Committee on Ways and Means, Subcommittee on Social Security, 6-11-02 Testimony

Statement of David C. Stapleton, Ph.D., Director,
Cornell Center for Policy Research

Before the Subcommittee on Social Security,
House Committee on Ways and Means

Hearing on Social Security Disability Programs' Challenges and Opportunities

June 11, 2002

This briefing paper is largely based on the knowledge I gained through work that was performed under contracts from the Social Security Administration to The Lewin Group and its subcontractors, Cornell University and Pugh Ettinger McCarthy Associates, LLC. I gratefully acknowledge the contributions of individuals in those organizations to the information contained herein and to my thinking on this subject. The views expressed in this briefing paper and my oral remarks to the Subcommittee are, however, my own, and do not represent the views of the Social Security Administration, Cornell University, or Pugh Ettinger McCarthy Associates.

Directions for Improvements to the Social Security Administration’s Disability Determination Process

I. Introduction

I am an economist, with strong training and experience in the conduct of quantitative evaluations of social service programs. The bulk of my professional career since 1993 has been devoted to work on projects concerning the Social Security Administration’s (SSA’s) disability programs, Social Security Disability Insurance (DI) and Supplemental Security Income (SSI), and other disability programs and policies. My knowledge of the disability determination process comes from assistance that I provided to the SSA in evaluation of three of its efforts to improve the determination process -- the Full Process Model (FPM), the Disability Claims Manager model (DCM), and the Prototype – and from a project to review and develop options for improvements to the Quality Assurance process for disability determinations. The product of the latter was a voluminous report, which included extensive background material, findings from our analysis, and numerous options for improvement.^[1] My Cornell colleague Gina Livermore co-led the first three of these efforts, and Michael Pugh of Pugh Ettinger McCarthy Associates, LLC co-led the last.

This paper focuses on the initial determination process for all types of applications and continuing disability reviews (CDRs) (i.e., everything up to appeals to the Office of Hearings and Appeals), reflecting the Subcommittee’s interest for this hearing. Some material concerning potential changes to appeals above this level is introduced because of its relevance to the initial determination process.

I begin by briefly summarizing my own conclusions about the performance of the disability determination process. I then describe challenges to improving the process that are posed by the nature of the determinations being made. It is critical to keep these challenges in mind as improvements are developed and implemented. I then describe five changes to the initial process that I believe are key to achievement of significant process improvements.

II. Performance of the Disability Determination Process is Poor

Performance of the process is poor. Mean processing times at the initial level are very long. Approximately one-third of all allowances are made to applicants whose applications are initially denied, including many applications that are allowed on the basis of the supporting evidence collected at the initial level. As a result, the administrative cost of appeals is very high. The cost to applicants whose awards are delayed, and who are often very vulnerable, is perhaps greater, although it is not a line item in the federal budget. Worse yet is the cost to an unknown number of applicants who should be found eligible, but whose claims are initially denied and who fail to appeal.

There is also substantial evidence that eligibility criteria are applied inconsistently across state Disability Determination Services (DDSs), and even across disability examiners within DDSs, although SSA does not currently have a good system to document it. We found one convincing indicator by looking at the spread of initial allowance rates across examiners within four DDSs. In each DDS, the examiners included were experienced, made several hundred decisions per year, and were drawing applications randomly from the same queue. We used statistical methods to calculate that the luck of the draw would lead to a spread in allowance rates across examiners of about 10 percentage points within each of the DDSs, but we found that the actual spread in allowance rates was 10 to 19 percentage points greater. Thus, it appears that the examiner in a DDS that happens to get assigned to a claim could affect the chance of that claim’s allowance by 10 to 19 percentage points. It is not hard to imagine that differences across DDSs are much greater, because of differences in management, training, examiner qualifications, expert consultants, the health care system in the state, and other factors.

Other evidence indicates that decisions are significantly influenced by incentives. Dramatic evidence of this occurred in the early 1980s. At that time, SSA, at the direction of Congress, started conducting pre-effectuation reviews of a large share of initial awards in response to low allowance accuracy identified through the quality assurance process; denial accuracy was much better. Over three years, allowance accuracy increased substantially, but denial accuracy fell by essentially the same amount. The most obvious explanation for this reversal in the two accuracy rates is that the pre-effectuation review process made it more costly for examiners to allow error-prone cases than to the deny them, tipping the balance in favor of denials. Although cause and effect cannot be proved, I have not heard a more convincing explanation.

People involved in the process are disgruntled about it. Applicants, adjudicators, managers, union leadership and others all have complaints that are difficult to dismiss. DDSs complain about the SSA field offices, and vice versa. Morale is low, leading to high turnover in many offices.

Some would say that the process is also very costly, but what is the benchmark against which we can say costs are high or low? In 1999, as near as I can tell, SSA spent about $1,400 per application adjudicated – including appeal costs, quality assurance costs, and any other costs that SSA associates with applications. The actuaries estimate that the present value of the average SSDI award, including Medicare benefits, is on the order of $100,000. Is $1,400 too much to be spending when this much money is at stake? One private disability insurer told us they were spending $2,400 per application when the amount at stake for the average claim was much lower. Putting more resources into the current process in a reasonably judicious manner, without any other changes, would clearly improve other aspects of performance, but it also appears that substantial improvements in performance could be achieved more efficiently by other means.

III. The Nature of Disability Determination Poses Significant Challenges to Performance Improvements

Several features of disability determinations are significant challenges to improving process performance. The first is the highly complex and diverse nature of the medical, vocational, legal, and financial issues involved in making determinations. An impressive array of knowledge is required to make appropriate decisions in all cases.

The second is the high level of subjectivity involved in many decisions – assessments of severity of functional limitations, the credibility of evidence about pain and other symptoms, the weight to be given to source evidence versus a medical examiner’s evidence, and the value of seeking medical evidence of record from certain providers, are all examples. It appears that there is room for well-trained, well-intended examiners to disagree on the “correct” decision in a substantial number of cases.

The third challenge is that the applicant has a substantial incentive – the potential benefit award - to mislead the adjudicator. The applicant’s supporter, including the applicant’s provider, might share that incentive.

The fourth challenge is that the program exists within an agency that, despite its fairly recent independence, is subject to the vicissitudes of the political process. Congressmen and governors attempt to influence the program in the interests of their constituents. While this can be a positive force for performance improvement in some circumstances, it can undermine improvement in others. Unions, professional organizations, and other groups can be expected to resist changes that are not in the interest of their members.

The fifth challenge is that major change takes time. The number of years needed is likely longer than a Commissioner’s six-year term. Major change also takes extra resources during the transition period, and is likely to crowd out other agency priorities.

Perhaps the most significant challenge is that the concept of disability in the programs is out-of-step with current thinking. Advocates, researchers, policymakers, and others have embraced the idea that the inability to work results from the interaction of impairment, or functional limitations, with a person’s environment, yet the program is required to make determinations on the basis of the old notion that disability is “medically determinable.” Policy is changing in ways that reflect the new thinking, albeit very slowly, and the determination process, as well as the broader operations of the disability programs, will need to change with it.

IV. Five Key Changes to Achieve Significant Process Improvements

There is much that can be done to produce small improvements to the initial determination process, but significant improvements require major change. I have developed a list of five changes that I think are key to making significant process improvements.

1. Appoint a disability czar: A single person who reports to the Commissioner must be given lead responsibility for all aspects of the disability programs. The recommended change follows from the maxim that “If it’s everybody’s responsibility, it’s nobody’s responsibility.” Currently, responsibility for the program is divided among the Deputy Commissioners, each Deputy Commissioner has responsibilities that go beyond the disability programs, and all interact with each other as equals. The Agency needs a single person whose sole responsibility is the disability programs, who has the authority needed to improve cooperation and coordination among the offices, and who is accountable to the Commissioner for all aspects of disability program performance.

2. Develop a modern performance management system. The thrust of our quality assurance report is that SSA needs to develop a modern performance management system for the disability programs, including the determination process. Such a system will require improved information technology, improved performance measurement, consistent use of performance score cards in everyday decisions, introduction of significant performance incentives, and use of management techniques that build a “quality culture” and support continuous improvement. The job of the disability czar would be to develop and lead this system.

3. Assign responsibility for each initial determination to a single office: This recommendation change also follows from the maxim that “If it’s everybody’s responsibility, it’s nobody’s responsibility.” Currently, SSA Field Offices and state DDSs share responsibility for disability determinations. The result is a system in which no office, let alone individual, takes ownership of an application, each blames the other for process problems, and inter-office communications delay the process, use significant resources, and result in duplication of effort.

The substantial success of the Disability Claims Manager model demonstrates the gains to be made from single-office responsibility. I think the DCM went too far, by giving a single individual responsibility for adjudicating all aspects of all adult applications, but in so doing, it consolidated responsibility into a single office and gave a single individual ownership of each application. SSA’s evaluation of the DCM test concluded that it substantially reduced processing time, increased claimant satisfaction, and improved employee satisfaction, but at a somewhat greater cost than the current process. My interpretation of the evidence from that evaluation is more positive than SSA’s; it appears to me that the DCM is cost neutral, and that it reduced processing time by more than the report indicates.

That SSA has decided not to pursue the DCM, despite the considerable success of the test, reflects the most vexing question about assignment of responsibility into a single office: Whose office? There are three options, and all have strengths and weakness. The first is to federalize the entire process, so that all decisions are made in field offices. This approach is favored by American Federation of Government Employees (AFGE), and likely has substantial support among SSA managers, but it is probably the most costly and it is also difficult to imagine a federal bureaucracy managing such a large internal process well. The determination process for Veterans’ benefits provides a useful model.

The second approach is to give the responsibility to the states. This might be the least expensive approach, would provide a process that is more tailored to the needs of the state, and make state governments accountable to their voters for process performance. SSA’s past management of the state DDSs has, however, been hampered by lack of political will; SSA does not have sufficient power to hold states accountable. Perhaps a deal that gives the entire process to the states in exchange for provisions that will make it possible for SSA to hold states accountable could be developed. Such a system already exists in the Food Stamp program, although it is imperfect.

The final approach is to contract the work to the private sector. Many would oppose such a change because of a fundamental distrust of entities that are driven by the profit motive. SSA might, however, be in a stronger position to ensure performance of local entities that are outside the political process than local entities that are either owned by states or internal to SSA. This approach also has the advantage that a successful firm could make determinations in multiple states. Every participating firm would be continuously threatened by competition from firms that are operating in other states. The Center for Medicare and Medicaid Services uses this approach in its effort to improve health care quality, with some success.

I’m not ready to endorse any one of these approaches over the others, but I think one of these approaches will have to be tried if significant progress is to be made.

4. Abandon the “one-adjudicator-fits-all” model. Currently, each medical determination is the responsibility of just one disability examiner. While expert consultation is available, and required in some cases, the decision rests on the shoulders of an individual. This approach is probably efficient for a large majority of cases, but the complexity of the medical, vocational, and legal issues involved in a significant number of cases requires a team approach. Such an approach is used for complex determinations conducted by private disability insurers, as well as for many other complex determinations, including medical ones.

5. Address the problems with the appeals process. Currently, the appeals process has significant performance problems of its own. Those problems undermine the initial process. The fact that the program is represented only by conflicted Administrative Law Judges (ALJs), and that performance management is limited, make it nearly impossible for SSA to obtain useful information from the appeals process and use it to improve the initial process. Instead, defenders of the initial process allege that ALJs are “out of control” and are undermining the initial process. Others allege that applicants, with the help of their representatives, game the system by withholding evidence at the initial level so they can present it to an ALJ, and by continuing to file whole new applications until they find an ALJ who will give them an award. Given the lack of information about the decisions at the appeals level, it is hard to know where the truth lies, but it is clear that the initial process is being undermined by the lack of trust in the appeals process.

[1] See D.C. Stapleton and M.D. Pugh, Evaluation of SSA’s Disability Quality Assurance (QA) Processes and Development of QA Options that will Support the Long-term Management of the Disability Programs, to the Social Security Administration, 2001. http://www.quintiles.com/products_and_services/specialty_consulting/ the_lewin_group/lewin_publications/detail/1,1278,213,00.html.