjustment   deaths     deaths     confidence
method                             intervals)

APACHE II    35         29.0917    1.20 (0.95 - 1.45)
(n = 91)

PRISM        -----      -----      -----
(n = 0)

SAPS II      32         30.4773    1.05 (0.82 - 1.28)
(n = 93)

MPM II0      32         25.3513    1.26 (0.99 - 1.53)
(n = 93)

MPM II24     21         24.7363    0.85 (0.61 - 1.09)
(n = 66)
                                             ICNARC 2001
Case mix     Mortality ratio (95% confidence intervals)
adjustment
method
             Last cycle              This cycle


APACHE II    1.36 (1.15 - 1.57)      1.20 (0.95 - 1.45)



PRISM        -----                   -----


SAPS II      1.16 (0.99 - 1.33)      1.05 (0.82 - 1.28)


MPM II0      1.44 (1.23 - 1.66)      1.26 (0.99 - 1.53)


MPM II24     1.18 (0.97 - 1.39)      0.85 (0.61 - 1.09)

                                                          ICNARC 2001
Admission   APACHE II         Equivalent probabilities (%)
number      probability (%)

                              SAPS II       MPM II0          MPM II24



Hospital
survivors
19990230    70.54             48.39         63.11            31.14




Hospital
non-
survivors
19990159    9.67              9.67          19.96            38.98
19990204    17.57             22.96         18.71            57.64
19990210    17.89             10.64         41.35            65.99
19990219    18.57             5.22          6.74
19990243    22.55             61.93         18.86            61.69
19990214    23.18             18.10         65.66            21.69
19990231    28.74             19.64         32.30            60.51
                                                             ICNARC 2001
Bed occupancy versus patient throughput for all admissions to your unit in
this cycle from 01/07/99 to 29/12/99 (n=103) compared with the CMPD


      Mean daily bed occupancy (%)
100                                                       the x axis represents mean daily
                                              Your unit   patient throughput
                                              CMPD
 80                                                       the y axis represents mean daily
                                                          bed occupancy
 60
                                                          the relationship between patient
                                                          throughput and bed occupancy for
 40                                                       your unit is compared with those
                                                          for the units in the CMPD
 20
                                                          bed occupancy and patient
                                                          throughput are based on the
  0                                                       number of beds registered by your
      0    0.1    0.2     0.3    0.4    0.5               unit with ICNARC (4)
          Mean daily patient throughput




                                                                              ICNARC 2001
Case Mix Programme impact:

• Survey (1999)




                             ICNARC 2001
Uses of reports – comparison…

• “…to indicate how we compare to other units…”
• “…investigating outcomes for specific groups…”
• “...feedback to unit staff and user groups on
  issues such as outcomes, occupancy, case mix
  and review of practice…”
• “...bench-marking - comparison of mortality with
  other units…”
• “…analysis of unexpected survivors/non-
  survivors…”
• “…outcomes for patients transferred out of
  base hospital due to shortage of beds…”
                                         ICNARC 2001
Uses of reports – informing change…

• “…to tighten up ICU admission policy…”
• “…as evidence for the need for an HDU…”
• “…as evidence to set up a Medical Emergency
      Team in hospital…”
• “…for development of operational policy…”
• “…to evaluate safety and workload…”
• “…planning of future paediatric services…”




                                        ICNARC 2001
Uses of reports – monitoring change…

• “…to monitor actual changes in consultant cover
  and an increase in the number of beds…”
• “…to assess the impact of opening up an
   HDU…”
• “...looking at the effect of opening an extra
  bed…”




                                          ICNARC 2001
Case Mix Programme impact:

• Survey (1999)
• Correspondence (mild enquiries to threats…!)




                                        ICNARC 2001
Case Mix Programme feedback:

•   Comparative data analysis reports
•   Ad hoc specific analyses
•   Annual Meeting
•   Publications




                                        ICNARC 2001
Case Mix Programme Database:

• High quality clinical database
• 75,621 admissions to 102 units
  (138,545 admissions to 136 units as of 27/3/01)
• Median size 5 beds (range 2-22)
• Representative




                                       ICNARC 2001
                                                          Is there an ideal scoring system
       True or False?                                         from those available for
 Pulmonary artery catheters                                    adult intensive care or
 in patient management are                                   can we design a better one?                                         Teamwork and safety.
    associated with lower                                                                                                         Are they important?
     mortality and cost.                                  To develop and validate the optimum method for risk adjustment of        A national survey.
                                                                 hospital mortality for adult intensive care in the UK.
                                                                  Commenced: Sept. 1999. Duration: 3 years. (Ph.D.)

An evaluation of the clinical and cost-effectiveness of                                                                       Teamwork and safety attitudes in intensive care.
Pulmonary Artery Catheters in patient Management                                                                               A National Questionnaire survey of UK ICUs.
           (PAC-Man) in intensive care.                                                                                         Commenced: Aug. 2000. Duration: 2 years.
    Commenced: July 2000. Duration: 3 years.


    Which outcome measures                                                                                                          How can we
    are best for intensive and                                                                                                        establish
     high dependency care?                                                                                                       research priorities
                                                                                                                                 for intensive care?
          Outcome measures for adult critical care:           ICNARC’s
                    a systematic review.
             HTA Report published: Oct. 2000.
             Publication submitted: Nov. 2000.
                                                            funded research                                                    Use of consensus development to establish
                                                                                                                              national research priorities in critical care.
                                                                                                                              Published: BMJ Apr. 2000, ICM Oct. 2000.
                                                              programme
 Can large databases such
     as the Case Mix
                                                                     April 2001
                                                                                                                        Optimisation of scoring
  Programme Database                                                                                                    systems for paediatric
                                                                  How does a unit’s
   facilitate research on                                                                                                   intensive care.
                                                                  organisation affect
  clinical interventions?
                                                                   patient outcome?
                                                                                                                        The development and assessment of the optimal
                                                                                                                            risk adjustment methods for outcomes
  Evaluation using an high quality clinical database.
                                                                                                                            in paediatric intensive care in the UK.
                                                                                                                                          ICNARC 2001
  Commenced: Oct. 1999. Duration: 4 years. (Ph.D.)            The organisation of work in intensive care.
                                                           Commenced: Oct. 1999. Duration: 3 years. (Ph.D.)              Commenced: July 2000. Duration: 3 years.
Thanks…!




           ICNARC 2001