Introduction

This concept gives a brief description of the methodology used at MCHP to create an episode of home care. Over time, based on a change in the collection of data for home care services in Manitoba, the methodology fo creating an episode of Home Care has changed.

Over time, two different datasets have been used to collect home care administrative and service information:


1. the Home Care Manitoba Support Services Payroll (MSSP) Data and
   
2. the Home Care Task Management Module (TMM) of the Home Care Minimum Dataset Assessment (MDS-HC) Data. Please see the provided links for more information.

Definitions

The following definitions are provided for reference within this concept:


• Home Care - health services, such as personal care assistance, home support, health care, family relief, respite care and supplies and equipment, provided free-of-charge to residents of all ages within their homes (Martens et al., 2003).
  
  
• Home Care Episodes - are defined by the start and end dates of services being provided to a home care client. Home care episodes have been defined differently in MCHP reports by Roos et al. (2001) and Doupe et al. (2011) due to different data being available in the Home Care Manitoba Support Services Payroll (MSSP) Data and the Home Care Minimum Dataset (MDS) Assessment Data, to create the home care episodes. The databases and the methodologies used to create home care episodes in MCHP research are described below.

Home Care Data

The Home Care data available in the Manitoba Population Research Data Repository consists of two major datasets. These are described below.

1. Home Care Utilization / Manitoba Support Services Payroll (MSSP) Database

The Home Care Manitoba Support Services Payroll (MSSP) Database) was developed in 1988 by Manitoba Health. It serves as a master file for Manitoba Home Care Program clients providing information on clients, employees, and services. Historically, the MSSP was used to measure home care use in Manitoba residents, as in Roos et al. (2001).

SAS Code (Macro) for Home Care Episodes from the MSSP Data

Use the MCHP SAS macro get_mssp_homecare to create home care episodes from the Home Care MSSP data. This macro will generate episodes of care across all time in the MSSP data based on the defined acquire date. An example call of the macro is:

get_mssp_homecare(acqdt_year=2011)

An example of the get_mssp_homecare macro code is available from the SAS code and formats section below (internal access only). More detailed documentation is available at the top of the macro program code. A current version of the macro code can be found in the MCHP system macro library.

2. Home Care Minimum Data Set (MDS) Assessment Database

The Home Care MDS Assessment Data is collected by the Winnipeg Regional Health Authority (WRHA) and only provides information on Winnipeg residents. It contains case activity files, and assessment files of home care service provision.

Within this database, the Home Care task management module (TMM) stores all home care case activity data including:


1. users of home care,
   
2. start and end date(s) of service provision, and
   
3. reason(s) for the start and end of service provision.

Because the TMM provides several start and end reasons for each home care record, one time users can be distinguished from multiple time users. Figure A1.1 - Schematic Outlining the Relationship between Home Care Clients and Records in the Home Care Task Management Module in Doupe et al. (2011) depicts the relationship between home care clients and records within the Home Care MDS Assessment data.

The TMM also implements the Resident Assessment Instrument - Home Care (RAI-HC©) to collect information on a client's functional ability. As this data is not used to create home care episodes to measure home care use, it will not be described in this concept. It is important to note that it is possible to have a case activity file without any assessment files for that same person as home care services can be initiated and provided without an assessment.

SAS Code (Macro) for Home Care Episodes from the MDS Data

Use the MCHP SAS macro get_mds_homecare to create home care episodes from the Home Care MDS Assessment data. This macro will generate episodes of care containing the start and end of MDS Home Care episodes. An example call of the macro is:

get_mds_homecare(acqdt_year=2011)

An example of the get_mds_homecare macro code is available from the SAS code and formats section below (internal access only). More detailed documentation is available at the top of the macro program code. A current version of the macro code can be found in the MCHP system macro library.

For more information on the get_mds_homecare SAS macro, please read the Data Analyst Meeting Summary Notes for November 27, 2013 (available internally only).

Comparative Use of the Home Care Datasets (MSSP vs MDS-HC)

As mentioned above, both the MSSP and the MDS-HC datasets have been used to investigate and evaluate home care use in MCHP deliverables: Roos et al. (2001) and Doupe et al. (2011).

Roos et al. (2001) reported that the MSSP underestimates the number of home care users by 14 percent. Reporting differences can be somewhat attributed due to the limitations in the MSSP dataset. First, there is inconsistency in the recording of home care service provision (type and direct hours) in MSSP. Secondly, unlike the MDS-HC database, the MSSP does not contain any reasons for the start or end of service, therefore complicating the identification of one time users versus multiple users. As a result, data from the TMM provides "different (and likely more accurate) patterns of home care use compared to MSSP"; however TMM does not capture all home care clients until April 2005 (Doupe et al., 2011, p.23). In Doupe et al. (2011), they recommend that in order to most accurately evaluate patterns of home care use over time, use only the MSSP data until March 31, 2004, and then use both MSSP and TMM at least until March 31, 2006.

Method to Define a Home Care Episode of Care

This section describes the different methods used to define an episode of Home Care in two different research projects: Roos et al. (2001) and Doupe et al. (2011).

1. Roos et al. (2001)

In Roos et al. (2001), Home care start and end dates within the MSSP data were used to identify the length of time of home care use. Hospital, personal care home and vital statistics records were used to verify and update the end dates within the MSSP data. This methodology did not identify the number of or frequency or type of services that were received. The length of the home care episode also included days spent in hospital during the time home care services were being provided. However, hospital stays could be excluded when desired. Please see Section 2.3 Conceptual Issues in Roos et al. (2001) for more information on how definitions of Home Care clients and Home Care episodes, and measures for describing the population's use of Home Care Services were developed and operationalized in this project.

2. Doupe et al. (2011)

Home care case activity files from the TMM were used to create home care episodes in order to:


• identify when a client started and completed using home care services;
  
• count the average duration of home care use; and
  
• define clients who used home care services once versus multiple times.

The following rules were used by Doupe et al. (2011) to create home care episodes from case activity files. Following these rules, some exclusions and clean-up steps are described. For more information, see:


• Natalia's presentation in the Links section below; and
  
• SAS code associated with this methodology in the SAS code and formats section below (To Be Uploaded when available) .

MAIN Rules

Case activity record(s) were considered to be one unique home care episode if:


• it is a new client (OPEN REASON = "0"), or a re-opened case (OPEN REASON = "1");
  
• it is the first or only record of PHIN; and
  
• it is a transfer > 30 days after previous case was closed (OPEN REASON = "2"; GAP >= 30 days)

  * The gap is the time between closed date on previous record and open date on current record.

A record is a continuation of previous episode if it is not the first record and the OPEN REASON = "2" (transfer).

NOTE: These rules are very specific to the research by Doupe et al. (2011) and may not be suitable for other MCHP population-based research.

Other Rules

Other rules (old and suggested new) have been described within Natalia's presentation to provide instruction on the management of multiple home care records when creating home care episodes from case activity files. The new rules have been created because new data is available since the last published research and when the "old" rules are applied to new data, there are much broader implications in terms of how many people are being affected by each rule (personal communication, Natalia Dik, September 14, 2012). The suggested new rules have yet to be approved by researcher(s) and other analysts (personal communication, Natalia Dik, September 14, 2012). Table 1.0 provides a brief look at the old and new rules.

Table 1: Old and New Rules for the Creation of Home Care Episodes using Case Activity Files

OLD Rules	Suggested NEW Rules
(4) - Record not closed (#1) and followed by another record (#2): If record #2 is transfer - remove #2; keep person If record #2 is new or re-open - delete these people completely	(4) - Record not closed (#1) and followed by another record (#2): If record #2 is transfer - remove #2; keep person If record #2 is new or re-open - flag these people and their records as problematic
(5) Overlap (record #2 starts before record # 1 is closed): If record #2 is transfer - count as one episode If record #2 is re-open and difference is </= 30 days, then change open date of record #2 to day of record #1 closure	(5) Overlap (record #2 starts before record #1 is closed): If record #2 is transfer - count as one episode If record #2 is re-open and difference is </= 30 days, then change close date (closure #1 on day of 2nd open) If opensystems date on #2 = closedate on #1 - use opensystem date on #2 instead of opendate (See Notes below).
(6) If other complex reasons - delete people	(6) If other complex reasons - flag these people as problematic

NOTES:

• The Open Date is a date entered manually. It can be missing or an invalid date, or greater than Close Date, for example. The Open System Date is generated by the computer automatically on the date the record is being created. As such it can be used as a proxy for Open Date with some limitations (could be years apart from actual date) (personal communication, Natalia Dik, September 14, 2012).
  
  
• "Delete" or "keep" indicates whether a person is or is not part of the study from this point on. In some cases only selected records per person are being used; a person may still part of the study, but not be all their records are being used (personal communication, Natalia Dik, September 14, 2012).

Exclusions

The following were excluded by Doupe et al. (2011) prior to the creation of home care episodes with the case activity files:


• Fake facilities
  
• Fake PHINs
  
• "non-admissions" (i.e. close reason =1)
  
• Missing open date
  
• Invalid dates
  
• Duplicate dates (* methods different if you need to know facilitates or if you only care about dates)

Clean-Up

Doupe et al. (2011) cleaned up the records by changing missing close dates and dates on records with negative LOS, or removing select clients and/or records. It is briefly described below. For more information, please see the section on Preparing TMM Data for Use in Appendix 1 of Doupe et al. (2011).


• If CLOSE REASON = "not admitted to home care":
  
  1. and only 1 record, then client and record removed from raw data
     
  2. and >/= 2 records, the record only removed from raw data
  
  
• If CLOSE DATE precedes OPEN DATE:
  
  1. Replace case activity open/close dates with system-generated dates
  
  
• If OPEN/CLOSE DATES identical:
  
  1. Removed record to avoid over-counting home care days open
  
  
• If records partially overlapped, the strategies used to modify/remove these depend on the degree of overlap and the OPEN REASON of the non-first record.