Discussion

This analysis identified 7 vulnerable DAs in the North Shore of Greater Vancouver with limited social and physical access to Lions Gate Hospital services (Vulnerable Populations). Conversely, no DAs were identified as vulnerable based on physical access (i.e. travel time) to chronic disease services, due to fact that no DAs exceeded 15min in travel to any CD service.

These findings suggest that more attention should be given to these DAs, that face both physical and social access limitations. The fact that they are spatially clumped could even mean that they could be served in unison (Vulnerable Population Zoom), perhaps by specific out reach programs. However, the fact that these 7 DAs are less populated (average = 246 people/DA; see Vulnerable Populations Table in Results) compared to the North Shore DA average of 588 people/DA means that attention may not be as urgent in the grand scheme of health care – especially considering we are discussing only 15min of travel.  Nonetheless, this remains an issue of equity vs. efficiency: is it possible to provide an efficient model of health care delivery while keeping it equally accessible for all populations?  Most likely not.  That is not to say though that we should overlook these populations, instead we should look at alternatives and possible solutions for underserved areas.

There are a few limitations that should be addressed in this study. First, this study assumes that each individual owns or has access to a car and does not incorporate the option of public transportation (a crucial component for future analysis).  Furthermore, though our network analysis model does take into consideration speed limits and road length of routes, other ‘costs’ such as intersections and delays due to traffic and traffic lights were not incorporated into our model. Thirdly, choosing the incident point as the centroid point of the DA may not be very accurate as the point where most people are actually living and departing from to access services. This could be a major limitation in areas of greater sized DAs, but may not play as big a role in our analysis considering that most DAs only spanned a few blocks.  Similarly, there are always assumptions when describing census level data and consequently there is a failure to reflect true individual demographics,  commonly known as the ecological fallacy (Openshaw 1984) Lastly, because our geographic region (North Shore) was not very large, and travel times rarely exceeded 15min, it is somewhat ambiguous to define >15min of travel time as ‘long(er)’. In other contexts where travel times are much longer, a one or two hour cut-off may be more appropriate, especially with respect to emergency services and the ‘golden hour’ of urgency (Peleg 2004).

This study also demonstrates that capacity of which GIS can play in health research and policy. In further research it would be interesting to compare these results to other municipalities, both urban and rural. Perhaps this approach presented here could be used provincially (or even nationally) to derive a standardized score based on physical and social access. In the past, standardized scores have compared the regional availability of services by comparing the ratio of population to services (and number of GPs), (Joseph 1984), but these methods don’t account for the cross flow between specific regions (e.g. Census Tracts) that are assigned to create the score. Nor does this approach address physical travel time or distance decay effects: utilization of services are likely to diminish the further you are away form them.  This brings up another key point in our study and future research: utilization. We have demonstrated access, but it would also be very beneficial to relate the level of access (and travel time) to actual utilization of services seeing as it is possible that individuals may not use the closest facility when faced with an option. In one English study, only 56% of the population of 3 counties in rural Eastern England were registered with their closest general practitioner (Haynes 2003). More importantly they found, as expected, that utilization rates decreased with increase in travel time. This was also demonstrated in BC in another GIS based study (Lin 2004). Lastly, it would also be extremely important to consider health outcomes as a function of access – both physical and social.

The future potential of GIS to help develop and plan equal (and efficient) access to health care services is perhaps closer than we think. After all, access to GIS is right in front of us.