Finance II
Room: Crimson Clover
Wednesday, April 22, 2015, 9:00 to 10:30 AM
Moderator: C.B. Mamaril, PhD
William Livingood, PhD
Using Cost Analysis to Identify Cost Saving Strategies by Local Health Departments in Delivering STI Services
Co-Investigator(s): Bonita Sorensen, MD, MBA & Lori Bilello, PhD, MBA
Background: Sexually transmitted diseases (STD) continue to be a major health problem in the U.S., particularly in Florida. The costs of delivering services across counties in Florida vary extensively across County Health Departments (CHD), even in a state with centralized policy and administrative services for local health departments. Research Objective: To identify the unit costs of delivering STD prevention/control services, and examine the effects of variations in delivery system characteristics on costs. Data Sets and Sources: CHD STD Expenditure Data - Florida Financial Information Reporting System (FIRS) Gonorrhea, Chlamydia, and Syphilis counts and rates - Florida Bureau of STD Prevention and Control Demographic data – US Census American Community Survey CHD STD service delivery practices – electronic survey tool Study Design: This mixed method study combined secondary data (2012) to identify variations in costs of services for all Florida counties, in depth interviews to develop plausible explanations for the variations, and survey data to examine the presence of factors causing the variations in STD services expenditures. The survey tool collected data from all 67 CHDs and examined how STD screening, investigation and partner notification were performed in each county. Analysis: General linear regression models were performed using the best subset selection method based on the R-square statistic to assess the best predictive model. Models examined the effect of county level factors such as total population, STD rate/100,000, population density/square mile, population with income below 200% FPL, total CHD revenue/capita, CHD county tax revenue/capita, percent of population 24 years old and under, and percent black/other population on CHD STD expenditure per case. Service delivery factors examined include cross jurisdictional sharing, types of staff providing services, prioritization of patient populations and partnering with community organizations. Principal Findings: Wide variability in STD rates and cost for STD services are found across the state. Two variables were found to be significant in the model that impacted the cost per case: CHD county tax revenue per capita (β=15.03, p=.023) and partnering with community organizations (β=-78.59, p=.087). Other variations in practices potentially impacting cost were so extensive that few trends were statistically significant using regression analysis. Conclusions: The availability of local tax funding for CHDs is statistically linked to higher CHD STD expenditure per case. Those CHDs that partnered with community organizations to perform screening and treatment saw a reduction in the cost per STD case. Statistically significant differences in costs per services between two variations in practice were observed, such as cross-jurisdictional sharing, but were not present in regression modeling with the relatively small numbers included in this practice. Implications for Public Health Practice: Revenue sources for STD prevention and control, particularly local tax support, appear to influence how CHDs respond to controlling these diseases. Partnering with community organizations or cross-jurisdiction sharing also appears to reduce costs. Understanding the nature, mutability and impact of variations in expenditures on STD service delivery is critical to improve the effectiveness and efficiency of STD service delivery, and perhaps public health services in general.
Kristina Rabarison, DrPH
1-2-3 Pap Intervention: Cost Analysis and Cost-Effectiveness Analysis
Co-Investigator(s): Rui Li, PhD; Connie Bish, PhD; Robin Vanderpool, PhD; Richard Crosby, PhD; & Mehran Massoudi, PhD
Background: Cervical cancer places a substantial economic burden on our healthcare system. The 3-dose human papillomavirus (HPV) vaccine series is a cost-effective intervention to prevent HPV infection and resultant cervical cancer. Despite its efficacy, completion rates are low among females ages 18-26. The 1-2-3 Pap health communication intervention – focused on improving HPV vaccination completion rates among 344 young adult females in rural Kentucky – showed that women who viewed an educational video were 2.44 times more likely to complete the HPV vaccine series than those who did not. Research Objective: The purpose of this study is to provide practitioners with objective measures and information about program cost, cost-effectiveness, and scalability. We tested the hypothesis that programmatic costs for 1-2-3 Pap would decrease and cost-effectiveness would improve if the program was expanded. Data Sets and Sources: Cost and efficacy data are from the Rural Cancer Prevention Center (RCPC)’s budget and expenditure reports, and the 1-2-3 Pap efficacy study. Study Design – This study used a retrospective study design, using RCPC cost and 1-2-3 Pap efficacy data from September 2010 – December 2012. Analysis: A cost analysis of the 1-2-3 Pap trial was conducted and cost was assessed for 1-2-3 Pap expansion through a hypothetical adaptation scenario in which the program was expanded to the entire Kentucky River Area Development District (KRADD). A cost-effectiveness analysis (CEA) was conducted to provide insight into program implementation costs and benefits such as healthcare costs avoided from prevention of cervical cancer as a result of the 1-2-3 Pap program, by using a HPV transmission model developed by the Centers for Disease Control and Prevention. Principal Findings: Our preliminary findings showed that in the adaptation scenario the intervention would cover a thousand more females ages 18-26 and almost four times as many women would complete 3-dose series than in the original efficacy study. 1-2-3 Pap’s total cost would decrease from $988 per 3-dose series completion to $536. Furthermore, the cost of the video would decrease from $214 per completed 3-dose series in the efficacy study to $68 in the rural adaptation scenario. CEA results will be completed by April 13, 2015. Conclusions: Determination of 1-2-3 Pap program cost is vital for program expansion; results suggest that the cost would reduce substantially if the program was scaled up. The CEA findings will provide further support of the scalability of the 1-2-3 Pap program, if it is found to be cost-effective. Implications for Public Health Practice and Policy: Considering the substantial annual cost of cervical cancer screening and treatment ($30,775 to $52,731 per case in 2010), an estimated $536 investment per completed vaccine series may be worthwhile. These results: 1) provide economic evidence and an adaptation estimate for other communities that may want to implement the 1-2-3 Pap intervention and 2) provide additional insight into 1-2-3 Pap program implementation costs and benefits.
Adam Atherly, PhD, MA
The Economic Cost of Communicable Disease Surveillance in Colorado
Co-Investigator(s): Melanie Mason, MS & Lisa VanRaemdonck, MPH
Background: Although many different national organizations, including the Institute of Medicine, have called for the development of a minimum package of public health benefits, little is known about the cost providing essential population-based public health services. This lack of knowledge results in an inability to make a clear financial case for public health and limits the amount of informed decision-making that can be done by public health leaders. In our project, we identify the economic costs associated with Communicable Disease (CD) Monitoring/ Surveillance in local public health agencies (LPHA). This research project measures the variable and fixed estimated costs of delivering core public health services in Colorado and identifies services and delivery characteristics with economies of scale and scope. We then compare the costs to the number of cases investigated to calculate a cost per case. Research Objectives The primary objective of this analysis is to estimate the fixed and variable costs associated with communicable disease monitoring by LPHAs in Colorado. Study Design and Analysis: To estimate the cost of the services, we use a micro-costing approach. Colorado local public health employees engaged in CD work logged their time spent on CD surveillance for two different two-week periods: one in the spring of 2014 and the other in the fall of 2014. Forty-three of the 54 local public health agencies in Colorado participated in the data collection. To calculate marginal costs, we used daily activity logs recording activities during each 15 minute interval of work over a two week time period. Based on this result, we will calculate the per-hour cost of each staff member type, including both wages and benefits. We also collected data on indirect costs. Methodologically, we estimate a statistical cost function using cost (from the microcosting data) as the dependent variable and key dependent variable as the number of reported investigable diseases during the matched two week period. We also controlled for other independent variables, including case mix, characteristics of the agency, the community and services provided. Principal Findings: Mean time spent per day on CD surveillance was 205 minutes per day, while median time was 58 minutes. Activities with the highest mean time were Routine Investigation (61 minutes), checking the CD database (47 minutes) and Communication (37 minutes). Activities with the highest median time were checking the CD database (27 minutes), Communication (14 minutes) and Routine Investigation (12 minutes). The differences between mean and median time reflect differences between large and small agencies. Seventeen counties had no cases assigned during the logging period, but spent a mean of 71 minutes on CD surveillance. Regression models show strong non-linarites in cases, with both Cases (b=567.7, p<.001) and Cases Squared (b= -22.7, p<.001) strongly statistically significant. Despite strong, practice-based hypotheses, we found no evidence of seasonality. Conclusions and Implications: Agencies took an average of 54 minutes per day to do CD monitoring, regardless of the number of cases. Time per case decreased as cases increased, suggesting economies of scale.