Health Affairs

Editor’s Note: This is the third and last installment of a series of blog posts by Ron Klar offering suggestions on how to make the Medicare Shared Savings Program a more viable vehicle for the creation of accountable care organizations. You can read the first installment here and the second installment here.

In this posting I will discuss and suggest alternatives regarding (1) evaluating quality, including the measures, performance scoring, and standards, (2) determining shared savings (and losses) rates and limiting payments (and losses), and (3) sharing CMS data and information.  These three areas have generated serious concerns among provider groups considering qualifying as accountable care organizations (ACOs) for the Medicare Shared Savings Program (MSSP) since the proposed rule was released on March 31.

The MSSP provides an extraordinary opportunity to improve the efficiency and effectiveness of the Medicare fee-for-service program, and it is important for it to be “successful”.  However, as I discussed in my Health Affairs blog of April 7, the rule needs a “major makeover” or the MSSP “is likely to have few participants”.  My specific suggestions are intended to improve the likelihood of a “four-part success”: (1) meaningful participation by providers; (2) widespread acceptance by beneficiaries; (3) enhancement of quality performance; and (4) demonstrable earning of savings.

As background, the following highlights some suggested changes from Parts One and Two:

Quality Measures

Provider groups have expressed widespread complaints that the “sixty-five” measures included in the proposed rule require too much of an administrative burden.  [Actually, there are only sixty-three, as two of them (#35 and #52) are composites of other individual measures (diabetes and coronary artery disease), although another (#24) is a composite of multiple health care acquired conditions.]

Some “prototype” groups, including the Billings Clinic, one of the Physician Group Practice (PGP) Demo sites, have suggested that the annual cost per measure for collection and reporting averages $30,000, i.e., a total cost of about $2 million.  However, as eleven of the measures are derived from claims data, presumably extracted and analyzed by CMS at their expense, this total should be less; however, the amount less would probably not be much, as most groups would still want to monitor their ongoing performance on these measures even if calculated by CMS.  In fact, the PGP Demo clearly evidenced that when performance on quality measures affects the amount of shared savings, provider groups will implement specific methods to monitor their performance and “score well” on them.

What is interesting about these complaints, or at least the reporting of them, is that very little is being said about the appropriateness of the measures themselves.  Maybe this is because groups understand that they can “perform to the test”, whatever the measures are, the only issue being the amount of effort.  This also raises the issue of “grade inflation”: the PGP experience with 32 measures, all of which are proposed in the rule, had scores clustering between 90 and 100 percent.

Another issue is the actual ability of each measure to “assess the quality of care” as required by the statute, not simply to be computable.  To me, the selection of measures proposed seems to be either: (a) a “smorgasbord” intended to have something for everyone, and/or (b) a collection cobbled together by reaching-into the existing process measure “parts bins”.  We must, and can, do better.

Suggestion N: There Should be Two Sets of Measures for the MSSP:

(A) A “Performance” Set of Claims-Based Measures, Focused on “Utilization”, With Comparative Quality Standards, To Determine Eligibility for and Amount of Payment, and

(B) A “Reporting” Set of Measures, Already Used for Other Existing Programs, for Monitoring Purposes Only

Performance Measure Set (PMS)

The burden (in time and expense) expressed by many provider groups for the collection and monitoring of chart-based and survey-based measures is legitimate.  In addition, the questionable benefit expressed by some analysts to detect any actual improvements in care effectiveness and efficiency derived from these measures is also legitimate.  Furthermore, the five-year experience of the PGP Demo reveals that there may have been too much attention on the measures and too little on the care improvement processes, i.e., high quality scores but not much improvement in utilization or outcomes.

Therefore, I suggest that the performance measures should instead:
.

• (# N1) be derived from claims data that require no additional collection or reporting effort or expense by an ACO.
• (# N2) focus on “outcomes” for which reductions are clearly suggestive of higher “quality”, i.e., improved care effectiveness or appropriateness results.

I offer for consideration the following twelve performance measures that satisfy both of these suggestions and are aligned with achieving the “three-part aim”:

(1)  Measure #8 in the proposed rule, “the rate of readmissions within 30 days of discharge … risk-standardized, and all condition” clearly meets the N1/N2 criteria and should be retained.

(2)  Measure #9 in the proposed rule, “30 day post discharge physician visit”, also should be retained, although I suggest (# N2a) that this be modified to within 14 days (more clinically desirable), and is applicable only if discharged to “home” and seen by a “billing” practitioner for an office or home visit.

Seven other measures in the proposed rule, #12 to #18 dealing with admissions/discharges for ambulatory care sensitive conditions, would seem to also meet the N1/N2 criteria.  However, for six of the conditions (other than congestive heart failure (CHF #15)), the rates are very small:

(3)  Therefore, I suggest (# N2b) that a single composite measure of the conditions in #12 to 14 and #16 to 18 be included in PMS.

(4)  For CHF, I suggest (# N2c) that a separate measure be included, a composite of #15 (admissions/discharges) and a new measure of the rate, per 100,000 population, of hospitalized days, to enhance capturing different permutations of undesirable possibilities.

Also, I suggest (# N2b′ and # N2c′) that the rates for each of these conditions be risk-standardized (with average CMS-HCC scores) to facilitate comparison of different populations.

Besides these, there are other measures not in the proposed rule that I believe meet the N1/N2 criteria and should be in PMS.  Most importantly, it is certain that reducing hospital days is demonstrative of enhanced quality, improved health, and reduced cost.  Therefore, I suggest the following additional claims-based outcome measures for the applicable populations:

(5)  # N2d – the percentage of (a) all-cause re-hospital days within 30 days of discharge, and  (b) all-condition, risk-standardized (with MS-DRG weights) initial-hospital days.

(6)  # N2e – the percentage of (a) all-cause re-hospital days from a post-acute facility, and (b) all-condition, risk-standardized (with MS-DRG weights) initial-hospital plus post-acute facility days.

(7)  # N2f – a composite of the rate (per 100,000 population) of hospital days for beneficiaries of selected “chronic condition categories” (e.g., CHF, COPD, CAD, CRF), risk-adjusted (with CMS-HCC score), who were not hospitalized for that condition in the prior year.

(8)  # N2g – same as # N2f, but for beneficiaries who were hospitalized for that condition in the prior year.

(9)  # N2h – the percentage of (a) all hospital days for discharges with DRG “recoding” for “complications”, and (b) all hospital days for all discharges, risk-standardized (with MS-DRG weights).

Many stakeholders believe that a “generalist” (family or general practitioner, general internist, or geriatrician), as a “primary care physician” (PCP), can assist care coordination and thus improve quality.  Therefore, I suggest that the following two measures also be included in the PMS:

(10)  # N2i – the percentage of (a) beneficiaries with a “new patient” non-institutional evaluation and management (E&M) service from a PCP following at least one “established patient” service prior to it during the same or previous year, and (b) all beneficiaries with at least one non-institutional E&M service from a PCP during the performance year [this is a measure of undesirable changing of physicians, either by beneficiary choice or provider circumstances].

(11)  # N2j – the percentage of (a) beneficiaries (other than those in # N2i) with “established patient” non-institutional E&M services from two or more PCPs, and (b) all beneficiaries with at least two non-institutional E&M services from PCPs (other than those excluded in (a)) during the performance year [this is a measure of lack of “continuity”, even for physicians in the same “practice”].

Finally, virtually every commentator identifies the use of the emergency department (ED) of a hospital, other than when an admission or “observation unit” follows, as both inefficient utilization and evidence of prior poor quality management.  Therefore, I suggest including in the PMS an additional measure:

(12)  # N2k – the percentage of (a) beneficiaries with an ED visit, not resulting in an admission or transfer to an observation unit, and (b) all beneficiaries during the performance year; this measure should also be risk-standardized with the population CMS-HCC risk score.

There are many other claims-based measures in use, but these do not meet suggested criteria N2 because they are only process measures.  For example, proposed measure #65, “monthly INR for beneficiaries on warfarin”, is merely a process measure of testing, rather than an outcome measure, such as the average percentage of monthly INRs between 2.0 and 4.0.  Until such results are universally captured on claims, as with specific “Category II” CPT modifier codes, this measure and others like it, e.g., hemoglobin A1c results, should not be included in the PMS.

Reporting Measure Set (RMS)

Section 1899(b)(3)(D), “other reporting requirements”, authorizes but does not require the Secretary to:

incorporate reporting requirements and incentive payments related to the physician quality reporting initiative (PQRI) under section 1848, including such requirements and such payments related to electronic prescribing, electronic health records [i.e., HITECH], and other similar initiatives.

Suggestion # N(B) above is consistent with this reporting authority.  While section 1899(b) deals with the “eligibility” for ACOs, section 1899(d) deals with “payments and treatment of savings”.  As a consequence, the statutory intent was to keep these referenced “reporting” measures separate from the measures that will include the required “quality performance standards”, i.e., the PMS measures.  In order to avoid any different reporting burden on ACOs than other provider groups, suggestion # N(B) limits RMS measures to only those “already used for existing programs”, e.g., the Group Practice Reporting Option (GPRO) measures for PQRI.

I believe that patient and/or caregiver opinion measures, obtainable only by surveys, can provide useful information to CMS regarding the experience of care from providers, and also be useful to beneficiaries just like other “consumer” surveys.  However, I do not believe that any of the seven measures proposed in the rule (#1 to #7) are sufficiently valid or reliable or that CMS can establish the quantitative “quality performance standards” required to be part of the savings and payments determination processes.  Until such measures become available, I suggest (# N(B)1) that any survey measures CMS chooses should be (a) collected at CMS expense, (b) include a “control group” of non-ACO fee-for-service beneficiaries in the ACO’s service area, and (c) used for program monitoring and “public information” only.

Quality Performance

After measures are specified, it is necessary to establish the basis for assessing performance for these measures.

Suggestion O: Quality Performance Should be Assessed as the Differences in Measure Values Between the Populations of (a) ACO Assigned Beneficiaries and (b) Comparable, Local-Area, Non-Assigned, Fee-for-Service Beneficiaries, i.e., the “Quality Benchmark”

Just as CMS suggests that we should view the financial benchmark “as a surrogate of what … the expenditures would otherwise have been in the absence of the ACO” for determining financial performance, I believe we should view the quality values for comparable beneficiaries as the equivalent “surrogate”, or “quality benchmark”, for determining quality performance.  As for financial performance, being below the benchmark amounts will be considered better.

For each of the twelve PMS measures identified above, I suggest (# O1) that there be a “quality performance score” (qpsⁿ) computed as (qpsⁿ = value° – valueª), with “value°” for the comparable other population and “valueª” for the ACO’s population.  For example, for suggested measure (12), the percentage of all beneficiaries with ED visits (not counting those with an admission or observation): if value° = 12 percent and valueª = 10 percent, the qps¹² = 12 – 10 = +2 percent.

To facilitate interpretation and shared savings rate applications (see below), a “positive” score evidences “higher quality”, as when the value for the ACO is lower than that for the comparable other.  A lower value is desirable because all of the measures deal with utilization outcomes that are to be avoided or reduced.

As the values for all of the twelve measures are percentages or ratios (which multiplied by 100 yields a percentage), one could sum all of these differences to compute a “composite quality performance score” (Cqps), i.e., Cpqs = ∑qpsⁿ.

Regarding identifying the “comparable population” of beneficiaries, I suggest (# O2) that (a) if the “local service area” (to be defined by CMS and/or FTC) has a larger number (N°) of non-ACO fee-for-service (FFS) beneficiaries than the number of ACO assigned beneficiaries (Nª), then these are the comparable population, or (b) if it has a smaller number, the comparable population would be a weighted average of (i) the “local service area” non-ACO FFS beneficiaries (weighted N° ÷ Nª), and (ii) a “regional” sample (to be defined by CMS) of all non-ACO FFS beneficiaries (weighted (Nª – N°) ÷ Nª).

Quality Performance Standards (QPSs)

Section 1899(d)(1)(A) specifies that

a participating ACO is eligible to receive payment for shared savings … if  (i) the ACO meets quality performance standards established by the Secretary … and (ii) the ACO meets the requirement … “ [regarding performance year and benchmark expenditures]

Suggestion P: The QPSs Should be Composite Quality Performance Scores (Cqps) of Minimum Positive Amounts

Focusing on the composite score, rather than individual measure quality performance scores, would (a) provide ACOs flexibility in meeting the QPS, and (b) facilitate determining the shared savings (and losses) rates (see below).  However, to provide an added incentive to ACOs to perform on each measure at least as well as their comparable population, I suggest (# P1) that any negative qps, i.e., when the ACO value is higher than the “benchmark” value, be doubled before it is summed to compute the composite score.

So what should be the specific minimum positive amount(s) for the “composite quality performance score(s)”, i.e., the “quality performance standard(s)”?

I suggest (# P2) that the QPS for the SS Model be 0% for PY¹, 2.5% for PY², and 5.0% for PY³.

I suggest (# P3) that the QPS for the SS&L Model be 0% for PY¹. 5.0% for PY², and 10.0% for PY³.

These two suggestions start with a QPS at least equal to the non-ACO comparable population.  [Remember, this will effectively be higher quality than the comparable because of suggestion # P1 that doubles any negative measure score.]  This would mitigate my challenge of a position in the proposed rule, expressed along with several other statutory guidance issues in my Health Affairs blog of April 7, to consider only “complete reporting” of the specified quality measures during PY¹ as sufficient to meet the QPS.

Subsection 1899(b)(3) mandates two separate and distinct provisions: “(B) Reporting Requirements” and “(C) Quality Performance Standards”.  The proposed rule can only reasonably be interpreted as having no QPS for PY¹.   In order to evidence from the outset that maintaining and improving quality is the prerequisite to any shared savings in the MSSP, obviously essential to beneficiaries and their advocates, there should be QPSs (as suggested) that apply at the start and increase over time.

Suggestions # P2 and # P3 would therefore also be in full compliance with the statutory requirement at section 1899(b)(3)(C) to “seek to improve the quality of care furnished by ACOs over time by specifying higher standards, new measures, or both …” [italics added for emphasis].

Determining Shared Savings (and Losses) Rates and Limits

Subsection 1899(d)(2) specifies that “subject to performance with respect to the quality performance standards … a percent (as determined appropriate by the Secretary)” of the difference between performance year and benchmark expenditures “may be paid to the ACO as shared savings”.  This payment “percent” is referred to in the proposed rule as the “net sharing rate”.

The net sharing rate in the proposed rule is “effectively” (to assist the explanation) derived by the multiplication of the base “sharing rate” by its proposed composite quality “performance scoring” percentage, plus a possible special addition.  Different base sharing rates are proposed for “Track 1” (SS model for two years and SS&L for the third) and “Track 2” (SS&L model all three years): 50.0 percent for Track 1 and 60.0 percent for Track 2.   The possible special addition depends on the percentage of assigned beneficiaries with visits to a federally-qualified or rural health center (FQHC/RHC): 0.5 to 2.5 percent for Track 1 and 1.0 to 5.0 percent for Track 2.

The proposed rule composite quality performance scoring percentage is “effectively” computed as the unweighted average of the percentage of possible points for each of five quality measure “domains”.  Each measure within the domains can receive between 0.0 and 2.0 points depending on “performance” relative to either the percentiles of FFS and Medicare Advantage values or the computed percentages for the measures.  The minimum possible proposed “earned” points for each measure is 1.1 for achieving at least the 30^th percentile or 30 percent (out of 100); the maximum of 2.0 points would be “earned” for achieving or exceeding the 90^th percentile or 90 percent.  The median percentile or percent performance for each measure would earn 1.4 points, or 70 percent of the possible 2.0 points.

It would seem highly unlikely that any ACO would be able to earn the full 2.0 points for each of the 65 measures proposed.  Therefore, the effect of the quality performance scoring percentage methodology as proposed would be to always reduce the 50.0 percent (Track 1) and 60.0 percent (Track 2).  To me, this undesirably (a) treats quality “performance” as a negative “modifier”, rather a positive one once exceeding the QPS, and (b) eliminates any true “base” savings rate known in advance, which would be helpful in understanding “the deal” and resource planning.

Suggestion Q: For Both the SS and SS&L Models, the “Final” Sharing Rate Should be (a) a Pre-Set “Base” Sharing Rate, Plus (b) the “Composite Quality Performance Score”, Not to Exceed (c) a “Maximum Sharing Rate” (Before Any FQHC/RHC “Special Addition”)

This suggestion emphasizes quality performance as a positive modifier.  In addition, it assures “ACOs that meet quality performance standards established by the Secretary are eligible to receive payments for shared savings” [section 1899(a)(1)(B), the preamble] of a known, minimum rate.

I suggest (# Q1) that for the SS model, the “base” sharing rate be 40.0 percent, and that the “maximum sharing rate” be 52.5 percent.  By contrast, in order to encourage ACOs to strive to progress to the SS&L model, I suggest (# Q2) that for the SS&L model, the “base” sharing rate be 50.0 percent, and that the “maximum sharing rate” be 75.0 percent.  [Because of the yearly increases in the QPSs suggested above (# O4 and # O5), the “base” sharing rates would effectively be 42.5 and 55.0 percent for PY² and 45.0 and 60.0 percent for PY³, for the SS and SS&L models, respectively.]

These suggested amounts allow 24 percent [(52.5 – 40.0) ÷ 52.5] of the “maximum sharing rate” for the SS model, and 33 percent [(75.0 – 50.0) ÷ 75.0] for the SS&L model, to be earned only by improving quality.  I believe these are important and appropriate balances for the MSSP to embrace, ones that should be considered desirable by all stakeholders.

I agree that the possible special additions for FQHC/RHC proposed in the rule of 0.5 to 2.5 percent for “Track 1” and 1.0 to 5.0 percent for “Track 2” should also be allowed.  With these, the “maximum possible sharing rates” would be 55.0 percent for the SS model and 80.0 percent for the SS&L model.

Suggestion R: The “Losses Sharing Rate” for the SS&L Model Should be 50.0 Percent for PY¹, 55.0 Percent for PY², and 60.0 Percent for PY³

The proposed rule includes a sharing rate for losses (for “Track 2”) of 100.0 minus the “final sharing rate” for savings.  It states that “in order to recognize these factors [quality performance and FQHC/RHC adjustment] appropriately … these factors must operate as decreases in the ACO’s shared loss rate, rather than as the increases … in … the shared savings rate”.

Unfortunately, this proposal guarantees an asymmetry of gains and losses (only the improbability of an exact 50.0 percent yielding symmetry), which tends to make no conceptual or analytic sense.  In addition, this would seem to be contrary to the basic concept of “two-sided risk” which spawned the “Track 2” option.

Instead, this suggestion R would create symmetry between the losses sharing rates and the effective base sharing rates for each performance year.  Of course, depending on the extra sharing percentage amounts that an ACO might earn because of their higher quality performance and/or FQHC/RHC participation, the final sharing rate for savings could always be higher.  I believe this is the outcome all stakeholders, and CMS, can support.

Losses less than a “minimum loss rate” (MLR) would not be subject to any sharing.  Per my suggestion L discussed in Part Two, the MLR would be equal to the “minimum savings rate” (MSR), which I suggest should vary between 1.0 and 2.0 percent depending on the number of assigned beneficiaries.

Limiting Payments and Losses

The statute requires the Secretary to “establish limits on the total amount of shared savings that might be paid to an ACO”.  The rule proposes “maximum shared savings caps” of 7.5 percent of the benchmark expenditures for Track 1, and 10 percent for Track 2.

Suggestion S: The Limits on the Shared Savings Paid Should be 7.5 percent for the SS Model and 15 Percent for the SS&L Model

The reason for increasing the differential between the limits is to provide a greater incentive to progress to the SS&L model.  There would seem to be no disadvantage to this suggestion.  Also, considering 60.0 percent as a reasonable expectation for an SS&L ACO’s final sharing rate (using the methodologies suggested above), the total savings below the benchmark necessary to reach this limit would be 25 percent.  This amount of potential total savings comports with the amount many analysts believe may be achievable.

The rule also proposes a “maximum shared loss cap” for Track 2 of 5 percent of the benchmark expenditures for PY¹, 7.5 percent for PY², and 10 percent for PY³.  [For Track 1 ACOs, which shift to the SS&L model in PY³, it proposes 5 percent as this is the first year of possible losses.  My suggestion A is that there be a “pure” SS model, not a “hybrid” as Track 1.]

Suggestion T: The Limits on the Shared Losses Repaid Should be 10 percent for Each Performance Year for the SS&L Model

The reason this suggestion is a deviation from perfect symmetry for savings and losses is to minimize the disincentive to participate, knowing that it is human nature to be more concerned about the equal possibility of a loss than a gain.  Furthermore, to facilitate cash-flow and continued participation, I suggest (# T1) that only one-half of any computed losses for PY¹ or PY² be repaid when invoiced, with (a) the other half “carried forward” to the next performance year(s) to be offset by potential future savings, and (b) an agreement with the ACO to continue to participate and continue the obligation to repay any net loss at the end of the agreement period.

For similar reasons, I suggest (# T2) that there not be included in the final rule a “prospective” 25 percent “withhold” of any shared savings (to offset future losses) as in the proposed rule.

Data and Information Sharing By CMS

The rule proposes to share data and information with ACOs in three ways:
.

1. a list identifying assigned beneficiaries (only by name, date of birth, sex, and health insurance claim (HIC) number) “at the beginning of the agreement period, at the request of the ACO”;
2. annual and quarterly (latest twelve months) claims-based composite summary reports, by categories of patients (e.g., conditions), likely similar as provided under the PGP Demo; and
3. monthly transfer of claims records, in a standardized form containing most data elements, including identifiers of the beneficiary and of all providers of services and suppliers, if requested by an ACO following a communication by an ACO PCP with a beneficiary who does not choose to “opt-out” of such data transfer.

Pursuant to my suggestions in Part One regarding assignment, for the SS&L model a list of “aligned beneficiaries” should be shared with the ACO, but only after those “tentatively aligned” beneficiaries, following notification by CMS (# E1), have chosen not to “opt-out” of the program (# E2).  Also, pursuant to suggestion D, there should be no identification of “attributed beneficiaries” to an ACO in the SS model.  [Please see Part One for a complete discussion of these suggestions.]

I agree with the objectives of providing claim-based summary reports to monitor an ACO’s progress over time, especially if their status relative to an all-ACO comparison is included.  Selected elements of such information should probably be posted on the “ACO Compare Website” I suggested (# A3) in Part One.  However, it must be recognized that there will be a time-lag between the data summarized and the receipt of the report, and thus, such information will have virtually no value for contemporaneous monitoring of an ACO’s activities or caring for beneficiaries.

Suggestion U: There Should be No Transfer of Beneficiary Claims Records to an ACO Under Any Circumstances

I strongly disagree with the proposed monthly transfer of beneficiary claims records, for three reasons: (1) relevance, (2) privacy, and (3) cost.   Although the rule notes the interest on the part of stakeholders in securing “timely data” on their patients, claim-based data simply cannot be timely.  By the time a claim for a service is (a) submitted, (b) processed and adjudicated, and (c) compiled and extracted, several months will have elapsed.  By the time the monthly transfer is (a) received and properly “loaded” on an ACO’s system, and (b) analyzed by the ACO’s or their consultant’s staff, several more months will have elapsed.

So what relevant purpose can be served with such claims data?  Certainly nothing to do with the care of a patient “in real time”.  This most relevant purpose (patient care) can only be served with clinically-precise, record-based data, especially when in digital form.  It will be extraordinarily difficult to succeed as an ACO without a shared electronic health record (EHR) containing such immediately-retrievable data and concurrent decision-support tools.

Providing claims data to ACOs could undermine the movement to EHRs if ACOs instead invest in free-standing programs to analyze claims data.  In the interim, ACOs without EHRs should focus on better communications among their providers and either (a) establishing some central repository for copies of all clinical records, at least for current episodes of care and high-risk individuals, an/or (b) encouraging their patients to participate in one of many commercially-available “personal health record” (PHR) services.

Some providers have argued that even with an EHR they need claims data in order to obtain information about their patients when they receive services from non-ACO professionals and providers.  Again, such data can only be available “after-the-fact”, and thus can have little, if any, relevance to the episode “at hand”.  More importantly, technology and regional cooperation are advancing rapidly, offering far more effective and efficient solutions than claims data to secure such information.

For example, PHRs tied-in with EHRs allow a patient to authorize any provider to transmit (i.e., “push”) their data to their PHR/EHR, or authorize their ACO provider to receive (i.e., “pull”) their data from non-ACO providers.  In addition, regional health information exchanges (HIEs) are forming to facilitate access to area-wide data from all providers, especially from hospitals.

Regarding the subject of privacy, the prerogatives of individual patients and providers, protected by various laws, must be respected.  I agree that giving a beneficiary the opportunity to “opt-out” of the claims data transfer would be sufficient; however, I am concerned that some Medicare beneficiaries may not opt-out simply to “go along” with what their doctor asks for or to “get along” with her or him.  If the claims data transfer proposal is included in the final rule, this should be closely monitored by CMS.

The rule proposes no equivalent opt-out opportunity to individual non-ACO professionals or providers who are identified in the beneficiary claims, and therefore, their prerogatives might not be respected.  The rule may be incorrect in its interpretation of the “routine uses” exception in the Privacy Act of 1974 not to need consent.  As I wrote in the April 7 blog:

This exception applies when the disclosure ‘is compatible’ with the purpose of the original collection.  A good case could be made that the data was obtained solely for the purpose of adjudicating a unique claim for payment.  Therefore, sharing … information that is not essential to the administration of either the Shared Savings Program or the Medicare program, can hardly be considered ‘compatible’.

If the claims data transfer proposal is included in the final rule, CMS may need to de-identify the non-ACO individual providers of service to be in compliance.

Finally, there is the issue of costs.  It will be a non-trivial expense for CMS to develop and operate, either directly or with contractors, the systems and processes to compile, transfer, administer, and monitor ACO-requested beneficiary claims data.  I am equally concerned with the extraordinary effort and expense on the part of an ACO to do anything meaningful with such data once they get it.  I believe any ACO would be much better served in investing the comparable effort and expense to enhance their data informatics capacity on their EHR-based analytic systems.

Conclusion

I continue to believe that the concept of accountable care organizations to improve the effectiveness and efficiency of services to Medicare fee-for-service beneficiaries, at least as embraced by the statutory guidance, is a valid one with tremendous opportunity for success.  However, the MSSP has already shown itself to be an extraordinarily complex endeavor, with many “simultaneous moving parts”, which if not perfectly synchronized will not function successfully or possibly not even “get off the ground”, i.e., little participation.

In response, I have offered specific suggestions for a “major makeover” of the key design issues.  These are all synchronized to achieve not only “meaningful participation”, but also the other elements of a “four part success”.  These suggestions deal with: Part One, the sharing model(s) and beneficiary assignment; Part Two, the benchmark, adjusting performance-year expenditures, and determining savings (and losses); Part Three, evaluating quality, determining and limiting payments (and repayments), and specifying CMS data and information sharing.  My sole objective for these blogs has been to expand the debate among both stakeholders and CMS.  Thank you.

 Email This Post  Print This Post

This entry was posted on Monday, June 6th, 2011 at 11:03 am and is filed under All Categories, Medicare, Payment, Policy, Quality, Spending. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.