Tobacco Documents Online

SHORT REPORT Advice. to. Red~uce Pl',asm~ Chof, esterOt ~R,C,.,~,O F. HEI.LE,.'~L, GH D. TUNSt~LL PEbOE. AND ~GEOFFREY.ROSE~ A slmplc eight-item questionnaire wa~ used to assess the salumte~ fat content of the diet of participants in a medlifactorial trial of the prevcnlion of coronary hea~ disease. queqlh~nnaire- was found Io have go~ repeat;~bility, and a score d~ri,ved from h correlated Hgn~ficantly with a more lengthy dietary ~ecall me~hgd (the 3-day dial) in assessin~ the amount ofs~tlurated fal in 1h¢ diet: high scorns reflected a low salura¢¢d ~t intake. The score was higher among Ihose advised Io reduce their salu~tled ~at inta[e (suggesting tha:t they had laken the advice) than among thos~ who hnd been given no advice. Men given personal counseling had higher ~ores than thoqe who ~er¢ only sent letters and leaflets containing _ c~.~cnfial!y lh~ ~ame information h3" pt~st, Among tmadvised conffo~ meq there was qo co~elalion between the score nnd the plasma choleslcrol level: but among ~he ~en given ndv~ce, the h;gher their score Ihe ~rcater w~s the £==]l in lhe]~ plasma cholesterol, This highlighls the parado~ that an, individuzff~ current.saturated ~at intake ¢anno~ predic( his p~sma.cholesterol level, yet a change in =hi~ level is readilT" predicted by even a crude a~ssment ~F change in dietary fat, • INTRODUCTION The Heart Disease F'revention Proj'ect is a controlled trial in middl¢-~ed men working in British industry (6). Its purpose is to discover the acceptability and effectiveness of advice on modifyin8 coronary heart disease (CHD) risk factors. One major part of'this advice concerns a reduction in saturated fat intake aimed at lowering the level of plasma cholesterol. Some 12- I5'~ of the men working in the 12 factories which had been randomly assigned to the intervention side of the trial were separated into a "'hi~h-risk'" ~roup. based on a composite score derived from blood pressure and cholesterol levels, age, cigarette smoking, over'weight, and physical inactivity as determined by an initial screening examination and questionnaire. These high-risk men were recalled for personal advice by the fac- tory doctors and were seen on averase four times in the first year (and no~ system- atically recalled for advice Jn later years), while the other~ non-hi8h-risk men initially received essenlially the same advice only by letter or leaflet, Approxi- mately one-third of the non-high-risk men had a single counseling session with a ' Current addrc<s: Depa~lment of Community Medicine. St. Thomas's Hospital Medical School. London ~EI ~EH England : Curvet add~: London School nf Hygiene & Tropical Medicine. Keppcl Street. London Engl=md " 1')091-7435.'g t'030304-07502.00'0 3~4 0037 NOTICE: THIS MATERIAL MAY BE PRO]ECI'ED ~£ .COE~GH[ LAY/ (TITLE 17. U. S. CODE) T105280107

ASSESSING THE EFFECT OF" DIETARY ADVICE 365 ~ nurse at a later ~nn,ual examination. A I0~ random sample of men working in the [tem~t.ue~firh:n~:ii~e w~,~.~fl~,o ~e~ ad:herei.~eeqo-~e~-~y advice at the anniversary examination. This report assesses the value of the questionnaire both as a measure of saturated fat intake and as a measure of compliance with advice. In addition, the questionnaire has been used to correlate changes in dietary fat intake wi, h changes in plasma cholesterol levels. METHODS , At the fourth anniversary of the initial examination, on the intervention side of the ~rial, a 5% random sample of man-together w4thall those inifiMly assigned to the high-:r~sk group were i~nvited for further r~examt.nation tas were men in the oril~i~n~l. I..~ sa~.~!~l'e on. ~'~he control side), l.n addison to ~the regular examin~a~J'en~ pro.cedu~res, the d~et questionnaire was given to those attending-the examination in 9 out ofthe 12 intervention and 9 out of the 12 control factories. We report here on -those men who completed a.diet questionnaire and had a cholesterol measurement (98% of ~h_e lotal .examined at this anniversary at these fact0ries): they comprise .... ~- : 220 non-high-risk men in the 5~ random sample and 575 high-risk men on the i .... intervention side, .as_~well as)511, men in the contro!.sample... The overall response : . rate at this anniversary examination was 93% on the intervention and 89% on'lhe .~ - The diet questionnaire (see Appendix) includes simple questions about the ust~ai use of butter or margarine; milk, egg, and sausage intake; whether fat is usually cut off meat; and oil or solid fat used for frying. These foods were chosen as they formed the basis of the dietar3, advice given to the me;a: advice about meat and baked foods was limited. A score was assigned to cacti answer and then summed to produce a total diet adherence score• A high score indicated good adherence to the dietary advice, the top score possible being 21. Plasma cholesterol levels were assessed on capillary blood by an Autoanatyser using a colorimetric method; the laboratory was standardized against the WHO reference laboratory (2). Questimmaire repeatability. A sample of 22 men in two factories who had been recalled for a separate study were asked to complete the diet questionnaire be- tween 3 and 4 months after completing one during the anniversary screening examination, There were only small differences between occasions (20 of the men had a difference of 2 points or less) and the mean scores for the group were identical on each occasion. Questiommire validatiom On a later occasion a random sample of 68 of the trial 15articipants was asked to complete the diet adherence score questionnaire and subsequently to make a detailed dietary diary reco~'d over 3 days (5), There was a significant negative correlation between the diel adherence score derived from the diet questionnaire and, the estimated mean saturated fat content for the 3 days derived from the diary record (r = -0.3, P < 0.05). The estimated mean saturated fat content for lhe 34 men with diet adherence scores of 15 or less was 53.4 g, compared with 41.2 g in the 34 men with scores of 16 or more (P < &011. T105280108

36(~ HELLER, TUI'~STALL PEDOE, AND ROS~E RE~.ULTS Distribtttion qf the Diet Adherence Scores trial, with a particularly marked sh~fi towards high level's in the high-risk men. A number of [he entry characteristics of the high-risk men were examined to see if it was possible [o discover what son of person would take the profe~ed dJeta~ advice (Table [). "'Advice taking" was clearly assoqJated with being a non- smoker: those who did not smoke pt enlry and those who stopped during [he 4 years were significantly more likely to have a high diet adherence score than those who continued to smoke. Older men and those who had managerial status or previous heart trouble wer~ slightly (and nonsignificandy) more likely to ~ake advice. P/et~s,t~t.a. ~hole,vtero~ Change and Diet Adherence Score Among the control men there was no relation between the diet adherence score and the change in cholesterol over 4 years. (The latter tended to increase in all groups.) In both the intervention samples, however, there was a marked relation- ship between the score and'the change in cholesterol levels (Table 2), In the high-risk men. the higher the score the greater the fall in cholesterol. Among the non-high-risk men the same gradiehtwas seen in that the expected increase in cholesterol levels over the 4 years ~'as attenuated in those with higher diet adher- ence scores. In the high-risk me~ (Table 2) those with higher diet. scores ha~ higher initial cholesterol levels, perhaps reflecting stronger advice given to those with higher cholesterol levels. To account for the secular trend over 4 years and regression to the mean, a regression equation was derived from the unadvised control sample reposed he~ by ordering their initial cholesterol distribution into decries and compa~ng the mean cholesterol levels of each of these deciles initially with the level after 4 20 .15. ]0 ~-6 7-9 J0-12 Lx-|5 ]6-1a ]9"21 DIET SCOIE FIG. I. Distribution of diet adherence score at fourth anniversary,. "[105280109

oO ASSF.$SING THE EFFECT OF DIETAR~ ADVICE Age 40-49 50- 59 Married Other Manager Other Heart trouble in past - Heart trouble in relati'..e- Evertold BP r, dsed . ~ C~gare~te__- smoker Yes Yes No Yes No - YeS.in} Still - smoking at fourth unni_versary; {b) Stopped smoking at fourth anniversary No 173 32 404 39 512 37 65 35 105 41' 472 36 51'6 36 155 38 422 36 71 38 506 36 297 27~ 110 45 170 49 * Proportion scoring 19- significantly lower in continuing smokers than in those who stopped smok- ing during lhe 4 years and those who were nonsmoker~ on entry {P < 0.001L years. This regression equation was then used to predict the expected cholesterol change of each of the groups in the absence of dietary advice. Table 2 shows that in both intervention samples the higher the dietary adherence score, the greater is the difference from the expected cholesterol level. It was also possible to compare the diet adherence scores and cholesterol change among the high-risk men in each of the nine intervention factories where the diet questionnaire was used. There was a significant rank order correlation between the factories ranked.according to the mean reduction in cholesterol and the proportion of men with a high diet adherence score (Spearman's rank order con'elation r~ = 0.80, P < 0.05). DISCUSSION This simple questionnaire was first designed'as a counseling aid. The score appears to be normally distributed in a control population, to have good repeat- ability, and to reflect broadly the actual amount of saturated fat eaten as assessed by other means. The same questionnaire used in a sample of British teenagers T!05280110

TABLE 2 DIF.F ADHI'REP, CE 5C()RI~, AND CtlOI.Fsrr, RoI. Percentose u h~_n~i Actual P~'~d" Diet adherence score Plasma c holcslcrol I ms/10"0 ml) N lnillnl 4 Y~rs Con(rul Low 182 I nte rmeditlte 14 I 112-131 High 188 114-211 Inlcrvemion Low 83 non-high-risk (5-131 Inlermcdiale 70 114-161 High 6"/ 117-211 221.3 215.6 216.4 266.9 213.3 247.7 Inlervenlion Low 184 243. I - 1.9 high.risk (6-15) • Inlermcdiale 179 259.4 241.7 -6.8 116- 181 High 212 ' 264.l 119-211 P~dicled from cholesterol chanFe in whole control sample~dlows for increase in cholesterol levels over4 ycilrs ;tnd Icxlh 229.9 223.0 +3.4 225.5 *4.2 218.6 211.7 ~2.3 211.5 -11.8 llri

ASSESSING THE EFFECT OF DIETAR~ ADVICE 369 produced a significant correlation of -0.6, between the diet score and saturated flit of its simpl,~c~t3" and ea~re of con~plet-ion. ~,hen compa~,rd! with. more l~Bgl~hy tionnaire, diary, or weighing methods (I. 5). This particular questionnaire may not work as well in situations where lhe customary dietary components are different or where the dietary advice is different. However, we suggest that a simple ques- tionnaire' such as this with appropriate modifications may be useful in different situations. It would appear that personal advice to people who have been told they are at special risk of developing CHD produces a greater behavior change than postal advice to men whose screening results we~'e reassuring. As expected, those pre- pared to rake one item of advice a~e more likely to take a, nother: both initial" nonsmokers and. t.hose who stopped smoking during the study had higher scores than initial smokers who continued to smoke. Of particular interest is the finding that such a simple score can correlate with actual cholesterol change. We have shown here that stated dietary change can be reflected by cholesterol change. This contrasts with the finding that in the control population there is no correlation between diet score and actual cholesterol level,_ repeating many p~'evious such observations (3, 4, 7). This highlights the paradox that an in.dividual's cholesterol level cannot be predicted by his current fat intake, - - -~;eichang~e in level is readilypredicted even-by a .crude assessment of change in dietary intake. APPENDIX: QUESTIONNAIRE ITEMS~ 1. Do (I) (0) (0) "2. (a) you believe that certain foods can really be bad for the heart? Yes No Not sure Do you usually use butter or margarine? (0) Always or usually butter (l) Always or usually margarine (b) If you use margarine, which brand? (2~ Tub marked "'high in polyunsaturates,'" e.g., Flora (1} Other s'aft margarine (0) Hard margarine (score 0 if"butter" in (all How many eggs do you eat in an average week? (3) Not more than three a week (1) At least one aday (2) More than three a week but less than one a day. How much milk do you drink (including on cereals~'? (3) Less than ~a pint a day (2} Between ½ and I pint a day (!) More than 1 pint aday Score given in parentheses. T105280112

370 HF-LLER. TUNSTALL PEDOE. AND ROSE (3)'Twice a week or less (1~ At least once a day ,{2~ ~e~ ~th~ iwiee a ~eek b~t less than O~c¢ a day ~=~z ~.~z .................................................. (0~' More than four a week 7. Do you usually cut the visible fat off your meat and bacon or do you usually eat it? (3) I usually cut if off {0) I usually eat it 8. (a) Do you (or does your wife) usually fry or grill food? (0) Usually fry (3) Usually grill {]) Sometimes fry. sometimes grill (O) Don't know (b) I~ ~ou. f~ food, what so~ of fat or oil is us~ally used? (I~ Cooking oil (score 0 if "usually grill" in (a)) (0) Lard or dripping (solid fat) (0) Don't know REFERENCES 1. Anderson, J.T.. Jacobs, O, R.. Foster,~ N.. Hall. Y.. Moss, D..Mojorinier. L.. a~d Blackbunt-H, Scoring systems for evaluating dietary pattern effect on serum_cholesteroL 525-5.':17 H979). 2. Grafnetter. D. World Health Organizalion co-ordinaled quality control in the lipid laborator3,'. Giorn. ArterioscL 2. 113-17'8 ( 19771. 3. Kahn. H. A.. Medalic. J. H.. Neufeld. H. N., e~ al. Serum cholesterol: its distribution and associ- ation with dietary and other variables in a survey of 10.1300 men. Isr. J. Med. Sci. 5. I 117-1127 (1969). • ~. Jacobs. D. R.. Anderson. J. T.. and Blackburn, H. Diet and serum cholesterol: Do zero correla- tions negate the relationship? Amer..L Epidemk~l. I10, 77-87 (19791. 5. Keen. H.. Thomas, B. J.. Jarrett. R. J.. and Fuller. J. H. Nutrient intake, obesity and diabetes. Brit. Med. J. 1. 655-658 (1979). 6. Rose. (3.. Heller. R. F.. Tunstall Pedoe, H. D.. and Christie. D. G. S. Heart Disease Prevention Project: A randomized controlled trial in industry. Brit. Meal. 2. 1,747-751 (1980) 7. Shaper. A. G.. and Mart. J. W. Dietary recommendations for the community towards the post- ponement of coronary heart disease. Br~t. Med. 3. 1,867-871 (1977~. Tt05280113