Organizational methods used in an epidemiological study in Leningrad*

Bulletin of the World Health Organization 59 (2): 281-284 (1981) Organizational methods used in an epidemiological study in Leningrad* A. G. KATRUSENKO1 & D. B. SESTOV 2 This paper describes the organization ofan epidemiological survey, carried out by the Lipid Centre ofthe Institute ofExperimentalMedicine in Leningrad, particularly the ways in which subjects were recruited. The different methods ofrecruitment resulted in a totalpopu- lation response of 78.14%. A time index is introduced to estimate the efficiency of each method ofrecruitment, and on the basis ofresultsfrom this study, a generalizedschemefor the planning of other surveys is proposed. The second half of this century has seen a marked increase in the prevalence of ischaemic heart disease (IHD) in developed countries. IHD affects younger population groups every year (1-4), and is often followed by serious complications, particularly in young men (5-6). In the last decade, epidemiological studies have been carried out in many countries in an attempt to examine the prevalence of IHD and to understand the factors that influence the development of the disease (7-10). One of the main problems associated with such studies is the efficient recruitment of sufficient numbers of volunteers. This paper describes the organization of one such project in Leningrad, and outlines a scheme for efficient distribution of work- load throughout the period of the study. STUDY METHODS A study on the prevalence of hyperlipoprotein- aemia and ischaemic heart disease was carried out in two lipid research clinics in Leningrad and Moscow, in the USSR, as part of the Soviet-American agreement in the field of cardiovascular diseases. The material presented here is based on the experience obtained in organizing the epidemiological study at the USSR Academy of Medical Sciences Institute of Experi- mental Medicine in Leningrad (11). The study group was chosen from 10 082 males born between 1916 and 1935, and resident in the Petrogradskij district of Leningrad. A random sample of 5000 subjects was taken, representing 50% of the * From the Atherosclerosis Department, Institute for Experi- mental Medicine, Leningrad, USSR 197002. 1 Senior Scientific Worker. 2 Chief, Laboratory of Epidemiology. population. The group was divided into 20 subpopu- lations of 250 persons each. The survey was conducted in two stages: 1) A primary screening of the whole sample; 2) A second detailed examination of 15% of the subjects who passed the first screening, plus all patients whose cholesterol or triglyceride levels were above certain age-specific limits (12). Three means of contact were used for recruitment: - up to three written invitations were sent by post to each person - subjects were contacted by telephone - subjects were visited at home. The letters were usually sent out once a week. Thus, for the first three weeks the recruitment work con- sisted of posting the invitations. The total number of letters posted per hundred persons invited was, on average, 230. If no response was obtained after three letters had been sent, the persons were telephoned up to 5 times during the third and fourth weeks. Those who could not be contacted by telephone were visited by the medical personnel. The recruitment and primary screening examin- ations were carried out between 1 February 1975 and 6 July 1977. The results of the data analysis are presented below. ANALYSIS The length of time available for the study was limited, and it was hoped to achieve an 80/o response rate within 2.5 years. In fact, 3907 persons (78.14%) were examined within 29 months, giving a response 40S6 -281- A. G. KATRUSENKO & D. B. SESTOV Table 1. Time, in days, required to achieve various levels of response for the best, average, and worst subpopulations Response level (percentage) Subpopulation 20 40 50 55 60 65 70 Best 3 9 15 16 23 27 43 Average 5 14 32 116 167 193 281 Worst 9 29 83 199 291 352 401 rate that is consistent with results from other epidemiological studies (3, 4, 7, 13-15). The duration of work with each subpopulation was between 90 and 511 days, and the response rate varied from 70.4% to 85.6%. There was therefore a con- siderable variation in the time necessary to achieve an equal level of response in different subpopulations. Although only 43 days were necessary to reach a 70/o response rate in the best case, up to 401 days were needed in the worst. Table 1 shows the time necessary to achieve various levels of response, for the best, average, and worst case. In order to evaluate the effectiveness of work with the population at different stages of the study, a time index was introduced. This index permits the calcu- lation of the time necessary for a further increase of % in the response rate at any given level of response. The time index is calculated by the following formula: T TTT- T,~Tl -p , where Ti, = the time index; Tu= the time necessary to achieve the upper response level, in days; T, = the time necessary to achieve the lower response level, in days; Pu= the upper level of response (%); Pu = the lower level of response (o). For example, to calculate the time index in the interval 55-65%, given that: T6e = 193 days, and T55 = 117 days, then substituting the values into the formula: 193 - 117 = 76 T55 5 = = 7.6 65 -55 10 Thus, about 8 days are necessary for each 1 % increase in response in the interval 55-65%. By this method, it is possible to compare the effectiveness of different methods of population recruitment, and also to develop an optimal system model for the epidemiological study. All the stages of work with the population were analysed using the time index. Calculation of recruit- ment time was based on the minimal total response rate achieved in all 20 subpopulations- namely 70%o. The work with the first thousand subjects proved to be the most economical in terms of response level obtained, time necessary for investigation, and specific recruitment time, and is considered close to optimal. A PROPOSED SCHEME OF RECRUITMENT On the basis of the completed experimental epi- demiological study, a scheme is presented for the organization of a survey of a population of 2500 men, aged 40-59 years. The scheme consists of a preparatory period of 30 days, followed by 20 days for posting the letters of invitation, and 15 days for contact by phone, while visits to the homes of the unresponsive part of the population are made throughout the remaining period of study. The total population is divided into 10 subpopu- lations, each consisting of 250 persons. An approxi- mate schedule of admission of the subpopulations to the survey was made (Table 2) on the basis of the average response shown in Table 1. In order to provide an evenly distributed work load for the medical staff the first posting consists of 500 invitations (2 subpopulations). The second one is made two weeks later (3rd subpopulation); the fourth, fifth, and sixth subpopulations are then invited at weekly intervals. There is then a two-week delay before the seventh subpopulation is invited, followed by the remaining groups at one week intervals. Such a recruitment schedule can provide a 50% population response within 14 weeks if the daily and weekly examination rates are strictly observed. The response rate is expected to be 14% of the population in the first month, 16% in the second, and 17% in the third. Thus, the rate of work increases as time goes on. After the fourteenth week, the main effort must be directed at the recruitment of the passive part of the population and at this stage, it is necessary to increase the number of personnel visiting the subjects at home. The recommended organization of the recruitment system allows the work with each subpopulation to be completed within 310 days, thus requiring 390 days for 282 ORGANIZATION OF AN EPIDEMIOLOGICAL STUDY Table 2. Suggested schedule for recruitment and examination of subjects in an epidemiological survey No. of Cumulative No. of persons examined per week persons total of invited persons invited 1 2 3 4 5 6 7 8 9 10 11 12 13 14 500 500 100 65 35 20 14 10 6 4 3 3 2 3 2 1 250 750 50 32 17 10 7 5 3 2 2 1 2 1 250 1000 50 32 17 10 7 5 3 2 2 1 2 250 1250 50 32 17 10 7 5 3 2 2 1 250 1500 50 32 17 10 7 5 3 2 2 250 1750 50 32 17 10 7 5 3 250 2000 50 32 17 10 7 5 250 2250 50 32 17 10 7 250 2500 50 32 17 10 Total no. screened Cumulative total Percentage of population 100 65 85 102 113 119 72 93 110 119 100 165 250 352 465 584 656 749 859 978 4 7 10 14 19 23 26 30 34 39 123 77 48 32 1101 1178 1226 1258 44 47 49 50 the complete survey. It is planned to use this scheme for a further epi- demiological study in Leningrad, to evaluate its applicability. DISCUSSION During the last two decades there have been many improvements in epidemiological techniques for the study of a number of non-communicable diseases in developed countries. Cardiovascular disease is a major health problem in these areas, and as such, has received a great deal of attention. In the USSR, the first studies of this type were carried out in the early 1960s, using objective methods of evaluation of the cardiovascular system, approved and recommended by the World Health Organization, and producing data comparable with those from studies in other countries (7, 10, 16, 17). The material presented here is concerned with an epidemiological study of a random sample of a section of the male population in an area of Leningrad. More than 50%o of all subjects were recruited by means of written invitations, and the response did not depend on the age of the subjects. Careful checking of each subpopulation and personal contact with the invited subjects ensured a high response rate. The recruitment of the passive group took 10 times as long as the recruitment of the active one, and concentrated efforts are needed with this group in order to increase the effectiveness and efficiency of such studies. In order to increase the motivation of potential subjects, the medical personnel working with the population must use all the available techniques of health edu- cation to emphasize the social significance of the project. An approximate operating schedule for the organiz- ation of an epidemiological study has been suggested. SUMt METHODES D'ORGANISATION APPLIQUEES A UNE -TUDE EPIDEMIOLOGIQUE EXtECUTtE A LENINGRAD Pendant les deux dernieres decennies, bien des ameliora- tions ont e apportees dans les pays developpes aux methodes appliquees aux &udes epidemiologiques de la prevalence de diverses maladies non transmissibles. Les affections cardiovasculaires, qui posent actuellement un probleme majeur dans ces pays, tiennent une place speciale dans ces etudes. En URSS, les premieres etudes de ce type ont et executees au debut des annees 1960, et on s'y est servi des methodes objectives d'evaluation du systeme cardio- vasculaire approuv&es et recommandees par l'OMS; les r& sultats obtenus ont ete comparables avec ceux qui avaient ete obtenus dans d'autres pays. 283 284 A. G. KATRUSENKO & D. B. SESTOV Le present article concerne une etude epidemiologique qui, portant sur un echantillon al6atoire de 5000 hommes nes entre 1916 et 1935, visait a determiner la prevalence des dyslipoprot6inemies et leurs rapports avec les cardiopathies ischemiques. L'echantillon a W divise en 20 sous-echantillons de 250 individus. Le recrutement s'etait fait par lettre et par contact personnel. Plus de 5001 des sujets ont ete recrutes au moyen d'invitations ecrites, et le taux de r6ponse s'est rev6le sans rapport avec l'fge des sujets. Un taux de reponse eleve a pu etre obtenu grfce a un contrble attentif de chacun des sous- echantillons et a des contacts personnels avec les sujets A leur domicile. Le recrutement de la fraction passive de la population enquetee a pris dix fois plus de temps que celui de la fraction active. Dans une grande mesure, une augmentation de l'effi- cacite et de l'economie de telles 6tudes depend d'une intensi- fication du travail sur la fraction passive de la population. Le personnel travaillant en contact avec la population doit se servir de toutes les techniques d'education sanitaire dis- ponibles pour renforcer la motivation des sujets et obtenir un taux de reponse maximal. On suggere, sur la base d'une analyse des donnees obtenues, un schema operationnel approximatif pour l'organisation d'une etude 6pid6miologique portant sur 2500 individus et qui peut etre achevee en 390 jours. REFERENCES 1. BOTIG, S. ET AL. Deutsche Gesundheitswesen, 25: 1048-1049 (1970). 2. WHO Chronicle, 23: 355-367 (1969). 3. GLAZUNOV, I. S. In: Epidemiologija arterialnoj giper- tonii i koronarnogo ateroskleroza, Moscow, Medicina, 1969, pp. 33-45. 4. FEIFAR, L. Acta cardiologica (Brussels), 15: 7-36 (1972). 5. BAUBINENE, A. V. & MIsJUNENE, N. B. In: Vnezapnaja smert'pri ostroj koronarnoj nedostatocnosti, Moscow, Medicina, 1968, pp. 12-14. 6. STAMLER, J. Medical clinics ofNorth America, 57: 5-46 (1973). 7. KANNEL, W. B. & FEINLEIB, M. American journal of cardiology, 29: 154-163 (1972). 8. PREJMATE, E. J. Kardiologiya, 12: 48-52 (1972). 9. RHOADs, G. G. ET AL. New England journal of medi- cine, 294: 293-297 (1976). 10. STROMBERG, J. ET AL. Social science and medicine, 8: 275-286 (1974). 11. Collaborative US-USSR study on the prevalence of dyslipoproteinemias and ischemic heart disease in American and Soviet populations. American journal of cardiology, 40: 260-268 (1977). 12. LIPOVECKIJ, B. M. ET AL. Kardiologija, 19: 71-75 (1979). 13. EPSTEIN, F. H. ET AL. Annals of internal medicine, 62: 1170-1187 (1965). 14. METELICA, B. I. & MAZUR, N. A. Epidemiologija i profilaktika isemiceskoj bolezni serdca, Moscow, Medicina, 1976, pp. 11-15. 15. METELICA, B. I. ET AL. Epidemiologija serdecno- sosudistyh zabolevanij, Moscow, Medicina, 1977, pp. 57-59. 16. ROSE, G. A. & BLACKBURN, H. Cardiovascular survey methods, Geneva, World Health Organization, 1968 (Monograph Series, No. 56). 17. Conference on the prevention and control of major cardiovascular disease, Brussels, 1973, Copenhagen, WHO Regional Office for Europe, 1974.