Using Administrative Data to Develop Indicators of Quality Care in Personal Care Homes Malcolm Doupe

1
MANITOBA CENTRE FOR HEALTH POLICY
Using Administrative Data to Develop Indicators
of Quality Care in Personal Care Homes Malcolm
Doupe, Marni Brownell, Natalia Dik, Charles
Burchill, Anita Kozyrskyj, Matt Dahl, Carolyn
DeCoster, Dan Chateau, Aynslie Hinds, Jen
Bodnarchuk
2

• The Manitoba Centre for Health Policy
• University of Manitoba
• Department of Community Health Sciences, Faculty
  of Medicine

3
Administrative Health Care Data
anonymous individualized data from Manitoba
Health
Manitoba Population
How often I am seen by a doctor for a bed sore,
fracture, fall, etc?
Am I taking any high risk drugs?
4
Goals of the study

• Using Administrative Data
• 1. Develop indicators of quality care
• Make comparisons between PCHs

• 2. Why do QIs occur more often in some PCHs?
• Resident Characteristics people in some PCHs
  are more sick!
• PCH (facility) Characteristics some PCHs have
• more beds, have more staff, etc

5
Goal 1
Developing QIs making comparisons between PCHs.
A. Diagnostic QIs

• Rates of hospitalization and physician visits
• Accidental falls
• Skin ulcers
• Non-hip fractures
• Hip fractures
• Respiratory infections
• Electrolyte imbalances

Physician visits
6
How many QIs are Acceptable?
10 of PCHs where QIs occurred MOST frequently
90th
NORMAL RANGE
10th
10 of PCHs where QIs occurred LEAST frequently
7
Rates of ACCIDENTAL FALLS, by PCH Data adjusted
for differences in resident characteristics
8
Rates of SKIN ULCERS, by PCH Data adjusted for
differences in resident characteristics
9
Rates of Non-hip Fractures, by PCH Data adjusted
for differences in resident characteristics
10
Summary of PCHs ABOVE THE 90 Threshold (combined
data for 6 diagnostic QIs)
Potential problems may exist in these PCHs!
Represents PCH ID (Manitoba Health, Annual
Statistics Report)
11
Summary of PCHs BELOW THE 10 Threshold (combined
data for 6 diagnostic QIs)
Good news stories also exist!!
Represents PCH ID (Manitoba Health, Annual
Statistics Report)
12
Goal 1, Continued
B. Drug-related QIs

• of PCH residents who were dispensed
• 9 drugs (polypharmacy)
• Benzodiazepines
• Antipsychotics
• Beers criteria medications

Measured just before after people were
admitted to a PCH
13
of residents taking 9 DIFFERENT DRUGS, by
PCH Data adjusted for differences in resident
characteristics
14.4, 22.6, (57.2 ?)
21.0, 18.6, (11.5 ?)
14
of residents taking ANTIPSYCHOTICS, by PCH Data
adjusted for differences in resident
characteristics
21.0, 44.3, (110.8 ?)
26.5, 44.1, (66.5 ?)
15
Summary of PCHs ABOVE THE (90 Threshold
(combined data for 4 drug QIs)
Potential problems may exist in these PCHs!
16
Goal 2
Why do QIs occur more often for some PCHs?
17
Trends in Results
Diagnostic QIs
Drug-related QIs
Facility-level

• PCH size
• PCH type
• Staff-to-resident ratios
• Contact bias

• PCH type
• Staff-to-resident ratios

Variable influenced how often QIs occurred
Resident-level

• close to admit/death
• Age
• Sex
• Level of care
• Marital status
• Dementia
• Chronic illness

• Age
• Sex
• Level of care
• Dementia
• Chronic illness
• 2 doc prescriber

18
Key Messages

• Identify PCHs where QIs occurred most
  frequently.
• Often WRHA proprietary PCHs for diagnostic QIs.
• Often rural PCHs for drug-related QIs.
• Good news stories also exist!!
• PCHs ranked lt 10th for 2 QIs.

• Knowledge of when QIs are more likely to occur.
• Mostly influenced by resident- level risk
  factors (GOOD NEWS).
• Similar QI rates for most Facility-level
  risk factors
• - Exception (some) WRHA proprietary PCHs.

19
Manitoba Centre for Health Policy
http//www.umanitoba.ca/centres/mchp/reports.htm
20

• Chapters 2 and 4
• Fig 2.1 Aging trends
• Fig 2.2 Map PCHs Non-Wpg RHAs
• Fig 2.3 Map PCHs WRHA
• Fig 4.1 Guidelines diagnostic QIs
• Fig 4.2 Strategy/ Reporting Rx related QIs
  before/after PCH admit

21
Chapter 5

• Table 5.1 Distribution PCH facilities and beds
• Table 5.2 Beds per PCH
• Table 5.3 Differences in PCH type by RHA
• Fig 5.1 Staffing hours worked per resident-day
• Table 5.4 PCH residents, admission, and deaths
  over 5 yrs
• Fig 5.2 residents admitted and died
• Fig 5.3 PCH admitted by age category
• Fig 5.4 PCH residents by age category

• Fig 5.5 Female to male ratio of PCH residents by
  age category
• Fig 5.6 admit PCH residents by level of care
• Fig 5.7 total PCH residents by level of care
• Fig 5.8 residents diagnosed with dementia
• Fig 5.9 residents with 2 or more chronic
  diseases
• Fig 5.10 residents 75 who were married
• Fig 5.11 Residential fees paid by PCH resident

22
Chapter 6

• Fig 6.1 Format of figures for diagnostic QIs for
  RHAs
• Fig 6.2Format of figures for Diagnostic QIs for
  PCHs
• Table 6.1 Diagnostic QIs by RHA
• Table 6.2 PCHs ranked below 10th percentile for
  1 or more QIs
• Table 6.3 PCHs ranked above 10th percentile for
  1 or more QIs
• Fig 6.3 Hip fractures by RHA
• Fig 6.4 Hip fractures by PCH
• Fig 6.5 Non-hip fractures by RHA
• Fig 6.6 Non-hip fractures by PCH

Fig 6.7 Accidental falls by RHA Fig 6.8
Accidental falls by PCH Fig 6.9 Skin ulcer rates
by RHA Fig 6.10 Skin ulcer rates by PCH Fig
6.11 Respiratory infections by RHA Fig 6.12
Respiratory infections by PCH Fig 6.13 Fluid and
electrolyte imbalances by RHA Fig 6.14 Fluid and
electrolyte imbalances by PCH
23
Chapter 7

• Table 7.1(pg 1) RHA summary results
  Polypharmacy
• Table 7.1 (pg 2)RHA summary results
  antipsychotics
• Table 7.1 (pg 3) RHA summary results -
  benzodiazepines
• Table 7.1 (pg 4) RHA summary results Beers
  criteria
• Table 7.2 PCHs ranked below 10th percentile for
  1 or more Rx related QIs
• Table 7.3 PCHs ranked above the 90th percentile
  for 1 or more Rx related QIs
• Table 7.4 PCH residents dispensed drugs prior
  and after admittance

Table 7.6 Antipsychotics dispensed 91-190 days
after admittance Table 7.7 Beers criteria
medications Table 7.8 Beers criteria
medications 91-190 days after admission Fig 7.1
Format of figures for Rx related QIs Figure 7.2
9 or more different drugs before and after
admittance by RHA Figure 7.3 9 or more different
drugs 91-190 days after admittance by PCH Table
7.5 Benzodiazepines dispensed 91-190 days after
admittance
24
Chapter 8

Table 8.1(pg 1) Diagnostic QIs risk factors
summary PCH level characteristics Table 8.1 (pg
2) Diagnostic QIs risk factors summary
Resident level characteristics Table 8.2 (pg 1)
Rx related QIs risk factors PCH level
characteristics Table 8.2 (pg 2) Rx related QIs
risk factors Resident level characteristics Tabl
e 8.3 Diagnostic QIs summary of results of
multivariate analyses Table 8.4 Dispensing of QI
drugs summary of results of multivariate analyses
Chapter 7, continued

• Fig 7.4 Benzodiazepine dispensing before and
  after admittance by RHA
• Fig 7.5 Benzodiazepine dispensing 91-190 days
  after admittance by PCH
• Fig 7.6 Antipsychotics dispensed before and
  after admittance
• Fig 7.7 Antipsychotics dispensed 91-190 days
  after admittance by PCH
• Fig 7.8 Beers criteria meds dispensed before
  and after admittance
• Fig 7.9 Beers criteria meds 91-190 days after
  admittance

25
Appendix C
Appendix D

• Table D.1 Modelling steps to predict variation
  in diagnostic QIs
• Table D.2 Crude count of diagnostic QI events
• Table D.3 Trends in results for diagnostic QIs
  skin ulcers
• Table D.4 Trends in results for diagnostic QIs
  hip fractures
• Table D.5 Trends in results for diagnostic QIs
  non-hip fractures
• Table D.6 Trends in results for diagnostic QIs
  accidental falls
• Table D.7 Trends in results for diagnostic QIs
  respiratory infections
• Table D.8 Trends in results for diagnostic QIs
  fluid electrolyte imbalances

Table C.l (pg 1) Ambulatory care visit rates per
PCH by Assiniboine RHA Table C.1 (pg 2)
Ambulatory care visit rates per PCH by Brandon
and Central RHA Table C.1 (pg 3) Ambulatory care
visit rates per PCH by Interlake and North
Eastman RHA Table C.1 (pg 4) Ambulatory care
visit rates per PCH by Parkland and South Eastman
RHA Table C.1 (pg 5) Ambulatory care visit rates
per PCH by WRHA proprietary PCHs Table C.1 (pg
6) Ambulatory care visit rates per PCH by WRHA
Non-Proprietary PCHs
26
Appendix E

• Table E.1 Modelling steps used to predict
  variation in drug-related QIs
• Table E.2 Proportion or residents who were
  dispensed QI-drugs
• Table E.3 Results of multivariate modelling
  polypharmacy medications
• Table E.4 Results of multivariate modelling
  Beers criteria medications
• Table E.5 Results of multivariate modelling
  benzodiazepines
• Table E.6 Results of multivariate modelling -
  antipsychotics