Lessons from HBO's 'Weight of the Nation'

This past week HBO debuted a four part documentary called The Weight of the Nation on the obesity epidemic in the US.  The clip above from part four talks about how geography can have a big effect on one's health due to the socioeconomic factors which surround these areas.  The fourth episode which deals with public health challenges can be seen below and is relevant to much of the research I have been doing for PUSH-Healthcare for All PA on Pennsylvania's uninsured.  The episode can be seen below.  All four parts can be seen at the above link in italics.  I'll review this episode in particular.

The program does a good job of presenting the data and issues related to the obesity epidemic in the US.  The impacts of their actions, intentional or not, are discussed at length including those on health care costs.  Various solutions to the problem are discussed such as ending farm subsidies, creating more park space in inner city areas such as Philadelphia County, and adding more bike trails.  While all of these are good things which I fully support, how much does the obesity epidemic really contribute to the high cost of health?  According to The Incidental Economist only around $25 billion in extra health care spending in 2004 can be attributed to health problems related to obesity because other non obesity related diseases such as prostate cancer are just as prevalent in the US relative to other countries with universal care such as Japan, Germany and the UK as can be seen in the graph below.  Diseases below the horizontal line in the graph such as Hepatitis B and Bladder Cancer are more prevalent in those countries.  You can see more cost analysis at this page.

In the opening credits of the episode above we can see that one of the sponsors of this documentary is Kaiser Permanente which was skewered for its profiteering practices in the film Sicko by Michael Moore.  The practices of the health insurance, pharmaceutical, and agribusiness industries to maximize profits often overlap.  I credit the filmmakers for skewering the food industry.  Is the Kaiser Permanente using this documentary to distract individuals from their own practices?  Congress only turned on the tobacco industry when the costs to the health care system became clear.

**Related Posts**

Evergreening

 

Moving Backward 

 

Unbelievable Promises Monopolized Care—UPMC

WaPo Interactive International Cost Graphic

County Health Rankings

While we're all waiting for the Supreme Court to rule on the Affordable Care Act, the 2012 County Health Rankings were released yesterday for all counties in the US.  An interactive map for Pennsylvania can be seen with all of the county rankings for an overall measure that considers morbidity and mortality above.  Ranked first is Union County(abbrev. UN) and last out of 67 counties is Philadelphia (PH).  These counties were the same in 2011.  Both counties were similar in the percent uninsured in 2009, the most recent year Census Bureau estimates are available, with the 2nd and 3rd highest rates in the state as can be seen in the table below.  The graph shows that in 2011 both were below the statewide median household income of $50,702 with Philadelphia having $37,090 and Union having $45,545.  

While overall rankings are interesting and make for interesting press articles such as "Are Philadelphians eating too many cheesesteaks?" they can gloss over important information such as income, the uninsured and gender.  I haven't yet looked at the 2012 data to see how it differs from last years but expect to find more info on how the recession is impacting the health of Pennsylvania.  Union and Philadelphia Counties caught my eye at first glance.  

Teasing apart cause an effect relationships is a lot more difficult.  Uninsured status (and underinsured status which is a lot harder to measure), gender and median household income are just two of the many possible confounding variables on health status.  This does not mean that one should not try to find these relationships.  It's better to rely on raw measures like the ones below than constructed ones like in the map above for these relationships.

Top 10 County Uninsured Rates in 2009 Overall & by Gender

**Related Posts**

Racial and Gender Differences in Pennsylvania's Uninsured 

 

Correlating PA County % Uninsured Rates with Other County Level Measures

 

Correlating PA's Uninsured with Sen Pat Toomey's 2010 Vote

Latino Rates in Pennsylvania's Uninsured

In 2009 when President Obama was presenting the affordable care act to Congress, Rep. Joe Wilson famously called out "you lie" when he stated that illegal immigrants would not be covered.  He wasn't lying.  On Feb 25 I did a presentation on census data for Pennsylvania's and Texas' uninsured as it relates to their respective Hispanic/Latino population.  In 2006 the city of Hazleton PA passed a law making English the official language of the town and making it illegal to rent apartments to illegal immigrants and to hire them (Their mayor at the time Lou Barletta is now a US Congressman).  The slides below are not necessarily with respect to immigrant status but the analysis I did reveals some patterns which I believe warrants further investigation.  Dr. Patricia Documet had some good comments on my presentation and data. 

Latino rates in pennsylvania’s uninsured

View more presentations from University of Pittsburgh.

In the Americas, Cuba, Chile, Costa Rica, and Canada provide universal coverage and are close to the United States in life expectancy as can be seen in this link for an online graph from Gapminder for life expectancy and income www.bit.ly/y205Ix (the graph cannot be embedded here.  It is interactive and can the labels can be arranged to make more readable by dragging with the mouse).  Other Latin American countries in the graph are moving towards universal coverage and are catching up to the US.  A similar pattern can be seen in the graph for infant mortality www.bit.ly/wFffeF.

You can see a program on health issues in Pittsburgh's Latino community at the link below with Dr Diego Chaves and Dr. Patricia Documet.

WQED Multimedia: TV :: Horizons

**Related Posts**

Italian Americans and Todays Immigrants

 

Rick Perry's Efforts to Save Us All

 

A Statistical Profile of the Uninsured in Washington, DC, New Mexico, and Texas

 

Racial and Gender Differences in Pennsylvania's Uninsured

Correlating PA's Uninsured with Sen Pat Toomey's 2010 Vote

The 2010 election was a referendum on the Affordable Care Act as the coming election and Supreme Court Session will be.  The 2010 Pennsylvania the Senate race between Joe Sestak and Pat Toomey was a contrast on economic issues.  It was won by Toomey with 51% of the vote.  In the CNN exit poll health care was the 2nd most important issue for PA voters behind the economy.  The table below shows how the voters went for the issues and candidates in the exit poll.

Important Issue	Sestak	Toomey
War in Afghanistan (7%)	N/A	N/A
Health Care (22%)	56%	44%
Economy (63%)	43%	57%
Illegal Immigration (5%)	N/A	N/A

When the voters were asked specifically about the health care law (but not Single Payer) the voters broke down this way.  Note even in this election those who liked the law plus those who wanted it expanded out number those who wanted it repealed.  It wasn't asked how many voters were uninsured.

What Should Congress Do With New Health Care Law?	Sestak	Toomey
Expand It (35%)	88%	12%
Leave It As Is (17%)	70%	30%
Repeal It (45%)	10%	90%

There is no county level exit poll data for Pennsylvania but I can correlate the percentage of the vote for Toomey in each county and the corresponding % uninsured.  The relationship shown above while statistically significant is not as strong as the ones discussed in my previous post, accounting for 6.9% of the variability.  Philadelphia is a strong outlier where Toomey had 16% of the vote and the county had 16.3% uninsured. 

One should not assume that the uninsured were likely to vote for Toomey outside of Philadelphia.  I am unaware of voter turnout rates in the uninsured.  The graph above suggests that a higher percentage of Toomey voters are more likely to be found in counties with higher percentages of uninsured (outside of Philadelphia).  Other variables are more strongly correlated with the % uninsured, such as % rural and % illiterate and explain this relationship better.  There was no relationship between the % illiterate and the % of vote for Toomey in PA counties.

I would've made a great exit poll question to see how many Toomey voters were uninsured.

**Related Posts**

Making Sense of the Pat Toomey-Joe Sestak Senate Race 

 

POLL: Dislike of healthcare law crosses party lines, 1 in 4 Dems want repeal - TheHill.com (But Doesn't Ask Why)

 

Correlating PA County % Uninsured Rates with Other County Level Measures

 

Correlating PA County % Uninsured Rates with Other County Level Measures

Now that the PUSH website is officially launched I've looked into the updated Census county level data to see how the estimates correlate with other county level data to see what information they could provide about the state's uninsured.  I downloaded data from the Robert Wood Johnson foundation and the census bureau on the state's 67 counties and computed correlation coefficients to determine if there is an association.  For those who are unfamilliar with correlation it is a number that tells you if there is an association between two variables.  The demographic variables I correlated the % uninsured with were % illiterate, diabetes, rural, female, no English, Hispanic, Native American, and African American.  Of these % illiterate, rural, female, Asian, and Native American had statistically significant correlations (which means that we can be confident that as one county level of these variables changes, the % uninsured in that county is likely to change).  Of all of these associations, the one with % illiterate was the strongest accounting for 68.2% of the variability in the data.  The data for this relationship are presented in the graph below. 

If the correlation were perfect, all of the counties would form a perfect straight line and it would account for 100% of the variability in the data.  Philadelphia County had by far the highest illiteracy rate at 22% and the third highest uninsured rate at 16.3%.

The second strongest association was between the % rural and % uninsured accounting for 26.4% of the variability.  The graph below shows why this relationship is not as strong as the one between illiteracy and the uninsured with a lot more scatter than the one above.  Philadelphia County with a high rate has a 0% rural population while four counties, Forest, Fulton, Potter and Sullivan, have 100% rural populations and high uninsured rates.

The other variables that are also correlated with % uninsured are also highly correlated with these other two variables and when they are accounted for they disappear.  One should always be careful in concluding that a cause and effect relationship exists between two correlated variables.  There are always possible third variables that can explain the relationship.