Concept: Discrepancies Between MCHP and Manitoba Health Published Reports

Problem:

Using PHIS (Population Health Information System) or annual medical services data at MCHP, the total fee and/or N-services counts show discrepancies from those given in annual Manitoba Health Reports.

Discussion:

There are several interrelated factors that contribute to this problem:

1. Orientation/Intent

Manitoba Health Insured Services Branch is generally reporting financial information for business purposes: the cost/payouts occurring between April 1 and March 31 of the given fiscal year regardless of the date a cost was incurred. They use normal accounting practices. MCHP refers to this orientation as "fiscal processing year".

Other reports developed by Financial Services at Manitoba Health wait for the majority of services to be processed and tend to be similar to MCHP fiscal service year data.

MCHP and PHIS datasets have been modified to investigators' specifications so that a fiscal year includes all services provided between April 1 and March 31 regardless of the eventual date on which Manitoba Health processed the claim. This is an utilization point of view and is known as the "fiscal service year".

2. Filters applied to data

Manitoba Health does not eliminate claims based on dates even though a claim/cost may be incurred up to five years in the past. Each cost is a current payable and is processed as a current entry (including negative adjustments).

MCHP has removed claims dated prior to the specified fiscal service year. The common assumption has been that any data removed will be balanced by an equivalent amount of data added during later periods. This concept is referred to as 'symmetric adjustment' (below). If it is true, then the figures will be approximately the same (but never completely in agreement). In fact, the number of physical records is usually similar in such circumstances, but the fee and N-services counts are considerably different for entirely different reasons than the choice of period/intent/overrun.

3. Net Value Factors

In the Medical Services data from Manitoba Health, the variables NSERV and FEE can be negative values due to the need to adjust accounting information, cheques or implement revised retroactive negotiated fee schedules. For Manitoba Health's files, each fiscal file may contain claims dated prior to April 1 which may show negative values. The matching positive value claim being adjusted may have been processed in a prior year and reported then.

Three outcomes are possible:


1. If the negative claim is intended to "deny" a fee previously paid out, then the prior report is over counting "true" fees, and the current report will show a reduction in annual costs or services not entirely consistent with current services and fees. From a financial perspective, this is correct.
   
   
2. If the claim is being retroactively adjusted (+/-) then there will be another positive claim in the current file which shows the new total fee/nservices. Since the original positive claim was not present in the current file, the NET effect of this change is to show zero net services (+1,-1) and a small NET change in total fee (usually increase but not always).
   
   
3. If the original claim is also present (regardless of date) then the result shows one or more NET services and a complete NET fee (same as a single positive claim). Remember, Manitoba Health is reporting cost/payout data and showing an estimate of number of services associated with that cost. The incomplete data from NET adjustments is a current cost averaged over all "complete" services (net=1+) for the year. The actual fee and N-services involved cannot be repeated since they were reported in a prior fiscal accounting period.

MCHP fiscal service years MUST include the original claim, and when an adjustment is found there should also be the supporting claim for the new full fee. This means that all of the information that MCHP attributes to the fiscal service year has corrected NET values (at least as of the date the fiscal files were produced). The net number of services is also corrected. The difference between MH and MCHP tables is pronounced since the idea of "symmetric adjustment" is no longer true. The MH data includes partial or negative fees for early claims, often with a net N-services of zero. MCHP substitutes correct full-fee claims dated late in the fiscal period, usually with a positive N-services. The impact of the differences on a given set of reports may be two-fold. MH tends to under-report the fiscal service year (not all claims) and minimize the impact of the delayed payment claims. MCHP tends to maximize the cost and services data (more claims; more services, higher costs) due to the availability of more claims relevant to the fiscal service year.

Strict attention to the orientation of the audience/research design is essential in attempting to compare utilization results to the essentially financial-based information contained in annual reports.

4. Number of Services versus Number of Claims

Manitoba Health accounting practices. tend to report number of services rather than number of paper or computer claims records. Each paper form can generate several computer records (one per unique service tariff) and each of the computer records can represent several repetitions of the same service.

MCHP files can reproduce the same details. Many investigators visualize the computer record as a single event/service/contact.

This may be reasonable for APV's but is incorrect for in-hospital services/visits and many ancillary services. Investigators should always account for the sum of NSERV for each TARIFF rather than the simple frequency of TARIFF. Other factors such as visits, diagnoses, physician specialties may be better presented using simple frequency measures. Exceptions exist, especially for services that report UNITS (i.e. fifteen minute blocks for anaesthesia or psychiatric visits) as well as NSERV and TARIFF.

5. Fee Schedules and Retroactive Adjustments

Manitoba Health routinely updates completed fiscal years to show retroactive changes in the negotiated fee schedule. This causes confusion among the claims, profiles and report contents. For example:


• 1990-1991 - MH adjustment Sept/93 - too small to justify producing new file for MCHP
• 1991-1992 - obtained Aug/92; replaced July-June/92 in Sep/93 replacing April-June/91 in Fall/1994 to correct true fiscal year. Total increase > $4 Million.
• 1992-1993 - obtained Aug/93; replaced Sep/93; replaced again, Fall/1994
• 1993-1994 - obtained Aug/94; to be replaced in Nov/94.
• 1993-94 profile not available until Nov/94

For 1992 to present, the changes being added retroactively are more complex. There was a net decrease in fees, which is being retroactively recovered from the physicians billing system. The reduced fees are applied regardless of the date of the claim so prior fiscal service year data may be affected by the changes.

When MH processes a complete revision to the fee schedule(s) they rebuild the entire medical dataset and replace the fee values directly. This means that the Physician Practice Profile reports are replaced and the physician payment system is updated to reflect the adjusted balances for each physician. MH does not insert the negative/positive adjustments in this situation since that would greatly increase the size of the Permanent Statistical File (medical fiscal year).

MCHP has a system of tagging medical claims with a code that defines fiscal process year and revised fee schedule designation (or "generation"), if any. Using this system users should be able to trace the history of each claim through all NET and analysis activities.

Therefore ...

... When Comparing Manitoba Health Reports to MCHP Results You Must:


1. Have the appropriate twelve months of process data, using the correct fee schedule and positive/negative rules; You must, therefore, also keep abreast of the ongoing revisions to fee schedules and files at MH.

   Note: At MCHP, please remember to keep the variables which define fiscal year-month-generation (on medical claims) to allow users to explain, repair or extend medical claims abstracts following changes in the master data.

2. Count N-services where appropriate. MH seldom reports the number of paper and/or computer records.
   
   
3. Net all fees/services. Include all entries having negative or zero fee and/or NSERV values BEFORE AND AFTER the net-claim process. All analysis-ready SAS files at MCHP have already been prepared; new or revised medical files require net value and fiscal service year adjustments.
   
   
4. Be aware of specific exclusion/inclusion rules such as "exclude chiropractic services". MH reports generally clearly state such exclusions. MCHP investigators frequently eliminate a variety of services including chiropractic services, Long Term Care patients, non-resident postal codes and so on. This decision tends to be project-specific and comparing results to existing MH reports may be inconclusive.
   
   
5. Be aware that some MH reports are based on accounting principles rather than medical services/fiscal service year rules. In some reports MH also may have access to information not available to MCHP.