 |
Web Hosting by Netfirms | Free Domain Names by Netfirms |
|
|
|
Individual Papers
Statistics: Back to Home
Motorcycle Fatalities: Comparison of State Helmet Laws as a Predictor of Motorcycle Fatalities
MAJ Charles J. Sizemore
U.S. Army - Baylor University
Statistics for Health Care Administration
20 September 2002
Abstract
The purpose of this research was to study the relation of motorcycle fatalities to present helmet laws in the 50 states, of the United States. The three demographic variables analyzed were whether a state law required a helmet for adults 21 years of age and older, or not; the percent of helmeted motorcycle fatalities per state; the percent of motorcycle fatalities per state; the percent of helmeted motorcycle fatalities per state; and the percent of non-helmeted motorcycle fatalities per state. The goal was to determine whether the predictor of required helmet wear significantly impacted motorcycle fatality rates, and to what extent. Statistical analysis of data from 2000 reveals that all of variables are significant predictors of helmet requirements impacting motorcycle fatalities. The results in order of their significance were, percent non-helmeted fatalities r = .746, and t(49) = 7.770, p = 4.91; percent helmeted fatalities r = .725, and t(49) = 7.298, p = 2.57; and percent motorcycle fatalities r = .289, and t(49) = 2.089, p = .042. State governments and health department leaders could use these three variables to analyze their present helmet laws and develop effective policies and programs to reduce the number of motorcycle fatalities on their highways and roadways. These programs may not necessarily involve new legislation, but may include initiatives such as educational programs and required rider instruction prior to licensing.
Section Page
1. Introduction...................................4
Literature Review
2. Statistical Methods............................5
Ten Steps for Hypothesis Testing
3. Results........................................9
Descriptive Statistics
Graphic Comparison of Prediction Equations
Inferential Statistics
4. Discussion....................................16
References........................................18
Appendices (SPSS*)................................20
A. Computer Data Files
B. Descriptive Statistics
C. Frequency Distributions (Histograms)
D. Correlation Analyses
E. Regression Analyses
Motorcycling has been on the increase in recent years, with over 4 million motorcycles registered in the United States. Motorcycles are popular as a mode of transportation due to fairly low purchase costs, good fuel efficiency, and their overall appeal as a pleasure vehicle. Unfortunately, motorcycle fatalities that occur on our nation’s highways and roads represent approximately 7 percent of all highway fatalities each year (National Highway Traffic Safety Administration, 2002). This is prefaced by the fact that motorcycles represent only approximately 2 percent of registered vehicles on the road. Motorcycle fatalities had actually been declining since the 1980s, however, have unfortunately, been on the rise since 1998 and continue to do so(Motorcycle Safety Foundation, MSF, 1999).
Literature Review
Naturally, motorcycles are less stable and visible than automobiles and larger vehicles. Additionally, motorcycles typically have greater performance capabilities, especially when weight is compared to horsepower. Where automobiles have greater bulk, door beams, a roof, safety belts, airbags, four wheels, and an enclosed environment, to enhance the safety of passengers, a motorcycle possesses none of these attributes. A motorcycle, however, due to its smaller size and maneuverability does exhibit greater ability to stop quickly and agility to avoid potential harmful encounters (MSF, 1999).
There are five major crash types that account for approximately 86 percent of fatal motorcycle accidents. These types include incidents when the motorcyclist runs off of the road, incidents when the motorcyclist or another motorist disregards a traffic control measure, such as a stoplight, and head on accidents with other vehicles. Additionally, these types include incidents when another motorist turns in front of the motorcyclist, and incidents, when the motorcyclist merely loses control in the roadway and crashes, such as due to inclement weather or excessive speed (Insurance Institute for Highway Safety, 2001).
According to the U.S. Department of Transportation’s Fatality Analysis Reporting (FAR) system, 2,793 motorcyclists died in crashes in 2000, with head injury attributed to the leading cause (U.S. DOT, 2001). It is estimated that a motorcyclist involved in a crash has a 40 percent probability of sustaining a fatal head injury, if not wearing a helmet. Additionally, the NHTSA estimates that motorcycle helmets are 29 percent effective in preventing fatal injuries to motorcyclists (NHTSA, 2000).
A ten-step quantitative analytic process (Finstuen, 1996) was used for the examination of demographic data. All demographic data was acquired from the National Highway Traffic Safety Administration and the Motorcycle Safety Foundation. Variables analyzed in this study were the percent of a state’s motorcycle fatalities, compared to overall vehicle fatalities; percent of a state’s motorcycle fatalities in which the rider was wearing a helmet, compared to overall vehicle fatalities; percent of motorcycle fatalities in which the rider was not wearing a helmet, compared to overall vehicle fatalities; and whether the state requires motorcyclists, 21 years of age and over, to be helmeted. The ten-step process was used to observe, describe, explain, predict, test, and evaluate hypotheses associated with the relationship between the state’s requirement for wear of a motorcycle helmet for riders age 21 and over, and the three other demographic variables. The ten steps were applied to the data set listed in Appendix A.
Step 1. Define beings, objects, or events, used. The data set consisted of n=50 states, minus the District of Columbia.
Step 2. Determine measures taken and units (operational definitions and data coding). The independent variable (X) was the states to require motorcycle helmet wear, or not, a binary variable. The independent variable (X) was coded 1 if the state requires motorcycle helmet wear for riders 21 years of age and older, and 0 otherwise. Multiple dependent variables (Y1-Y3) were examined for the sample. Y1 was the percentage of motorcycle fatalities compared to overall vehicle fatalities per state, a continuous variable. Y2 was the percentage non-helmeted motorcycle fatalities, compared to overall vehicle fatalities per state, a continuous variable. Y3 was the percentage of helmeted motorcycle fatalities, compared to overall vehicle fatalities per state, a continuous variable.
Step 3. Delineate the hypothesized functional relationship and specify the context. Motorcycle fatalities in accidents vary as a function of required wear of a helmet. Dependant upon the outcomes, and whether there exists a relationship between the wear of motorcycle helmets in the 50 states and the number of motorcycle fatalities, state governments may wish to reevaluate their current helmet laws.
Step 4. State the formal alternate and null hypotheses in terms of a difference model versus a no difference model.
H0: Y¹f(x), Motorcycle fatalities are not related or significant to motorcycle helmet wear.
Ha: Y=f(x), Motorcycle fatalities are related and significant to helmet wear.
Both of these hypotheses were applied to all three of the studies.
Step 5. Tentatively select alpha, the critical probability level, as a baseline decision rule. Alpha probabilities were set at the p = .05 level for all of the dependant variables.
Step 6. Compute descriptive statistics summary (1- & 2-way frequency distributions, means, standard deviations, and correlations.) Graph the results in Functional Form. Note the direction and magnitude of X and Y. The data set for the sample is included as Appendix A. Appendix B contains various descriptive statistics, to include the sum, mean, and standard deviation for each variable. One-way frequency distributions, or histograms, are shown at Appendix C. Appendix D contains computations for correlations. Regression equations were also computed for each independent variable (see Appendix E). Two-way distributions are included in the Results section of the paper as Figure 2. The two-way distributions indicate negative trends for the dependent variables: percent of motorcycle fatalities per state, and the percent of non-helmeted fatalities per state with in relation to required helmet wear. Conversely, the independent variable of percent of helmeted fatalities per state indicated a positive trend.
Step 7. Select and calculate the appropriate inferential statistical test, proper degrees of freedom (df), and the probability of the test results. Student’s t test was used to determine the statistical significance of the correlation results (see Appendix E). This statistical test is also a test of the probability of the regression slope being zero.
Step 8. Evaluate the computed test result for statistical significance. Exact probabilities are computed in Appendices D and E. Since 49 df does not appear on the Areas for t Distributions table (Sanders & Smidt, 2000), the critical values for 45 df were used to evaluate the test results. For the dependent percent of motorcycle fatalities (Y1), t (49) = 2.089, critical values for a two-tailed t test with 45 df are 2.014 (alpha at .05 level), and 2.412 (alpha at .01 level), so the null hypotheses was rejected. For the dependent variable percent of non-helmeted motorcycle fatalities (Y2), t (49) = 7.770, critical values for a two-tailed t test with 45 df are 2.014 (alpha at .05 level), 2.412 (alpha at .01 level), and 3.281 (alpha at the .001 level), so the null hypotheses was rejected and alternate hypothesis Ha was accepted. For the dependent variable percent of helmeted motorcycle fatalities (Y3), t (49) = 7.298, critical values for a two-tailed t test with 45 df are 2.014 (alpha at .05 level), 2.412 (alpha at .01 level), and 3.281 (alpha at .001 level), so the null hypotheses was rejected and alternate hypothesis Ha was accepted.
Step 9. Write a narrative for both descriptive and inferential results. Interpret the results; note the direction and magnitude of the test, and the effects of the X upon the Y. Generalize the results to the population and relate the results and implications of the findings to the stated context. This step appears in the results and conclusions sections of this paper. An abbreviated version of this step is included in the abstract of this paper.
Step 10. Write the result in standard form for statistical results. For the dependent variable percent of motorcycle fatalities (Y1), r = .289, t = 2.089, p < .05, significant results. For the dependent variable percent of non-helmeted motorcycle fatalities (Y2), r = .746, t = 7.770, p < .001, significant results. For the dependent variable percent of helmeted motorcycle fatalities (Y3), r = .725, t = 7.298, p < .001, significant results.
Results are displayed in Table 1 on the next page. A full set of descriptive statistics is included at Appendix B.
Table 1
Descriptive Statistics for Percentage of State Motorcycle Fatalities, Percentage of Non-helmeted Fatalities, and Percentage of Helmeted Fatalities (n=50)
_________________________________________________________________
Variables Average Std. Dev Pearson Correlation
_______________________________________________________________________
% Motorcycle Fatalities 7.22 .50 -.289
% Non-helmeted Fatalities 49.61 27.90 -.746
% Helmeted Fatalities 47.29 27.63 .725
_______________________________________________________________________
The three dependent variables were measured on in comparison to whether the stated required motorcycle helmet wear or not, binary data coded 1 requiring a helmet and 0 otherwise, so their averages are directly comparable to one another. The Pearson correlation coefficient for each independent variable is listed in Table 1. The correlations for motorcycle fatalities and non-helmeted motorcycle fatalities were negative, indicating that as these variables decrease as percent of a state’s motorcycle fatalities with helmet wear requirements. The opposite can be said for the variable percent helmeted fatalities. This indicates that as states’ require helmet wear, helmet wear fatalities will be directly proportional to increases in fatalities. The correlations were interpreted using the shared variance technique. Figure 1 (page 12) displays the Venn diagrams for each variable’s correlation. Out of the three dependent variables, percent non-helmeted fatalities accounted for the most variance in relation to helmet requirements. Second was percent helmeted fatalities, and thirdly, percent motorcycle fatalities accounted for the least variance in relation to helmet requirements. The disparity between the percent motorcycle fatalities and the other two was quite large, with percent non-helmeted and helmeted fatalities approximately six times greater than that of percent motorcycle fatalities.
Graphic Comparison of Prediction Equations
Regression results
are displayed in Table 2. The “b” value in each of the equations is the slope
of the regression line, and the plus or minus sign preceding it indicates the
direction of the line. For the variables percent motorcycle fatalities and
percent non-helmeted fatalities the sign is a minus sign, so the regression line
moves downward as the state requirement for helmet wear increases. For the
other variable, helmeted fatalities, the sign is a plus sign, so its regression
lines move upward. The two-way distributions shown in Figure 2 (page 15) show
the regression lines for the three dependent variables. The regression
equations are directly comparable due to the binary data of whether states
require helmet wear or not, and the percentages are in relation to overall
vehicle accidents in all three studies.
Panel a
Panel b
Panel c
Figure 1.
Percent of variance in percent of motorcycle fatalities compared to states’
overall vehicle fatalities (Panel a), % motorcycle fatalities (Panel b), %
non-helmeted fatalities (Panel c), % helmeted fatalities.
Table 2
Regression Results and Inferential Hypothesis Tests of Motorcycle Helmet Wear Requirement and Demographic Indicators
_______________________________________________________________________
Variable Regression Equation* Pearson’s Student’s Exact
(Y1-Y3) Y’= a + b X r t df p _______________________________________________________________________
% Motorcycle Y’= 8.05 - 2.083 X .289 2.089 49 .042
% Non-Helmet Y’= 66.44 – 42.072 X .746 7.770 49 4.91
% Helmet Y’= 31,09 + 40.495 X .725 7.298 49 2.57
_______________________________________________________________________
* where a is the Y intercept or regression equation constant, and b is the slop of the regression line.
In the two-way distributions a clear difference can be seen in the slopes of the regression lines. The regression line for percent motorcycle fatalities is much flatter than the other regression lines, indicating that a change in this variable has a much less effect on the percent of motorcycle fatalities than any of the other variables. The variables percent non-helmeted and helmeted fatalities have almost the same slope, however non-helmeted is negative and helmeted is positive. The horizontal line in the two-way distributions is the mean of the Y variable, and is included as reference point to which the tilt of the regression line can be compared.
Results for each of the hypothesis tested are also presented in Table 2. As shown, the results for percent non-helmeted motorcycle fatalities were the most significant, with t(49) = 7.770, and an exact probability of p = 4.91. This far exceeds the alpha level of p = .05. Percent helmeted motorcycle fatalities (t(49) = 7.298 and exact p = 2.57), and percent motorcycle fatalities (t(49) = 2.089 and exact p = .042), also exceeded the alpha level of p = .05. For these three variables, the null hypothesis is rejected and the alternate hypothesis, motorcycle fatalities are related to mandatory helmet wear, is accepted. Additionally, the probability associated with percent below non-helmeted and percent helmeted fatalities also exceed the p = .001 level. This indicates that these results are highly significant and the null hypothesis is badly false for these two variables.
Figure 2. Two-way distributions with regression line and mean line
The findings from this study indicate that the demographic variable of whether a helmet is required by law in a given state can be used as an indicator of a states’ percentage of motorcycle fatalities. The percent motorcycle fatalities and percent non-helmeted fatalities are the most useful indicators of the three variables tested. Percent helmeted fatalities is also a significant indicator, however, it primarily reflects that states with helmet laws have higher fatality rates. The relevance of these findings is demonstrated when we consider the political climate with regards to the federal government’s allowing states to determine their own helmet laws and the numerous special interest parties lobbying for and aginst mandatory helmet laws. These factors have direct impacts on our community health and the ability to reduce not only deaths on our highways and roadways, but also the potential for increased numbers of patients in permanent vegetative states. These significant variables can be utilized as a basis for further studies into other variables impacting motorcycle fatalities, and their interrelations.
It is difficult to predict the absolute impact these variables have on motorcycle fatalities across the nation, because there are many other factors that may serve as impacts. Some of these additional impacts are weather, speed, alcohol and drug use, and fatalities caused by other motorists. Regardless, motorcycle fatalities are on the rise, and will likely continue in this trend, since motorcycles seem to be increasing in popularity. Additionally, states may wish to evaluate their riding instruction requirements for licensing and rider education programs. These studies reflect fatalities, thus, the outcomes is death, with or without a helmet. The findings in this study indicate that more effective policies and education programs must be developed, and more resources allocated to researching all of the possible factors that impact motorcycle fatalities, if motorcycle fatality trends are to be halted or possibly reversed.
References
U.S. Department of Transportation, National Highway Traffic Safety Administration (2001). Traffic Safety Facts 2000. [On-line] Available: http://www.nhtsa.dot.gov
Motorcycle Safety Foundation(2001). Cycle Safety Information. [On-line] Available: www.msf.usa.org
Finstuen, K. (1996). Ten steps for hypothesis testing: Functional form statistics. Unpublished manuscript, U.S. Army-Baylor University, Army Medical Department, Ft. Sam Houston, TX.: .
Insurance Institute for Highway Safety (2001). Fatality Facts. [On-line] Available: www.highwaysafety.org
U.S. Department of Transportation, National Highway Traffic Safety Administration (1999). Motorcycle Safety. [On-line] Available: http://www.nhtsa.dot.gov
U.S. Department of Transportation, National Highway Traffic Safety Administration (2001). Motorcycle Helmet Use Laws. [On-line] Available: http://www.nhtsa.dot.gov
U.S. Department of Transportation, National Highway Traffic Safety Administration (2000). Motorcycle Helmet Law Repeal Evaluated in Texas and Arkansas. Traffic Tech. [On-line] Available: http://www.nhtsa.dot.gov
U.S. Department of Transportation, National Highway Traffic Safety Administration (2002). Motorcycle Helmets The Facts of Life. [On-line] Available: http://www.nhtsa.dot.gov
Branas, C., Knudson, M. (2001). State Helmet Laws and Motorcycle Rider Death Rates. LDI Issue Brief. [On-line] Available: www.upenn.edu/ldi
Bikersrights.com (2002). States & Helmet Laws. [On-line] Available: http://www.bidersrights.com/states/1national/states.html
Sanders, D., Smidt, R. (2000). Statistics: A first course (6th ed.). New York: MaGraw-Hill.
A. Computer Data Files
B. Descriptive Statistics
C. Frequency Distributions (Histograms)
D. Correlation Analyses
E. Regression Analyses
*SPSS: Statistical Package for the Social
Sciences Version 11
Appendix B. SPSS Descriptive Statistics
Appendix D. SPSS Correlation Matrices
U.S. Health:
Pharmacy Benefit Management Companies: Do they add value?
Charles J. Sizemore
Baylor University
U.S. Army-Baylor University Graduate Program in Health Care Administration
INTRODUCTION
The intent of this paper is to provide a summarization of three peer-reviewed, professional journal articles relating their underlying themes to reach a conclusion as to whether pharmacy benefit management companies add value to our present health care system, or not. These articles were chosen because of their relevance to the question, our present studies in the classroom, and relationship to our text, as well as their continuity of theme. The first article, Prescription Drug Plans, authored by Burton T. Beam Jr. and Kenn B. Tacchino gives a broad overview of prescription drug plans and the evolution of pharmacy benefit management companies. The second article, Promising strategies help employers integrate pharmacy and medical programs—and reap cost, quality advantages, authored by Mr. Wayne Miller, expands on the themes of the first article and offers insight into two viable, but somewhat opposing views on approaches to managing prescription drug costs. Finally, the third article, Pharmacy Benefit Management Companies: Dimensions of Performance, authored by Helene L. Lipton, David H. Kreling, Ted Collins, and Karen C. Hertz, provides continuity of themes with insight to actual realized, quantifiable impacts of pharmacy benefit management companies.
SUMMARIZATION
Prescription Drug Plans (Beam, Tacchino, 1998) presents a broad overview of the prescription drug plan environment, to include discussion on costs, the impact of new drugs entering the market, the evolution and roles of pharmacy benefit management companies (PBM), and a good introduction to the overall nature of plans in general. The later discusses aspects, such as federal and state requirements, parameters or limitations commonly found in plans, billing and payment methods, and potential impacts of pursuing prescriptions outside of one’s plan.
The article begins with the perspective that although most major medical insurance plans cover the cost of prescription drugs, those with specific and/or separate prescription drug plans have historically reaped additional benefits. The primary benefit has been the ability to leverage for cost savings in the market place. This has become increasingly important due to the increase of claims for prescription drugs. It is estimated that over 10 percent of the cost of all medical claims is for prescription drugs. The demand is and will continue to increase, due to two significant factors. The first factor is the fact that new, expensive medications are readily available and are aggressively injected into the market place by pharmaceutical companies. The second factor, an increase in use, is two fold, in that it stems from the fact that the population is living longer and requiring more prescription drugs with the passage of time. The second stimulus is the fact that medical treatment facilities have decreased inpatient stays to the greatest extent possible, thereby increasing the amount of outpatient services. This lends itself to an increased, more intense utilization of prescription drugs.
The advent of new, costly prescription drugs into the market has resulted in the development of managed care techniques in attempts to control costs. A primary technique is the implementation of formal utilization review programs. This overall managed care approach has served to expedite the growth of pharmacy benefit management companies. The focus of these companies has evolved from merely attaining discounts with participating pharmacies and mail order suppliers to an all-encompassing role in the quality, provision, and costing of prescription medications. Pharmacy benefit management companies increased by the mid-1990s to accommodate approximately 50% of the market share in providing drugs, with fewer than 10 major companies providing coverage to over 80% of health plan participants. These companies may be independently owned, subsidiaries of pharmaceutical companies, associated with insurance companies, and managed care organizations. They primarily administer prescription drug plans on behalf of employers, insurance companies, health maintenance organizations (HMO), preferred provider organizations (PPO), other managed care organizations and third party payers.
The authors then describe the expanding roles of pharmacy benefit management companies, beyond managing prescription drug plans, to include drug utilization reviews, formulary development, physician profiling and education, pharmacy profiling and education, and possibly most important, patient profiling and education. In addition to these, pharmacy benefit management companies are also branching out and developing disease management programs that are intended to assist in identification, development of treatment guidelines, education, and data/outcome measurement of specific diseases. Drug utilization reviews serve to capture patient drug use data in order to identify trends in over utilization, underutilization, drug interactions, refill trends, and the duplication of therapies. The physician, pharmacy, and patient profiling and education programs are all initiatives geared to capture the demographics of each of those subsets, especially with regards to utilization. Once this data is compiled, the PBM will be able to track or monitor trends, establish benchmarks or norms, and implement education initiatives to address and/or prevent significant variances from the established benchmarks. The overarching goal is to maximize efficiency and minimize costs in administering pharmaceutical treatments. Formulary development has been a “cornerstone” of the PBM. Formularies are developed by committees of pharmacists and physicians outlying the preferred medications for specific medical conditions and then provided to health care providers to assist in their prescription writing. These formularies enable the provider to have a reference for the appropriate use of the drugs, when prescribing generic drugs or therapeutic substitutions is acceptable, and the particulars of new drugs that have recently entered the market. Typically, prescription drug plans limit their enrollees to what is provided in the formulary for complete coverage. There may be a financial incentive offered by the PBM to encourage compliance with the formulary, if the plan is not entirely restrictive.
The last portion of the article covers a broad overview of the nature of pharmacy benefit plans. Prescription drug plans normally only cover the cost of drugs that are required by the state or federal government to be dispensed by prescription. There are a few exceptions to this rule, but it is entirely dependant upon the BPM and the plan that is in effect. Plans also typically do not accommodate utilization of drugs in excess of the specified amount, and may or may not cover maintenance drugs, such as high blood pressure medications. Regardless of whether a physician prescribes these medications, if they are not required by law to be prescribed, and adhere to the normally prescribed amount, the plan will probably not cover their cost. Additionally, plans usually do not cover the cost of extraneous supplies for the administering of the drugs, such as hypodermic needles and bandages.
With regards to billing and payment, most plans require a deductible or co-payment. These must be paid by the enrollee in the plan, and usually range from $3 to $5 per prescription. The two common methods of payment are executed through a reimbursement approach or a service approach. In the reimbursement approach, the enrollee purchases the prescription out of their own funds and an electronic claim is submitted by the individual, the provider of benefits, such as an employer, or the pharmacy for the reimbursement, less any co-payment. The most common method, however, is the latter approach. In the service approach, the enrollee acquires his or her medications at a “participating pharmacy”, and the pharmacy, after securing payment for any co-payment, bills the provider of the coverage for the balance. The provider may be a PBM, an HMO, a Blue Cross-Blue Shield association, an insurance company, or other third party payer. Most companies, however, use pharmacy benefit management companies to execute their prescription drug plans. Lastly, the authors note that if prescriptions are filled at a nonparticipating pharmacy, the costs may be covered, but on a reimbursement basis, and the reimbursement, less co-payment, is usually less than the total cost of the medication. Many of the broad underlying themes that the authors outlined in this article, such as the expanding role of PBMs and new, promising quality and cost savings initiatives are expanded upon and reinforced in the following article by Wayne Miller.
Promising strategies help employers integrate pharmacy and medical programs—and reap cost, quality advantages (Miller, 2001) presents a couple of key alternatives to managing costs from the benefit provider perspective. The article begins with setting the stage as to the characteristics of the present environment of providing prescription drug plans and the role of pharmacy benefit management companies. The author then discusses a proven method of reducing costs and provides some insight as to the pros and cons of this rather “quick fix” option. He then elaborates on some new and innovative methods of providing cost savings, which equates to an entirely new perspective regarding cost savings and builds upon some of the themes we touched on in the first article.
The present environment of providing prescription drug plans is becoming a spiraling cost situation with ever increasing utilization and costs associated with providing benefits to enrollees. Unfortunately, these costs have increased from 18 – 20% due to the increased demand, the increased amount of new, expensive drugs hitting the market.
The key for benefit providers in our present market environment is to develop effective methods to not merely manage pharmacy costs, but to strive to manage overall medical costs as well. Mr. Miller promotes the perspective of integrated strategies that consider both pharmacy plan strategies and overall medical care plan strategies together, resulting in increased effectiveness and optimal outcomes. The integrated plan approach should manage employee satisfaction, improve quality, improve outcomes, and improve the value of the plan to employees, as well as the management of costs.
The quick avenue to the management of costs piece, that Mr. Miller outlines, has been proven effective, and is a very tempting, easy approach to combating the rising costs of prescription drugs. This approach consists of options, such as increasing co-pays, reducing benefits, instituting restrictive formularies, and instituting a three-tier system. The benefit provider has the latitude to implement any combination of the aforementioned initiatives to combat prescription drug costs, primarily through impacting utilization. For example, an employer could realize savings of 20% or more by merely increasing co-pays from $10 to $20 per prescription. Three-tier systems segregate both brand and generic drugs into a list or formulary that has three tiers of co-pays. The first tier may be a primarily generic drug with a copy of $5, the second tier could be brand drugs with co-pays ranging from $10 - $20, and the third tier may include non-formulary and/or non-preferred drugs with co-pays ranging from $25 - $30. According to Mr. Miller, this latter strategy has resulted in cost savings in excess of 10%. Since these savings in cost are, in essence, realized by shifting greater costs to the enrollees of a plan, there arises the question as to whether these strategies serve to improve employee health and outcomes and reduce overall medical costs in the long term. Mr. Miller presents the argument that indeed these measures will not improve the overall quality of benefit programs in the long term, primarily because they are not comprehensive enough to remain viable. He proposes that plans that are not more comprehensive in nature, including elements such as physician education and disease management initiatives, may very well end up costing the benefit provider more in terms of overall medical costs, in the long term.
Pharmacy benefit management companies are, however, developing and implementing new initiatives that take a more comprehensive approach to managing costs and improving quality, some of which are: analysis of medical and pharmaceutical claims, development of targeted quality improvement initiatives and interventions, initiatives to promote compliance and persistence, identification of pre-catastrophic patients through predictive modeling, implementation of generic use programs, and pharmacoeconomic approaches to clinical program development. Mr. Miller suggests that benefit providers need to ask some critical questions when assessing the effectiveness of their programs. Some of these questions are, “What is happening with our program in the real world? How are employees using the benefit? For what? What is working? and What isn’t?” (Miller 2001). Additional, expanded roles of pharmacy benefit management companies are compiling data through claims analysis that includes information such as the level of compliance rates per drug (adherence and persistence); therapy completion rates, levels of drug switching, dosage levels to achieve desired outcomes, and whether certain drugs encourage compliance or have higher levels of complications and side effects. However, from a cost savings perspective, the ability to determine where and why additional drug costs are being incurred is an incredibly valuable aspect of claims analysis.
Next, the article deals with quality improvement initiatives and interventions, and Mr. Miller uses the example of how a pharmacy benefit management company could assist in the cost savings of antidepressants, which is, for many employers, the most common prescription medication used by employees. Furthermore, his example broaches the overall medical and societal costs of the treatment of depression. He provides his perspective that not only pharmacy costs should be considered in the treatment of depression, but also the societal costs as related to missed workdays, missed schooldays, overall medical expenses, and premature death, which equate to $43 billion annually. Mr. Miller purports that a pharmacy benefit management company, under a quality improvement initiative, could work in conjunction with the physicians of a health plan to identify utilization and prescribing patterns, misdiagnosis, and develop optimal prescription decisions. The most important elements of a quality improvement initiative are the provision of information to physicians, enabling them to make the most appropriate, informed decisions, and the education of the patients with regards to the importance of compliance and persistence. Mr. Miller does point out in his article that initial pharmacy costs may increase, upon the implementation of a quality improvement program, however, this should be offset by lower overall medical costs realized through decreases in utilization.
Compliance and persistence are at the heart of most of these improvement initiatives, for if the patient is not properly informed to his or her part, the outcomes will be less than optimal and possibly worse. Patients must be informed and take a proactive role in this overall equation for success to be realized. According to the article, less than half of all people taking prescriptions follow their physician’s instructions, or they may not comply with the therapy regime. Unfortunately, they are not realizing the full benefit of the drugs, when compliance and persistence are not adhered to, and the disease or illness may actually get worse. Naturally, this could very well result in greater medical costs in the long term. Thus, patient education and commitment is absolutely critical.
Predictive modeling is a potentially powerful tool designed to estimate illnesses that may impact the plan in the future. This is accomplished through the use of neural networks to build mathematical models that can predict the likelihood of future diseases and/or illnesses in patients. Thus, the PBM may be able to head off catastrophic, high cost illnesses such as cancer, diabetes, or heart disease through proactive programs. Ideally, these diseases and illnesses may be prevented or, at least, their onset delayed.
Generic use programs are another vital method to reduce the costs of pharmacy benefit programs by assisting in controlling the cost of new more expensive medications. The difficulty here is combating the aggressive marketing of pharmaceutical companies to utilize the “latest and greatest” drugs. Realistically, there are numerous generic drugs that are already on the market that are clinically recognized as being the most appropriate “first-line” treatment. The problem arises through the inappropriate prescription of the newer, more expensive “second-line” drugs prematurely. Ideally, an effective generic use program will promote appropriate prescribing so that patients receive best overall treatment, avoiding over-treatment, and cost savings are realized as well. The critical element, in this initiative, is the education and “buy in” of physicians.
The advent of pharmacoeconomics is serving to assist employers in maximizing their benefit programs for their employees, with limited resources. Pharmacy benefit management companies are assisting managers in making the most appropriate, informed decisions taking into account the clinical and humanistic factors, as well as the economic factors. Many of these decisions primarily concern the high cost, high utilization, high impact drugs, where therapies may be extremely varied. The example of pharmacoeconomics utilization in the article is given two enrollees in a plan with identical conditions, but different pharmacy patterns and outcomes. Pharmacoeconomics may be able to compare and contrast the details and help develop more uniform and beneficial guidelines.
The article closes with discussions and recommendations that pharmacy benefit management companies and their vast data gathering capability and ample technology could also serve to identify and decrease or prevent medical errors through education. Additionally the article promotes the cost savings through mail order, and the benefits of real time audits of claims and identification of errors at retail pharmacies. The capstone of this article, however, is employee education. The article promotes not merely initial education, but, rather, the continuous distribution of information to enrollees through newsletter articles, letters, or other means. Mr. Miller ends the article promoting an overall comprehensive approach to cost savings and quality improvement, as opposed to the easier, quick approaches that are mostly concerned with short-term economics. This more comprehensive approach to cost savings and quality improvement versus merely cost saving and market share initiatives is further addressed in the following article about the dimensions of performance of pharmacy benefit management companies.
Pharmacy Benefit Management Companies: Dimensions of Performance (Lipton, Kreling, Collins, Hertz 1999) is an all encompassing article, including most every aspect of the two previous articles. The authors also cover the history, evolution, and growth of PBMs; and the roles and functions of PBMs in the health care system. However, it is the area of impacts of PBMs, within the article that I will concentrate on for this paper, pages 386 - 389.
In particular, the article addresses the impact of pharmacy benefit management companies on drug costs. Although there are no independent assessments as to the extent to which PBMs control prescription drug costs for their clients, the available research does suggest that, overall, PBMs are reducing the rate of drug cost increases for their clients. Quantifiable data is difficult to attain, since most PBMs consider performance specific data as proprietary. BPMs have reported that the realized savings are gained through reduced pharmacist payment levels, patient copayments, restricted or managed formularies, and mail service. The Government Accounting Office (GAO) did a study of three of its major prescription benefit plans and concluded that the BPMs saved them over $600 million in 1995. This was accomplished primarily through obtaining manufacture and pharmacy discounts and through managing drug utilization. Additionally, it was estimated that mail and retail pharmacy discounts accounted for the greatest proportion of savings.
The impact of pharmacy benefit management companies on quality of care is also addressed in this article. At the publishing of this article, the competition among PBMs for clients primarily centered on price as opposed to quality. Additionally, HEDIS (Health Plan Employer Data and Information Set) indicators contained only one pharmacy related measure, the recommendation to use beta-blockers after acute myocardial infarction at that time. The article did acknowledge that HEDIS and other quality of care measures were becoming more sophisticated and that PBMs may have to respond to future drug related HEDIS measurements. The article does outline one peer-reviewed study on PBM sponsored intervention to improve prescribing patterns. The study evaluated provider prescribing for and elderly population of mail service patients, and involved potential inappropriate drug use. The article reports that 15% of the alerts regarding inappropriate use of prescribed drugs, (8% of all alerts generated) resulted in an immediate therapy change. Additionally, 9% of the alerts (5% of all alerts generated) resulted in the physicians’ intentions to the therapeutic alternative at the patients’ next visit. Physicians were successfully contacted for 56% of the alerts, and the article reported an overall successful outcome of intervention of 9%. Thus, the article indicates that, although there may be an improvement in physician prescribing with the intervention of PBMs, the actual overall effect may be smaller than initially indicated.
DISCUSSION
All three of these articles are closely related in that they discuss the current environment of benefit provider pharmacy benefit plans, rising costs, the development of pharmacy benefit management companies and their functions and expanded roles. The last article expands more into the actual impact of the PBMs. Attributing managed care techniques, which are consistent with our textbook, Delivering Health Care in America; A Systems Approach (Shi, Singh pp. 325-330), to the management of these pharmacy plans is the underlying theme throughout all three of the articles. The primary technique that is incorporated is the attainment of cost control through impacting the utilization of the plans. However, the second article by Mr. Miller strongly purports a more comprehensive, holistic approach.
Do they add value? In my opinion, yes, they do add value. I believe that pharmacy benefit management companies can and are playing a vital role in the management of prescription drug costs, and the improvement in the quality of pharmacy therapies. Unfortunately, it is difficult to accurately, quantifiably assess the magnitude of the impact, due to the lack of objective studies. Although there is a benefit from the BPMs being realized in our health care system, their underlying motivation is still to function in a most efficient, manner, maximizing their profits as well.
In the second article by Mr. Miller, I believe that he expresses the correct approach to pharmacy benefit plans, in that they should be comprehensive and well integrated with the overall medical plans that are provided. I agree wholeheartedly with the more holistic approach that he recommends and the many potential benefits that may be provided by PBMs, such as claims analysis, quality improvement initiatives, predictive modeling, pharmacoeconomic program development and encouraging education across the board. These concepts are consistent with the holistic model of health care presented in our text (Shi, Singh pp. 40-41), which stress the importance of comprehensive health, to include the aspects of physical, mental, social, and spiritual health. However, these initiatives take time, resources, and money, and that is what I believe is the Achilles heel to the more comprehensive approach. Realistically, I believe that there will continue to evolve a combination of the two approaches.
Benefit providers are concerned with the viability of their organizations at present and the significant piece that both medical and pharmacy plans plays in their viability is of a greater impact in the near-term, as opposed to the long-term. Thus, a happy medium will have to be realized between implementing managed care principles in the near-term, such as managing utilization, and promoting initiatives that will ensure overall healthier enrollees and lower costs in the future. Again, I do believe that pharmacy benefit management companies are adding value to our present health care system, but their full potential and benefit is yet to be realized. It is my impression that the largest hurtle that will have to be overcome in our society, to realize the potential that Mr. Miller so strongly promotes, is the “the bottom line” on the financial statements.
Managed Care:
The Present and Future State of Disease Management Programs
Charles J. Sizemore
Baylor University
U.S. Army-Baylor University Graduate Program in Healthcare Administration
Abstract
Disease Management in the present managed care environment continues to make headway, despite some formidable obstacles. It is the purpose of this paper to highlight the key characteristics of disease management programs, their requirements for success, and the obstacles to the development of disease management programs. Additionally, I will discuss some of the current initiatives that are being implemented by organizations to overcome the current environmental obstacles, as well as present some current developments with disease management and their implications on its future.
Introduction
Disease management (DM) as defined by the National Disease Management Association (NDMA), and supported by definitions by the National Committee for Quality Assurance (NCQA), and the Centers for Medicare and Medicaid Services (CMS), is as follows: “Disease Management is a strategy of delivering health services using interdisciplinary clinical teams, continuous analysis of relevant data, and cost-effective technology to improve the health outcomes of patients with specific diseases” (Disease Management Now, 2002). This approach includes the entire continuum of health care and focuses on the natural course of chronic diseases and their affected populations. Some of these chronic diseases include asthma, AIDS, cancer, chronic renal failure, and congestive heart failure (Kongstvedt, 2001).
Disease management is fundamentally different than traditional case management, as illustrated in the differences between goals, emphasis, setting, timing, and guidelines. Traditional case management’s goal is to streamline components, while DM’s goal is to integrate components. Likewise, the differences between emphases in the treatment of sickness vs. the prevention and education of sickness are of paramount importance. Ideally, a DM program enhances patient care and decreases benefit costs by identifying and selecting chronically ill persons who are traditionally high cost users of medical care and linking them with appropriate providers and outpatient interventions (Kongstvedt, 2001). DM interventions should address one or more of the following goals:
- Improving patient self-care through education, patient monitoring, and communication.
- Improving physician performance through feedback on patients’ progress in compliance with protocols.
- Improving communication and coordination of services between patient, physician, the DM organization, and other providers.
- Improving access to services, such as prevention services and prescription medications (Disease Management Now, 2002).
The common features of disease management programs to provide a comprehensive approach to the management of patients’ care are:
- Identification of patients
- Use of evidence-based practice guidelines
- Supporting adherence to evidence-based medical practice guidelines by providing medical treatment guidelines to physicians and other providers, reporting on patient’s progress in compliance with protocols, and providing support services to assist the physician in monitoring the patient.
- Services designed to enhance patient self-management and adherence to his/her treatment plan.
- Routine reporting/feedback loop.
- Communication and collaboration among providers and between the patient and his or her providers.
- Collection and analysis of process and outcomes measures (Disease Management Now, 2002).
Disease management programs should optimize clinical and economic outcomes by facilitating proper diagnosis, maximizing clinical effectiveness, maximizing the efficiency of the overall delivery of health care, and by continuously improving (Kongstvedt, 2001).
Discussion
Primary obstacles to DM validity and evolution in the managed care environment
There are several obstacles/barriers that have served to impede the progress and validity of disease management programs and or organizations. Some of these barriers are the existence of a fragmented delivery system, reimbursement concerns, and information system incompatibility (Kongstvedt, 2001). Regardless of the obstacles, they must be overcome to the satisfaction of the physician/health care deliverer, the HMO, and the consumer, in order for disease management to be effective and accepted as viable. One of the largest obstacles is the conflict between physicians and HMOs with DM programs. Physicians generally desire to deliver the highest quality of care within cost-effective parameters. However, the time consumed with respect to DM demands from multiple managed care organizations is surmounting. These demands are manifested in volumes of faxes, e-mails, requests for home health care authorization, phone calls from case managers, and requests from pharmacies to change prescriptions (Leider, 2001). Additionally, physicians may perceive that with DM programs, there will be fewer primary care office visits, and fee for service revenues may be negatively impacted (Kongstvedt, 2001).
Another enormous, and possibly the most critical obstacle is the fact that many health plans are demanding proof that DM will actually save them money, and oh, by the way, they wish to see the hard numerical data to back up the cost savings claims. The positive financial outcomes that are being touted by many DM organizations are based purely on cost avoidance strategies. The validity of these claims is what is in question. Many MCOs are feeling pressure from accreditation groups to show commitments to improving long-term health, and a DM program can be an effective tool to accomplish this. The options are for MCOs to develop their own DM programs, or to contract out the services. There are pros and cons to both, and both require an investment of time and money. It is the reluctance of investment that must be overcome, through proven positive outcomes, for DM programs to thrive (Carroll, 2000). This return on investment (ROI) proof is presently difficult to attain, since there have been few external studies and they have been fairly short-term studies, at that. Disease management has really only been of significance since the mid-1990s, and many of the programs have only been in place since the late 1990s, not lending to long-term studies of greater than 10 years.
The key to ROI is the development of a valid baseline for costs associated with patients, with which to conduct adequate comparisons and analyses. This is difficult, since health care data is enormous, and often maintained on various, incompatible systems. A compounding effect is that health care is in a constant state of change, and data may be skewed. Delays in claims, definitions of disease, and the rapid turn over of plan membership are all states of flux that can skew data and decrease its quality (Carroll, 2000). Also, most DM programs only focus on one or two co-morbidities, as opposed to numerous. This is sometimes perceived as unrealistic, since the majority of patients have multiple co-morbidities (Carroll, 2000). Thus, for DM to rise to the next level, there must be a greater focus on data, there must be more peer-reviewed material on its effectiveness, there must be consolidation of a broad cross-section of programs covering many co-morbidities, and there must be objective, external studies supporting effectiveness and cost efficiency.
Medicare is embracing a new round of trial disease management demonstration projects, and the agency expects the results to validate DM. Additionally, under a mandate from the Benefits Improvement And Protection Act of 2000, Medicare is planning to create a new demonstration project that combines DM programs with outpatient prescription drugs for advanced-stage congestive heart failure, diabetes, and coronary heart disease. In this initiative, DM companies will earn a premium for coordinating care and be reimbursed for the cost of medications. However, any participating DM company must offer the government a defined set of savings, and post a bond to guarantee performance (Carroll, 2002).
LifeMasters, a DM company, who was an alpha test site during the National Committee for Quality Assurance’s (NCQA) development of the disease management standards and assisted in the formulation of those standards, has received full NCQA accreditation in five diseases. They received a three-year full accreditation to cover congestive heart failure, coronary artery disease, chronic obstructive pulmonary disease, diabetes, and asthma. They were judged by six standards in each category: patient service, practitioner service, measurement and quality improvement, program content, and clinical systems and program operations (PR Newswire, 2002). It appears as though the future of disease management programs, as well as disease management companies/organizations, is somewhat secure for the present. However, if the Medicare demonstration projects are not successful in validating DM programs, the future could take a turn for the worse. Although DM programs are more than logical, the crux of the matter is cost efficiency in addition to increased health outcomes.
References
Carroll, John (2000). Health Plans Demand Proof That DM Saves Them Money. Managed Care 2000, from http://www.managedcaremag.com/archives/0011/0011.dm_roi.html
Carroll, John (2002). DM and Medicare: A Marriage Made in Heaven? Managed Care 2002, from http://www.managedcaremag.com/archives/0206/0206/medicare_dm.html
Carroll, John (2002). DM Vendors Start To Address Costs Created by Comorbidities. Managed Care 2002, from http://www.managedcaremag.com/archives/0203/0203.dmconsolidate.html
Dazell, Michael D. (2000) Show Me The Outcomes! Managed Care 2000, from http://www.managedcaremag.com/archives/0001/0001.dmpac.showme.html
Disease Management Now (2002). What is Disease Management? from http://www.dmnow.org
Downey, Charles (2001). Disease Management Uses Web To Net Savings. Managed Care 2001, from http://www.managedcaremag.com/archives/0107/0107.online_dm.html
Kongstvedt, Peter, R. (2001). Essentials of Managed Health Care (4th ed.). Aspen Publishers, Inc.
Leider, Harry, L (2001). HMOs Need To Share Gains of DM Programs. Managed Care 2001, from http://www.managedcaremag.com/archives/0107/0107.leider.html
PR Newswire (2002). LifeMasters Receives Full NCQA Disease Management Accreditation in Five Diseases – Confirmation of Excellence for Health Plans and Other Payors. PR Newswire, from http://library.notrhernlight.com/FD20021125090000019.html
Jurisprudence:
In this paper, I will discuss the guidelines for expert witness testimony in medical malpractice litigation, as outlined by the American Academy of Pediatrics. The source article, Guidelines for Expert Witness Testimony in Medical Malpractice Litigation was printed in the journal of the American Academy of Pediatrics, Pediatrics, ISSN: 0031-4005; Vol. 109 No. 5; p. 974+, on May 1, 2002, by Charles H Deitschel, Jr; Jerome M. Buckley; Geoffrey Evans; and John J. Fraser, Jr., with the intention of informing the physician community, patient advocates, private citizens and pediatricians.
The American Academy of Pediatrics (AAP) published their first guidance in 1989 and has refined this outline for giving expert witness testimony in the litigation process, to include pretrial, trial and post trial proceedings. Courts need the technical expertise provided by physicians to determine whether or not medical negligence has in fact occurred in a specific case. The jury must be informed to be able to make a determination as to whether the case is one of malpractice, in that an adverse event was caused by negligence or “bad care”, or it is a case of maloccurrence, the adverse event is a “bad outcome”. The physician is expected to establish and explain the appropriate standard of care and make an assessment whether or not there occurred a breach of the established standard of care. He or she is then expected to provide an opinion expressing whether or not the breach of care is the most likely cause of injury. His or her testimony should be clear, understandable, and consistent with the standards applicable to the specific case. It is advised that the physician, upon evaluating the most appropriate, applicable standard of care not eliminate other possible treatment options that may be addressed during cross-examination.
The causes of action for liabilities in medical malpractice cases may include intentional misconduct, breach of contract, defamation, divulgence of confidential information, insufficient informed consent, failure to prevent foreseeable injuries to third parties, or negligence. Thus, medical malpractice cases may involve tort or contract law. In determining the standard of care, with reference to negligence, it is understood “that reasonable and ordinary care, skill, and diligence as physicians and surgeons in good standing in the same neighborhood, in the same general line of practice, ordinarily have and exercise in like cases.” Since bad outcomes can be realized as a result of the proper treatment being rendered, as well as improper, it is crucial that expert testimony give the jury enough information to discern between actual negligence or mere unfortunate, bad outcomes. The burden of proof falls solely upon the plaintiff to convince the jury by a preponderance of the evidence, at least 51%, as to the validity of their allegation. The physician’s role in pretrial expert testimony is to evaluate and review the particulars of the case to determine the merit of proceeding with legal action. Depending upon the state, he or she may be required to testify that the standard of care was, in fact, breached. This may result in a deposition, in which the physician’s testimony is recorded under oath and subject to cross-examination. This is possibly the most important part of identifying all of the facts that are pertinent to the case. It is absolutely imperative that the testimony be truthful, reliable, objective, and accurate. The integrity of the expert witness testimony is of paramount importance; it must be unbiased and completely ethical.
With regards to improving the quality of expert testimony, the AAP recommends various programs. Peer reviews to critique the content of the testimony and sanctioning that may impose disciplinary actions upon physicians who commit breaches in ethics or provide inaccurate, incomplete or unscientific testimony are two of the most popular methods. Additionally, advocacy and education programs to establish minimal qualifications are excellent initiatives to improve the overall quality of expert testimony.