Tobacco Documents Online

_~red, and based on od¢l con- r~ented in sses from submodei guarantee lecdons. a that will placed by (a) Actual se intoler- tion of the tt~rmation n--flhmental :stimate of • emprise .this • Tacker, and profile: The ~pport of the lustrial Engi- es in Support "., of Industrial ~plementation ~ of Industrial in cardiovas- PREVF-I~FI'|V~ MEDICIN~ t 87--291 (I981) Evaluation of the U.S. Air Force HEART Progmm, iD~rn,on,st~ation,1 /~,Louls J. COTE Department o~_ Statistics, ~urdtle Universir.~. ~esr Lafayette. Indiana 47907 iand bi0g~aph~al, d~G-f~0~-~th6 p~tieipan~s, some Of which is used in conducting the HEART program activities. The other type of data, relating to the performance of these activities, includes offie~ records and results of some studies designed to furnish detailed information about some of the program activities. We omline plans for summarizing the data and presenting it in a form that will be useful to those who will plan the Air Force-wide program. INTRODUCTION The HEART program dcmon.st.ration is. now being conducted in somewhat dif- ~forent-modes on four Air l~oreg Bases. !rts purpose is to give data on how w¢l.l the dii~erent a~ti.vifi~s__ .6$~.~h-e~ ~IEAReI" prog.~m work together aad how effective they adhfeving ~'f~tim~ of ~he i~'~g.rah~. In this papea" we ~utline the plans for taking data and for usl.ng it to assess the performance of the program demonstra- tion. We describe part of the work done by a research group at Purdue University to assist the United States Air Force School of Aerospace Medicine. More detailed information is reported elsewhere (2, 4, 5). RISK REDUCTION EVALUATION - The ultimate aim of the-HEART program is to reduce the incidence of cardio-~ vascular disease in the Air Force. The nature of the disease makes it impossible to observe the effect of preventive measures in a period as short as that of the 18:month demonstration. Therefore, in assessing the effectiveness of the demon- stration in achieving the program's aims, we are necessarily limited to short-term aims, i.e., to reduce the values of risk factors in the Air Force population. Risk factors for cardiovascular disease have been identified in the Framingham study (2). This study also developed a formula which uses values of an individual's risk factors, age, and sex to estimate his probability of having cardiovascular disease within 6 years. We refer to the probability calculated from this formula as the Framingham risk. Changes in all of an individual's risk factors over a period of time can be summarized as a change in his Framingham risk. At present it is not known how changes in actual risk (or incidence) of cardiovascular disease are related to changes in the Framingham risk. It seems clear, however, that if the ' Research funded under Contract F33615-80-C-0615 for ~he U.S. Air Force. School 6f Aerospace Medicine. 0035 287 009 !-7435/81/030287-05502.00/0 All rigkts of r©lmXt~ti~a itt ~ny form rtscrved. NOT'ICEe Tills MAT~R!~L P!'Y B; PX371~Clr'O BY COPYRi.~al t:'.. ( .... := ,t/.. U. S.. ~ T!05280100

LOUIS J. COTE Framingham risk does not decrease, then the actual risk has not decreased. That is to say, the HEART program must achieve the short-terra aims in order to achieve its ultimate aim. Two activities of the HEART program are devoted to reducing values of risk faetors~ These ~ve i~l~erv.~nti;on and~ general education., dcscff~bed in de~ta~! in an a 2 ~ ~ Facton~' experiment wtth t~ese acuv~t~es as factors wtth the levels, pres- ent-absent. Each of the four t~atment combinations is used at one Air Force B~e as in.all, tee ~n ~. personnel are given two e~fions, iB the ~:ra~mingham risk or h~ ia~dividu~l risk factors can be analyzed as an an~ysis of variance. The expedment departs from an ideal factorial in several ways. The sample unit to which the treatments, Intervention and Gene~l Education, are applied is the Air Force base. Each treatment combination has only one sample unit so that there are no degrees of freedom for error. Therefore, any differences among the Air Force ba~s.es in the demonstration will be confounded with the effects of the treatments. ~hese d~f~erences may be differences in morale, or differences d,ue to differenl ~t.ti~des. o'ft~e base commanders, etc. The individual pa~ci~ants are su~aits. T~re are am~Ldegr~e~es of freedom for error among thd~e bat they ea~t be u~.d~ for t~fi~ ~e effects of the treatments. Re~fic~tio,n ~r Force ~B~ was ~o~-"feas~l~ ~use o~ the expense. To reduce t~e amo~l o~ con- fonnding, the bases w~re c~osen to have app(oximately the same sizes, age dis- tributions, and types of mission. The effects of the factors cannot be tested for statistical significance. Their statistical reliability will be judged from sources outside the experiment. A matter which does not affect the experimental design, hot must be taken into account in discussing the results, is the absence of general education. All fou~ bases.have two screening examinations for_ each participant which take place {n- the HEART Program Office, The o~ce is prominently marked with the HEART Progr~ logo. The participant completes a questionnaire, which asks detailed questions about smoking and dietary habits. Questions raised by the pz~icipant can be answered during conversation with the HEART Program personnel during the screening examinations. This process provides a broader background ofgen eml education than exists at nonexperimental Air Force bases. Nevemheless, it is Present N. H~unp. NOo Car. Texas Texaa Absen~ General Education Fro. I. HEART program demonstration. uniform of Air Ft cardiova bases. The ex is one la Carswell' screenin~ screenin~ Althou only a srr at each ~, cholester hypertrot frequenc" fasting b~ ante, for Chang~ eombirtal similarly. In addi taken int, graphic, :. in Table regressio or chang, regressio the ehan categoric as "' alum analysis hope to ~ regressio listed in ~. analysis terms of: among th We ha working this anal.~ begin the T105280101

That is achieve of risk s, pres- r Force :.pie unit • d is the so that ~ong the s of tlae s due to ants.: .are i.r ~.~¢~c age sted for sourCeS ken into All four place in _ HEART detailed ~icipant ~1 during t of gen. ess, it is THE LI.S. AIR FORCE HEART PROGRAM 289 uniform in the experiment so it will not affect comparisons. Moreover. if any kind of Air Force-wide HEART program is adopted, the background of education in e'~tio, wa~eular .disea, s~ preven~ig~ ~!| probabl~ rise ,to. this ~evcl at al.~ Air Fore bases. The .experimen'ta~ results Will probcbly ~e ~ffeeted by 'the type of control, which ~'s..._o:~|}~l~ ,_of ~i~' ir~ ~a~ ,~0,:l~d~ttet. Th~ ~p.art~cip~an~s at the control screening. At the other three bases each participant is informed soon after the first sc~esaaing o£his..~:.a.mi~,~ham ~i~sk nnd ~lso w~het:her it i.s high. or low. The eon,trol Alth'ongh a great deal' Of'[nformat.i0n is collected in the screening examinations, only a small part is used in the evaluation of the program. This includes the values at each screening of the risk factors; age, systolic blood pressure, total serum cholesterol, fasting blood sugar, and cigarette smoking. We omit left ventricular hypertrophy (LVH) because of the difficulty and expense of measuring it. The frequency of true positives for LVH is extremely small in the Air Force. The fastin~ blood sugar levels are converted to a zero-one variable, glucose intoler- a;~t¢~ tb~.r use in, th~ F\r_a~g~.i~ngham ri~sk f~mula. --'~,~0ges in: t~e'?lar-~"m~gh~m ~'~k ~v$~i; be measured for e~ch of the treat~fient ¢o, "~_b~fi~s, Ig.gd~i°tion eg~h~,of~he s~- l~.arate risk factor changes will be a~nal~yzed In, addition to the treatment co:mb~fiations, several concomitant variables will be t~eu into ac~oum. These are grouped for convenience into the categories, demo- graphic, physical characteristics, and hs~ factor changes. The variables are listed in Table 1.. These variables will. be used as independent variables in multiple- regression analyses whose dependent variables ~e changes~in. Framingham risk, or changes in one of the risk factors. The risk factorchanges will not be used in the regression of Eramingham risk; but in a regression on the change in a risk factpr, the changes in Other risk factors will ~ used as independent variables. The categorical variables, sex, rank, etc., including the treatment variables areentered as "dummy variables" (see (6)). Age, also, is categofzed for the regression analysis because the effect of age on change of risk is probably nonlinear. ~e hope to gain some understanding of the effect from the results of this study. The recession will also have other v~iables created by taking products of the ones listed in Table 1. The coefficients of the products co~espond to interactions in the analysis of variance. We will interpret the estimated regression coefficients in te~s of an analysis of variance. Insofar as possible we will adjust for differences among the bases in the distributions of demographic and other variables. LOGISTICS EFFECTIVENESS ANALYSIS We have used the term logistics ~ffectiveness for the smooth and efficient- working of the different operations of the HEART demonstration. The elements of this analysis are many and diffuse in contrast to the analysis of risk reduction. To begin the work we organized the elements into three principal divisions: aspects T105280102

29O LOUIS J. COTE TABLE I REGRESS|O~ VAR|A~ L~ Dependent Y Risk exponent d~crease, risk factor decrease D~moBraphic D'~ Sex (male. female)= (0~I) -~. Rank (enlisted, officer) = .{0, l) ,w,,,~ :' ~'" D~-~~ Fl'y~ng Status '('n~mF,~ = ¢0,1') Base (treatment) Bt Base-wide education {no, yes) = (0,1) Pz, (<35, ~35) = (35 ~ ag~ < 50, other) = (0,1) Pa Systolic blood pressure (BP) P, Serum cholesterol (CH) Ps Fasting blood su~r (GI) Cigarette smoking (not SM, SM) = (0,I) % lean ~dy mass (LBM) Risk factor change R~ BP change Rz CH cha~¢ Ra S~ change (Ol,ll,~,lO) = (-2,-I,0,I) TABLE 2 LOGISTICS E FFECTIVENI~SS--ELEMENTS ...... I- Participant related ....... A First and second ~creenings on each base ............... 1, Measures of queueing parameters 2, Manpower task loading 3. Scheduling and administration B Intervention (for each Intervention Base) L Participation in each focal group 2. Use of hot lines, taped telephone messages, drop-in hour 3. Scheduling and administration C General education (for each G. E. Base) 1. Type and t'requeney of activities 2. Appraisal of effectiveness or popularity of activities II Equipmenl0 supplies, services A Frequency of failure, amount of down time B Supplies and consumables, rates of usage, incidence of exhaustion C Services other than repair and maintenance of equipment D Accuracy of measurement equipment II1 Slaff related A Air Force base staff B HEART staff related t, operating pertainin: offices o~ power 1o Force Da Operaffng is adopte For these It is d'e program includes samples 1 on. Quesl The ass soon afie. plan for tl and coml: been mad analysis c analysis c - formance rial for ot results of I. Arnold. progr 2. Kannel, Fram 3. Petersm HEA neeri~ 4. Peterse~ of the Engir 5. Peterset of the Engin 6. Wesolo, York. 3"105280103

THE U.S. AIR FORCE HEART PROGRAM related to the participants, to equipment, supplies, and services, and to the operating personnel. An outline of this organization is shown in Table 2. Data pe~-t~i,n[ag ,to, mo~t of ~he .elem~nts..~'li~ :be .avai~lab~e ~m the HE~T ,~g,mm offices on the four bases. Exceptions are the queuei~ pa~meters and the man- power loading of the scree~g .operation., and the el'em~nts related to the Air ,~; ~o~ ~ ~ ~ ~ 9~o~m, is adopted Air Force-wide will probably include screening and risk assessment. For th~_o re~oo~,~ de~ed s~ud~ ofthi~ o~eration wig h~made. A pl,an, for work prog~m by Air ~ase personnel: not directly connected with the program. This includes scheduling of participants for screening, laboratory analyses of blood samples from the screening, work related to general education activities, and so on. Questionnaires will be prepared to obtain information of this type. CONCLUSION The assessmem of the ~EA~T pros~m demonstration must be completed very seen, a~r the ~n;~] da{a £rom ~e~ demonstratiea becomes a~fl~ble.. A derailed ~&~ ~e£ ~e m6uetion ~[ ~h¢ ~ ~8,Res~e~ta~,[65 of ~¢suks 5aS been ~:~ e.o~ut~ ~r~g~m~ ~i]I ~e r~d,y ~he~ the .data is re.cej~ed. No p]a,~ ~ave b~*~ ~'~ ~II~9 ~r~ on ~ne.stlons that arfse ~om the ~sgl~s ~ ~ a~o~e analysis or from other aspects of the demonstration data. We have s¢leeted for analysis only that pa~ of the screening data which is directly related to the per- -formance of the demonstration. The remaining screening data Will provide mate- rial for other related studies. Follow-up work on questions that arise from the results of this analysis will undoubtedly present an interesting challenge. .... i. Arnold. C. B., and Jaeobson, L~-~isk reduction in the U.S, Air For~e ~i~r~-prevention HEART program, Prey. Med, 10, 270-276 (1981). 2. Kannel. W. B., McGee, D., and Gordon. T. A general cardiovascular disease risk profile: The Framingham Study. Amer. J. CardioL 38, 46-51 (1976). 3. Petersen. C, C., Ravindran, A.. Sweet, A, L., and Cote. L.J. "Analyses in Suppo~ of the HEART Program;" Projecl Repo~ ~94-53-1287. Purdue University, School of Industrial Engi- neering, Lafayette, Ind., Septem~r 1978. 4, Petersen, C. C,. Ravindran, A,, Sweet, A. L., and Cote, L. J. "'Modeling and An~yses in Suppo~ of the HEART Program;" Project Repoa 0153-53-1287. ~rdue University, School of Indust~al Engineering, ~fayette. Ind. September 1979. 5. Pete~en, C. C.. Ravindran, A., and Cote. L. J. "'Planning for Data Analysis and Implementation of the HEART Program;" Project Repo~ 0195-53-1287. Purdue University, School of Industrial Engineering, Lafayette, Ind., September 1979. 6. Wesolowsky, G. O. "Multiple Regression and Analysis of Varlanee,'" pp. 103- ! 14. Wiley, New York. 1976. 3-105280104