• No results found

ENROLLMENT RATES

In document IN TO (pagina 46-61)

HEALTH DIFFERENCES

2.2.5 ENROLLMENT RATES

. The data collection for the base-line measurement started in March 1991, following a publicity campaign in the local newspapers and other media. The postal questionnai-re was mailed in a personally addquestionnai-ressed envelope, accompanied by an introductory let-ter signed by the project leader and the director of the municipal public health service.

A stamped envelope was added to facilitate an easy response. Three reminders were sent:

after 1 week (a simple postcard), after 3 weeks (a letter with another copy of the postal questionnaire), and after 6 weeks (a very urgent letter), The design of this procedure was guided by Dillman IS recommendations29

Table I. The base~line measurement in the Longitudinal Study on Socio~Economic Health Diffe;ences (LS-SEHD)

Educational level of partner Family income

The response rates were quite satisfactory (Table 2). The over-all response rate was 70.1%, slightly lower than the expected 75% but still rather high for a postal question-naire. Differences in response rates between different subgroups of the sample were modest in size: women, elderly people, the better-off, and country-dwellers responded a little morc frequently than did their respective counterparts.

Table 2. Response rates: postal questionnaire

Numbers approached Numbers responding (abs) Numbers responding

[al [bl (%)

Gender

men 13.583 9.207 67.8

- women 13.487 9.766 72.4

Age

15·34 years 7.083 4.762 67.2

35·54 years 10.088 6.977 69.2

55·74 years 9.899 7.234 73.1

Post code group [el

1 (well-ta-do) 6.805 4.960 72.9

2 3.829 2.727 71.2

3 4.537 3.232 71.2

4 4.163 2.853 68.5

5 (deprived) 7.615 5.134 67.4

Degree of urbanization

1 (rural) 213 160 75.1

2 2.681 1.969 73.4

3 4.462 3.268 73.2

4 3.639 2.521 69.3

5 [big city) 16.075 11.055 68.8

Total 27.070 18.973 70.1

[al net sample. i.e total sample (n=27.278) minus:

questionnaires which were returned because the address was wrong (n=124); persons who had died (n=30); persons who were absent for a long time (noo18); nursing home residents (n=7); mentally handicapped (00=29).

[b] i.e. those who returned a completed questionnaire

[c) classification based on commercial post code segmentation data;

unknown for 121 persons in the net sample and for 67 responders respectively.

The data collection for the oral interview started in April 1991, and lasted until the end of June. A personal letter was sent to announce the interviewer, who visited the address a maximum of three times. The over-all response rate was 79.4% (Table 3), with

even smaller differences between subgroups of the sample than in the case of the postal questionnaire.

239 non-responders to the postal questionnaire were approached for a brief oral interview. Of these, 64 (26.8%) completed this interview.

Table 3. Response rates: oral interview

Numbers approached Numbers responding (abs) Numbers responding

(a] (b] (%)

Gender

men 1,718 1.372 79.9

- women 1.811 1.430 79.0

Age

15·34 years 912 732 80.3

35·54 years 1.295 1.033 79.8

55·74 years 1.322 1.037 78.4

Post code group (e]

1 (well-to-do) 981 796 81.1

2 507 412 81.3

3 591 469 79.4

4 452 348 77.0

5 (deprived) 981 764 77.9

Degree of urbanization

1 (rural) 27 22 81.5

2 335 277 82.7

3 597 480 80.4

4 476 394 82.8

5 (big city) 2.094 1.629 77.8

Total 3.529 2.802 79.4

[a] net sample, i.e total sample (n=3,637) minus: persons whose addresses were wrong (n= 18); persons who had moved (n=50); persons who were absent for a long time (n"'40).

Persons who had not sent back their postal questionnaire, but were selected accidently for the interview, are excluded from the sample.

[b] i.e. those who returned a complete questionnaire

[c} classification based on commercial post code segmentation data; unknown for 17 persons in the net sample and for 14 responders respectively.

2.2.6 DISCUSSION

The LS-SEHD is the first large-scale longitudinal study of socio-economic ine-qualities in health in the Netherlands. It represents a conscious attempt to translate recent insights and hypotheses on the possible causes of socia-economic inequ'alitics in health into an appropriate and cost-effective research design. The conceptual frame-work of the study (Figure 1) reflects the complexities of the phenomenon: causality is probably bidirectional, multiple factors are involved in the 'causation' mechanism, and the distribution of these ('letors across socia-economic groups is partly determined by circmnstances and experiences in early life. The usc of postal questionnaires and admi-nistrative data from public and health care administrations, in addition to the more conventional oral interviews, enabled us to combine a large sample size with an ade-quate data collection efron.

The response rate of the postal questionnaire llsed for the base-line measurement actually is higher than that obtained in large-scale oral interview procedures in the Netherlands: surveys of the Netherlands Central Bureau of Statistics, including the Health Interview Survey, currently have response rates of ca. 550/030As there is no reas-on to suppose that the validity of respreas-onses to postal questireas-onnaires is lower than that of responses to oral questionnaires:H, we believe that the data collection procedure adop-tcd for the LS-SEHD will prove to be a good choice.

A comparison of the design of the LS-SEHD with that of other studies investigating the explanation of socio-economic inequalities in health suggests some interesting simi-larities and differences32Table 4 summarizes the design of the 15-SEHD on the one hand, and that of a number of frequently cited British studies on the other hand.

The 'OPCS Longitudinal Study' is rightly famous for its tremendous contributions to the debate on socio-economic inequalities in health, especially mortality, both in the United Kingdom and internationally. Of the 4 British studies mentioned in Table 4, it is by far the largest in terms of sample size, and it is also much larger than the 15-SEHD. Its advantage in statistical power is, however, counterbalanced by the relatively limited number of variables on which information was collected in the base-line meas-urement (i.e. the 1971 census). Its stronghold therefore is description, not explanation.

The other 3 British studies mentioned in Table 4 clearly offer many more opportu-nities for explanatory analyses. The !National Survey of Health and Development' exemplifies the 3 birth cohort studies which are currently undenvay in the United Kingdom, and which permit extremely interesting analyses of life histories. This is important for the explanation of socia-economic inequalities in health, because the dis-tribution of risk factors across socio-ecOllOmic groups is mediated by factors which find their origin in early life (childhood environment, cultural factors, psychological fac-tors)(Figure 1). Birth cohort studies enable researchers to disentangle the time-order of events in these areas, and thereby provide insight into the causality of associations.

Table 4. A comparison benveen the Longitudinal Study ofSocio-Economic Health Differences (LS-SEHD) and selected other longitudinal studies of socio-economic inequalities in health

Starting year (t"'O)

20-07

In addition, the effect of health in childhood on 'intergenerational social mobility' can be studied in a prospective way. The LS-SEHD was not constructed to permit such analyses: we start with a cross-section of age-groups in the range 15-74 years. The objec-tives of the LS-SEHD are more closely comparable to that of the 'Whitehall (I)-study' and the 'West of Scotland 20-07-study' (Table 4), Data collection at the base-line measurement has been quite extensive in both studies, with an emphasis on biomedical measurements in the \X1hitehall-study and on social factors in the 20-07-study, The Whitehall-study's sample size is much larger than that of the 20-07-study, but it is res-tricted to men in the age-range 40-64 years, The 20-07 -study intends to document health effects of social factors in three distinct age-cohorts: those 15, 35 and 55 years at base-line respectively, The comparison in Table 4 shows that the LS-SEHD has the sample size of the \X1hitehall-study, but the emphasis on social factors of the 20-07 -study. A large sample size is necessary to detect socia-economic inequalities in the incidence of e.g. specific conditions Of mortality from the largest causes. \V'e did not focus on specific age-groups: perhaps the explanations of socio-economic inequalities in health differ between generations, but if they do, the sample sizes of the generations in the study would have to be quite large to detect sllch differences. On the other hand, a comprehensive analysis of the mechanisms and factors involved in the explanation of socia-economic inequalities in health requires an emphasis on social f.:1CtofS, as is also evident from the data collected in the 'Whitehall II-study''', Which does not imply that we would not have liked to include biomedical measurements, both to validate some of the self-reports in the LS-SEHD (e.g. on body mass index, on the prevalence of cluo-nic conditions) and to provide information which is impossible to obtain with question-naires (e.g. on serum cholesterol and blood pressure). The absence of such measure-ments is probably the main weakness of our study.

Although there are many differences between the LS-SEHD and the other studies mentioned in Table 4, as well as longitudinal studies carried out in other countries32, a comparison of the results of studies performed in different countries may still be worth-while. International comparisons of socio-economic inequalities in health have shown that the size of these inequalities differs between countries%-59. Actually, as these socie-ties differ in many respects, the contribution of different mechanisms and factors to ine-qualities in health is probably also different, A comparison of the results of different longitudinal studies offers interesting opportunities for an exploration of this issue.

NOTES

l.Power calculations were made for the raref outcome measures (incidence of speci-fic chronic conditions, cause-specispeci-fic mortality), using Breslow and Day's reference values"'. With a

=

0,05 (one-sided) and 1 - ~

=

0,80, a minimum of 300 new cases of ill health during follow-up is necessary to detect a Relative Risk of 1.5 of the lowest versus the highest socio-ecOIlOmic group, assuming 5 socia-economic groups of equal size and the use of a test for linear trend. As the number of person-years of follow-up necessary to find 300 flew cases heavily depends on the age-composition of the cohort at the start offollow-up, we decided to increase the number of 45-74 year olds at the expense of the 15-44 years olds (the study cohort was to have 65% of people in the age-range 45-74 years, as compared to 350/0 in this age-range in the Dutch population as a whole). Using national hospital admission rates by diagnosis we calculated that in order to find 300 new cases of ischemic heart disease 32,000 person-years of follow-up will be necessary (cerebrovascular disease: 83,000; lung cancer: 143,000; respiratory disease:

42,000).

Using national mortality rates by cause of death we calculated that in order to find 300 cases of death due to ischemic heart disease 99,000 person-years of follow-up will be necessary (cerebrovascular disease: 284,000; lung cancer: 228,000; respiratory disease: 416,000). \Ve therefore chose a cohort size at the start of follow-up of approxi-mately 20,000 persons: with 80-95% completeness of follow-up this should generate enough cases in 10 years time to study incidence or mortality for some specific conditions at least.

REFERENCES

I. Mackenbach)P. Socia-economic health difterences in rhe Netherlands: a review of recent empirical findings. Soc Sci Med 1992;34:213-26

2. Gunning-Schepers L), Spruit IP and Krijncn ]H. Socia-economic inequalities ill "ealth:

qlltstiom 011 trends lind expland/iom. Rijswijk: Ministerie van \Velzijn, Volksgezondheid en CUhUUf (the Netherlands), 1989

3. Mackenbach JP. Sodo-economic health differences: proceedings of fl symposium held 011 Febmmy 1st, 1991 ill Rotterdam, the Netherlands. Ministerie van W'eizijn, Volksgezondheid en

CUltUUf, Rijswijk (the Netherlands), 1991

4. Stronks K, Mheen H van de and Mackenbach JP. Achtergronden van sociaal-economische gezondheidsverschillcn: een overzicht van de literawur en cen onderzoeksmodel. (The background to socio-economic inequalities in health: a review of the literature and a concep-tual model) Tijdschr Soc Gezondheidsz 1993;71 :2-10 [in Dutch]

5. Townsend P and Davidson N. Inequalities in hettlth. The Black Report. Harmondsworth:

Penguin, 1989

6. Fox, AJ, Goldblatt PO and Adelstein AM. Selection and mortality differemials. J Epidemiol Community Health 1982;36:62-79

7. \'7ilkinson RG. Socio-economic diHerences in mortality: interpreting the data on their size and trends. In: Wilkinson RG (ed.). Clllss and health. London: Tavistock Publications, 1986 8. \'{fadsworth MEJ. Serious illness in childhood and its association with later-life achievement.

In: RG. \VIlkinson (cd), Clms Ilnd health. London: Tavistock Publications, 1986

9. Power C, Fogelman K & Fox AJ. Health and social mobility during the early years of life.

Qart! Soc AffI986;2:397-413

10. Lundberg O. elms positiolllll1d health: social CtlflSatiolJ or social selection? Stockholm:

University of Stockholm, 1988 [dissertation]

11. Marmot MG, Kogevinas M and Elston MA. Social/economic status and disease. Amz Rev Pub! Health 1987;8:111-135

12. Macintyre S. The patterning of health by social position in comemporary Britain: directions for sociological research. Soc Sci Med 1986;23:393-415

13. \'{Ihitehead M. The Health Divide. In: Townsend, P, Davidson N & \'{!hitehead M.

Inequalities ill hertlth. London: Penguin, 1988

14. Green G, Macintyre S, \'{fest P & Ecob R. Like parent like child? Associations between drin-king and smodrin-king behaviour of parems and their children. BrJ Addict 1991;86:745-58 15. Barker DJP & Osmond e. Inequalities in health in Britain: specific explanations in three

Lancashire towns. BM! 1987;294:749-752

16. Blaxter M. Social class and health inequalities. In: Carter CO & Peel J. Equalities lind ine-qUlllities in health. London: Academic Press, 1976

17. Lazarus RS & Folkman S. StreSi, appmisal and coping. New York: Springer, 1984 18. \'7est P. Rethinking the health selection explanation for health inequalities. Soc Sci A1ed

1990;32:373-84

19. Mascie~Taylor CGN & McManus Ie. Blood group and socio-economic class. Nature 1984;309:395-6 [Commentary]

20. Power C, Manor 0 & Fox J. Health and clllSi: the ertrly yean. London: Chapman & Hall, 1991

21. Yelin EH, Kramer JS & Epstein \YfV. Is health care use equivalent across social groups?

A diagnosis-based study. Am! Pub! Hlth 1983;73:563-71

22. Mackenbach JP, Stronks K & Kunst AE. The contribution of medical care to inequalities in health: differences between socio-economic groups in decline of mortality from conditions amenable to medical intervention. Soc Sci Med 1989;29:369-76

23. Berkel-van Schaik AB van and Tax B. Mlflr een stallddartioperatioJlalisatie van sociaal-ecoIJo-mische stattlS tloor epidemiologirch en sociaal-medisch onderzoek. (Towards a standardized operationalisation of socio-economic status in epidemiological and social-medical research) Rijswijk: Ministerie van \Velzijn, Volksgezondheid en Cultuur (the Netherlands), 1990 [in Dutch]

24. Centraal Bureau yoor de Statistiek. Netherlands Health Interview SII1'lley 1981-1985. The Hague: Staatsuitgeverij, 1988

25. Dirken JM. Het meten vrtll stress in illdllStrifle sitl/atiff. (Measurement of stress in industrial settings) Groningen: \Volters, 1967 [in Dutch]

26. Hunt SM, McEwen J & McKenna SP. Measuring health stlltllS. L:lOdon: Chroom Helm, 1986

27. Essink-Bot ML, Van Agt HME & Bonsel GJ. NHP of SIP: cell vergelijkend onderzoek onder chronisch zieken. (NHP or SIP: a comparative research among chronically ill people)

Tijdscbr Soc CeZIJlldbeid,z 1992;70: 152-9 [in Dutch]

28. Mc\'{1hinnie JR. Disdbility indiClitor for medsl/ring well-being. The OECD social indicators development programme. Special studies no.5. Paris: OEeD, 1979

29. Dillman DA. Mail aud telephone surveys: the total design method. New York: \Viley, 1978 30. Central Bureau of Statistics. Netherlands Hmlth Interview Survey 1981-1991. The Hague:

SDU-publishers, 1992.

31. O'Toole BI, Battistutta D. Long A & Crouch K. A comparison of costs and data quality of three health survey methods: mail, telephone and personal home interview. Am] Epidemiol 1986;124:317-28.

32. Mheen H. van de and Mackenbach ].P. Een overzicht valllongitudint1t11 ollderzoek na((r soci-lIal-eco1Jomische gfzol1dheidsverschillen. (A review of longitudinal research on socio-economic inequalities in health) Rijswijk: Ministerie van \Veizijn, Volksgezondheid en Cultuur (the Netherlands), 1990 [in Dutch]

33. Blaxter M. Longitudinal studies in Britain relevant to inequalities in health. In: \Vilkinson RG. (ed.). Clms lind Health. London: Tavistock Publications, 1986

34. Douglas J\'78. Health and survival of infants in different social classes. Lancet 1951;2:440-6 35. Atkins E, Cherry N, Douglas )WB, Kierman KE, Wadsworth ME). The 1946 British

cohort: an account of the origins. progress and results of the NSHD. In: Mednick SA, Baert AE (ed.). Prospective Longitudinal Reasearch. Oxford: Oxford University Press, 1981 36. \VadS'tvorth ME]. Follow-up of the first national birth cohort: findings from the Medical

Research Council National Survey of Health and Development. Paediatr Perinat Epidemiol 1987;1:95-117

37. Kuh D, \Vadsworth M. Parental Height: Childhood Environment and Subsequent Adult Height in a National Birth Cohort. Int] Epidemio/1989;18:663-8.

38. Reid DD, Breitt GZ, Hamilton P]S, Jarrett RJ, Keen M, Rose G. Cardiorespiratory disease and diabetes among middle-aged male civil servants. Lancet 1974;1:469-73

39. Rose G, Hamilton PJS, Keen H, Reid DO, McCartney Po Jarrett RJ. Myocardial ischaemic risk factors and death from coronary heart disease. Ltlncet 1977;1:105-9

40. Marmot MG, Rose G, Shipley M, Hamilton PJS. Employment grade and coronary heart disease in British civil servants.] Epidemiol Community Health 1978;32:244-9

41. Rose G, Marmot MG. Social class and coronary heart disease. Br Heart J 1981;45:13-9 42. Marmot MG. Shipley MJ, Rose G. Inequalities in death-specific explanations or a general

pattern. Lallcet 1984; 1: 1003-6

43. Fox AJ, Goldblatt PO. OPCS-Longitudinal Study: socia-demographic mortality difl-eren-tials. Series IS JlO. 1. HMSO. London, 1982

44. Moser KA, Fox AJ, Jones DR. Unemployment and mortality in the OPCS Longitudinal Study. In: \Vtlkinson RG (ed.). elms lind Health. London: Tavistock Publications. 1986 45. Moser KA, Goldblatt PO, Fox AJ, Jones DR. Unemployment and mortality: comparison of

the 1971 and 1981 longitudinal study census samples. BM] 1987;294:86-90

46. Barker R, Roberts H. Social clmsifimtion scheme for women. LS \"'{lorking paper no. 51. Social London: Statistics Research Unit City University, 1987

47. \Vest P. 7'llf study o/)'outh and health. \Vest of Scotland 20-07 Study \'\forking paper no. 2.

Glasgow: Medical Research Council Medical Sociology Unit, 1986

48. Ecob R. An evalttt1tioll o/possible stlmplingfim}les for all age defined cohort study. \'{fest of Scotland 20-07 Study \"'{lorking paper no. 5. Glasgow: Medical Research Council Medical Sociology Unit, 1987

49. Ecob R. The St1JJlplillg Scheme, Fmme and Procedures for the Cohort Studies. \Vest of Scotland 20-07 Study \"'{lorking paper no. 6. Glasgow: Medical Research Council Medical Sociology Unit, 1987

50. Macintyre S. Hftllth ill the community. The S1frvey's background & mtiol1ale. \"'{lest of Scotland 20-07 Study \"'{lorking paper no. 7. Glasgow: Medical Research Council Medical Sociology Unit, 1987

51. Annandale E. Qllt1lity cOlltrol tape recording exercise. \Vest of Scotland 20-07 study working paper no. 9. Glasgow: Medical Research Council Medical Sociology Unit. 1987

52. Macintyre S, Annandale E. Ecob R, et al. The \"'{lest of Scotland 20-07 Study: healdl in the community. In: Martin c.J .• McQueen D.V. (ed.). Readings for anew public health.

Edinburgh: Edinburgh University Press, 1989

53. \Vest P, MacIntyre 5, Annandale E, Hunt K. Social class and health in youth: findings form the west of Scotland 20-07 study. Soc Sci 1I1ed 1990;30:665-73

54. Macintyre S, Sooman A. A comparison of health measures between urban localities in the Wi'st of Scotland. \Vest of Scotland 20-07 Study \'{forking paper no. 30. Glasgow: Medical Research Council Medical Sociology Unit, 1991

55. Marmot MG, Davey Smith G, Stansfield S, et al. Health inequalities among British civil servants: the \'\fhitehall II Study. The Ltlncet 1991;337:1387-93

56. Leclerc A, Left F, and Fabien C. Differential mortality: some comparisons between England and \Vales, Finland and France, based on inequality measures. JntJ Epidemiol

1990; 19: 1 00 1-1 0

57. Lahelma E, Valkonen T. Health and social inequities in Finland and elsewhere. Soc Sci Med 1990;31:257-65

58. Kunst AE, Mackenbach JI~ An international comptlrison ofsocio-economic inequalities in mor-tality. Rotterdam: Department of Public Health and Social Medicine (Erasmus University Rotterdam), 1992

59. Kunst AE, Geurts J. Berg J van den. international/!flriatioll ill socio-economic inequalities in se/freported benlth. The Hague: SOU-publishers, 1992

60. Breslow NE, Day NE. Statistical methods ill Cflnar research, vol.2. Tbe desiglland I1l1alysis of cohort studies. IARC Scientific Publications no. 82. Lyon: IARC. 1987

FOLLOW-UP,

DESIGN AND ENROLLMENT RATES

In document IN TO (pagina 46-61)