State
Barbour
Berkeley
Boone
Cabell
Calhoun
Clay
Doddridge
Fayette
Gilmer
Grant
Greenbier
Hampshire
Hancock
Hardy
Harrison
Jackson
Jefferson
Kanawha
Lewis
Lincoln
Logan
McDowell
Marion
Marshall
Mason
Mercer
Mineral
Mingo
Monongalia
Monroe
Morgan
Nicholas
Ohio
Pendleton
Pleasants
Pocahontas
Preston
Putnam
Raleigh
Randolph
Ritchie
Roane
Summers
Taylor
Tucker
Tyler
Upshur
Wayne
Webster
Wetzel
Wirt
Wood
Wyoming

Appendix A
Appendix B

West Virginia 2004 County Health Profiles
Introduction

Overview | Selected Causes of Death | Premature Death - LPLL | Percent Change in Mortality Rates | Leading Causes of Death | Incident Rates for Leading Causes of Cancer | Behavioral Risk Factor Prevalences | Births, Infant and Fetal Deaths | Percent Change in Birth Statistics | Census Data | Significance Testing |

This 2004 edition of the West Virginia County Health Profiles updates the previous editions published in 1992, 1997,, and 2000. While a few changes have been made to this latest edition, the majority of the presentation has been kept consistent in able to allow comparisons among data in the earlier editions. We have attempted to incorporate suggestions from County Health Profile users around the state to make the updated version even more useful to researchers, health planners, health educators, grant writers, policymakers, and other health data consumers.

Overview

The linked health profiles compose an overview of the health status of West Virginia residents on the state and county levels. The first page of each county’s profile provides a snapshot of that county’s health status through a comparison of the county with the United States on the basis of 33 selected health indicators, including mortality rates, behavioral prevalences, birth statistics, and prenatal care measures. Each indicator has been tested for statistical significance in relation of the United States (see Appendix A for testing methodology) and classified as falling within a range of values (percentages or rates) that is lower than, the same as, or higher than that found for the nation as a whole. Based on these standards, county indicators were then categorized as better than, similar to, or worse than the United States. For example, in Barbour County, the teen fertility rate was found to be significantly lower than the national rate, the prevalence of obesity was found to be neither significantly lower or higher than the U.S. median, and the heart disease mortality rate was significantly higher than the national average. While the homicide rate in Barbour County was similar to the U.S. rate, the actual number of deaths over the study period was too small for valid significance testing (indicated by a double asterisk [**]).

This information is provided to allow counties to more accurately identify specific problems among their residents. This method of presentation is not meant to infer that indicators falling in ranges similar to or even lower than those for the nation do not pose serious problems to county residents or should not be addressed in health planning. However, a knowledge of a county’s worst health problems is necessary in determining the priority of community interventions.

Selected Causes of Death: 1992-2001

County and state data on selected causes of death from 1992 through 2001 are presented on the second page of each profile. The total number of deaths from 1992-2001 are given, followed by the county’s crude mortality rate (deaths among county residents per 100,000 population) and adjusted mortality rate (adjusted by age to the 2000 U.S. standard million) for each cause. The 1997 U.S. adjusted rate is provided for comparison purposes, together with the percentage difference between the county adjusted rate and the U.S. adjusted rate. (The percentage differences presented in the profiles cannot be reproduced in all cases from the rates given due to rounding.) Finally, to provide comparability among counties, each county is ranked according to its adjusted rates, from highest rate (1) to lowest rate (55). As an example, in Barbour County, the 1992-2001 adjusted rate for deaths due to heart disease was 15.1% higher than the 1997 U.S. rate (321.0 deaths per 100,000 population vs. 279.0 deaths per 100,000 population). Statistically significant differences are noted with an asterisk (*) in the significance indicator (SI) column. (Barbour County’s heart disease mortality rate was significantly higher than the national rate.) Barbour County ranked 38 th among the 55 counties in heart disease mortality.

Premature Death - Years of Potential Life Loss

The second page of each profile also presents data for each county on premature mortality, or years of potential life lost before the age of 75 (YPLL), calculated as the difference between the age of 75 and the age at death occurring before age 75. For example, a person dying of lung cancer at the age of 55 contributes 20 years to the total YPLL for lung cancer (75-55=20 YPLL). The sum of YPLL for a specific cause of death is the total YPLL from all persons dying from that cause before the age of 75. YPLL is an important tool for emphasizing and evaluating causes of premature death. (Previous editions of the County Health Profiles have used the definition of premature death as that occurring before the age of 65. The change to age of 75 was made in order to conform to the definition now used by the U.S. Centers for Disease Control and Prevention.)

The total number of YPLL for selected causes of death from 1992-2001 are presented in column 1 of the YPLL table, followed by the crude YPLL rate (number of years of potential life lost by county residents per 100,000 population) and the adjusted YPLL rate (adjusted by age to the 2000 U.S. standard million). The 1997 adjusted YPLL rate for the United States is provided, as well as the percentage difference between the county adjusted YPLL rate and the U.S. adjusted rate. Significance is indicated in the next column by an asterisk (*). The county’s rank for each cause from highest (1) to lowest (55) is found in the last column. In Barbour County, our example, there were a total of 3,240 YPLL due to cancer from 1992-2001, yielding a crude rate of 2,241.9 YPLL per 100,000 population in the county. Barbour County’s adjusted rate for YPLL due to cancer was 2,087.8, which was 147.5% higher than the 1997 U.S. rate of 843.7, a statistically significant difference. Based on its adjusted rate, Barbour County ranked 12 th in premature deaths due to cancer among the 55 counties.

Percent Change in Mortality Rates: 1987-1996 amd 1992-2001

The table on the third page of each profile presents and illustrates changes in the numbers and rates for selected causes of death between 1987-96 and 1992-2001. Both the number of deaths and the adjusted rate for each cause are provided in the table, followed by the percent difference and an indication that the difference is statistically significant, if applicable. Using Barbour County once again as an example, the adjusted rate for all cancers increased from 210.4 deaths per 100,000 population in 1987-96 to 227.3 deaths per 100,000 population in 1992-2001, an 8.0% increase. This increase, however, was not statistically significant when tested. (Due to space limitations, 1987-96 and 1992-2001 are referred to by their midpoints, 1991 and 1996, respectively, in the table’s title.)

Leading Causes of Death: 1992-2001

The table on the fourth page presents the total number of reported cases and the incidence rates for each county for all cancer sites, as well as for the four leading causes of cancer mortality, i.e., lung, colorectal, female breast, and prostate cancer, for each year from 1996 through 2000. Data on cancer cases diagnosed among West Virginia residents (with the exception of basal cell and squamous cell carcinomas of the skin and carcinoma in situ of the cervix) have been collected and reported by the West Virginia Cancer Registry (WVCR) since 1993. The Registry adjusted the 1996-2000 incidence rates to the 2000 U.S. Census population to permit comparisons with national data published by the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program. Inquiries involving cancer incidence rates should be directed to the WVCR, 350 Capitol Street, Room 125, Charleston, West Virginia 25301-3715, (304) 558-5358, (800) 423-1271.

In Barbour County, there were a total of 419 diagnoses of cancer reported from 1996 through 2000, representing an average annual age-adjusted rate of 458.8 per 100,000 population. The average annual rate of cancer of the lung and bronchus was 76.2 diagnoses per 100,000 population. The county’s average rate for colorectal cancer was 55.5 per 100,000 population, the rate for prostate cancer was 154.2 per 100,000 male population, and the rate for female breast cancer was 101.5 diagnoses per 100,000 female population.

Incident Rates for Leading Causes of Cancer: 1996-2000

Table five presents the total number of reported cases and the incidence rates for each county for all cancer sites, as well as for the four leading causes of cancer mortality, i.e., lung, colorectal, female breast, and prostate cancer, for each year from 1996 through 2000. Data on cancer cases diagnosed among West Virginia residents (with the exception of basal cell and squamous cell carcinomas of the skin and carcinoma in situ of the cervix) have been collected and reported by the West Virginia Cancer Registry (WVCR) since 1993. The Registry adjusted the 1996-2000 incidence rates to the 2000 U.S. Census population to permit comparisons with national data published by the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program. Inquiries involving cancer incidence rates should be directed to the WVCR, 350 Capitol Street, Room 125, Charleston, West Virginia 25301-3715, (304) 558-5358, (800) 423-1271.

Barbour County, there were a total of 419 diagnoses of cancer reported from 1996 through 2000, representing an average annual age-adjusted rate of 458.8 per 100,000 population. The average annual rate of cancer of the lung and bronchus was 76.2 diagnoses per 100,000 population. The county’s average rate for colorectal cancer was 55.5 per 100,000 population, the rate for prostate cancer was 154.2 per 100,000 male population, and the rate for female breast cancer was 101.5 diagnoses per 100,000 female population.

Behavioral Risk Factor Prevalances

Behavioral risk factor prevalences are found on the fifth page of each profile. Prevalences are given for nine behavioral risk factors that have been determined to affect health adversely: physical inactivity, obesity, hypertension, diabetes awareness, cigarette smoking, smokeless tobacco use, binge drinking, uninsured status, and difficulty in seeing a doctor because of the cost. County prevalences from 1997 through 2001 have been aggregated and compared with (1) the estimated 1999 U.S. prevalence for each of the nine risk factors and (2) aggregated county prevalences from 1992 through 1996.

For both the 1992-96 and 1997-2001 prevalences, 24 counties had aggregated sample sizes large enough to yield individual prevalence calculations. Samples from the 31 counties that had sample sizes too small to stand alone were combined with samples from other less-populated counties into 12 groupings, or multicounty regions. (A map depicting the final 36 groupings is found in Appendix B, as well as state trends for selected risk factors from 1984-2001.) A single behavioral prevalence was then calculated for each grouping. This prevalence was provided as the prevalence for each county included in the grouping. Counties are ranked from highest to lowest on the basis of the final 36 groupings, with counties within each grouping sharing a rank.

The table presents prevalences from 1992-96 and 1997-2001 as well as estimated 1999 U.S. prevalences for additional comparisons. Statistical significance between county and national rates is indicated. The last column contains the county’s rank for each risk factor. A rank of 1 denotes the highest prevalence of a risk factor, while 36 denotes the lowest prevalence. (Due to limited space, 1994 is used to refer to aggregated data from 1992-96; 1999 is used to refer to aggregated data from 1997-2001.)

Continuing our example, the prevalence of physical inactivity in Barbour County was 47.1% in 1992-96, decreasing to 39.5% in 1997-2001 (not a statistically significant difference). The county 1997-2001 rate of 39.5% was 35.8% higher than the estimated 1999 U.S. prevalence of 29.1%, a statistically significant difference. Barbour County’s grouping ranked 11 th among the 36 groupings in the rate of physical inactivity.

Births, Infant Deaths, and Fetal Deaths

Aggregated birth data from 1997-2001, found on page six of each profile, are included to provide health educators, researchers, and planners with numbers of births, age-specific proportions of births, and age-specific fertility rates (the number of births to women in a specific age category per 1,000 women in that age category) by county. Comparable U.S. fertility rates are provided for 1999, followed by the percentage difference between the county and national rates, as well as significance indicators where applicable. Finally, for each age group, counties are ranked from highest fertility rate (1) to lowest fertility rate (55). In Barbour County, there were a total of 808 births from 1997-2001 among women aged 15-44, yielding a fertility rate of 50.0 (births per 1,000 women aged 15-44). This was significantly lower than the comparable 1999 U.S. rate of 64.1. Barbour County ranked 43 rd among the 55 counties in fertility rates among women of childbearing age (15-44).

Age-specific fertility rates provide a more accurate measure of the problem of births to teen mothers than does the more traditional method of comparing the proportion of all births that occur among teenagers. Statewide, and in most counties in the state, the proportion of total births that are to teenagers is higher than that in the nation as a whole. This does not take into account, however, the drastic decline in both teen (15-19) and, in particular, older (20+) births occurring in West Virginia since 1980. Because women in the state over the age of 19 are having fewer babies, the proportion of teen births has remained high, while statewide, and in many counties, the teen fertility rate is actually significantly lower than the national average.

Selected birth certificate data collected over the five-year period from 1997 through 2001 have also been included for each county. Information is provided on the number and percentage of low-birthweight births, births to unwed mothers, early (1 st trimester) vs. late (3 rd trimester/no care) prenatal care, mother’s use of tobacco during pregnancy, and mother’s educational level (i.e., percent of mothers having less than a high school education). Comparable U.S. data from 1999 are presented, as well as the percentage difference between county and national statistics and significance indicators where applicable. The counties are ranked for each category from the highest percentage (1) to the lowest percentage (55). In the example of Barbour County, there were 76 low-birthweight births from 1997-2001, representing 9.5% of all births. This percentage was 23.5% higher than the 1999 U.S. rate of 7.6% but was not a significant difference. Barbour County ranked 11 th in percent of low-birthweight births.

Total infant and fetal deaths from 1997 through 2001 are provided with accompanying rates. Infant mortality is broken down into neonatal deaths (infants less than 28 days old) and postneonatal deaths (infants 28 days to one year of age). Fetal deaths are those deaths that occur before birth at or after 20 weeks of gestation. Both infant mortality rates and fetal death ratios are calculated as the number of occurrences per 1,000 births to county residents. The 1999 U.S. rates for infant and fetal mortality are provided, followed by the percentage difference between the county and the nation and statistical significance indicators where applicable. Counties are ranked from highest (1) to lowest (55) rate. There were 10 infant deaths in Barbour County from 1997 through 2001, yielding an infant death rate of 12.3 deaths per 1,000 live births to county residents. This rate was higher than the 1999 U.S. rate of 7.1; however, the number of deaths was too small for significance testing. Barbour County’s infant death rate was 8 th highest among the 55 counties.

Percent Change in Birth Statistics: 1992-1996 and 1997-2001

The table on the seventh page of each profile presents changes in the numbers and rates of selected natality characteristics between 1992-96 and 1997-2001. For example, in Barbour County, the percentage of total births that were low birthweight was 18.7% higher in 1997-2001 than in 1992-96 (7.9% in 1992-96 vs. 9.4% in 1997-2001), but this was not a statistically significant difference. Both the number of births and the rates are provided in the table, together with the percentage difference between the time periods and significance indicators. (Because of limited space, 1992-96 and 1997-2001 are referred to by their midpoints, 1994 and 1999, respectively, in the table’s title.)

2000 Census Population Distribution

The eighth page of each profile contains county population distributions by age, sex, and race based on the 2000 Census. Each county’s population is broken down by five-year age groups and sex, with comparable breakdowns by race (white, black, and other races). These data can be utilized in the calculation of age-, sex-, and race-specific mortality and birth statistics, information that is useful in determining the most effective intervention strategies on the individual county level.

Appendix A

Significance Testing

The two-tailed test was used to determine statistical significance by constructing 95% co the two intervals were found to overlap, there was not a statistically significant difference nfidence intervals around each two values (rates, ratios, or prevalences) being compared. If between the two values, i.e., any difference between them could be due to chance. If the intervals were not overlapping, the difference between the two values was determined to be statistically significant, i.e., there is a 95% certainty that the difference is not due to chance.

Using Barbour County as an example, the 1992-96 weighted prevalence of obesity for the Barbour/Taylor County grouping was 16.4% (0.164), while the 1997-2001 prevalence of obesity was 21.3% (0.213). To determine if the 1992-96 value is significantly lower than the 1997-2001value, the following calculations were made:

Calculate the standard error for each prevalence, using the formula

SE = square root of (p * [1-p]/N),

where p is the proportion (prevalence) and N is the number of people in the survey. For 1992-96, p=0.164, N=181; for 1997-2001, p=0.213, N=193.

SE (1 st value) = square root of (0.164 * [1 - 0.164] / 181)
SE (1 st value) = 0.028

SE (2 nd value) = square root of (0.213 * [1 - 0.213] / 193)
SE (2 nd value) = 0.029

Calculate the confidence range for each prevalence, using the formula

R = p + or - (1.96 * SE)

R (1 st value) = 0.164 + or - 0.28
R (1 st value) = 0.136 through 0.192

R (2 nd value) = 0.213 + or - 0.029
R (2 nd value) = 0.184 through 0.242

In this example, the two ranges (0.136 - 0.192) and (0.184 - 0.242) overlap. There is therefore no statistically significant difference between the two values (16.4% and 21.3%).