oversight

Social Security: SSA Could Save Millions by Targeting Reviews of State Disability Decisions

Published by the Government Accountability Office on 1990-03-05.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

                United   States   General   Accounting   Office
                Report to Congressional Committees
-GAO

March   1990
                SOCIAL SECURITY
                SSA Could Save
                Millions by Targeting
                Reviews of State
                Disability Decisions




GAO/HRD-90-28
Human Resources Division

B-238212

March 5, 1990

The Honorable Andy Jacobs, Jr.
Chairman, Subcommittee on Social Security
Committee on Ways and Means
House of Representatives

The Honorable Daniel Patrick Moynihan
Chairman, Subcommittee on Social Security
  and Family Policy
Committee on Finance
United States Senate

This report, prepared at our own initiative, is an assessment of the Social Security
Administration’s (w’s) effectiveness in reviewing disability determinations made by state
agencies.

Section 221(c) of the Social Security Act requires SSAto review 65 percent of the state agency
decisions that award or continue disability benefits. SSAcould substantially improve its
performance by using a targeted sample rather than its current random approach. Moreover,
reviews of continuances are relatively unproductive, and these administrative resources
could be better used reviewing initial awards. This would require a revision of section 221(c)
to exclude continuances from the universe of cases that SSA is required to review. The report
also illustrates that targeting the reviews would enable SSAto correct more incorrect awards,
while reviewing fewer cases.

This report is being sent to the Secretary of Health and Human Services and to interested
congressional committees and subcommittees.

Please contact me at (202) 275-6193 if you or your staffs have any questions concerning this
report. Other major contributors to this report are listed in appendix III.




Franklin Frazier
Director, Income Security Issues
  (Disability and Welfare)
Executive Summary


                   The Social Security Administration (%A) spent about $29.2 million in
Purpose            fiscal year 1988 reviewing disability decisions made by state disability
                   determination services (DDSS). These reviews are done primarily to (1)
                   measure DDS performance in meeting accuracy standards and (2) correct
                   as many erroneous benefit allowances as possible. This report evaluates
                   SSA'S effectiveness in accomplishing the latter objective.


                   SSA administers two disability programs under the Social Security Act:
Background         the Disability Insurance program under title II and Supplemental Secur-
                   ity Income for disabled and blind persons under title XVI. For both pro-
                   grams, SSArelies on state agencies (DDSS) to make initial disability
                   determinations on individual claims. The DDSSalso (1) reconsider unfa-
                   vorable decisions if requested by claimants and (2) periodically review
                   the medical condition of persons on the disability rolls (continuing disa-
                   bility reviews) to determine if they are still disabled. SSAfunds the DDSS,
                   provides guidance to them, and reviews a sample of their decisions.

                   This report discusses only SSA’Sreviews of title II decisions made by
                   DDSS.SSAreviewed about 409,000 of the 1.5 million title II decisions DDSS
                   made in fiscal year 1988. About 51,000 of these reviews were classified
                   as quality assurance (QA) reviews, done to determine whether DD~S are
                   meeting standards of accuracy. (See p. 10.) The QA reviews covered both
                   favorable and unfavorable DDS decisions. The remaining 358,000
                   reviews covered only favorable decisions. They were done to satisfy a
                   1980 legislative requirement that SSAreview at least 65 percent of
                   favorable DDS decisions. All reviews are done before the claimant is noti-
                   fied of the decision, but SSAcommonly refers to only the legislatively
                   required reviews as preeffectuation reviews (PERS).


                   SSAselects all review cases randomly. While this is appropriate for the
Results in Brief   &A sample that measures DDS accuracy, the PER sample could produce
                   better results if SSA targeted it to categories of cases most susceptible to
                   incorrect DDS decisions. SSAknows from its QA data that some types of
                   decisions (such as allowances of claims involving back injuries or
                   chronic lung disease) are more difficult for DDSSthan others. If SSA
                   focused its sample on the more difficult (error-prone) types of cases, it
                   could correct more erroneous decisions than it does using a random
                   approach, even with a lower volume of reviews.

                   The current PER reviews of DDS continuances (resulting from continuing
                   disability reviews) change very few DDS decisions. If the resources spent


                   Page 2                        GAO/HID-90-28   SSA Reviews   of State Disability   Decisions
                    Executive   Summary




                    on those reviews were made available for targeted reviews of initial DDS
                    allowances, substantially more incorrect benefit awards would be identi-
                    fied and reversed, with future benefit savings.


                    GAO estimated that SSA'S reviews of favorable DDS decisions in fiscal year
GAO Analysis        1988 will result in long-term net savings of about $69 million-about      $6
                    million from QA reviews and $63 million from PER reviews. PER reviews
                    of initial DDS allowances will result in savings of about $55 million, or
                    $5.18 in reduced benefit payments for each $1 .OOspent reviewing cases.
                    GAO calculates that targeting fiscal year 1988 PER reviews of initial
                    allowances could have increased SSA’Ssavings to $87 million. Because
                    such targeting would involve reviewing the more difficult types of cases,
                    it would require an increase in reviewer and physician resources of
                    about $2.1 million. (See p. 17.)

                    GAO estimated that SSA'S fiscal year 1988 PER reviews of DDS continu-
                    ances will save $0.6 million, or $1.09 in future benefit payments for
                    each $1.00 spent. If SSA had used the resources spent on these reviews
                    for targeted reviews of initial DDS allowances, additional savings of
                    about $33 million could have been obtained. (See p. 18.) However, SSA
                    would need legislative authorization to exclude continuances from the
                    universe of cases that SSA is required to review. Without such authoriza-
                    tion, SSA could reduce its reviews of continuances and shift some
                    resources to its reviews of initial allowances. GAO did not estimate the
                    fiscal impact of such an adjustment.


                    GAO recommends that the Secretary direct SSA to use a targeted sample
Recommendation to   for its PER reviews of initial DDS allowances. While this would require
the Secretary of     some additional review and medical staff, costs would be far exceeded
Health and Human     by the reductions in future benefit payments resulting from the targeted
                     reviews.
Services

                     The Congress should revise section 221(c) of the Social Security Act to
Recommendation to    exclude continuances from the universe of DDS decisions SSA is required
the Congress         to review. SSAcould then limit its reviews of continuances to a quality
                     assurance sample and transfer administrative resources to a more cost-
                     effective targeted review of initial DDf3allowances.




                     Page 3                      GAO/Mu190-28   SSA Reviews   of State Disability   Decisions
                  Executive   Summary




                  The Department of Health and Human Services (HHS) provided written
Agency Comments   comments on a draft of this report. HHS agreed that the PER reviews of
                  initial allowances can be targeted to be more effective, and that the PER
                  reviews of continuing disability reviews are only minimally effective.
                  HHS said that, although it agreed that the relative effectiveness of the
                  PER process could be improved, the PER review should not be limited to
                  initial cases only. HHS added that any legislation in this regard should
                  provide maximum flexibility to direct resources where they are most
                  needed.

                  GAO'S recommendation would increase SSA'S flexibility     by eliminating the
                  current legislative requirement to include continuances in the universe
                  of cases that SSAis required to review. It would not prohibit SSAfrom
                  reviewing continuances or any other types of cases if SSAthought this
                  was needed to improve the accuracy of decisions. Such reviews would be
                  in addition to the legislatively required reviews of initial and reconsider-
                  ation allowances.




                  Page 4                       GAO/IiRD-SO-28   SSA Reviews   of State Disability   Decisions
Page 5   GAO/HlUMO-28   SSA Reviews   of State Disability   Decisions
contents


Executive Summary
Chapter 1                                                                                                        8
Background               Quality Assurance in the Disability Program                                             8
                         Objectives, Scope, and Methodology                                                     11

Chapter 2                                                                                                       13
Targeting Samples and    SSA’s Reviews Have Some Positive Results, but Leave                                    13
                              Many Erroneous Decisions Uncorrected
Reallocating Resources   Insufficient Medical Staff Resources May Reduce the                                    14
Could Improve                 Effectiveness of PER Reviews
Effectiveness of SSA’s   PER Reviews Would Be More Effective If the Sample                                      15
                              Were Targeted
PER Reviews              Resources Used to Review CDR Continuances Could Be                                     18
                              Better Used Reviewing Initial Awards
                         Conclusions                                                                            20
                         Recommendation to the Secretary of Health and Human                                    21
                              Services
                         Recommendation to the Congress                                                         21
                         Agency Comments                                                                        21

Appendixes               Appendix I: Methodology and Tables                                                     22
                         Appendix II: Comments From the Department of Health                                    31
                             and Human Services
                         Appendix III: Major Contributors to This Report                                        33

Tables                   Table 1.1: SSA’s Preeffectuation Reviews (Fiscal Years                                 10
                              1981-88)
                         Table 1.2: SSA Reviews in Fiscal Year 1988                                             11
                         Table 2.1: Costs and Future Benefit Reductions From                                    13
                             SSA’s Fiscal Year 1988 PER Reviews
                         Table 2.2: SSA’s Correction of DDS Decisions in Fiscal                                 14
                             Year 1988
                         Table 2.3: PER Return Rates for Fiscal Year 1988                                       15
                             Compared With QA Return Rates
                         Table 2.4: Net Future Benefit Savings From Alternative                                 17
                             Sizes of Targeted PER Samples
                         Table 2.5: Net Future PER Savings From Alternative Sizes                               20
                             of Targeted Samples Using Reallocated CDR
                             Resources



                         Page 6                      GAO/IiRB96-28   SSA Reviews   of State Disability   Decisions
          Contents




          Table 1.1: Results of GAO’s Study of Claimants Denied                              23
              Benefits by PER Reviews (Jan.-June 1987)
          Table 1.2: Potential Results of Targeting PER Sample                                25
          Table 1.3: Potential Results of Targeting PER Sample                                26
               Using Reallocated CDR Resources
          Table 1.4: Calculation of Future Benefit Savings From                               27
               Targeting PER Sample
          Table 1.5: Calculation of Future Savings From Targeting                             28
               PER Sample Using Reallocated CDR Resources
          Table 1.6: Costs and Net Future Benefit Savings From                                29
               Targeting PER Sample
          Table 1.7: Costs and Future Savings From Targeting PER                              30
               Sample Using Reallocated CDR Resources

Figures   Figure 2.1: Benefit Savings From Targeting PER Sample                               16
               of Initial Allowances
          Figure 2.2: Savings From Targeted Sample Using                                      19
               Reallocated CDR Resources




           Abbreviations

           CDR       continuing disability review
           DDS       disability determination service
           D&B       Disability Quality Branch
           GAO       General Accounting Office
           HHS       Department of Health and Human Services
           PER       preeffectuation review
           Q-4       quality assurance
           SSA       Social Security Administration


           Page 7                     GAO/HRDS@28   SSA Reviews   of State Disability   Decisions
Chapter 1

Background


                         The Social Security Administration (SSA)administers two disability pro-
                         grams under the Social Security Act: the Disability Insurance program
                         under title II and Supplemental Security Income for disabled and blind
                         persons under title XVI. As required by law, SSArelies on state disability
                         determination services (DOSS) to make initial determinations on individ-
                         ual claims. DDGS also handle claimant requests for reconsideration of
                         initial denials and periodically review the status of persons on the disa-
                         bility rolls to determine if medical improvement has occurred. These lat-
                         ter determinations are known as continuing disability reviews (CDRS).

                         SSA funds the DDSS, provides guidance to them, and reviews a sample of
                         their decisions. SSA classifies some of its reviews as quality assurance
                         (QA) reviews, which it uses primarily to measure performance in meeting
                         accuracy standards. Other reviews, called preeffectuation reviews
                         (PERS), are done to satisfy a 1980 legislative requirement that !%A review
                         at least 65 percent of DDS title II favorable decisions (allowances and
                         continuances). SSAcounts the QA reviews of favorable DDS decisions as
                         part of the 65-percent requirement.     l




                         .%Auses a three-tiered quality assurance process to foster accuracy and
Quality Assurance in     consistency in the disability program. DDGS are required to have internal
the Disability Program   quality assurance programs. SSA’S regional Disability Quality Branches
                         review DDS decisions, and ss~ headquarters staff review a sample of the
                         cases examined by the regional branches.

                         States may vary their approaches to quality assurance to suit their par-
                         ticular needs. We visited the Ohio and Indiana DDsS to discuss their inter-
                         nal QA programs. Ohio was randomly reviewing all types of decisions.
                         Indiana was also randomly reviewing all decisions except reconsidera-
                         tions and cases involving mental impairments. Indiana officials said
                         they had stopped reviewing the latter because they were getting few
                         returns of these decisions from SSA.Both DDSS used their internal QA
                         reviews to give accuracy ratings to examiners and examiner units.

                         SSA’Sregional branches review decisions to assign accuracy rates to each
                         DDS. The reviewers are ss~ employees, while SSAcontracts with physi-
                         cians to provide medical consultation to the reviewers. The regional
                         branches return cases to DOSS if they believe the decisions are incorrect
                         or the supporting documentation inadequate. If a DD6 disagrees with
                         !%A’sreasons for returning a case, it may attempt to rebut ~3~‘s position.
                         If the DDS agrees that its decision was deficient, it changes the decision



                         Page 8                       GAO/HIUHO-26   SSA Reviews   of State Disability   Decisions
                            Chapter 1
                            Background




                            or obtains additional evidence to support its original decision. About
                            one-half of the QA returns result in a change of DDS decisions.

                            SSAuses the QA results to determine whether DDSSare at least go-percent
                            accurate in deciding claims and are properly documenting their deci-
                            sions. (Many documentational errors are corrected without changing the
                            decisions.) If a DDS fails to meet the standards for two consecutive
                            quarters, ss~ may conduct a management review and require corrective
                            actions. SSA’SOffice of Disability Program Quality also monitors consis-
                            tency among the regions by reviewing a sample of the QA cases from the
                            regional branches.


Reviews Required by the     In the early and mid-1970s, the title II disability program grew rapidly,
                            straining the resources of the disability trust fund. The program’s
1980 Amendments             growth was a result of high numbers of disability applications, a high
                            approval (allowance) rate by DDSS,and benefit increases. In addition, the
                            advent of the title XVI disability program in 1974 greatly increased DDS
                            caseloads.

                            In an effort to improve the quality and uniformity of DDS decisions, the
                            Congress (in P.L. 265) amended section 221 of the Social Security Act to
                            require SSA to review at least 65 percent of favorable title II DDS deci-
                            sions (allowing or continuing benefits) before the decisions took effect
                            (hence the name preeffectuation reviews, or PERS).


SSA’s Implemen .tation of   %A originally conducted PER reviews on a targeted basis but discontin-
Preeffectuation Reviews     ued targeting once the program was fully implemented. The law pro-
                            vided that PER reviews could be phased in: at least 15 percent in the first
                            year, 35 percent in the second year, and 65 percent thereafter. SSA began
                            in fiscal year 1981 by reviewing certain types of allowances that were
                            more error prone, such as those involving vocational considerations as
                            well as medical conditions. When expanding its sample in fiscal year
                            1982, SSA added certain types of disabilities, such as back ailments, that
                            were considered difficult to evaluate. SSAalso began doing its QA reviews
                            before the decisions took effect, and thus began counting them toward
                            the PER review requirement.

                            At the 15- and 35-percent review levels, SSA was targeting its samples at
                            the more error-prone types of cases. In September 1981, SSA recom-
                            mended to the Congress that the 65-percent level be deferred until it
                            could be determined whether the targeted 35-percent review could
                                                   .

                             Page 9                      GAO/HlUMK&28   SSA Reviews   of State Disability   Decisions
                                   Chapter 1
                                   Background




                                   achieve the results that the Congress was seeking at the 65-percent
                                   review level. When the Congress made no legislative changes, SSA aban-
                                   doned targeting and went to a random 65-percent sample. SSA officials
                                   told us they abandoned targeting because (1) selecting the sample was
                                   becoming burdensome for DDSs, (2) it was easier to estimate regional
                                   staffing needs with a random sample, and (3) targeting was less effec-
                                   tive at the 65-percent review level.

                                   Table 1.1 summarizes the results of W'S PER reviews from 198 1 Lhrough
                                   1988. It includes reviews of title II allowances and continuances. The
                                   table shows that the percentage of DDS decisions reversed by PER reviews
                                   has steadily declined. SSAofficials cited the following reasons for this:
                                   (1) targeted sampling was replaced by random sampling, (2) DDS accu-
                                   racy for allowances has improved, and (3) PER reviews of continuance
                                   decisions have yielded fewer reversals since implementation of the med-
                                   ical improvement standard required by 1984 legislation.’

Table 1.1: SSA’s Preeffectuation
Reviews (Ftscal Years 1981-88)                                             Returned            Percent         Decision             Percent
                                   Fiscal year            Review@            to DDS           returned         reversed           reversed
                                   1981D                    -73,738             6,413                8.7           3,725                 5.1
                                   1982b                     182,824           12,792                7.0           7.294                 40
                                   1983                      285,584           12,299                4.3           6,586                 2.3
                                   1984"                     282,261            8,525                3.0           4,846                 1.7
                                   1985"                     221,983            6,054                2.7           3,297                 15
                                   1986                      262,418            6.566                2.5           3.559                 14
                                   1987                      309,202            6,483                2.1           3,565                 12
                                   1988                      386,150             7,663               2.0            3,540                09
                                   %eglnnlng in 1982, SSA counted QA reviews of favorable declslons In meeting the 65percent
                                   requirement.

                                   bSSA targeted PER samples In 1981 and 1982.

                                   ‘SSA suspended contlnulng disability reviews for part of 1984 and all of 1985




Reviews in Fiscal Year             SSAreviewed 409,172 DD!3 title II decisions in fiscal year 1988, at a cost
1988                               of about $29.2 million. These included about 51,000 &A reviews of both
                                   favorable and unfavorable decisions and 358,000 PER reviews of
                                   allowances and continuances. (See table 1.2.)


                                    ‘Under Public Law 98-460 (Social Security Disability Benefits Reform Act of 1984), disability benefits
                                    cannot be terminated unless substantial evidence shows that there has been medical improvement in
                                    an individual’s condition allowing the individual to engage in substantial gainful activity.



                                    Page 10                                GAO/HRD-90-28     SSA Reviews    of State Disability   De&&s
                                          Chapter 1
                                          Background




Table 1.2: SSA Reviews   in Fiscal Year
1988                                                                                         DDS title II      Reviewed    by          Percent
                                                                                              decisions                  SSA         e’aviewed
                                          Favorable decisions                                   628,756              386,150      __~___     61”
                                          htial allowances                                       331,738              197,143                59
                                            QA sample                                                                  10,477                  3
                                            PER sample                                                                186,666                56
                                          Reconsideration
                                          ~__             allowances                               39,724              35,646                90
                                            QA sample                                                                    2,291                 6
                                            PER sample                                                                 33,355                 a4
                                          CDR continuances                                        257,296             153,361                 60
                                             QA sample                                                                  15,073                 6
                                             PER sample                                                               138,288                 54
                                          Unfavorable decisions                                  895,049               23,022                  s
                                          Initial denials (QA)                                    606,940               11,128                 2
                                          Reconslderatlon denials (QA)                            254,620                2.929                     1
                                          CDR cessations (QA) -                                    33,489                8,965                27
                                          Total                                                1 S23.807             409.172                  27
                                          Source: SSA State Agency Operations Reports, QA Reports, and PER Reports.
                                          5SA offlclals believe they achieve the required 65percent revlew because some of their reviews cover
                                          people who have flied simultaneous claims as a disabled worker and disabled widow or disabled adult
                                          child.



                                          This report analyzes how USA can be more effective in reviewing DDS
Objectives, Scope, and                    decisions. SSAhas the following objectives for its reviews:
Methodology
                                          1. To measure the accuracy of disability determinations and to deter-
                                          mine DDS performance accuracy, as required by SSA regulations.

                                          2. To detect and correct erroneous disability allowances.

                                          Our work focused on SSA’Seffectiveness in meeting the second objective.

                                          SSAhas said that it could be more effective if it were not required to
                                          review as many decisions. It has proposed several times to reduce the
                                          review requirement to 50 percent of allowances and 25 percent of con-
                                          tinuances SSA would then target its sample to the more error-prone
                                          types of decisions. We considered this and other alternatives in assess-
                                          ing ssA’s program.

                                          During our review, we interviewed SA officials in the Office of Program
                                          Integrity Reviews, Office of the Actuary, Office of Disability Operations,



                                          Page 11                                GAO/HRD-90-28     SSA Reviews   of State Disability   Decisions
Chapter 1
Background




During our review, we interviewed SSAofficials in the Office of Program
Integrity Reviews, Office of the Actuary, Office of Disability Operations,
and Office of Management and Budget. We visited DDSSin Ohio and Indi-
ana, interviewing management officials and claim supervisors. We vis-
ited SSAregional offices in Chicago and San Francisco, interviewing case
reviewers, supervisors, and medical directors about their workloads and
the impact a targeted PER review sample would have on their work.

We developed a cost-benefit model to estimate the fiscal impact of fiscal
year 1988 reviews and various alternatives. Our model shows the poten-
tial effects of targeting the PER sample of initial DD!3allowances to error-
prone types of cases, in comparison with SSA’Srandom approach. We
used SSA’Squality assurance data from fiscal years 1987 and 1988 to
construct a model for targeting the reviews by disability code and basis
of DDS decision. Although we did not verify SSA’Sdata, we did review its
procedures and controls for collecting and reporting the data. Appendix
I presents a detailed discussion of our cost-benefit model and our meth-
odology for estimating the effects of targeting the PER sample.

We did our review from December 1988 through June 1989 in accord-
ance with generally accepted government auditing standards.




                        .

 Page 12                      GAO/HIUMO-28   SSA Reviews   of State Disability   Decisions
Targeting Samples and Reallocating Resources
Could Improve Effectiveness of SSA’s
PER Reviews
                                         SSA’S review program saves more in future benefit payments than it
                                         costs in administrative funds. However, SSA’S reviews could be more
                                         effective in identifying and correcting erroneous DDS decisions if (1) SSA
                                         had more regional medical staff to review PER cases, (2) SA targeted its
                                         PER reviews to error-prone types of cases rather than using a random
                                         sample, and (3) SA reduced its reviews of continuances and used those
                                         resources on a targeted review of initial DDS allowances.


                                         We estimate that SSA spent about $29.2 million reviewing DDS decisions
SSA’s Reviews Have                       in fiscal year 1988, of which about $7.2 million was for QA purposes and
Some Positive Results,                   $22 million for PER reviews. We used a cost-benefit model (see app. I) to
but Leave Many                           estimate the savings in future benefit payments (disability and Medi-
                                         care) resulting from those reviews that changed DDS awards to denials.
Erroneous Decisions                      We estimated that total net savings (after deducting costs) from the
Uncorrected                              reviews of favorable decisions in fiscal year 1988 would be about $69
                                         million, $5.6 million from QA reviews and $63.2 million from PER
                                         reviews.

                                         Table 2.1 breaks this estimate down into reviews of initial allowances,
                                         reconsideration allowances, and continuances. It shows that reviews of
                                         allowances are much more cost-effective than reviews of continuances.
                                         Table 2.1 also distinguishes the QA review results from the PER results.

Table 2.1: Costs and Future Benefit
Reductions From SSA’s Fiscal Year 1988   Dollars in millions, except per-dollar amounts
PER Reviews                                                                             Initial     Reconsideration                       CDR
                                                                                 allowances              allowances             continuances
                                         QA reviews:
                                         Future savings (benefit and
                                           Medicare reductions)                           $4.8                     $1.4                     $2 8
                                         Review costs                                        1.2                    03                       2.1
                                         Net future savings                               $3.6                     $1 2                     $0.8
                                         Future savings per dollar
                                           spent                                         $4 00                   $5.42                     $1 36
                                         PER reviews:
                                         Future savings (benefit and
                                           Medicare reductions)                          $68.5                     $9.6                      $7 0
                                         Review costs                                      13.2                     2.4                       6.4
                                         Net future savinqs                              $55.3                     $7.3                      $0 6
                                         Future savings per dollar
                                           soent                                         $5.18                   $4.08                     $1 0s

                                         Note Some totals may not add because of rounding. This table does not include QA revlews of unfavor-
                                         able DDS declslons




                                         Page 13                                 GAO/ElRIHO-28     SSA Reviews   of State Disability   Decisions
                                     Chapter 2
                                     Targeting Samples and Reallocating
                                     Resources Could Improve Effectiveness      of
                                     SSA’s PER Reviews




                                     As noted in chapter 1, one of SSA’S goals for its review program is to
                                     detect and correct erroneous disability allowances made by the DDSS.
                                     Using SSA’S QA results, we estimate that DDSSmade 42,342 incorrect deci-
                                     sions in fiscal year 1988, of which over 31,000 were incorrect denials.
                                     Through its PER and QA reviews, SSA changed an estimated 4,459, or 10.5
                                     percent, of the incorrect decisions. As table 2.2 shows, most of the incor-
                                     rect DDS denials were left uncorrected. DDSSdeny nearly twice as many
                                     claims as they allow. The error rate on denials is twice as high as on
                                     allowances. However, some of these claimants can be expected to
                                     receive benefits through appeals.

Table 2.2: SSA’s Correction of DDS
Decisions in Fiscal Year 1988                                                Estimated    number of                             Percentage
                                                                              incorrect   decisions’         Changed               changed
                                     Initial allowances                                          4,976           2,352           ---    47 3
                                     Reconslderatlon %owances                                    1,112             495                  44.5
                                     CDR contmuances                                             3,345             693                   20.7
                                     Initial demals                                             22,457             409            --_____ 18
                                     Reconsideration denials                                     8,912             101           -.___-- 1 1
                                     CDR cessations                                              1,540             409                   26.5
                                     Total                                                     42,342           4,459                    10.5
                                     aProlected based on SSA’s CIA data



                                     SSA'S PER reviews do not detect all the estimated erroneous decisions. SSA
Insufficient Medical                 officials said this is because not enough physicians are under contract to
Staff Resources May                  review all the cases selected for review.
Reduce the
                                     Compared with the results of &A reviews, PER reviews do not return as
Effectiveness of PER                 many deficient DDS decisions. Since the QAsamples are statistically ran-
Reviews                              dom samples with a margin of error of 0.3 percent, they should reliably
                                     estimate the percentages of deficient DDS decisions expected in the large
                                     random PER samples. As table 2.3 shows, the QA sample produced a
                                     higher return rate of cases with errors for all categories of reviews in
                                     fiscal year 1988 than did the additional PER sample.




                                      Page 14                                GAO/HRD!W28       SSA Reviews    of State Disability   Decisions
                                         Chapter 2
                                         Targeting Samples and Reallocating
                                         Resources Could Improve Effectiveness        of
                                         SSA’s PER Reviews




Table 2.3: PER Return Rates for Fiscal
Year 1988 Compared With QA Return                                                                    Percent of cases returned            to DDSsa
Rates                                                                                                      QA                                   PER
                                                                                                      sample                                 sample
                                         Initial allowances                                                3.1                                   2.2
                                         Reconsideration allowances                                        44                                    29
                                         CDR continuances                                                     2.9                                 i3
                                         aAs explained In chapter 1, not all demons   returned are reversed


                                         Most of the SSA personnel we interviewed said that this difference
                                         resulted because regional medical staff physicians review nearly all &A
                                         cases, but can review only about 33 percent of the PER cases. They said
                                         that physicians identify some deficiencies in the DDS development of
                                         cases that SSA’S reviewers do not. SSA’S Office of Program Integrity
                                         Reviews has asked regional branches to try to achieve 40-percent physi-
                                         cian review of PER cases, but the goal is not being met. Administrative
                                         budget cuts have affected regional medical staff budgets, and physician
                                         review of PER cases fell to about 30 percent in late fiscal year 1989. In an
                                         attempt to make the best use of limited resources, ss~ has directed that
                                         regional reviewers automatically refer certain error-prone categories of
                                         cases, such as back ailments and lung diseases, to the medical staffs.


                                          One of SSA’S objectives for PER reviews is to detect and correct erroneous
PER Reviews Would                         disability allowances. The reviews could be more effective in accom-
Be More Effective If                      plishing this objective if SSA targeted its sample to cases more likely to
the Sample Were                           be incorrect rather than selecting cases randomly.
Targeted                                  To more fully demonstrate the effects of targeting, we analyzed results
                                          using targeted samples ranging from 5 to 100 percent of initial
                                          allowances. We used QA data from fiscal years 1987 and 1988 to identify
                                          the most error-prone types of decisions. (See app. I.) At the 5-percent
                                          level, only the most error-prone types of decisions would be reviewed.’
                                          For each 5-percent increase in sample size, the types of cases added to
                                          the sample would be slightly less error prone than in the smaller sample.
                                          Thus, the number of reversals grows more slowly as the sample is
                                          enlarged.

                                          Our analysis shows that if the fiscal year 1988 reviews of initial
                                          allowances had been targeted, SSA would have identified and reversed a

                                          ‘Fifty-five disabilities would be involved, including chronic lung disease, anxiety, alcoholism, and
                                          multiple sclerosis.



                                          Page 15                                 GAO~HRD!3O-28      SSA Reviews    of State Disability     Decisions
                                   Chapter 2
                                   Targeting Samples and Reallocating
                                   Resources Could Improve Effectiveness     of
                                   SSA’s PER Reviews




                                   larger percentage of incorrect allowances than it did with its random
                                   sample. Using our cost-benefit model (described in app. I), we estimate
                                   that with the same number of reviews, net future savings could have
                                   increased by $32 million. (See fig. 2.1.)


Figure 2.1: Benefit Savings From
Targeting PER Sample of Initial    120         Mwotm at Dollan
Allowances
                                   110
                                    100
                                     90
                                     so
                                     70
                                     60
                                     50
                                     40
                                     30
                                     20
                                     10
                                      0
                                          ;I




                                    Note: Based on revlewlng the same number of cases SSA reviewed in fiscal year 1988 (186,666)


                                    Table 2.4 shows our analysis of future savings that would be generated
                                    by targeted samples of other sizes. We present these data to show what
                                    could be expected from alternatives to the current 65-percent require-
                                    ment. Table 2.4 excludes &A reviews because we believe these should be
                                    done on a random basis in order to measure DDS accuracy.

                                    We assume in this analysis that SSA’Sregional medical staffs, in addition
                                    to reviewing &A cases, could review up to 20 percent of all PER initial
                                    allowances, which is about the number they have been reviewing. This
                                    would require some additional medical staff to cope with the higher
                                    number of error-prone cases in the sample.


                                    Page 16                                GAO/HRIMO-28      SSA Reviews   of State Disability   Decisions
                                         Chapter 2
                                         Targeting Samples and Reallocating
                                         Resources Could Improve Effectiveness      of
                                         SSA’s PER Reviews




Table 2.4: Net Future Benefit Savings
From Alternative Sizes of Targeted PER   Dollars in mhons                                                                         - .-_____-
Samples                                  initial allowances                      Percent of total                                      Net future
                                         reviewed                                       reviewed         Cost of reviews                 savings
                                          53,008                                             16.5                      $6.7                  $58 1
                                          66,180                                             20.6                       8.4                   71.7
                                          79,763                                             248                        9.3                   74.9
                                          97,021                                             30.2                      104                    77.9
                                         112,441                                             35.0                       11 3               802
                                         128,504                                             40.0                       123                82 4
                                         145,531                                             45.3                       132        ________--842
                                         159,988                                             49.8                       14.0               856
                                         176,694                                             55.0                       14.8               866
                                         195,969                                             61.0                       15.8               878
                                         207,856                                             64.7                       16.3               883


                                         Ten of the 12 regional reviewers we interviewed said targeting the PER
                                         sample would be a good idea, because they now spend part of their time
                                         reviewing cases that are obvious allowances with little or no chance of
                                         error. However, a targeted sample would be more time consuming
                                         because the case mix would be more difficult to review and more cases
                                         would be returned requiring written rationale for the disagreement.

                                         We believe %A would have to increase review and medical staff to do a
                                         targeted review if the volume of cases remained at current levels. Also,
                                         productivity expectations, in terms of the number of cases reviewers
                                         must complete, would have to be reduced. After reviewing SSA’S data on
                                         productivity of regional review and medical staffs and calculating cost-
                                         per-case averages, we concluded that SSAwould have to spend an addi-
                                         tional $2.1 million to review the same number of initial allowances it
                                         reviewed in fiscal year 1988. (See p. 24.) This calculation is included in
                                         our analysis in table 2.4.




                                         Page 17                                 GAO/HRlMO-28       SSA Reviews   of State Disability   Decisions
                     Chapter 2
                     Targeting Samples and Reallocating
                     Resources Could Improve Effectiveness      of
                     SSA’s PER Reviews




                     The PER reviews of DDS continuances, done by W’S Office of Disability
Resources Used to    Operations and program service centers, change very few DDS decisions.
Review CDR           By our estimate, PER reviews of continuances produced only $1.09 in
Continuances Could   future benefit savings for each $1 .OOspent in-fiscal year 1988, com-
                     pared with $5.18 from SSA'S PER reviews of initial DDS allowances. (See
Be Better Used       table 2.1.)
Reviewing Initial
                     ss~ could do somewhat fewer continuance PER reviews if it chooses.-!
Awards -             However, SSA would need legislative authorization to exclude continu-
                     ances from the universe of cases that it is required to review. If the
                     resources used to review continuances were available to the regional
                     review branches for a targeted review of initial DDS allowances, substan-
                     tially more could be saved in future benefit payments. As explained ear-
                     lier, SSA'S PER reviews of initial allowances would be more effective if
                     review and regional medical staffs were larger. This would be especially
                     true with a targeted sample of initial allowances, which would focus on
                     the more time-consuming cases.

                     If the fiscal year 1988 PER reviews of initial allowances had been
                     targeted and the funds now spent reviewing continuances had been
                     used, PER reviews could have identified cases with a potential for about
                     $120 million in net future benefit savings. This would have been an
                     increase of $64 million over the estimated $56 million saved by SSA’S
                     actual PER reviews of initial allowances ($55.3 million) and continuances
                     ($0.6 million). (See fig. 2.2.)




                     “We did not estimate the fiscal impact of such adjustments.




                     Page 18                                 GAO/HRD#Mf3       SA   Reviews   of State Disability   Decisions
                                    Chapter 2
                                    Targeting Samples and Reallocating
                                    Resources Could Improve Effectiveness     of
                                    SW’s PER Reviews




Figure 2.2: Savings From Targeted
Sample Using Reallocated CDR        Milllons ot Dollars
Resources
                                    120
                                    110
                                    100
                                     90
                                     so
                                     70
                                     so
                                     !I0
                                     40
                                     30
                                     20




                                            I         Cost of review
                                                      Net savings
                                     Note: Based on revtewlng the same number of cases SSA reviewed in fiscal year 1988 (186,666)


                                     Table 2.5 shows net future benefit savings (disability and Medicare) for
                                     alternative sample sizes in the same way that table 2.4 did, except that
                                     table 2.5 assumes that over $7 million (mostly from the continuance
                                     reviews) could be made available for a more intensive review of initial
                                     allowances. This would make it possible for regional medical staffs to
                                     review up to 50 percent of all initial allowances. (See table I.7 for sample
                                     sizes from 5 to 100 percent.)




                                                                       .

                                     Page 19                                GAO/HRIhW28       SSA Reviews   of State Disability   Decisions
                                         Chapter 2
                                         Targeting Samples and Reallmting
                                         Resoorces Could Improve Effectiveness      of
                                         SSA’s PER Reviews




Table 2.5: Net Future PER Savings From
Alternative Sizes of Targeted Samples    Dollars in mdlions
Using Reallocated CDR Resources          Initial allowances                      Percent of total                                     Net future
                                         reviewed                                       reviewed         Cost of reviews                savings
                                          53,008                                             16.5                      $6.7        --__~ $58.1
                                                                                                                                         -__
                                          66,180                                             20.6                       8.4                  71 7
                                          79,763                                             24.8                      10.0                  82 0
                                          97,021                                             30.2                      12.0                  92.0
                                         112,441                                             35.0                      13.8                  998
                                         128,504                                             40.0                      15.6                 1073
                                         145,531                                             45.3                      17.5                 113.4
                                         159,988                                             49.8                      19.1                 1183
                                         176,694                                             55.0                      19.9                 1193
                                         195,969                                             61.0                      20.9                 120.6
                                         207,856                                             64.7                      21.4                 122.5




                                         SA'S PER reviews could be more effective if the sample were targeted
Conclusions                              according to error-prone characteristics rather than selected randomly.
                                         This would, however, produce a sample of cases that are more difficult
                                         and time-consuming to review. If no change is made in the legislative
                                         requirement to review 65 percent of favorable DDS decisions, ss~ would
                                         have to provide its regional review branches with more resources to
                                         adequately review a targeted sample. Our analysis shows that the addi-
                                         tional administrative spending would be justified by the substantial
                                         increase in future benefit savings generated by the reviews.

                                         The current reviews of DDS continuances yield little in future savings. A
                                         shift of resources from these reviews to a targeted review of initial
                                         allowances should be beneficial because it would permit a more intense
                                         review of such allowances with more medical staff reviews. This would
                                         enhance the effectiveness of a targeted review of error-prone
                                         allowances.




                                         Page 20                                 GAO/HRD!3@28       SSA Reviews   of State Disability   Decisions
                    Chapter 2
                    Targeting !%uuples and Reallocating
                    Resources Could Improve Effectiveness     of
                    SSA’s PER Reviews




Recommendation to   its PER reviews of initial DDS allowances. While this would require some
the Secretarv of    additional review and medical staff, costs would be far exceeded by the
                    reductions in future benefit payments resulting from the targeted
Health and l&man    reviews.
Services

Recommendation to   exclude CDR continuances from the universe of DDS decisions SSA is
the Congress        required to review. SSA could then limit its reviews of continuances to a
                    quality assurance sample and transfer administrative resources to a
                    more cost-effective targeted review of initial DDS allowances.


                    On January 22, 1990, the Department of Health and Human Services
Agency Comments     (HHS) provided us with written comments on a draft of this report. (See
                    app. II.) HHS agreed that the PER review of initial allowances can be
                    targeted to be more effective and that the PER review of CDRSis only
                    minimally effective. HHS also agreed that targeted PER reviews generally
                    require more administrative resources.

                     HHS said that while it agreed that the relative effectiveness of the PER
                     reviews can be improved by reducing the number of CDRS, it did not
                     believe that the PER review should be limited to initial cases only. HHS
                     said any legislation

                     “should provide the maximum flexibility to direct resources where they are most
                     needed. That would enable us to judge what types of cases are most error-prone at
                     any given time and expend resources to remedy the problem, whether it involves
                     allowances, denials or continuing disability review cases. Legislation either specify-
                     ing percentages and types of cases SSA is to review or excluding types of cases from
                     review severely restricts SSA's ability to manage the disability program
                     effectively.”

                     Our recommendation would increase SSA’Sflexibility by excluding con-
                     tinuances from the universe of cases that ss~ is required to review. It
                     would not prohibit SSA from reviewing continuances or any other types
                     of cases if SSA thought this was needed to improve the accuracy of deci-
                     sions However, such reviews would be in addition to the legislatively
                     required reviews of initial and reconsideration allowances.




                     Page 21                                GAO/IiRD-!X%28   SSA Reviews   of State Disability   Decisions
Methodology and Tables


                         We developed a cost-benefit model to estimate (1) the effect of SSA'S PER
                         reviews on the disability and Medicare trust funds and (2) the effect of
                         targeted sampling at sample levels from 5 to 100 percent. We also used
                         the model to estimate the effect of shifting resources from reviewing
                         continuances to a targeted review of initial allowances.


                         Our model used the following cost-per-case averages:
Cost-Benefit Model
                     l $63 for regional Disability Quality Branch (DQB) QA reviews of initial and
                       reconsideration decisions,
                     l $77 for regional DQB QA reviews of continuing disability decisions,
                     l $53 for regional DQB PER reviews,
                     l $54 for regional medical staff reviews,
                     . $38 for program service center and Office of Disability Operations PER
                       reviews of continuances, and
                     l $63 for central medical staff reviews of continuances.

                         We derived these averages from cost and workload data supplied by the
                         relevant components of .%A.We allocated regional DQB costs to the differ-
                         ent types of reviews based on productivity studies done by SSA'S Office
                         of Disability Program Quality.

                         To estimate future benefit savings due to PER reviews, we used the
                         number of cases reversed by SSA’Sreviews, reduced by an estimated
                         number who would be expected to file successful appeals or file success-
                         ful new claims in future years.

                         To determine the effects of appeals and new claim filings, we studied
                         the 1,192 disabled worker claims that were reversed to denials by PER
                         reviews from January 1 through June 30, 1987. This study, reported in
                         table I. 1, also determined the sex and average age of claimants who did
                         not successfully appeal their denials. We used the age and sex variables
                         in determining the present value of the benefit awards (including Medi-
                         care benefits) that would have been made to these claimants if not for
                         the PER reviews. We derived the average present value from SSA'S table
                         of present values by age and sex, which take into account that some
                         persons will die or recover from their disabilities before reaching age 65,
                         at which time they would be eligible to receive retirement benefits
                         rather than disability benefits.


                                                 .


                         Page 22                      GAO/HRD9@28   SSA Reviews of State Disability   Decisions
                                       Appendix I
                                       Methodology   and Tables




Table 1.1: Results of GAO’s Study of
Claimants Denied Benefits by PER                                                                                                   DDS
Reviews (Jan -June 1987)                                                                                             reconsideration
                                                                                           DDS initial awards                  awards
                                       Reversed by PER reviews                                     923                    269
                                       Benefits granted on appeal                                  373 (40.4%)            155 (57 6%)
                                       Benefits denied                                             550 (59.6%)            114 (42 4%)

                                       Status of denials (Feb. 1989):
                                       Receiving disability benefits from a new
                                       apphcatlon                                                  38                          7
                                         Filed 7/87-6188                                           27                           6
                                         Filed 7/88-IO/88                                          11                           1
                                       Receiving retirement benefits                              105                          16
                                       Retirement beneficlanes deceased                             5                           0
                                       Not receiving benefits                                     402                          91
                                       Average age of denied claimants                             49                          47




                                       We used SSA’SQA data to identify the most error-prone types of cases.
Effect of Targeted                     Our database was the 30,275 initial DDS allowances reviewed by SSA for
Sampling                               QA purposes in fiscal years 1987 and 1988. ss~ provided data showing
                                       the number reviewed and the number returned to DOSSfor each disabil-
                                       ity (primary diagnosis) code. We were also able to separate the cases
                                       that DDSSdecided using both medical and vocational factors from those
                                       decided on a medical basis only. The medical/vocational allowances gen-
                                       erally have a higher error rate than allowances based on medical condi-
                                       tion alone.

                                       Using these case characteristics, we ranked each type of case from those
                                       with the highest return (error) rates to those with the lowest. We then
                                       calculated the percentage of all erroneous cases that would be reviewed
                                        if SSA reviewed 5 percent of all DDS allowances, focusing on the types of
                                        cases with the highest error rates. We made this calculation for each 5-
                                        percent increase in sample size up to a loo-percent review. We then pro-
                                       jected these results to the universe of initial DDS allowances in fiscal year
                                        1988. This permitted us to compare targeted sampling at various levels
                                        with SSA’S actual PER results in fiscal year 1988. We excluded the QA
                                        cases from the universe of allowances, on the assumption that the QA
                                        sample would remain the same regardless of how SSA selected the PER
                                        sample.




                                       Page 23                             GAO/HRD-90-28   SSA Reviews   of State Disability    Decisions
Appendix I
Methodology   and Tables




Table I.2 shows the results of the above calculations. In this table, it is
assumed that the medical staffs would be able to review 20 percent of
initial DDS allowances, about the number they actually reviewed in fiscal
year 1988. Because the medical staff review all QA cases, QA results and
PER results are equal up to the 20-percent level. For additional PER cases
reviewed above a 20-percent sample, SSAwould be dependent on its non-
medical reviewers to identify incorrect or inadequately documented
decisions. Hence the PER results above a 20-percent sample are lower
than would be projected from &A results.

Table I.3 shows the effect of increasing the medical and review staffs up
to the point that 50 percent of initial allowances could be reviewed with
the same attention given to QA cases. These additional resources would
be approximately equal to what is now expended on PER reviews of con-
tinuances. The number of erroneous decisions identified increases con-
siderably in this illustration compared with table 1.2.

Tables I.4 and I.5 use our cost-benefit model to show the future benefit
savings from the various levels of targeted sampling compared with
SA’S fiscal year 1988 reviews of initial DDS allowances. Table I.5 shows
the increased benefit savings that would result from using additional
resources, as done in table 1.3.

Tables I.6 and I.7 include the costs of the reviews and subtract them
from the projected savings to obtain net savings. In estimating reviewer
costs, we made adjustments to reflect the difficulty in reviewing error-
prone cases. Based on SSA productivity data, we adjusted reviewer costs
at lo-percent sample intervals. The costs are highest for the first lo-
percent sample; they then decrease for each lo-percent increase in the
targeted sample. This takes into account that each increase in the sam-
ple contains fewer erroneous cases and is thus less time-consuming to
review. Table I.7 shows the increased costs that would be incurred and
the increased future benefit savings from reviewing initial allowances if
resources were shifted from continuance reviews.




 Page 24                      GAO/HRD90-28   SSA Reviews   of State Disability   Decisions
                                              Appendix I
                                              Methodology   and Tables




Table 1.2: Potential   Results of Targeting
PER Sample                                                                                     Percent of
                                              Initial                                          erroneous               Percent
                                              allowances                      Percent                cases             actually               Number
                                              reviewed                      reviewed            reviewed            identified’             identified
                                                16,706                              5.2                17.3                  17.3                1,745
                                                34,054                             10.6                27.8                  27.8                2,804
                                                53,008                             16.5                38.2                  38.2                3,853
                                                66,180                             20.6                47.2                  47.2                4,761
                                                79,673                             24.8                54.2                 49.6                 5,001
                                                97,021                             30.2                61.3                 52.0                 5.245
                                              112,441                              35.0                66.9                 53.9                 5.437
                                              128,504                              40.0                72.4                 55.8                 5,626
                                              145,531                              45.3                77.1                 57.4                 5,787
                                              159,988                              49.8                80.9                 58.7                 5,917
                                              i76,694                              55.0                84.2                 59.8                 6.030
                                              195,969                              61.0                87.8                 61.0                 6,154
                                              207,856                              64.7                89.7                 61.7                 6,219
                                              225,204----                          70 1                92.3                 62.5                 6,308
                                              240,303                              748                 95.6                 63.7                 6.421
                                              258.294                              80.4                97.5                 64.3                 6,487
                                              274,357                              854                 98.9                 64.8                 6,535
                                              290.099                              90.3                99.7                 65.1                 6,562
                                              306,162                              95.3               100.0                 65.2                 6572
                                              321,261                             100.0               100.0                 65.2                 6,572

                                              SSA’s fiscal year 1988 results:
                                              186.666                         58.1                     58.1                 40.3                  4.069
                                              aThls table assumes regional medical staffs can revlew only 20 percent of PER cases. Thus, at higher
                                              levels (above 20 percent), not all erroneous declslons are being returned to the DDSs




                                               Page 25                                GAO/HRtNO-28      SSA Reviews    of State Msability    Decisions
                                            Appendix I
                                            Methodology   and Tables




Table 1.3: Potential Results of Targeting
PER Sample Using Reallocated CDR                                                                   Percent of
Resources                                   Initial                                                erroneous              Percent
                                            allowances                            Percent               cases             actually                 Number
                                            reviewed                            reviewed            reviewed           identified’               identified
                                             16,706                                    52                  17.3                 17.3                     1,745
                                             34,054                                   10.6                 27.8                 27.8                     2,804
                                                                                                                                                 .__-~
                                             53.008                                   165                  38.2                 38.2                     3,853
                                             66,180                                   20 6                 472                  47.2                   4,761
                                                                                                                                                 ___.--
                                             79,673                                   248                  54.2                 54.2                   5,467
                                             97,021                            -__    30 2                 61.3                 61 3                     6,184
                                            112,441                                   35.0                 66.9                 66.9                     6,749
                                            128,504                                   40.0                 72.4                 72.4                     7,303
                                            145,531                                   45.3                 77.1                 77.1                     7,778
                                            159,988                                   49.8                 80.9                 80.9                     8,161
                                            176,694                                   55.0                 84.2                 82.0                     8,274
                                            195,969                                   61 0                 87.8                 83.2                     8,398
                                            207,856                                   64.7                  89.7                83.9                     8,463
                                            225,204                                   70.1                  92.3                848                      8,552
                                            240,303                                   74.8                  95.6                85.9                     8,665
                                            258,294                                   80.4                  97.5                865                      8,730
                                            274,357              .-____-              85.4                  98.9                870                      8,778
                                            290.099                 ___-              90.3                  99.7                87.3                     8,806
                                            306,162                        -          95.3                 100.0                874                      8.816
                                            321,261                                  100.0                 100.0                874                      8,816

                                             SSA’s fiscal year 1999 results:
                                            186.666                                   58.1                  58.1                40.3                     4.069

                                            aThis tableassumes      reglonal medlcal staffs can review 50 percent of PER cases




                                             Page 26                                         GAO/HRD9@28     SSA Reviews   of State Disability     Decisions
                                                Appendix I
                                                Methodology   and Tables




Table 1.4: Calculation   of Future Benefit Savings From Targeting PER Sample
                             Percent of     Erroneous    Decisions     Denials after         Tentative         Savings lost to
Initial allowances                   total   decisions    reversed          appeals  benefit savings               new claims            Adjusted
reviewed                       reviewed       returned           (-54)          W)            (76,384)                     (.32)          savings
 16,706                                 5.2        1,745           942           565     $43,185,986               $13,819.515      $29366,470
 34,054                               10.6         2,804        1,514            908        69,394,558              22,206,259          47 188.300
 53,008                               16.5         3,853        2,081          1,248        95,355,647              30,513x807          64,841.840
 66,180                               20.6         4,761        2,571          1,543      117,827,209               37,704,707          80,122.502
 79,673                               24.8         5,001        2,701          1,620      123,766,828               39,605,385          84,161,443
 97,021                               30.2         5,245        2,832          1,699      129,805,442               41,537,741          88267 701
112,441                               35.0         5,437        2,936          1,762      134,557,138               43,058,284          91.498 854
128,504                               40.0         5,626        3,038          1,823      139,234,588               44,555,068          94.679520
145,531                               45.3         5,787        3.125          1,875      143,219,083               45,830,107          97.388,977
159.988                               49.8         5,917        3,195          1,917      146,436,377               46,859,641          99,576,737
176.694                               55.0         6,030        3,256          1,954      149,232,948               47,754,544       101 478.405
195,969                               61.0         6,154        3,323          1,994      152,301,752                48.736,561 ___---103.565.191
207,856                               64.7         6,219        3,358          2,015      153,910,399                49,251,328-      104,659,071
225,204                               70.1         6,308        3,406          2,044      156,113,008                49,956,163       106.156846
240,303                               74.8         6,421        3,467          2,080       158,909,579               50,851,065       108.058 514
258,294                               80.4         6,487        3,503          2,102       160,542,975               51,373.752       log,169223
274,357                               85.4         6,535        3,529          2,117       161,730,899               51,753,888       109.977011
290,099                               90.3         6,562        3,543          2,126       162.399.106               51.967714        l'Q.431392
306,162                               95.3         6,572        3,549          2,129       162,646,590               52,046,909       110.599681
321 261                              100.0         6,572        3,549          2,129       162,646,590               52,046,909       110.599.681

SSA’s fiscal year 1988 PER results:
186,666                             58.1         4,069            2,197            1,318        100,701,305            32.224.418             68476.807
                                                 Note, QA cases are excluded on the assumption that SSA would continue to select them randomi,,




                                                 Page 27                               GAO/HRD-90-28     SSA Reviews    of State Disability    Drcisions
                                                 Appendix I
                                                 Methodology   and Tables




Table 1.5: Calculation   of Future Savings From Targeting PER Sample Using Reallocated CDR Resources
                             Percent of    Erroneous      Decisions    Denials after         Tentative    Savings lost to
Initial allowances                  total   decisions      reversed         appeals   benefit !i$$19~         new claims                            Adjusted
reviewed                      reviewed       returned            1.54)          1.601         .- -1--  -I
                                                                                                                     1.321
                                                                                                                     a---,
                                                                                                                                                     snvinaa
                                                                                                                                                     -- ---*a-
 16,706                              5.2          1,745            -94;       -     565        $43.185.986         $13,819,515
                                                                                                                   _                          $29,366,470
 34,054                             10.6          2,804           1,514             908         69.394.558              2; !,206,259               47,188,300
 53,008                             16.5          3,853           2,081           1,248         95,355,647              30,513,807                 64.841.840
 66,180                             20.6          4,761           2,571            1,543       117,827,209              37,704,707                 80,122,502
 79,673                             24.8          5,467           2,952            1.771       135.299.590              43,295,869                 92,003,721
 97,021                             30.2          6,184           3,339           2,004        153.044.205              48,974,145            i 04,070,059
112,441                             35.0          6,749           3,644           2,187        167,027,060              53448.659                 113,578,401
128,504                             40.0          7,303           3,944           2,366        180,737,682              57,836:058                122.901.624
145,531                             45.3          7,778           4,ZC 10         2.520        192.493.180              61,597,817                130,895,362
159,988                             498           8,161           4,407           2,644        201.971.823              &




225,204                             70.1          8,552           4,618           2,771        211.                     .      - ,-_-
240,303                             74.8          8,665           4,679           2,807        214.445.025              68.622.408



290,099                             90.3          8,806           4,755                            ,--   ,--
306,162                             95.3          8,816           4,761            2,856       218,182,035              69,818,251                148,363,784
321,261                            100.0          8,816           4,761            2,856       218,182,035              69,818,251                148363,784

SSA’s fiscal year 1988 PER results:
186,666                             58.1          4,069           2,197            1,318       100,701,305              32,224,418                 68,476.887

                                                 Note: QA cases are excluded on the assumption that SSA would contrnueto        select them randomly




                                                 Page 28                               GAO~HRD-90-28      SSA Reviews       of State Disability      Decisions
                                           Appendix I
                                           Methodology    and Tables




Table 1.6: Costs and Net Future Benefit Savings From Targeting PER Sample
Percent of
allowances                Number    Future benefit                      Medical                                                        Return per
reviewed                reviewed           savings   Review costs    staff costs                Total costs          Net savings      dollar spent
   5.2                      16,706      $29,366,470      $1,252,950     $902,124                 $2,155,074           $27,211,396            $13.63
  106                       34,054       47,188,300       2,554,050    1,838,916                  4,392,966            42,795,334             1074
 16.5                       53,008     64,841,840           3,880,830         2,862,432            6,743,262           58,098,578                9.62
 20.6                       66,180     80,122,502           4,802,870         3,573,720            8,376,59O           71,745,912                9.57
24.8                        79,763     84,161,443           53685,765         3,573,720            9,259,485           74$X,958                  9.09
 30.2                       97,021     88,267,701           6,807,535         3,573,720           10,381,255           77,886,446                8.50
 35.0                      112,441     91,498,854           7,732,735         38573,720           11,306,455           80,192,399                8.09
 40.0                      128,504     94,679,520           8,696,515         3,573,720           12,270,235           82,409,285                7.72
-45.3                      145,531     97,388,977           9,633,OOO         3,573,720           13,206,720           84,182,257                7 37
 49.8                      159,988     99,576,737          10,428,135         3,573,720           14,OO1,855           85,574,882                7.11
 550                       176,694    101,478,405          11,263,435         31573,720           14,837,155           86,641,250                6.84
 61.0                      195,969    103,565,191          12,227,185         3573,720            15,800,9O5           87,764,286                6.55
 64 7                      207,856    104,659,071          12,762,lOO         3,573,720           16,335,820           88,323,251                6.41
 701                       225,204    106,156,846          13,542,760         3,573,720           17,116,480           89,040,366                6.20
 748                       240,303    108,058,514          14,146,720         37573,720           17,720,440           90,338,074                6.10
 804                       258.294    109,169,223          14,866,360         3,573,720           18,440,080           90,729,143                5.92
 85.4                      274,357    109,977,Oll          15,428,565         3,573,720           19,002,285           90,974,726                5.79
 903                       290,099    110,431,392          15,979,535         3,573,720           19,553,255           9O,878,137                5.65
 95.3                      306,162    110,599,681          16,461,425         3,573,720           20,035,145           9O,564,536                552
100.0                      321,261    110,599,681          16,914,395         3,573,720           20,488,115           9O,lll,566                5.40

SSA’s fiscal year 1968 results:
 58.1                      186,666     68,476,887            9,893,298        3,326,388           13,219,686           55,257,201                5.18
                                            Notes:

                                            1. This analysis assumes that regbonal medical staffs can revkew cases up to the 20- percent level

                                            2. Reviewer costs are highest for the 10 percent of cases in the most error-prone categories, and
                                            decrease with each addItIonal 10 percent




                                                                              .

                                            Page 29                                 GAO/HRD-9W28 SSA Reviews of State Disability           Decisions
                                            Appendix I
                                            Methodology     and Tables




Table 1.7: Costs and Future Savings From Targeting PER Sample Using Reallocated              CDR Resources
Percent of
allowances             Number     Future benefit                  Medical staff                                                             Return per
reviewed             reviewed            savings   Review costs           costs                   Total costs         Net savings          dollar spent
  5.2                    16,706    $29,366,470            $1,252,950            $902,124           $2,155,074          $27,211,396                   $1363
 10.6                    34,054     47,188,300             2,554,050            1,838,916            4,392,966            42,795,334                  10.74
 16.5                    53,008     64,841,840             3,880,830           2,862,432             6,743,262            58,098,578                   9 62
 20 6                    66.180     80,122,502             4,802,870           3,573,720             8,376,590            71,745,912                   957
 24.8                    79763      92,003,721             5,685,765           4,307,202             9,992,967            82,010,754                   921
 302                     97.021    104,070,059             6,807,535           5,239,134            12,046,669            92,023,390                   8 64
 350                    112,441    113.578,401             73732,735           6,071,814            13,804,549            99,773,852                   823
 400                    128,504    122.901,624             8,696,515           6,939,216            15,635,731           107,265,893                   786
 453                    145,531    130,895,362             9,633,OOO            7,858,674           17,491,674           113,403,688                   748
 49.8                   159,988    137.340.840            10,428,135            8639,352            19,067,487           118,273.353                   7.20
 550                    176,694    139,242,508            11,263,435            8,639,352           19.902.787           119.339.721                   7 00
 61 0                   195,969    141,423,174            12,227,185            8,639,352          20,866;537            120,556.637                   6.78
 647      -             207,856    143,920,948            12,762,100            8,639,352          21,401,452            122,519,496                   6 72
 70 1                   225,204    145,822,617            13,542,760            8,639,352          22,182,112            123,640,505                   6.57
 74.8                   240,303    146,916,497            14,146,720            8,639,352          22,786,072            124,130.425                   645
 80.4                   258,294    1473724,285            14.866,360            8,639,352          23,505,712            124,218,573                   6.28
 85.4                   274,357    148,195,495            15,428,565            8,639,352           24,067,917           124,127,578                   616
 90.3                   290,099    148,363,784            15,979,535            8,639,352           24,618,887           123,744,897                   6.03
 95 3                   306.162    148,363,784            16,461,425            8,639,352          25,100,777            123,263,007                   5 91
100.0                   321,261    148,363,784            16,914,395            8,639,35i           25.553.747           122.810.037                   581

SSA’s fiscal year 1988 results:
 58.1                   186,666     68,476,887             9,893,298            3,326,388           13,219,686            55,257,201                   518



                                             1. This analysis assumes reglonal medical staffs can review cases uptothe      50.percent level
                                             2 Reviewer costs are highest for the 10 percent of cases in the most error-prone    categones     and
                                             decrease with each addItIonal 10 percent.




                                             Page 30                                  GAO/HRD!9@28      SSA Reviews      of State Disability   Decisions
Appendix II

Comments From the Department of Health and
Human Services



                    DEPARTMENT OF HEALTH & HUMAN SERVICES                           Olf~ceof lns,xclorGeneral

                                                                                    Washmgton,   0 C 20201


                                                        JAFi221990




              Mr.   Lawrence H. Thompson
              Assistant     Comptroller     General
              United    States   General
                Accounting      Office
              Washington,      D.C.     20548
              Dear    Mr.   Thompson:
              Enclosed     are the Department's          comments on your draft        report,
              "Social     Security:         SSA Could Save Millions       by Targeting      its
              Reviews of State Disability             Decisions."        The comments represent
              the tentative         position    of the Department       and are subject        to
              reevaluation       when the final       version     of this report     is received.
              The Department    appreciates    the opportunity          to comment on this
              draft report   before    its publication.
                                                      Sincerely     yours,



                                                      Richard     P. Kusserow
                                                      Inspector     General
              Enclosure




                     Page 31                            GAO/HRD9028      SSA Reviews    of State Disability     Decisions
     AppendixII
     CommentsFromtheDepartmentofHealth
     and HumanServices




COMMENTS OF THE DEPARTMENT OF HEALTH AND HUMAN SERVICES ON THE
aERAL      ACCOUNTING OFFICE'S DRAFT REPORT, "SSA COULD SAVE
KILLIONS BY TARGETING ITS REVIEWS OF STATE DISABILITY    DECISIONS"
ltfBP-go-201
General      Accountino     Office    Recommendations
--     That the Secretary          of the Department       of Health       and Human
       Services    direct      the Social       Security  Administration         (SSA) to
       use a targeted       sample for its preeffectuation               reviews    (PER)
       of initial     disability         determination    services      (DDS)
       allowances.        While this would require           some additional        review
       staff    and medical       staff,     costs would be far exceeded by the
       reductions      in future       benefit    payments resulting        from the
       targeted    reviews.
--        The Congress should revise             Section  221(c) of the Social
          Security      Act to exclude     CDR continuances        from the DDS
          decisions      SSA is required       to review.      SSA could then limit
          its reviews       of continuances       to a quality     assurance    (QA)
          sample,      and transfer     administrative      resources     to a more
          cost-effective       targeted    review of initial         DDS allowances.
penartment        Comment
We agree that the PER of initial          allowances       can be targeted      to be
more effective,     and that the continuing          disability        review (CDR)
PER is only minimally       effective.      We also agree that targeted
reviews   generally   require      more administrative          resources.
While we agree that the relative                    effectiveness         of the PER process
can be improved by reducing                 the number of CDR continuance
reviews,     we do not believe            that the PER review should be limited
to initial       cases only.          Any legislation          in this regard          should
provide     the maximum flexibility               to direct       resources        where they
are most needed.            That would enable us to judge what types of
cases are most error            prone at any given time and expend resources
to remedy the problem,              whether     it involves         allowances,        denials  or
continuing       disability      review cases.             Legislation        either
specifying       percentages         and types of cases SSA is to review or
excluding      types of cases from review severely                       restricts       SSA's
 ability    to manage the disability                program effectively.               However,
 if Congress insists          on retaining          requirements         regarding
 percentages       and types of cases to be reviewed,                      those contained      in
 the Administration's           bill     (Section       702 of "Social          Security
 Amendments of 1989") are superior                    to those contained             in present
 law.




                                       .



      Page 32                                  GAO/HRD-90-28      SSA Reviews of State Disability    Decisions.
Appendix III

Major Contributors to This Report


;
Human Resources       Cameo A. Zola, Assignment Manager
Division,
Washington, DC.

                      Daniel L. McCafferty, Regional Assignment Manager
Cincinnati Regional   Kenneth R. Libbey, Evaluator-in-Charge
Office                Ellen Soltow, Evaluator




 (105343)             Page 33                    GAO/HRD9&28   SSA Reviews   of St&e Disability   Decisions