CASE : U.S. GENERAL ACCOUNTING OFFICE
The U.S. General Accounting Office (GAO) is an independent, nonpolitical audit organization in the legislative branch of the federal government. The GAO was created by the Budget and Accounting Act of 1921 and has three basic purposes:
1. To assist Congress, its committees, and its members in carrying out their legislative and oversight responsibilities, consistent with its role as an independent, nonpolitical agency.
2. To audit and evaluate the programs, activities, and financial operations of federal departments and agencies and to make recommendations toward more efficient and effective operations.
3. To carry out financial control and other functions with respect to federal government programs and operations including accounting, legal, and claims settlement work.
GAO evaluators, in fulfilling the main role in the GAO, determine the effectiveness of existing or proposed federal programs and the efficiency, economy, legality, and effectiveness with which federal agencies carry out their responsibilities. These evaluations culminate in reports to the Congress and to the heads of federal departments and agencies. Such reports typically include recommendations to Congress concerning the need for enabling or remedial legislation and suggestions to agencies concerning the need for changes in programs or operations to improve their economy, efficiency, and effectiveness.
GAO evaluators analyze policies and practices and the use of resources within and among federal programs, identify problem areas and deficiencies in meeting program goals, develop and analyze alternative solutions to problems of program execution, and develop and recommend changes to enable the programs to better conform to congressional goals and legislative intent. To effectively carry out their duties, evaluators must be proficient in interviewing, data processing, records review, legislative research, quantitative methods, and statistical analysis techniques.
Impact of Services on the Well-being of Older People
GAO evaluators obtained data from a random sample of noninstitutionalized persons aged 65 and older living in Cleveland, Ohio. The health conditions of the sampled individuals in the 65- to 69-year-old groups were defined by the following three states:
- Best. Individual able to perform 13 identified activities of daily living without help.
- Next best. Individual able to perform the same 13 activities but required help for at least one activity.
- Worst. Individual unable to perform the same 13 activities even with help.
Using a 2-year period, GAO evaluators developed estimates of the year-to-year transition probabilities of individuals in the 65- to 69-year-old group. These estimates were then used to develop a transition probability matrix such as shown in the following table. Note that a death state has been added as an absorbing state.
|
Current
|
|
|
|
|
|
Year
|
Following Year Condition
|
|
Condition
|
Best
|
Next Best
|
Worst
|
Death
|
|
Best
|
P11
|
P12
|
P13
|
P14
|
|
Next Best
|
P21
|
P22
|
P23
|
P24
|
|
Worst
|
P31
|
P32
|
P33
|
P34
|
|
Death
|
0
|
0
|
0
|
1
|
Using the transition probabilities, a Markov process analysis can be used to determine the state probabilities for any number of periods (years) into the future. To verify the appropriateness of the Markov process model, GAO evaluators used the transition probabilities to determine the state probabilities for the 65- to 69-year-old age group 5 years into the future. The resulting state probabilities were compared with the health states of individuals in a known 70- to 74-year-old age group. No statistically significant difference occurred between the probabilities provided by the model and the actual state probabilities of the 70- to 74-year-old group.
Estimating the Likely Effects of Health Care Programs
He individuals in the original study were subdivided into two groups: those receiving appropriate health care and those not receiving appropriate health care. For the purpose of the study, a person was classified as receiving appropriate health care if the person was taking medication and/or treatment for each illness present and if the person was receiving help to perform each of the 13 activities of daily living as specifically needed. As a result, GAO evaluators developed two matrixes of transition probabilities: one for individuals receiving appropriate health care and one for individuals not receiving appropriate health care.
For any individuals not receiving appropriate help, the kind of additional help needed was determined, and the cost of that help was estimated. Then those persons were artificially aged using the transition probabilities to establish the likely benefits in terms of improved health states for the individuals. Over a 20-year period, a net savings occurred for the health care program provided all other factors remained equal. That is, the increased cost to provide sufficient appropriate help to all persons was eventually offset by individuals either improving their health state or by not spending as much time in a worse state. Although benefits of health care are often proclaimed theoretically, the Markov process model provided evidence that indicated benefits would be achieved with the health care program.
This type of Markov analysis was also conducted for economic, social, and life view status as well as for the health status. In some instances, the Markov model showed that additional help and/or programs did not result in net savings over time.
Questions
1. Suppose the transition probabilities in the table were as follows:
|
0.80
0.05
0.00
0.00
|
0.10
0.75
0.05
0.00
|
0.06
0.15
0.75
0.00
|
0.04
0.05
0.20
1.00
|
Assume that a particular city has 1000 individuals in the best state, 2000 in the next best state, and 500 in the worst state. Estimate how many of each of these individuals will be in each state 2 years from now.
2. How might health care programs affect the matrix of transition probabilities in Question 1? What effect would you expect to see in the distribution of individuals across the four states 2 years from now?