DenguEcon User Manual
Contents
1. death averted Cost of vaccination program Cases averted X treating a case deaths averted Where Cases averted Rate of cases per unit population X population per unit population X vaccine effictiveness Cost of vaccination program Total cost to vaccinate one person X population X percentage compliance Study perspective DenguEcon calculates the cost effectiveness ratios using the equations outlined above from the perspective of the health care system That is how much will the health care system pay for vaccinating a target population how much will the health care system save from such a vaccination strategy and what will be the net cost or savings per case or death averted This perspective was chosen because the intended audience is practicing public health officials working in countries e g member states of the Southeast Asian Region Office SEARO that typically have centralized government funded health care systems Broader societal considerations such as work days lost to dengue and gained saved due to dengue vaccination are not considered under this perspective This is important because the calculated cost effectiveness ratios will be underestimates 10 Time horizon considered Current data from vaccine trials suggest two doses of vaccine separated by a 6 month interval will most likely provide at least 10 years and possibly lifetime protection DenguEcon evaluates the costs and be2. DenguEcon Software to estimate the economics of a dengue vaccination program from the perspective of the health care system DenguEcon 1 0 Beta Test Version E ER HEALTHIER PEOPLE TABLE OF CONTENTS Page ACKNOWLEDGMENTS 5 098 yan iain gren QUOND pun RO dal MD UND E 1 AUO eiae E aki css als E ER 1 Sugossted eltablolias co eoee rra eo por e o teo ithe neetaeacan 1 DISCLAIMERS 4 ccepit eut D eek n ISI ORE AE YU PASO senses HUE QUEE ULI EXP TY ETEA N asa aair 2 SYSTEM REQUIREMENTS nainen pe o ausente eau dais uad eens 3 LOAD AND STAR eed sete omiies to anase e atanan nesi das 4 AUTORECOVER iot tU PUDE uetnto PRSE Motus states Ee EU 7 INTRODUCTION roesini nnns eian EUIS XX GRUB UAR IH UR A sienna 8 Background two guiding principles eeeseeeeeeeeeeneeeneen ennt 8 METHODS atawene 9 Equation Cost effectiveness ratio Non fatal outcomes eene 9 Equation Cost effectiveness ratio deaths ooo eee bl iD Hr Doa niter tent 9 SUITS 21101 5 AA aaa cosh cana obiecit etat Ate esM eere 10 DISCOUIL TALS od ascss ac eec oe aese testudo tac Cal a e 11 How to enier dala cosi iii Do eet deb etai aded E a ae 12 DATA ENTRY Page 2 or 13 Population 2 oe resa esa t Uo eon Ph i Sa ER TE i Nae Hee ve o 13 Page 3 of 13 Rates of non death outeomes oid eoe gan ep doe tes a eH ge Reged 14 Type of DULCOMIE OTT DTE OTT 14 Combined datas in mac e teo
3. Methods page 8 The graph on the left depicts to cost per case averted where a case may be inpatient only outpatient only or combined inpatient outpatient The definition of what type of case is depicted in the graph depends upon the data entered in Page 2 of 13 The graph on the right depicts the cost per death averted What do these results mean On the X axis along the bottom of each graph is a range of person vaccinated This range encompasses the range of costs of vaccinating a person entered by the user on Data Entry Page note that if the user changes the range of costs in Data Entry Page 8 of 13 the range in the graph may change On the y axis is the cost effectiveness ratio either case averted left hand graph or death averted right hand graph The solid line in each graph indicates the median cost effectiveness ratio whilst the two dotted lines represent the 5 and 95 confidence intervals In each graph any time any results a line or part of a line are above the 0 case averted or 0 death averted that means that dengue vaccination results in a net cost to the health care system In order for there to be a net savings per case or death averted to the health care system the line must be below the 0 case averted i e read as case averted persons vaccinated Displays the number of persons assumed vaccinated Total target population from Page 2 of 13 multiplied by t
4. copied and pasted to another Excel compatible program application e g for presentations or report writing using standard procedures Most of the time a page or table or graph can be copied and pasted by first selecting the area graph needed to be copied right clicking on the mouse to produce a menu and selecting copy from the menu The copied object can then be pasted in the new application as need by clicking the right button in the mouse and then selecting Paste For more information please consult Excel s online help or a manual for Excel Printing In order to print your results on a single page you may need to change the printing page setup to Landscape format To do so 1 Click File and then choose Page Setup 2 In the Orientation section change Portrait to Landscape 3 Click OK 35 CONTACT For additional help or feedback please email your comments or questions to Martin I Meltzer M S Ph D MMeltzer cdc gov 36
5. currently recommends using a discount rate of approximately 396 per year This rate is largely derived from the interest rate for 30 year U S treasury bonds after adjusting i e removing the inflation effect DenguEcon users in other countries may wish to consider using other rates commensurate with factors such as population growth rates and government bond market rates Also different perspectives such as private versus public can result in users selecting different discount rates Discount rate Sensitivity analysis We recommend that DenguEcon users as part of their sensitivity analyses explore the impact of using different discount rates 11 METHODS DATA INPUT HOW TO ENTER DATA In DenguEcon data should only be entered in the boxes cells outline with bold black borders DATA TO BE ENTERED Age group evaluated Enter the age group in years of the population being studied For example 9 to 16 year old children and adolescents If the user desires to compare different implementation strategies e g EPI alone versus a catch up campaign we suggest that you either open different copies of the model or rerun the model with the appropriate age groups costs doses and effectiveness rates A DenguEcon Supplement has been developed to compare these different options Number in target pop population Enter the number of people in the population being studied Study perspective The only perspective
6. doses Currently there is no approved vaccine so data on the effectiveness of a tetravalent dengue vaccine is limited Chanthavanich et al 2006 evaluated antibody titers several years after vaccination with a tetravalent dengue vaccine however antibody titers are a poor correlate of protection For this reason DenguEcon users can vary the level of effectiveness to determine the lower limit of effectiveness that remains cost effective by running the program multiple times varying the levels of effectiveness An example of the varying level of vaccine effectiveness is shown in the example below What does DenguEcon do with the data DenguEcon calculates the weighted average vaccine effectiveness This is calculated as follows Average vaccine effectiveness The effectiveness of the first dose X the percent who get only one dose the effectiveness of the second dose X the percent who get both doses Limitations of methodology The current methodology used in DenguEcon to evaluate all the user entered data concerning vaccine effectiveness is simple and direct However if does not necessarily fully utilize all the data contained in the probability distribution Future versions of DenguEcon may contain a more sophisticated methodology for incorporating the vaccine effectiveness data Another limitation is that vaccine effectiveness is assumed to be identical for all outcomes 22 DATA TO BE ENTERED Cost per person vaccinated Cos
7. ethane ee stead 14 Unit OF Tate esis o ences E E On RU EROR aos NY eer 14 LYS Ol OMNCOMC S ous cies E e pA TE cte Ea xU atas 14 Type of rate Excess Tales ou oem eere debetis usce tis 14 Date fates etes fue te vox reto oa amita pei a LA Rules for data entiy c s utate br eet Ma VE Reid Aedes 15 Page 4 of 13 Modeling the burden of dengue eeeeenee 16 Data Simulation Option tette ad etstbudt i itentistoreefite gibt etes iitesp tob Mn 17 Pros and Cons of Simulation sees 18 Pase SOf 13 Rates of deatlisco s stun e ete debt a e m tate E s ous 19 Page 6 of 13 Cost person We ale iaieoeois droite erit suede ri e esee ase boves tese tepuyua 20 Outpatients erno aE bro Ere Dacidise bra Od cd eoo 20 Tapa Ie tenei eo utex desk s tata a cekagt hae ates te tr fue 20 ii Of outpatient VISIS x oreet et E HR rt aiia Page 7 0E 13 Vaccine eiectiveness 44 2 2a oe eek eld ate 22 Example default values exi i hina ti re ret oiilne 22 Assumed probabilities ocio adu cesc eu pe eee ruere tet oit 2D Limitations of methodology sseesseseeeeees 22 Page 8 of 13 Cost person Vaectated coge e Uc el Pe ERRRER HR O BUCEA GRE cuoi de 23 Why Upper and Lowen e e er ro eaa y teoieaed SIGE CTICCIS oed de iba SE pA nas RA NPEAWE REL ROM or a SUR S 23 Assumed COMPIANCE 2 iori x ecd as co HE Ta IR maui 24 Pase 9 oL 13 Discount tate Lose ines ib qt icu ial E IA EROR RA EPA N a vaut 25 RESULTS d
8. methodology used in DenguEcon explicitly and deliberately allows for simultaneous variability in most of the input variables The mathematical principles used in the construction of DenguEcon are similar to those used in Meltzer MI Neuzil KM Griffin MR Fukuda K An economic analysis of annual influenza vaccination of children Vaccine 2005 23 1004 1014 Meltzer et al 2005 in order to allow for the variability in input data 1 e incorporate the first guiding principle used a Monte Carlo methodology That methodology samples probability distributions of input values i e mathematically incorporates allows for variability of input values However Monte Carlo methodology requires a great deal of data to construct reasonable probability distributions of input values as well as ownership and understanding of specialist statistical software In this software in order to satisfy both guiding principles instead of using Monte Carlo sampling methodology we used a bootstrap methodology This methodology re samples several hundred times a small dataset of input values e g 5 data points to build a larger data set A probability distribution of input data values is built from this larger dataset see Methods section for further technical details This building of a probability distribution of input values even when the input dataset is small allows for variability in input values Further the re sampling is automatically progra
9. studied DATA RATES In these cells a user enters the actual data see diagram below in the form of rates per unit population the unit having already being recorded see Unit of rate above Unit of rate 100 000 e g per 100 000 population under consideration DATA rates Warning A mi imum of 5 years of consecutive data is required 2560 1607 191 ur1v 2517 d fT 14 RULES FOR DATA ENTRY i For DenguEcon to work you need to enter a minimum data for 5 consecutive years If no cases are reported in a given year then enter 0 ii If you only have 5 years of consecutive data or less than 10 years of data then leave the remaining 5 years blank iii Years seasons entered here must match seasons entered into Deaths input sheet 15 DATA TO BE ENTERED It was assumed when constructing DenguEcon that many users will not have readily available 10 years of national dengue surveillance data However dengue is a disease of epidemic potential and therefore its impact varies greatly comparing outbreak and non outbreak years Outbreaks typically occur at 3 8 year intervals in endemic countries Incidence of severe disease can also vary based on the genotype of the prevalent virus serotype and prior exposure to dengue and the serotype of the previous infecting strain DEN 2 3 DEN 3 DEN 1963 1967 1971 1975 1979 1983 1987 1991 1995 1999 2003 Cases Rate 1000 Laboratory positive case
10. that DenguEcon uses is that of the health care system see earlier 13 DATA TO BE ENTERED Select one by entering Yes and N A for others Type of dengue Combined in and health outcome outpatients Outpatient onl Inpatient onl Background The data entered in this page records the rates of dengue related health outcomes such as outpatient visits and inpatient stays hospitalizations Type of dengue health outcome The user indicates selects the type of non fatal health outcome data that will be used in the analysis see figure below The user must enter yes without the quotation marks in one and only one of the three boxes In the other two boxes the user must enter n a without the quotation marks The abbreviation n a means not applicable Combined inpatient and outpatient data DenguEcon can accept data that combines the rates of outpatient and inpatient dengue related health outcomes For example a rate of outpatient visits of 550 per 100 000 a rate of hospitalizations of 10 per 100 000 gives a combined rate of 560 health outcomes per 100 000 Note that in order to combine outcomes correctly the user must first ensure that both rates are recorded in the same measure of unit population e g both rates record per 100 000 population or whatever is the most convenient rate Unit of rate Record here the unit of rate for the data that will be entered e g per 100 000 population being
11. 36 iii ACKNOWLEDGMENTS AUTHORS Martin I Meltzer Mark E Beatty and Rebekah Heinzen Borse 1Centers for Disease Control and Prevention U S Department of Health and Human Services Atlanta GA 2 Pediatric Dengue Vaccine Initiative International Vaccine Institute SNU Research Park San 4 8 Bongcheon 7 dong Kwanak gu Seoul Korea 151 919 SUGGESTED CITATION Martin I Meltzer Mark E Beatty and Rebekah H Borse DenguEcon 1 0 a manual to assist country and district public health officials in estimating the cost effectiveness of a dengue vaccination program Beta test version Centers for Disease Control and Prevention U S Department of Health and Human Services 2008 THANKS Charisma Atkins CDC Atlanta and Ole Wichmann PDVI IVI Seoul COVER PHOTO Electron image of mature Dengue 2 virus particles replicating in five day old tissue culture cells The original magnification is 123 000 times available at http www cdc gov ncidod dvbid dengue electron micrograph htm DATE of this draft May 25 2009 DISCLAIMERS The numbers generated through DenguEcon are not to be considered absolute forecasts of the cost effectiveness of a dengue vaccination program Rather they should be treated as estimates with degrees of uncertainty which may or may not be known The methodology findings and conclusions in this manual any accompanying appendices and results generated by the software DenguEcon are those of the author
12. a ready and rapid examination of the impact of changes in most of the input variables For example the user can readily make changes in vaccination costs rates of health outcomes vaccine effectiveness population under study by age risk groups size etc as well as changes in cost of vaccination What sensitivity analyses can a user do The user is encouraged to extensively explore the impact of making many changes in input values and is particularly encouraged to make several changes at the same time to different variables this is called multi variable sensitivity analyses A user can produce any number of multi variable sensitivity results by collecting the results produced by each new combination of input values from the detailed results tables in the Results sheet scroll down below the results graphs Example The sensitivity analyses graph on Page 13 of 13 On Page 13 of 13 there is an example of a graph produced by running DenguEcon several times This graph illustrates how step wise changes in vaccine effectiveness impacts the cost effectiveness ratio for two different costs of vaccination Essentially the graph illustrates how fixed increases in vaccine effectiveness can reduce the cost effectiveness ratio i e case averted becomes less costly as vaccine effectiveness increases The graph contains 3 different types of data i two costs of vaccination ii range of vaccine effectiveness and iii resultant cost effecti
13. ars HERE For Option 2 only user defines minimum number gears between peak gears 5 reas between peaks With the data entered in previous sheet this is what your two options look like OPTION 1 Use data as entered OPTION 2 Simulate 10 gears of dengue by rate Annual rates of dengue User must tes e we 3000 Data az entered no simulation chose between acca cate op E g 2500 one ofthese 8 2000 options in box 3 1500 below 5 E 1000 note If wished user E 500 can run model with one option and then 12 3 4 5 6 7 8 3 10 12 3 4 5 6 Tf 68 3 10 rerun model with Seasons Seasons other option Examplei Example 2 Example 3 CHOOSE ONE option 1 C Hint For Option 2 to see other ezamples from the set of 150 simulations hit the F97 key on your keyboard Choose no simulation Option 1 or simulation Option 2 HERE 17 Pros and Cons of Simulation The pros of such a system is that the simulation model amplifies small sets of data and provides a more stable probability distribution The cons of such a system is that the technique ignores possible biologically important factors that may determine the actual real life probability distribution For example in a given year the rate of health outcomes may have been influenced by the rate in the previous year s The degree of such year to year correlation and influence is unknown and thus it is unknown how importa
14. as entered for 10 years option 1 to estimate the cost effectiveness or use the boot strapping to simulate 10 years of dengue incidence data if 10 full years is not available option 2 If the user selects the simulation option option 2 DenguEcon randomly selects samples four years worth of data DenguEcon uses a system of sampling called sampling with replacement which means that it is possible but very unlikely that a single year will be selected 4 times It then calculates the average for those years and places that average in a separate column DenguEcon does this random sampling 150 times The result is a column of 150 averages With 150 data points DenguEcon then calculates the cumulative probability distribution of the excess health outcome rates including the 5 and 95 percentiles the median average minimum and maximum values If the user selects the simulation option the user must first enter the minimum number of years between peak or outbreak years Typically outbreaks occur in endemic areas every 3 8 years Then the user must select the option 1 or 2 In either case a graph displaying the annual incidence entered on page 2 of 13 is shown on the left If simulation is chosen the graph on the right displays the results of the simulation By hitting the F9 key the user can see other examples of the data using simulations If simulation Option 2 is to be chosen enter the minimum years between peak or outbreak ye
15. ccrued for dengue vaccination up to ten years in the future The Office of Management and Budget of the U S government uses a discount rate of 396 http www whitehouse gov omb circulars a094 a094 html but higher discount rates may be appropriate in developing and transitional economies What is a discount rate In a situation of limited resources societies and even individuals have a preference for consuming now versus delaying consumption to a future date That is there is a cost to delaying use of resources This time preference means that 1 available today is valued more than 1 available in say 10 years time In order to be able to directly compare 1 now versus 1 received ten years from now the 1 that will be received 10 years from now is discounted to the current year i e discounting allows an apples to apples comparison of costs and benefits that occur in different years It should be noted that all benefits and all costs should be discounted such as costs of future vaccinations or treatments or cases averted Entering 0 is also an option Note that a discount rate of 0 essentially indicates no discounting is done 25 RESULTS This page summarizes the input data These graphs are provided so that a DenguEcon user has a convenient method of displaying some of the critical input data These graphs can be copied and pasted into presentations and reports Age group evaluated Number in target populatio
16. cine effectiveness That is in terms of cost effectiveness and within the range of values examined cost of vaccination is more influential than vaccine effectiveness Of course entering in other data may change that conclusion 32 ADDITIONAL COMMENTS What about herd immunity Herd immunity was not included in this version of DenguEcon because without an approved vaccine there is no way to predict the level of herd immunity that will be achieved If included herd immunity would improve the cost effectiveness through the added benefits accrued through the prevention of disease in persons who were not vaccinated We therefore assumed a conservative stance and did not include the potential for herd immunity in this model Reduction in Vector Control Costs When considering the cost effectiveness of a proposed dengue vaccine the potential for additional saving through the reduced need for dengue vector control was considered If an effective dengue vaccine is introduced vector control will likely continue at least until adequate vaccine coverage in the population is achieved Even then equipment and personnel used for dengue vector control are often shared for the control of other vectors and therefore reduction of routine operating budgets seem unlikely to occur One could argue that excess budget spent during dengue outbreaks on vector control could represent a true savings however this benefit will occur until adequate vaccine coverag
17. e in the population is needed The time required to achieve that coverage may be greater than 10 years Again we adopted a conservative point of view and did not include the potential for saving from reduced need for dengue vector control or the prevention of outbreaks 33 REFERENCES Chanthavanich P Luxemburger C Sirivichayakul C Lapphra K Pengsaa K Yoksan S Sabchareon A Lang J Short report immune response and occurrence of dengue infection in thai children three to eight years after vaccination with live attenuated tetravalent dengue vaccine Am J Trop Med Hyg 2006 Jul 75 1 26 8 Meltzer MI Rigau P rez JG Clark GG Reiter P Gubler DJ Using disability adjusted life years to assess the economic impact of dengue in Puerto Rico 1984 1994 Am J Trop Med Hyg 1998 Aug 59 2 265 71 Meltzer MI Neuzil KM Griffin MR Fukuda K An economic analysis of annual influenza vaccination of children Vaccine 2005 23 1004 1014 SUGGESTED READINGS Haddix AC Teustch SM Corso PA eds Prevention effectiveness A guide to decision analysis and economic evaluation 2nd edition Oxford UK Oxford WHO Definition DALY http www who int healthinfo boddaly en The World Bank 1996 The disability adjusted life year DALY definition measurement and potential use Human capital development and operations policy working papers no HCD 68 1996 07 31 34 COPYING AND PRINTING Copying Any Page Table or Graph in DenguEcon can be
18. ed Total target population from Page 2 of 13 multiplied by the compliant with at least one dose of vaccine page 8 of 13 Total of vaccination effort 1 year Shows the estimated lower and upper costs of vaccination of the target population These estimates are calculated as follows Cost of vaccination campaign 1 year cost of vaccinating 1 person X number of persons vaccinated The costs of vaccinating 1 person are derived from Data Entry Page 8 of 13 29 Additional results By scrolling down below the graphs the user will find two tables in which additional more detailed results of the cost effectiveness analyses can be found Each table provides from the 5 to 95 percentile the number of cases deaths averted the costs saved by the health care system before considering costs of vaccination and the cost effectiveness ratios expressed in terms of the lower and upper limits of the cost of vaccination as defined in DATA Entry Page 8 of 13 30 Page 13 of 13 Results Sensitivity analyses Built in sensitivity analyses When user selects Option 2 on Page 4 of 13 to compensate for small data sets recording dengue related health outcomes DenguEcon is programmed to repeatedly sample 150 times the small set of the recorded health outcome rates producing a large number of possible average health outcomes This built in methodology is a form of sensitivity analysis Further the model allows a user to make
19. he compliant with at least one dose of vaccine page 8 of 13 Total of vaccination effort 1 year Shows the estimated lower and upper costs of vaccinated the target population These estimates are calculated as follows Cost of vaccination campaign 1 year cost of vaccinating 1 person X number of persons vaccinated The costs of vaccinating person are derived from Data Entry Page 8 of 13 27 Additional results By scrolling down below the graphs the user will find two tables in which additional more detailed results of the cost effectiveness analyses can be found Each table provides from the 5 to 95 percentile the number of cases deaths averted the costs saved by the health care system before considering costs of vaccination and the cost effectiveness ratios expressed in terms of the lower and upper limits of the cost of vaccination as defined in DATA Entry Page 8 of 13 28 RESULTS Page 12 of 13 Results with Simulated Rates Option 2 Net cost case or death averted This page provides two graphs depicting the cost effectiveness accrued over 10 years after vaccinating a cohort of persons in the target group defined on Page 2 of 13 against dengue in a single year when the user does not have a full 10 years of incidence data and chooses Option 2 on page 4 of 13 to use the bootstrapping method estimate a full 10 years of data based on the available data The equations used to calculate the data dep
20. icted in these graphs are provided earlier in this manual Methods page 8 The graph on the left depicts to cost per case averted where a case may be inpatient only outpatient only or combined inpatient outpatient The definition of what type of case is depicted in the graph depends upon the data entered in Page 2 of 13 The graph on the right depicts the cost per death averted What do these results mean On the X axis along the bottom of each graph is a range of person vaccinated This range encompasses the range of costs of vaccinating a person entered by the user on Data Entry Page note that if the user changes the range of costs in Data Entry Page 8 of 13 the range in the graph may change On the y axis is the cost effectiveness ratio either cost case averted left hand graph or cost death averted right hand graph The solid line in each graph indicates the median cost effectiveness ratio whilst the two dotted lines represent the 5 and 95 confidence intervals In each graph any time any results a line or part of a line are above the 0 case averted or 0 death averted that means that dengue vaccination results in a net cost to the health care system In order for there to be a net savings per case or death averted to the health care system the line must be below the 0 case averted i e read as case averted persons vaccinated Displays the number of persons assumed vaccinat
21. mmed into the spreadsheets and the user does not have own specialized statistical software Thus this methodological approach of bootstrapping meets the two principles guiding the design of the DenguEcon METHODS DenguEcon is written constructed in Microsoft Excel Visual Basic Equation Non fatal outcomes e g outpatient visits and or hospitalizations The cost effectiveness ratio in the model software is calculated as follows Equation 1 Cost case averted Cost of vaccination program Cases averted X cost of treating a case Cases averted Where Cases averted Rate of cases per unit population X population per unit population X vaccine effectiveness Cost of vaccination program Total cost to vaccinate one person X population X percentage compliance case here could be an outpatient only a hospitalized patient only or a combination of both population here refers to a given age and risk group defined by the user that is the focus of an analysis done using DenguEcon see also Methods Page 2 of 10 below per unit population refers to the standardized unit used to measure the health outcome s For example number of outpatient visits per 10 000 population When using equation 1 DenguEcon is coded to automatically adjust the population to the same unit used to measure rates Equation Deaths The cost effectiveness ratio in the model software is calculated as follows Equation 2
22. n Perspective Health outcome data Population unit of rate Modeling option Outpatients per inpatient Vaccine effectiveness Compliance receive the first dose Require more than 1 dose Effectiveness of 1 dose Effectiveness of 2 doses persons among those who receive the first dose receive a second doses Net Effectiveness Cost of vaccination Lower cost person Upper Cost person Discount rate Cases per unit rate reported by year Deaths per unit rate reported by year Do not enter data on fhis page Value entered Country Name 1 ta2 1 847 351 Health Care System Inpatient and Outpatient 100 000 Option 1 Use rates as entered no simulation 40 000 1 6 75 yes 41 90 75 7895 6 24 54 33 3 2560 1607 1751 1717 2517 0 42 0 233 0 31 0 25 0 368 This page summarizes user defined inputs 26 RESULTS Page 11 of 13 Results without Simulated Rates Option 1 Net cost case or death averted This page provides two graphs depicting the cost effectiveness accrued over 10 years after vaccinating a cohort of persons in the target group defined on Page 2 of 13 against dengue in a single year when the user has a full 10 years of incidence data therefore chooses Option 1 on page 4 of 13 no bootstrapping is needed to estimate a full 10 years of data The equations used to calculate the data depicted in these graphs are provided earlier in this manual
23. nefits of vaccinating a cohort over a 10 year period with vaccination occurring completed at the beginning of Year 1 and benefits 1 e disease averted accruing for 10 years Discount rate In a situation of limited resources societies and even individuals have a preference for consuming now versus delaying consumption to a future date That is there is a cost associated with delaying use of resources This time preference means that 1 available today is valued more than 1 available in 10 years time In order to be able to directly compare 1 now versus 1 received ten years from now the 1 that will be received 10 years from now is discounted to the current year i e discounting allows an apples to apples comparison of costs and benefits that occur in different years So at a 396 annual discount rate society is essentially indifferent or values equally 1 received 1 year from now or 0 97 received now It should be noted that all benefits and all costs should be discounted such as costs of future vaccinations or treatments or cases averted Note that a discount rate of 096 essentially indicates no discounting is done DenguEcon allows the user to use any discount rate desired What is an appropriate discount rate Appropriate discount rates can differ by perspective In analyzing for example a domestic project funded by the U S Federal Government the Office of Management and Budget http www whitehouse gov omb
24. nt it is to have such year to year correlation included in the construction of a bootstrapped probability distribution 18 DATA TO BE ENTERED This data entry page is the same in concept and has the same elements as Page 3 of 13 Rates of non death dengue related health outcomes The explanation of each of the data entry elements required is the same as for Page 3 of 13 with the difference that in this data entry page the rates entered must measure the excess rate of deaths due to dengue Unit of rate e g per 100 000 540 000 population i DATA rates Warning A minimum of 5 years of consecutive data is required 0 420 0 233 0 310 0 250 0 368 Calculated mortality rate per unit population 19 DATA TO BE ENTERED Page 6 of 13 Cost per person treated In and outpatients NOTE Costs are marked in terms However the user can enter cost data in any currency desired The only rule is that all cost data must be entered in the same currency Outpatient This column of data records the cost of an outpatient visit for dengue related illness The categories in this column cost visit cost drugs cost other are suggested cost elements that analysts and policy makers may be interested in recording However there is no absolute set rule as to what should be recorded in this column Technically what is needed is a number recording the cost to the health care system the perspective of Deng
25. oed A A 26 Page 10 of 13 Summary Input Graphs eese 26 Page 11 of 13 Net case or deathaverted without simulation 27 What do these graphs mean ioter decode te teeth ee eR e eae Epor eode 27 Total vaccination effort csocteonie con ne nhe tee TR AR Pei ixi 27 Additonal tOsults o ose eR o rexE ta over eA ou de APTE 28 Page 12 of 13 Net case or deathaverted with simulation 29 What do these graphs mean ssssssssssssreesrssserrssersseese 29 Total vaccination effort o eoe cereo ree tene tenriecores 29 Additional results Gaus pee e a Dore Raves Fade yh teet Uus 30 Page 13 of 13 Sensitivity analysis uei Pen ariete de ead anaes 31 Built in sensitivity analysis DX What additional sensitivity analyses c can a user r do cadeh 31 Example Graph of sensitivity analysis sss L Most influential E ACDRRNMRMNNMMME 1 ADDITIONAL COMMENTS ccccccccccccecessencccccecceecsssssensccecceccsssssencaceseccssessentaceseess 33 What about herd imfeinity7 uisi ccantecinsaerascaiaadesnnauaeoo Reduction in Vector Control Costs Lees 22 REFERENCES omstan Neat tec sea tbt M tu tute iae toam qu d ut UE 34 SUGGESTED READINGS ocsi 55 Gace saddles aea piden nage suid eeu PN pda Recon de iu d ene 34 COPYING AND PRINTING 4 citet ciate banca en a a ae vM GE UR COVERS ER RU Ee UNE 33 e OW GIN
26. ost of any negative vaccine side effects The methodological concepts that are embedded in DenguEcon therefore focus on how to combine these four data elements in as simple a manner as possible but still provide a realistic picture of the cost effectiveness cost per case averted or cost per death averted BACKGROUND TO METHODOLOGY TWO GUIDING PRINCIPLES First guiding principle Include variability of input data A model for calculating cost effectiveness of dengue vaccination must allow for the variability in the four types of data used Second guiding principle useful to intended audience Any model software developed must be sufficiently simple so that the intended audience practicing public health officials can readily use it and readily interpret input and output values The intended audience must be able to use the model without using specialized software programming skills or having advanced knowledge of statistics The methodology of combining these four types of data must allow for the fact that biologically the impact of dengue varies greatly from year to year or season to season with large outbreaks Furthermore an approved vaccine for dengue is not currently available Therefore the methodology must explicitly include allowances for the variability of the effectiveness of proposed dengue vaccines and the inevitable variability in costs both of treating dengue related illnesses and dengue vaccination Thus the
27. s Figure Variation in total cases incidence and laboratory positivity rate of dengue cases in Puerto Rico 1963 to 2003 Source J Rigau Americas Dengue Prevention Board Meeting Mexico City Jan 17 20 2008 Therefore a descriptive statistic such as the average rate of health outcome would often be misleading A single outbreak year could greatly increase the estimate of the average health outcomes Similarly the estimate of an average could be influenced by the inclusion of one or two years which are atypical and unusually mild due to environmental conditions or improved prevention activities Thus if a user has only five years worth of data the calculated average cost effectiveness ratio can vary greatly depending on which 5 years are measured 16 Conclusion Having less than 10 years worth of data could create an inaccurate probability distribution describing the rates of health outcome i e the small data sets that are likely to be available will not reliably record the probability of a given rate of health outcome occurring Thus in order to reduce potential biases in probabilities caused by having only 5 10 years of data DenguEcon uses a statistical technique called bootstrapping to amplify the number of measurements available to record the probability of health outcome rates Occurring Data Simulation Option Therefore on this page the user is given the choice to use the dengue incidence data
28. s and do not necessarily represent the views of the author s sponsoring agency This version is a beta test version As such it has not been officially cleared SYSTEM REQUIREMENTS DenguEcon uses the Windows operating system Microsoft Windows 2000 or higher and Excel Microsoft Office 2000 or higher We recommend using a computer with at least a 486 Pentium processor and at least 128MB RAM DenguEcon requires up to 4 1 megabytes of storage space on the computer s hard drive Microsoft Windows and Office are copyrighted products produced by Microsoft Corporation WA Use of trade names and commercial sources is for identification only and does not imply endorsement by the U S Department of Health and Human Services LOAD AND START Before loading and starting DenguEcon you must make sure Excel s security level is set appropriately You must first do the following steps 1 Open a blank Excel spreadsheet 2 Click Tools and then click Macro choose Security at Data Window Help Spelling F7 3 Research Alt Click Error Checking Shared Workspace Share Workbook Protection Online Collaboration Formula Auditing Macros Alt F8 Customize Record New Macro Options Security Data Analysis Visual Basic Editor Alt F11 v w 36 Microsoft Script Editor Alt Shift F11 Continued on next page 3 In the pop up box
29. sider altering the default value of 10 minutes between each AutoRecover To do so change the interval between automatic file back ups as follows 1 Onthe Tools menu in Excel click Options and then click the Save tab 2 Select the Save AutoRecover info every check box 3 In the minutes box specify how often you want your Excel program to save files e g 20 minutes We strongly suggest that users do NOT select the disable AutoRecover feature INTRODUCTION Estimating the economics of a dengue vaccination program requires combining four types of data First we need data recording who becomes clinically ill with dengue and the consequences of such illnesses i e epidemiological data We also need data that measures the effectiveness of the vaccine The effectiveness of the dengue vaccine is defined as the reduction in the incidence of dengue amongst people who have received the vaccine compared to the incidence in unvaccinated people As no vaccine is currently available we suggest using effectiveness estimates from similar vaccines such as the yellow fever vaccine We also need cost data recording the cost of treating patients who become ill with dengue infections and receive medical care and finally we need data recording the cost of vaccination and not just the cost of the vaccine in a vial In other words we also need the cost to pay the staff that administer the vaccine pay for any additional supplies and as well as the c
30. t person vaccinated Lower amp Upper In these two columns the DenguEcon user enters upper and lower estimates of the cost of vaccinating one person against dengue The guiding principle in this data entry sheet is that the cost of vaccinating a person is a great deal more than just the cost of the vaccine Indeed delivering the vaccine to the clinic and assembling all the elements to administer that vaccine physician and nurse time syringes clinic overhead etc may be more costly than the actual purchase of the vaccine As in the other cost data entry pages Page 6 of 13 the categories in this column vaccine physician nurse etc etc are suggested cost elements that analysts and policy makers may be interested in recording Again if a user only has at least initially a total estimate that incorporates all the costs without a breakdown the user can merely enter that total cost in the row labeled Vaccine and enter 0 for all other elements WHY two columns Lower and Upper Lower and upper estimates do not necessarily have to represent minimum and maximum range of potential cost However it is realistic to assume that over time and by locale costs will likely vary This potential for variability must be explicitly modeled Side effects Adverse events may occur with dengue vaccination i e patients can experience harmful side effects The vast majority of recorded side effects tend to be of a minor na
31. that will appear see below set Security Level to Medium 4 Click OK Continued on next page 5 Go to the drive location where you have saved the copy of DenguEcon e g C Drive Double click on the file which will open DenguEcon 6 As the computer is loading Excel if it is not already loaded and DenguEcon a window will pop up see figure below asking you if you want to Disable Macros or Enable Macros Select click Enable Macros DenguEcon will then fully load and you will be ready to start Security Warning C Documents and Settings Mark Desktop DenguEcon xls contains macros Macros may contain viruses It is usually safe to disable macros but if the macros are legitimate you might lose some functionality Disable Macros Enable Macros AUTORECOVER Changing Excel s AutoRecover time Excel software contains an AutoRecover feature which essentially automatically backs up at pre set intervals a copy of an open workbook This reduces the risk of work being lost should there be a problem with either the software or hardware that causes Excel to malfunction e g crash When the AutoRecover feature automatically starts all other functions in the workbook are temporarily frozen With large workbooks which can take a few minutes to back up this can mean an inconvenient halt to working To reduce the number of potentially frustrating halts in the flow of work users may wish to con
32. ture such as soreness at the site of vaccination However some side effects can be serious and very rarely life threatening Such serious side effects require medical attention and therefore represent a cost associated with vaccinating a patient The costs associated with treating these serious side effects no matter how rare should be explicitly accounted for In this page two types of side effects are accounted for mild side effects and anaphylaxis For either side effect the user enters the probability of such a side effect in column Probability and the cost of treating one case of that side effect in column treat DenguEcon then calculates the average cost per person vaccinated for treating either type of side effect as follows Cost per person vaccinated for treating side effect probability of side effect x treating person with side effect DenguEcon then sums up the three costs per person vaccinated and this sub total is then automatically entered into the columns totaling cost per person vaccinated both lower and upper estimate columns receive the same estimate for treating side effects 23 Assumed compliance Enter the assumed percentage of persons in the target population identified on Page 2 of 13 that will be successfully vaccinated against dengue in a given year The range of possible entry values is from 0 to 100 24 On this page the user is allowed to enter the discount rate for benefits a
33. uEcon see earlier of treating an outpatient for dengue Such estimates should follow good accounting practices and record all expenditures of delivering appropriate treatment to a patient Such costs would include physician and nurse time salary and wages salary for any clinic administrators cost of supplies medicines treatment and laboratory tests overhead to operate a clinic or hospital e g utilities building repairs and maintenance etc The detail entered into DenguEcon is dependent upon the user It may well be for example that a DenguEcon user can readily obtain an estimate of total cost but not a breakdown of costs In that case the user would simply enter the total cost in the cell cost visit and leave the other cells in the column as 0 A user may for example find it easier to estimate cost of a health care visits in a separate sheet and just enter the sum total in the cost visit cell Inpatient This column records all the costs of treating an inpatient for dengue related illnesses As in the outpatient cost column there is no absolute set rule as to what should be recorded in this inpatient cost column Technically what is needed is a number recording 20 the cost to the health care system the perspective of DenguEcon see earlier of treating an inpatient for dengue The detail entered into DenguEcon is dependent upon the user It may well be for example that a DenguEcon
34. user can readily obtain an estimate of total cost but not a breakdown of costs In that case the user would simply enter the total cost in the cell cost day per hospital bed and leave the other cells in the column as 0 outpatient visits It is realistic to assume that an inpatient may also have had an outpatient visits before and possibly after for follow up an inpatient stay However the user must ensure that if a number of outpatient visits associated with an inpatient stay are entered in this column then those outpatient visits are REMOVED from the rate of outpatient data Page 3 of 10 In many instances it is likely that a user will not be able to identify which outpatient visits if any are associated with an inpatient stay In such situations in order to avoid any double accounting the user MUST enter 0 visits in the outpatient visits in this column of costs for inpatients of outpatients per inpatient If in data entry Page 3 of 10 Rates of non death dengue related health outcomes you entered combined outpatient and inpatient data then you MUST enter the approximate number of outpatient visits per inpatient hospitalization 21 DATA TO BE ENTERED Effectiveness of dengue vaccine In this page a DenguEcon user enters the assumed probabilities of vaccine effectiveness Users are asked to enter the probability for thos who receive only one dose and the probability for those who receive both
35. veness ratios The data for the graph were obtained by first fixing the weighted average vaccine effectiveness on Data Entry Page 7 of 13 and the recording the resultant median 50 percentile cost effectiveness ratios from the detailed tables of results Results page 11 or 12 of 13 The process is repeated a number of times each time entering a different estimate of vaccine effectiveness Note that in order to keep the graph clean and simple I have chosen to omit any confidence intervals i e the 5 and 95 percentiles are not drawn in but they could be added if the user so wishes Interpreting the results determining which variables are most influential By looking at the graph in the sensitivity analysis page Page 13 of 13 and the graphs in the results page Page 11 or 12 of 13 a user can begin to see which variables within the range of values examined are most influential For example in the sensitivity analysis 3l drawn from results produced using the illustrative data on Page 13 of 13 we note that if there is a 6 fold increase in the cost of vaccination then the cost effectiveness ratio essentially doubles But as vaccine effectiveness increases from 5096 to 9096 an 8096 increase we note that the cost effectiveness ratio only at a much lower rate Thus a user can conclude that the results derived from the illustrative default data are more sensitive to changes in the cost of vaccination than changes in vac
Download Pdf Manuals
Related Search