Concept: Residential Mobility

Introduction

Residential mobility is defined as the movement of a Manitoba health insurance registrant within the province that results in a change of six-digit postal code or municipal code (a code assigned to each municipality in Manitoba for administrative and funding purposes) but does not affect continuity of one's eligibility for health insurance coverage. Residential mobility should be differentiated from migration. Migration is concerned with the movement of individuals at an inter-provincial or international level, whereas residential mobility focuses on intra-provincial movement. Information related to migration is available in the Follow-up Issues concept.

Residential mobility is usually defined at MCHP from the Manitoba Health Insurance Registry. The registry contains longitudinal histories for each individual registered with the Manitoba Health Services Insurance Plan since 1970, including residential postal code information. Changes in postal code are tracked over time from "snapshots" of registry files supplied to MCHP by the provincial Ministry of Health on a semi-annual basis (June and December) and linked via the PHIN. Given that these are snapshots, a maximum of 2 changes in postal code can be detected each year.

Residential mobility may be defined using other data sources, including:


• Case registries for patients of specific clinics/groups;
  
• Longitudinal health survey data (i.e., National Population Health Survey).

As well, Statistics Canada Census data can be used to define the aggregate, or overall, percentage of residential mobility of populations within enumeration/dissemination areas (also known as DAs).

Background

Mobility data allows for comparisons to be made between the morbidity and/or mortality of movers and non-movers. Exposure to a particular neighbourhood for a longer period of time may produce area-specific health effects that differ between long-term residents and those recently moving to the area. Furthermore, selective mobility out of a disadvantaged neighbourhood may leave behind 'unhealthy survivors' (Roos et al., 2004) .

The database of Registry snapshots maintained at MCHP record changes in residential postal code over time. This allows a study cohort to be divided into those who never moved over a given period and those who moved one or more times. Data on residential persistence (the percent of individuals living in the same postal code year after year), as well as the correlation between residential persistence and socio-economic status (SES), are presented in the Residential Persistence concept.

While residential mobility itself may not be a significant factor in the well-being of an individual - although frequent moves can be stressful - it has possibly undesirable consequences: disruption of friendships, especially among children, and a lack of continuity of schooling and health care provider.

Residential mobility is of importance to individuals assessing the planning and delivery of health care services and has been used in several studies. Changes in location among people with severe mental illness (SMI) can lead to discontinuity in health care treatment, increased social isolation, and a severance of ties to friends, neighbours and relatives. Individuals with SMI are prone to changes in residential location more frequently than that of the general population (this generalization, however, is not fully verifiable due to a lack of databases expressing person-specific information on health status and changes in location of residence).

Registry data can be combined with income data for further studies. The income data used is derived from the Dissemination Area files of the Canadian Census. A Postal Code Conversion File (PCCF) links DA information to postal codes within the Registry. Multiple DAs associated with a single postal code are resolved by taking a population-weighted average over all the DAs. The Census is taken every five years, so the average household income values attached to each postal code are taken to be constant between Census years. This is of course not true, neighborhoods change over time, but without more frequent sampling some assumption like this has to be made.

Combining Registry data with area-level income data from the Census allows a researcher to determine if a residential move is also a move to a neighborhood of significantly different average household income. Moving to a neighborhood of a different SES does not necessarily mean the family's financial picture has changed, but certainly indicates a change in neighborhood characteristics. Frequent changes in residence may also indicate financial instability. In this case, mobility can be used as a substitute for other unmeasured factors such as changes in employment or marital status, or even an improvement in a family's financial situation. For example, movers from the suburbs to the core are different than movers from the core to the suburbs. The inner core (low SES) has different characteristics than the outer core (higher income, but still older houses, dense population) (Lix et al., 2006).

Manitoba data, like the American Panel Study in Income Dynamics, show considerable persistence in neighborhood environments (Kunz et al. 2003). Fifty-six percent of the nine-year birth cohort remaining in the province had no residential moves between the ages of 8 and 17.5, while another 24 percent moved only once. Adopting the five year interval used by Kunz et al. (2003), 73.7 percent of those in our 1978-1987 birth cohorts and staying in Manitoba did not move (change postal codes) over the 1991-1995 period. The log mean income of their 1991 Manitoba census area correlated 0.937 with the log mean income of their 1992 census area and 0.840 with the 1995 census area. With the cohorts and their parents five years older, 79.7 percent of the cohorts did not move over the 1996-2000 interval.

Methodology

Residential mobility may be operationalized in a number of ways:


• At least one change in postal code in the specified time period may be used to distinguish movers from non-movers (i.e., individuals who had no change in postal code).
  
• The total number of changes in postal code in the specified time period. This information may be used to characterize infrequent movers (i.e., a single change in postal code) from frequent movers (i.e., more than one change in postal code).
  
• Change in regional health authority (RHA) or region of residence. For example, rural-to-rural, rural-to-urban, urban-to-urban, and urban-to-rural residential mobility may be defined on the basis of changes in postal code that correspond to movement across RHA or region boundaries (Lix et al., 2006).
  
• Change in neighborhood. For example, both individual's movement across neighborhoods or change in neighborhood may be monitored over time. This can be done on the basis of proximity, using postal codes within a census dissemination area.
  
• Direction of change in residence. For example, a change in postal code signifying a move from urban core to the suburbs may be monitored over time. See (DeVerteuil et al., 2007) for more information on this.

Strengths

• Population-based: Unlike survey data, this requires no special analytical techniques.
  
• Regular data updates: Note that, like a panel study, the Registry captures residence information for largely the same people every six months, but with a very low dropout rate. The overlap between June 2005 and December 2005 snapshots is 98.6%, not including births or deaths ineligible to be in both files.
  
• Duration of residence: Note that this variable can be determined; individuals can be distinguished as persistently living in poor neighborhoods versus transiently in such neighbourhoods.

Limitations

• Non-residential postal codes: Note that postal code may not always be location of residence (e.g., Public Trustees - homeless; because postal code is the address for head of household - older dependent children may live elsewhere).
  
• Variation in size of postal codes: Note that postal codes differ in size (i.e. rural vs. urban areas) and different scales of movement are recorded in different areas.
  
• Contact Bias: Persons with more health utilization may appear to move more simply due to the fact that they have more contact with Manitoba Health and more opportunities to update their address. Note that Census has a question about "moving in the last year." The results show 13% mobility, but MCHP data show 10-11%. Our rates may be slightly conservative.

SAS Code

The SAS code (internal access only) accounts for a specific kind of mobility (residential) by checking the Manitoba registry files every six months to determine whether postal code for residents has changed. It uses scrambled PHIN and records postal code changes from birth to age 18 for the group studied during a series of education runs, i.e. those born in Manitoba 1978-1985 (excluding 1983).