What other Variables Contribute to Nursing Home Mortality in PA?

 

	% of death in LTCF	Case Mortality in LTCF	Case Mortality in County	% over 66	% Rural	Median Household Income
% of death in LTCF	1.00
Case Mortality in LTCF	0.38	1.00
Case Mortality in County	0.02	0.26	1.00
% over 66	-0.19	-0.30	0.33	1.00
% Rural	-0.13	-0.21	0.11	0.62	1.00
Median Household Income	0.15	0.38	-0.12	-0.48	-0.49	1.00

Two weeks ago I looked at the correlation between the percentage of residents over the age of 66 in the county with the percentage of total county deaths in a nursing home and the case mortality rate in the nursing home.  This week I'm looking at the the additional variables of the % of the county in a rural area and the median household income in the county.  Of these, the relationship between case mortality and median income was the strongest accounting for 14.9% of the variability in the rates.

While this is a relatively weak (though significant) positive correlation there is a stronger correlation between that may moderate this relationship.  There a stronger relationship between % rural and % over the age of 66 of 0.62 (seen below).  Both of these have a relatively strong negative correlation of -0.49 and -0.48 with median household income respectively.    

The % over the age of 66 has a stronger positive correlation with the case mortality in the county (0.33) that it does with case mortality in a nursing home (LTCF) (-0.30 which is smaller in absolute value than 0.33).  This suggests that the elderly who live in more rural counties are more likely to die outside of a nursing home from corona virus.  The more rural and more elderly counties are also more likely to have lower incomes.

**Related Posts**

Why the New York COVID-19 Nursing Home Scandal Matters

A Look at Positive COVID-19 Testing Rates in Cambria, PA, and the US

The Number of Corona Virus Cases in Cambria County has Grown Exponentially While Health Behaviors Predict Cases in PA

Size of PA County Elderly Population Predicts Smaller Nursing Home COVID-19 Mortality

This week I thought I would take a look at mortality in nursing homes (or Long Term Care Facilities (LTCF)) in Pennsylvania.  Statewide, nursing home deaths account for 52.1% of the total coronavirus deaths.  Looking at the county level there is considerable variability in the rates where the 10 counties with the highest rates are seen below.  The rates for this group rage from 115.8% for Forest county to 67.2% for Delaware County.  Forest county actually has 3 more LTCF deaths than total deaths with one facility.  This suggests that at least 3 of their residents reside outside the county.  You can see the full list of counties here.

Facility County	Number of Facilities with Cases	Number of Cases Among Residents	Number of Cases Among Employees	Number of Deaths in LTCF	Total Deaths in the County	% of death in LTCF	Case Mortality in LTCF
FOREST	1	121	26	22	19	115.8%	15.0%
ERIE	35	1354	717	287	363	79.1%	13.9%
MCKEAN	9	372	110	47	60	78.3%	9.8%
CENTRE	16	675	146	155	203	76.4%	18.9%
BUTLER	29	1421	156	257	350	73.4%	16.3%
LACKAWANNA	35	1423	295	290	395	73.4%	16.9%
LYCOMING	19	922	283	165	232	71.1%	13.7%
MONTGOMERY	111	4723	186	1026	1456	70.5%	20.9%
CARBON	7	407	48	95	138	68.8%	20.9%
DELAWARE	59	3197	871	796	1184	67.2%	19.6%

The bottom 10 counties are listed below.  The percentage of deaths in the in LTCFs for this group rage from 35.7% in Sullivan County to 15.0% in Potter County.  Surprisingly, Philadelphia county is in this bottom tier with 26.1% of their deaths in LTCFs.  Another way of looking at mortality in LTCFs is the case mortality or the % of cases who have died.  Philadelphia ranks 11th in the state on this measure with Pike county ranking 1st at 27.0% and Cameron County ranking last at 4.3%.

Facility County	Number of Facilities with Cases	Number of Cases Among Residents	Number of Cases Among Employees	Number of Deaths in LTCF	Total Deaths in the County	% of death in LTCF	Case Mortality in LTCF
SULLIVAN	3	94	13	5	14	35.7%	4.7%
WAYNE	5	88	19	20	58	34.5%	18.7%
BEDFORD	5	269	29	43	126	34.1%	14.4%
FRANKLIN	26	799	94	106	313	33.9%	11.9%
CAMERON	1	36	10	2	7	28.6%	4.3%
SNYDER	2	124	31	20	73	27.4%	12.9%
PHILADELPHIA	85	4222	16	782	3001	26.1%	18.5%
ELK	3	122	50	8	34	23.5%	4.7%
CLINTON	6	112	48	9	53	17.0%	5.6%
POTTER	3	45	15	3	20	15.0%	5.0%

I looked at the percentage of the population in each county over the age of 65 and found some interesting patterns.  For the correlation with % of deaths in LTCFs, as the proportion of the population over 65 increases the % of deaths decreases as seen with the graph below.  Forest county is an outlier and is labeled green in the scatterplot.  With Forest county included, this relationship accounts for 6% of the variability in the data.  When Forest county is excluded, the relationship accounts for 11.6% of the variability in the data and is still negative.  I don't have a good explanation for this but will investigate further.

I then looked at the correlation between case mortality and the % of the county over 65.  This relationship was also negative accounting for 11.3% of the variability.  Pike county is a mild outlier.  With Pike excluded, this relationship accounts for 16.6% of the variability.  This pattern is similar to the one with the percentage of deaths in the county.  

 

Case mortality and % of deaths in LTCFs are positively associated but not strongly accounting for 15.3% of the variability.  There may be stronger relations ships among the variables not considered yet.  There also may be variables at the facility level to which I do not have access.  Investigating will continue.

**Related Posts**

A Look at Vaccinations in Pennsylvania: Montour County has a high Vaccination and Hospitalization Rate

Extreme Outliers in Pennsylvania: Beaver and Montour Counties

COVID-19 and County Health Rankings in PA: Which Variables Predict Cases and Deaths