Chronic Disease Trends in Arizona Counties
Introduction
Chronic diseases such as heart disease, stroke, lung disease, cancer, diabetes and asthma are among the most common, costly and preventable of all health problems in Arizona (azdhs.com). Chronic diseases are currently responsible for 7 out of 10 deaths in Arizona, and diagnoses of multiple chronic diseases are related to a higher mortality rate.
In this dashboard, we explore the hospital inpatient discharge data of Arizona maintained by the Arizona department of health services (ADHS). We analyzed 915,355 chronic disease related patient visits in hospitals across 11 AZ counties during time period 2012-16. The dataset is anonymized and each observation has information related to visit (visit cost, discharge status, payer type), patient (age, gender, marital status, ethnicity), procedures and diagnoses (diagnoses codes, body system category of diagnoses). Health conditions related to cardiovascular problems, pulmonary diseases, hypertension, diabetes, ulcer, obesity, depression are recorded in inpatient encounter records through the International statistical classification of diseases (ICD Version 9/10) coding standard. The prevalence of these health conditions for each time period and location is different. The difference in count may be due to variability of socio-environmental factors across time and location.
Information about time of patient visit is available at six month time periods (first half of 2012 or 2012-01, second half of 2012 or 2012-02, etc). We use the 3 digit zip code information provided for each patient visit to approximately map to the 11 counties in Arizona. Additional information about this dataset can be found at azdhs.com.
In the following sections, we present three dashboards to conduct population-level exploratory analysis of patient visit count, average visit cost, and a combination of both elements. We wish to detect patterns in patient visits and costs across different regions in AZ and different time periods.
Dashboard 1: Patient visit count of population
We explore the distribution of visit count across sections of the overall patient population in AZ. The bottom left graph shows the map of Arizona with counties colored based on average median income. Clicking or selecting regions, will filter the other visualizations for those regions. The other two graphs in this dashboard show the distribution of patient visit count across the 17 body system categories (Circulatory system, digestive systems, musculoskeletal system, neoplasms, etc) and patient age groups (0-9 years, 10-19 years, 20-29 years, etc). The visualizations shows that a majority of Arizonans were affected by Circulatory and Musculoskeletal System disorders. For a majority of disorders, over half of the affected gender was female. A majority of diseases affected those who identified as white in most regions. An exception to this was Apache county where the most affected ethnicity was American Indian, due in part to variability in ethnic backgrounds based on various counties across Arizona.
Readers are urged to use the filters and click on the interactive (linked) visualizations to explore other patterns of visit counts across different categories of age groups, body systems, time periods, counties, gender and ethnicity.
Dashboard 2: Average visit costs of population
This dashboard presents the variability in average patient costs of visits across diseases classified under different body systems. We further include information on payer type using color codes to compare visit costs across different payer types such as medicaid, health maintenance organizations (HMO), preferred provider organization (PPO), medicare, etc.
Visits with congenital anomalies related diagnoses have the highest costs followed by neoplasm and circulatory system related visits. Proportion of average cost expenditure by Arizona Health Care Cost Containment System (AHCCCS) Medicaid is higher than other payers for congenital anomalies related visits, whereas PPOs spend on average for parasitic and infectious disease related visits. Commercial indemnity plans have higher average spending on visits related to injury/poisoning and skin problems related visits. These patterns could be due to different patient segments (differentiated by age groups, income-levels, education levels, occupations, etc) having access to different insurance options and hence different payer types. We urge the reader to use the filters and click on the interactive (linked) visualizations to explore average visit cost patterns across different values of filter variables.
Dashboard 3: Visit count and cost across population
In this dashboard, we explore patterns in visit count and average visit cost simultaneously across two variables - Age groups and discharge status. Discharge statuses have been grouped into four categories - discharged home, transferred to another institution, expired and left against medical advice. Visit cost where patient has expired is 1.5 times higher than visit costs in other discharge status categories for all time periods. Across all time periods, the pattern of patient visit counts is fairly consistent across age groups; admission is highest for 60-69 and 70-79 age groups categories. The average cost for patients in age group 0-9 is consistently increasing with time, from $68,211 in the first half of 2012 to $100,689 in the second half of 2016. There is a data input issue for the second half of 2016 for discharge status and hence this particular set of observations can be ignored. The rate of increase in average cost and total patient visits has steadily increased from 2012-01 (i.e. first half of 2012) to 2016-02 (i.e. second half of 2016).
Readers are urged to use the filters and click on the interactive (linked) visualizations to explore other patterns across different values of filter variables.