Concept: Costing Using the 1999 Cost List for Manitoba Health Services

Introduction

This concept describes methods of costing resulting from the development and use of the 1999 Cost List for Manitoba health services. NOTE: For an overview of general costing methods (sources of data, types of costs, and approaches) and the methods of costing specific health services (hospitals, physicians, home care, personal care homes and pharmaceutical/prescription drugs) that have been used in MCHP research over time, please see the Costing Methods: An Overview of Costing Health Services in Manitoba concept.

In the report Cost List for Manitoba Health Services by Jacobs et al. (1999), MCHP developed a Manitoba Cost List for health services in Manitoba to easily integrate with the provincial administrative databases. The list includes hospital inpatient services and outpatient procedures, home care, long term institutional care, and diagnostic procedures.

The Manitoba Cost List has been tested on 200 patients from the Aging in Manitoba (AIM) study. Costing has been completed on physician claims, hospital claims, pharmaceutical claims and personal care home claims.

Costing has also been done using the cost list in the ' Direct Medical Costs of Low Birth Weight from Birth to Seven Years of Age in Two Canadian Provinces ' (Christine Newburn-Cook MCHP Reference 2001-006).

People working on costs should be aware that CMGs™ and associated RIWs™ are more commonly being used. See the Calculating Hospital Costs Using Cost Per Weighted Case (CPWC) / Cost of a Standard Hospital Stay (CSHS) Values concept for more information.

Note : The non-adjudicated pharmaceutical file contains all medications that are not covered by Pharmacare, The Nursing Home Drug Program or Family Services. These drugs may be paid for by a third party such as Blue Cross, Department of Veteran's Affairs (DVA, Government of Canada), Medical Services, or Winnipeg Social Services, or paid directly by the patient.

This concept also provides a link to relevant SAS code. Please see the SAS code and formats section containing the following link to: Costing - SAS Programs - (internal access only).

Details of Costing Method

A cohort of 200 Aging in Manitoba (AIM) patients were selected from the larger AIM study. The claims for this subset were drawn from the MCHP claims database for calendar year 1996. The AIM database contained information about the birthdate, place of residence (postal code) and sex of each patient in the cohort. By merging the AIM data with the MCHP claims database, utilization information and costing information was associated with each patient. Five different data sources were used: physician, hospital, personal care home, pharmaceutical, and home care.

For more information, see the costing files table.

Discharge Claims: Discharge claims are often used to select hospital claims for cost analysis. However,

• Individuals who are present in hospital in 1996 would be excluded from analysis if they were discharged in a later year.
• Claims made in 1996, when most of the treatment was provided in 1995 would be incorrectly included for analysis.

Our assumption is that these two conflicting effects will cancel each other out. However, differences in usage patterns from year to year could result in more cost being ascribed to certain services and less to others. There may also be issues of incomplete data due to late filing in medical claims files and other databases.

Generally in costing work you will want to collect all possible data for a given year. This usually means waiting until the next year or at least until the first quarter is available.

1. Medical / Physician Claims

A. Medical claims files contain information on a full fiscal year based on 27 months of data. The use of raw data is also possible.
• Key variables from this file: PHIN, service date and fee.
  • The service date allows the investigator to limit claims to the time period of interest.
  • The fee reflects all of the costs for the service, taking into consideration where the service occurs and what type of physician is providing the service (GP or specialist).
  
  
• SAS programming (internal access only)

B. The Physicians Claims data now allows billing under new agreements compensating physicians' for time on-call. Most of these on-call cases are for anaesthetists.
• Starting in 1998/99, such claims contain dummy values for scrambled PHIN, Registration Number (REGNO) / (REGNO_CODE), SEX, and BIRTHDATE. The associated costs cannot be attributed to actual patients but are of interest when reviewing physician costs.
  • These two scrambled PHINS must be explicitly dropped from studies that define a cohort or population directly from physician services claims.
    
    See the Physician "Dummy" Claims Data section of the Manitoba Health Insurance Registry / MCHP Research Registry - Overview concept for more information.
    
    Thanks to Deborah Malazdrewicz at Manitoba Health
  
  Further questions regarding physician costs and radiological services should be directed towards Carol Baskervielle, Manager, Complex Claims, Insured Programs and Pharmacare Services, Manitoba Health.

2. Hospital Claims

• Key variables from this file: PHIN, admission date, separation date, length of stay, RDRG® (Refined Diagnosis Related Group) and variables used by the RDRG® grouper, LOS (Length of Stay), CWC (Cost per Weighted Case), DPG (Day Procedure Group) weight and DPG.
  
  • There is a prepared file with the costing information associated with each RDRG® averaged from 1993-1995 data and another for average DPG values from 1993-1994.
    
  • The file containing the RDRG® cost information contains the following variables: average length of stay for typical cases (ALOS), trim, cost per case, average cost per day and marginal cost per day for each RDRG®.
    
  • The DPG cost information file contains the variables DPG and cost per case.
  
  Note : Special attention should be given to changes in RDRG® and DPG coding from year to year.
  
  E.g. from 1993-1994 to 1995-1996, an extra DPG was added (67 Electroshock Therapy (ECT)). From 1995-1996 to 1996-1997, the DPG 44 was changed from Chest Wall Procedures to Minor Bone Procedures. The claims from the hospital file were first divided into outpatient and inpatient claims and evaluated separately.
  
  A. Inpatient Claims The inpatient claims were evaluated according to the following formula : cost = average cost for RDRG® + ((LOS-ALOS) * marginal cost of stay per day)
  
  Lengths of Stay (LOS) longer than the trim date are considered to be outliers and will require special consideration. It is questionable whether the marginal cost of stay per day will apply in this case.
  
  • SAS programming (internal access only)
  
  B. Outpatient Claims Outpatient Claims are evaluated on the basis of their associated DPG. RDRG® average costs are usually determined from several years of costing data whereas DPG is calculated yearly. For analyzing the 200 AIM patients, the DPG cost values from 1993-1994 were used so it would be consistent with the cost values used for calculating RDRG® costs (1993-1995). When the more recent costing information was used to calculate costs from 1996-1997 data, the differences were significant and predictable.
  
  For further information on costing hospital claims, please refer to the Hospital Costing in the 1990's concept.
  
  When using the "Cost List for Manitoba Health Services," Jacobs et al. 1999 tried to identify any potential issues that should be considered. Here are a couple of comments.
  
  1. The weights that are assigned to RDRGs® were developed using calendar year 1991 and 1992 cost data from Maryland, adjusted to reflect Manitoba length of stay experience during fiscal years 1993/94 and 1994/95.
     
  2. The cost per weighted case (CWC) is based on 1993/94 costs for all Manitoba hospitals (and the above mentioned weights). The CWC excludes all physician, capital, depreciation, education and research costs--the CWC includes all direct nursing, diagnostic and therapeutic, supplies and drugs, and a share of overhead/administration costs.
     
  3. The estimated cost per case (as well as marginal daily cost) is therefore based on the practice and mix of cases in Maryland and Manitoba in the early 1990s, and on the average provincial cost for 1993/94.
  
  Given these facts, in applying the cost list to years other than 1993/94 it is assumed that there have not been material differences either in the mix of cases or the relative resource requirements for different types of cases. Because the weights that are assigned to cases are relative, changes in the mix of cases could affect the weights that are assigned to individual RDRGs. Secondly, if all types of cases have changed in resource requirements equally (e.g., an across the board 10% decrease in LOS), there is no problem because there would be no relative change in resource requirements. However, if some types of cases experienced significant reductions in LOS while others experienced little or none, static weights would not reflect this shift.
  
  Finally, the CWC is in 1993/94 dollars--there may be situations where it would be desirable to apply an inflationary factor to represent current rather than constant dollars. It is also worth noting that the CWC is the provincial average so does not reflect the variation in cost that is seen between different individual hospitals and types of hospitals.
  
  When using these data it will be important to consider the above assumptions, and their potential impact on the results. In a perfect world, weights for each RDRG® would be generated on an annual basis (or even every 2 years), and costs associated with the cases would be identified. Weights are produced on an annual basis for CMGs, but costs were last determined for 1997/98.

3. Pharmaceutical Claims File

• Key variables from this file: PHIN, drug identification number (DIN), the total drug cost and the professional fee.
  • The total cost for a prescription is the sum of the total drug cost and the professional fee.
  
  Note: The professional fee is zero in two situations:
  1. When the prescription is filled in drugstores such as Superstore where no professional fee is charged.
  2. Claims from Nursing Home file. These claims do not contain a professional fee because they pay a set fee to the pharmacy per month based on the number of beds in the nursing home. The rate for Winnipeg is $26.45 and $26.95 outside of Winnipeg. Some governmental programs that provide Nursing Homes with basic medical supplies that are not recorded in the pharmaceutical file. These costs must be separately considered when costing the pharmaceutical file.
  Note: The costs associated with drugs dispensed in the hospital are represented in the RDRG® and DPG costs. The out-of-hospital pharmaceutical costs can be found in the Manitoba Drug Benefits & Interchangeability Formulary documentation on the Manitoba Health web site (http://www.gov.mb.ca/health/mdbif/). For further information about the pharmaceutical files, see the Drug Program Information Network (DPIN) Data glossary term.
  
  For more information about drug costs and dispensing fees, please contact: Jack Rosentreter, Pharmaceutical Consultant, Insured Programs and Pharmacare Services, Manitoba Health.
  
  
  Non-adjudicated Pharmaceutical Claims File
  
  Non-adjudicated pharmaceuticals are difficult to price:
  • Pharmacare does not record the total drug costs since they do not cover it.
  • Inherent complexities in the unit value for medication makes calculating cost difficult even when the unit cost is available.
  • Some DINs are not covered by the drug interchangeability formulary. To price these non-essential items (i.e.: some diabetes supplies, bandages) the researcher must contact the pharmaceutical company directly to obtain a raw cost and add a percentage to the manufacture cost.
  
  Colleen Metge and Sandra Peterson developed a method to bypass this problem. They created a database with the mode average cost for DINs in the year of interest using the pharmaceutical file as the database. More information about costing using this method is available in the Pharmaceutical Use Project Documentation available in the concept dictionary.
  
  • SAS programming (internal access only)

4. Personal Care Homes (PCH)

• Key variables: PHIN, start date, termination date, and rate.
  • Cost associated with the personal care home file is determined by multiplying the number of days of stay within the personal care home by the rate.
  • Any start date prior to January 1, 1996 will be readjusted to that date in calculating the number of days of service in the 1996 calendar year.
  
  
• SAS programming (internal access only)
  
  For more information about Nursing Costs (95/96), please contact: Ed Golembioski, Acting Director, Financial Services, External Programs and Operations, Manitoba Health

5. Home Care

The cost list suggests that home care claims should be evaluated according to the health care provider title and the number of hours worked. The current home care claims provided by the AIM study do not allow for cost evaluations to be made.

Future MCHP plans include obtaining home care data directly from Manitoba Health in a form which supports cost analysis.

MORE INFORMATION

• Canadian Institute for Health Information. Resource intensity weights: Summary of methodology 1994/1995.

• Fetter R et al.(1990). RDRGs: Refined diagnosis related groups. Version 2.3: Definitions manual. Karen Schneider and Henry Dove, Health SystemsConsultants.

• Health Systems Consultants Inc.(1995) RDRGs: Refined diagnosis related groups Version 9.0/13.0: Definitions Manual.

• Institute of Health Economics: A National List of Provincial Costs for Health Care: Canada 1997/8 , July 10, 2000

• Jacobs P, Bachynsky J, Hall, EM (1995). A Manual of standard costs for pharmaco-economic studies in Canada: feasibility study. The Canadian Coordinating Office for Health Technology Assessment.

• Jacobs P, Hall EM, Bachynsky J (1996). An Alberta standard cost list for health economic evaluations. Edmonton: Health Economics Research Centre, mimeographed.