Transcript Cognitive Decline, Disability, and Other Unappreciated

Clinical Practice Guidelines:
Implications for Vulnerable Patients
Development of Geriatric Diabetes Guidelines
Arleen F. Brown, MD, PhD
Associate Professor of Medicine
Division of GIM and HSR
UCLA, Los Angeles, CA
Outline
• Challenges in developing and disseminating guidelines
meaningful for the care of vulnerable patients
• Example of geriatric diabetes guideline development
– Strategies for identifying and addressing limitations
of the literature
• Examples of RCTs that have been used to develop care
practice recommendations for vulnerable populations
• Recommendations for improving the “trustworthiness” of
clinical practice guidelines
Challenges in Developing / Disseminating
Clinical Practice Guidelines
Pertinent to Vulnerable Populations
• Lack of inclusion of “typical” patients in many RCTs and some high quality
observational studies
– Clinically dissmilar
• e.g., new onset disease; no/few comorbid conditions
– Demographically dissimilar
• Under-representation of vulnerable subgroups of patients
– Older persons
– Racial/ethnic minorities
– Low income / education / literacy
• Extrapolation from existing data is often required
– “Double” or “triple” extrapolation
• Where minority or low income patients receive care
Diagnosed Diabetes – Standardized Prevalence
Diabetes Affects Older Persons and Racial/Ethnic Minorities
Non-Hispanic
Whites
Non-Hispanic
Blacks
Mexican
Americans
• Men
5.5%
9.6%
11.0%
• Women
3.7%
12.7%
12.0%
• Men
14.3%
29.2%
25.6%
• Women
14.3%
28.0%
24.3%
40-59 years
> 65 years
* NHANES 1999-2002, Cowie CC et al.. Diabetes Care 29(6):1263-1268, 2006
Prevalence (%) of Diagnosed and Undiagnosed Diabetes and
Impaired Fasting Glucose (IFG)
Among Adults, Aged 65+ years*
16% Diagnosed
6% Undiagnosed
40% IFG
39% All others
* NHANES 1999-2002, Cowie CC et al.. Diabetes Care 29(6):1263-1268, 2006
~ 6 in 10
CHCF/AGS Geriatric Diabetes
Guideline Development Process
• Synthesized and evaluated results from randomized
controlled trials and observational studies
• Reviewed existing guidelines
• Rated the evidence and guidelines with validated
consensus panel methods
• Modified existing guidelines and developed new
guidelines specific to older persons with diabetes
• Peer reviewed
JAGS, 51:S265-S280, 2003
CHCF/AGS Geriatric Diabetes
Guideline Development Process
• Synthesized and evaluated results from randomized
controlled trials and observational studies
• Reviewed existing guidelines
• Rated the evidence and guidelines with validated
consensus panel methods
• Modified existing guidelines and developed new
guidelines specific to older persons with diabetes
• Peer reviewed
JAGS, 51:S265-S280, 2003
Development of Care Recommendations
Required Extrapolation
• Very little research directed at older, minority adults with
diabetes
• Required extrapolation from studies of:
– Older adults in the general population
– Younger persons with diabetes
– Minority adults with diabetes
– Older minority adults with diabetes
• Developed evidence tables that indicated
– whether older persons / persons with diabetes were included
in the original studies
– estimated the effect size / number needed to treat (NNT) for
older persons with diabetes
Randomized Controlled Trials that Included
Older Adults with Diabetes
Included
Subgroup
Presented
Age-Stratified
Older
Analyses by Age-Stratified
Results
Adults with
Age
Results
Presented by
Diabetes
Race/Ethnicity
Glycemia
4
0
----
----
Blood Pressure
9
5
2
0
Lipid
9
3
0
----
Aspirin
4
1
0
----
CHCF/AGS Guidelines, 2003
Why We Cannot Always Extrapolate RCT Findings
to Older, Minority Adults with Diabetes
• Clinical Heterogeneity
– Comorbid conditions – variation between racial/ethnic groups
– Functional status, Cognitive status
– Geriatric Syndromes more common in older adults with diabetes
• Polypharmacy: Drug-drug or Drug-disease interactions
• Depression
• Cognitive Decline
• Injurious Falls
• Life expectancy in relation to
– time to incidence or progression of \ complications
– time to expected benefit of intervention
• Factors that influence uptake of therapies among patients / clinicians
– Patient preferences / Cultural factors
– Socioeconomic factors
Diabetes Prevention Program (DPP)
• N=3234
• Mean age 50.6 years (10.7), 20% > 60 years
• White 54.7%; African American 19.9%; Latino 15.7%; American
Indian 5.3%; Asian / Pacific Islander 4.4%
•
Treatment effects varied by age, but not race/ethnicity:
Placebo
Metformin
Lifestyle
Modification
11.0%
6.8%
4.8%
Reduction in incidence (vs. placebo)
----
31%
58%
• 25-44 years
----
44%
48%
• 45-59 years
----
31%
59%
• > 60 years
----
11%
71%
Incidence of T2DM (% per year)
Knowler, NEJM, 2002
ACCORD Study
Action to Control Cardiovascular Risk in Diabetes
• 10,251 patients
– Mean age 62.2 years (33.9% > 65 years)
– 64.4% White, 19.7% Black, 4.9% Latino
• Conclusions:
– Intensive therapy (Goal A1c < 6.0%) for 3.5 years:
• No reduction in CVD events
• Higher all-cause mortality
• Higher rates of other serious adverse events
– Hypoglycemic and non-hypoglycemic)
• Findings did not vary by race/ethnicity or age
ACCORD Study Group, NEJM; 358:24.
BiDil
- BiDil (hydralazine+isosobide dinitrate)
- Not efficacious in V-HeFT Trials
- Post hoc subgroup analysis suggested greater efficacy in blacks
- A-HeFT - BiDiL reduced mortality in African-American patients
with advanced heart failure. No racial/ethnic comparison group.
- Controversial departure from usual practice
- FDA’s stated purpose was to reduce disparities
- Used disparities reduction to “create” an expensive “new”
medication
- Incorporated into the AHA/ACC guidelines for symptomatic African
American patients, with caveats that race is “imprecise concept” and
that others may benefit.
Recommendations for Improving the
“Trustworthiness” of Clinical Practice Guidelines
• Improve the quality and scope of the evidence
– Increased representation of racial/ethnic minority, older, and other
potentially vulnerable patients
– Rating (or weighting) recommendations to indicate the
representativeness of the RCT evidence
• Obtain evidence in “real world” settings to improve the feasibility of
implementing the guideline in heterogeneous clinical settings
• Assist clinicians with understanding the likely effect size (e.g. use of NNT)
of a proposed intervention for important subgroups
• Incorporate time horizon for different subgroups (e.g. time to benefit vs.
longevity)
• Address patient burden – disproportionate effect on vulnerable subgroups
– Cost, polypharmacy, competing demands
• Address patient preferences
• Address quality of life
Time Needed to Benefit
Microvascular
Complications
(Median Years)
Control of:
Glycemia
Blood Pressure
Lipids
4.5
4.5
--
Macrovascular
Complications
(Median Years)
10
3
3 to 6
Polypharmacy
• Several medications for diabetes + Additional medications for
comorbid conditions
• Polypharmacy may contribute to or exacerbate several other geriatric
syndromes such as depression, cognitive decline, and injurious falls
• Quality of life
• Costs of medical care may be prohibitive for elders on fixed incomes
Number of Prescription Medications
Used by Older Adults with Diabetes
25
20
15
12
10
5
12
13
13
14
12
8
6
5
3
0
1
2
3
4
5
6
7
8
Number of Prescription Medications
9
10+
Clinical Recommendations
• Screen for physical and cognitive disability
– Look for easily reversible causes of disability (e.g.
uncorrected visual impairment, untreated depression)
•
•
•
•
Treat hypertension first
Treat lipids second
Aspirin
Screen for evidence of microvascular disease
– For those with microvascular disease and good functional
status, apply the younger age targets for glycemia
– For everyone else, clinical judgment and patient preference
should drive choices in the absence of evidence
• Consider costs
Number Needed to Treat (NNT)
to Prevent One Event
DM
Endpts
DM
Deaths
MI
CHD
Events
CVA
Deaths
All-cause
Mortality
Glucose
Control1
31* 111
46
- 172
125
HTN
Treatment2
11* 20
29
27*
28
Lipid Rx (1o)3
6* - 49
Lipid Rx (2o)3
5* - 13*
1 UKPDS
* p<0.05
149*
33; 2 UKPDS 38; 3 RCTs of lipid management with diabetes subgroup analyses
32
Prevalence (%) of Diagnosed and Undiagnosed Diabetes and
Impaired Fasting Glucose (IFG)
Among Adults, Aged 65+ years*
16% Diagnosed
6% Undiagnosed
40% IFG
39% All others
* NHANES 1999-2002, Cowie CC et al.. Diabetes Care 29(6):1263-1268, 2006
~ 6 in 10