Have you ever wondered where weight loss centers are located? After I have had friends join weight loss centers like Jenny Craig, Weight Watches, and Lindora I wonder what factors lead to where these weight loss centers are located. These questions lead me to my research question of "In Los Angeles County, do higher income areas have a higher number of weight loss centers compared to that of lower income areas? Does a lower income area with a high obesity rate have access to a weight loss center compared to that of a high income area with a low obesity rate? What factors play a role in the placement of these weight loss centers?" After doing some research, the United States Department of Health and Human Science stated "There is no question that the rates of obesity and type 2 diabetes in the United States follow a socioeconomic gradient, such that the burden of disease falls disproportionately on people with limited resources, racial-ethnic minorities, and the poor." (1) When I read this, I knew that I was onto something and that there might be a correlation between the two and how could I show this using GIS.
Map 1
Map 2
Data
I obtained the obesity percentage shapefile and rates (OBS_PCT) from the Y drive and the LA_cnty_tracts_MHHINC layer from the Y drive. The weight loss center points I found through researching on Google maps. To conduct the location allocation, I used the roads layer from the network analyses lab data and my weight loss centers.
Methods
In order to create Map 2, I geocoded the weight loss centers and put them over the obesity percentages for each county. I performed a spatial join of weight loss centers to the health districts which gave me Map 3. Using the kernel density tool, I created Map which based on the weight loss locations and how clustered they are. Figure 1 was created using the Moran's I tool and by putting in the OBC_PCT to see if the obesity percentage rates were clustered. With my weight watcher points being spatial, I was able to create Figure 2 and run the nearest neighbor tool to see how clustered my points were and if I had a high amount of confidence in my data. For Map 5, I wanted to see how the points would fall when I place them into the median household incomes for 1999. The location allocation (Map 6) was created using the 72 facilities and the 26 health districts by using the feature points to polygons tool. These were the settings in the layer properties:demand to facility, minutes for impedance, maximize coverage, 72 facilities to choose and 10 for the impedance cutoff.
Results
Map 3
Figure 1
Figure 2
Map 5
Conclusions
From Map 2 and 3, it seems as though there is not a high number of weight loss centers in areas where there are high obesity percentages. In fact, if you look at top four highest obesity percentage, there are no weight loss centers there. My results from Map 3 and 4 show that there is a high clustering of Weight Watchers, Lindora's and Jenny Craig's in the West and West Vally health districts even though they have some of the lowest obesity percentages. In Figure 1, the Moran's I tool calculated the obesity percentages and found they are clustered and with a z-score of 3.63 of and a p-value of 0.00, there is a high confidence level for my results. The average nearest neighbor results (Figure 2) show that my weight loss centers are clustered and if you match up those results with Map 5, one could try and make a correlation that they are clustered in middle to higher income areas. My location allocation would offer a solution on how to lower the obesity percentages by giving a visual representation of the nearest Weight Watchers, Jenny Craig or Lindora within 10 minutes of each health district. It is hard to make a really bold conclusion in my results as to why the weight loss centers are not located in areas with the highest obesity percentages, but my guess is median household income. In the kernel density map and the median household income maps, the highest amount of weight loss center locations are in the west areas of Los Angeles and also where the highest median household incomes are. The location allocation map could be used in order to bring down these obesity percentages.
References
(1) http://www.ajcn.org/content/79/1/6.full
