The CAHPS Macro Program

Slide 1

Text Description is shown below image.

The CAHPS Macro Program

Alan Zaslavsky
Kayo Walsh, M.S. Data Manager/Programmer
Department of Health Care Policy
Harvard Medical School

Slide 2

Text Description is shown below image.

Basics of Macro

Analyzes multi-item composites and single items from CAHPS survey data.
Compares the performance of an entity to the average of all entities:
- Entity – hospitals, clinics, health plans, etc.
Accommodates weighted and stratified data
Produces (SAS):
- Scores for each entity
- Case mix adjustment/coefficients
- Standard errors and hypothesis tests
- Tabulations for bar charts

Slide 3

Text Description is shown below image.

Data Input and Output

Ex. Rating of Care item (0 – 10 scale), age and health status as adjusters

Plan name	Rating of Care	Age	Health status
A	8	65	4
A	7	77	3
A	9	74	5
B	10	66	4
B	8	64	4
B	7	70	3

Macro output

Plan name	Unadjusted mean	Adjusted mean	n used for analysis	Sample size	Variance of mean	Significance level (above mean= 3, below mean= 1)	Adjusted frequency (0 -6)	Adjusted frequency (7-8)	Adjusted frequency (9-10)
A	8.16	8.46	212	218	0.021212	2	15%	23%	62%
B	9.09	9.04	325	338	0.007142	3	7%	16%	77%

Slide 4

Text Description is shown below image.

Macro Setting and Options

Main Options	Description
VAR	Name of variable
NAME	Label of variable
VARTYPE	Type of variable (1 = 1/0, 2 = 0-10, 3 = 1-4, 4 = 1-3,. 5 = other
ADJUSTER	Casemix adjuster variables
IMPUTE	1 = impute casemix adjuster, 0 = no
WGTRESP	Name of weight variable
DATASET	Name of input dataset
OUTNAME	Name of output dataset

The list of the options can be found in 2015_Instructions_for_Analyzing_Data document.

Slide 5

Text Description is shown below image.

Macro Setting and Options

Example

%cahps(
VAR = rate_care,
NAME = Rating of Care,
VARTYPE = 2 ,
ADJUSTER = age ghs ,
IMPUTE = 1,
WGTRESP = plan_weight,
DATASET = Madata ,
OUTNAME = rate_care
);

Slide 6

Text Description is shown below image.

Imputation Option for Case Mix Variables

Images: Two sample tables showing Original data and Using &Impute option.The differences are in Adj1 and Adj2 variables for Plan A Question 1 4.

Slide 7

Text Description is shown below image.

Composite Calculation

Composite score = (2.5 + 3.33 + 2.63)/ 3 = 2.82

Image: Sample table showing how the composite score is calculated.

Slide 8

Text Description is shown below image.

Current Status – CAHPS Macro 4.1

Macro_cahps41.sas: http://ahrq.gov/cahps/surveys-guidance/cg/instructions/12monthsurvey.html
Improvement in 4.1:
- (Option) Automated weights reduction for items with disproportionately small samples
  - Composites – even weight option has been improved
- (Option) Variance smoothing
  - To avoid a very small standard error(almost 0) due to a small sample size with all equal scores

Slide 9

Text Description is shown below image.

Current Status and Development

Ongoing project:
- User comments and requests
- Bug fixes and code improvements
- Improved capabilities to support new CAHPS applications:
  - Include reliability calculation
  - Automatic calculation of smoothing parameters
  - Weighting options for between and within plan weights
  - R Macro version (free software)

Slide 10

Text Description is shown below image.

Questions?

Return to 2015 AHRQ Research Conference page