Journal of Epidemiology and Community Health 1997,51:67-73

Socioeconomic variations in hysterectomy:evidence from a linkage study of the Finnishhospital discharge register and populationcensus

Riita Luoto, Ilmo Keskimaiki, Antti Reunanen

AbstractStudy objective - To explore variations inrates for hysterectomy in relation to socialclass, education, and family income.Design - Retrospective analysis ofthe 1988Finnish hospital discharge register linkedindividually to the 1987 population census.Setting - Finland.Participants - All women living in Finlandaged 35 and over were the denominatorpopulation. The numerators were the 8663women who underwent hysterectomy in1988.Main results - The overall rate for hys-terectomy was 63.5/10 000 women aged 35and over. There was a marked positivecorrelation between disposable family in-come and hysterectomy rates even afterage, hospital catchment area, education,and occupational status were adjusted for.However, no linear trend for overall hys-terectomy rates was observed in relationto social class or education. Proceduresdue to myomas, accounting for 48% of allhysterectomies, were more frequentamong women of high socioeconomic sta-tus according to all socioeconomic in-dicators. Larger proportions of hyster-ectomies for myoma were also performedin patients in private hospitals and in paybeds in public hospitals than in women inworse off groups.Conclusions - Unlike the findings in earlierstudies from other countries, there was apositive correlation between income andhysterectomy rates as a result of the highnumbers of hysterectomies performed totreat myoma in the well off women. Thefindings are discussed in terms of socio-economic differences in the use of privategynaecological services, and factors, suchas parity and use ofhormonal replacementtherapy, that affect the growth ofmyomas.

(J7 Epidemiol Community Health 1997;51:67-73)

Hysterectomy, removal of the uterus, is one ofthe most common operations performed inwomen. In Finland, as well as in Great Britain,every fifth woman aged 45 to 64 years hasundergone hysterectomy.'21 However, clinicalcriteria for hysterectomy are still under dis-cussion. Several studies have shown wide

regional variations in the rates of hysterectomyboth internationally and nationally.45 A recentstudy6 on surgery rates in Denmark, Finland,Norway, and Sweden showed a greater thantwofold variation in hysterectomy rates betweenthese, in many respects similar, countries. Inthis study, the highest national hysterectomyrate was in Finland.6 At the same time, thevariation within the countries themselves wasfrom 2.6 fold in Denmark to over 10 fold inNorway.

In addition to regional variations, there aresociodemographic differences in the oc-currence ofhysterectomy (table 1). In previousstudies,7" 13-16 education has most often beenconsidered as an indicator of social status. Themain trend in those studies has been the higherprevalence of hysterectomy in lower edu-cational and income groups. The same wasalso true in the single earlier study concerningFinland, in which the proportion of women inthe highest education group who had under-gone hysterectomy was 13% and that in thelowest educated group was 22%.2 In thisstudy,2 women in the lower white collar grouphad a higher proportion of hysterectomy thanthose in the upper white collar group, butthe difference was not statistically significant.However, in previous studies there may havebeen methodological oversights, as the resultswere mostly based on past data for hys-terectomies or information on socioeconomicstatus was based on geographical proxies, noton individual data.We aimed to evaluate socioeconomic pat-

terns in clinical practices concerning hys-terectomy and to assess differences in accessto this operation. In this study we exploresocioeconomic variations in rates for hys-terectomy in Finland in relation to the in-dications, surgical approach, occupationalstatus, education, and family income. By usingan administrative hospital database individuallylinked to population census files we have triedto avoid the methodological problems of earlierstudies.

MethodsThis study was based on all 8663 hys-terectomies performed in 1988 in Finnish hos-pitals in women aged 35 or more. The data onhysterectomies were obtained from the Finnishhospital discharge register, which contains sum-

National Public HealthInstitute, Departmentof Epidemiology andHealth Promotion,Mannerheimintie 166,FIN-00300 Helsinki,Finland.R Luoto

National Research andDevelopment Centrefor Welfare andHealth, HealthServices ResearchUnit, Helsinid,Finland.I Keskimaki

National Public HealthInstitute, Departmentof Health andDisability, Helsinki,FinlandA Reunanen

Correspondence to:Dr R Luoto.Accepted for publicationJune 1996

67

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jech.bmj.com

/J E

pidemiol C

omm

unity Health: first published as 10.1136/jech.51.1.67 on 1 F

ebruary 1997. Dow

nloaded from

Luoto, Keskimaki, Reunanen

Table 1 Socioeconomic variation in hysterectomy - findings from previous studies

Reference (y) Sample Origin of sample Indicator of Adjusting vanrables Findingsize socioeconomic status

Bunker7 (1974) 7400 Selected professionals, Groups of selected Spouses of male physicians beforeCalifornia professionals and their hysterectomy 22.6 (1.3)%, spouses of

spouses professionals 17.3 (1.4)%, p<0.01Koepsell8 (1980) 1087 Random, two Washington Education Age, sociodemographic and Highest income quartile 28% before

counties Income on census reproductive factors, health hysterectomy, lowest 49%, p<0.001tract level conditions, health practices

Coulter' (1985) 4120 Random, Oxford Occupation Age, parity No social class difference, if privateEducation surgery included

Age at completion of education < 16 ybefore hysterectomy 19%, >19 y 14%

Meilahn"' (1989) 2137 Random, driver's license Education Age at birth of first child, Highest educated odds of priorlists, Pittsburgh age at menarche, education hysterectomy 0.69 (95% CI 0.56,

0.86) v lowest educated (ref)Schofield" (1991) 8896 Random, North South Education 83% of those who had hysterectomy

Wales, Australia Husband's were not educated beyond middleoccupational status high school. Husbands of 63% of

women in the lower occupationalstatus, 18% in the higher levels

Gittelsohn'2 (1991) 28891 Hospital discharge data on Income Age, area Lowest income group odds ofMaryland, income data hysterectomy 1.2 v highest incomefrom census group

Luoto2 (1992) 1713 Random, Finland Occupation Age Highest educated before hysterectomyEducation 13%, lowest educated 22%, p<0.001.

Vessey'3 (1992) 17032 Selected, British married Husband's social class Parity, indication for Significant trend towards higher rateswomen using the pill or an hysterectomy for hysterectomies for menstrualIUD as contraception problems and non-significant trend

towards lower rates for fibroids andendometriosis in lower social classes

Santow" (1992) 2547 Random, Australia Education Age, parity, reproductive University educated odds of priorfactors, education, area hysterectomy 0.58 v lower educated

(ref)Kjerulff's (1993) 12465 Random, 15 states in USA Education Age Lowest income quartile odds of prior

Income hysterectomy 1.6 (95%CI 1.2, 2.0) vhighest (ref). Lowest educated odds ofprior hysterectomy 2.1 (95% CI 1.7,2.7) v highest (ref)

Kuh'6 (1995) 1628 Random, Britain Education Age, parity Highest educated women had lowestOwn and husband's risk for hysterectomysocial class

mary information on each inpatient episodefrom all public and private hospitals since1969.'7 Diagnoses and dates of hospital ad-mission and discharge have been included inthe register since its inception, but recordingof surgical procedures started in 1986. Theregister also contains data on patients treatedas private (pay bed) patients in public hospitals.The reliability of the Finnish hospital dis-

charge register has been considered to be good.It has been estimated that the register covers90-95% of all surgical procedures, and prin-cipal diagnoses are identical with the diagnosisin the medical records in 95% of discharges. 18The discharge register data on hys-

terectomies were linked by the patients' per-sonal identification numbers with the 1987population census in order to obtain data on

the patients' occupation, education, familystructure, and income. The census was alsoused to derive data on populations at risk. Inorder to obtain the data on occupational statusfor retired women, the 1987 census files were

supplemented with the data on occupation be-fore retirement in the 1970, 1975, 1980, and1985 censuses. Women were classified ac-

cording to their own occupation or, if they werenot employed, their husband's occupation.

Social class was defined according to oc-

cupational group'9 as follows:1. Upper white collar employees: upper leveladministrative, managerial or professional em-

ployees,2. Lower white collar employees: lower leveladministrative or clerical employees,3. Blue collar workers: skilled or unskilled man-ual workers,

4. Farmers: farmer employers or farmers on

own account,5. Others: employers or self employed workers,students, and those with unknown occupation.

Education was divided into three categories:high, medium, and low. People with a universitydegree or a qualification requiring 13 yearsor more of education were included in higheducation group. Medium education was de-fined as secondary school and vocational train-ing, or high school matriculation requiring10-12 years of education. Low education cor-

responds to a basic education of fewer than 10years.The data on disposable family income were

based on the registries of taxation and welfarebenefits. Income was adjusted for family sizeusing the OECD equivalence scale, where thefirst adult of a family is weighed as 1.0, otheradults 0.7, and children less than 18 years old0.5.20

Since 1986, surgical procedures in the Finn-ish hospital discharge register have been codedaccording to the classification of procedures bythe Finnish Hospital League.2' In the clas-sification, hysterectomies were classified in re-

lation to the surgical approach as follows:1. Subtotal hysterectomy,2. Hysterectomy without oophorectomy, hys-terectomy with unilateral oophorectomy,3. Hysterectomy with bilateral oophorectomy,or4. Vaginal hysterectomy.The combination of subtotal hysterectomy andbilateral oophorectomy is extremely rare, andtherefore all subtotal hysterectomies are

summed together.

68

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jech.bmj.com

/J E

pidemiol C

omm

unity Health: first published as 10.1136/jech.51.1.67 on 1 F

ebruary 1997. Dow

nloaded from

Socioeconomic variations in hysterectomy

I+1I+1

Upper Lower Blue Farmerwhite white collarcollar collar

Social class

I

High Medium Basic

Education

I+1 1+ ±

Highest 2nd 3rd 4th Lowest

Disposable income

Figure 1 Age adjusted rates for hysterectomy in relation to social class, education, and disposable income.

Table 2 Hysterectomy rates per 1O00 women aged 35 or more adjusted for age in relation to socioeconomic status andindication for surgery, Finland 1988

Indications for hysterectomy

Myoma Endometriosis Prolapse Disorders of Other benign Malign Totalmenstruation indications indications

Socioeconomic statusRate (%) Rate (%) Rate (%) Rate (%) Rate (%) Rate (%) Rate

Social class:Upper white collar 33.6 (54) 5.5 (9) 5.7 (9) 3.4 (5) 10.1 (16) 3.6 (6) 61.9Lower white collar 33.6 (50) 7.5 (11) 5.0 (7) 4.5 (7) 11.7 (17) 4.7 (7) 67.0Blue collar 28.9 (44) 6.8 (10) 5.7 (9) 6.6 (10) 14.2 (21) 4.1 (6) 66.3Farmers 25.0 (42) 4.6 (8) 7.6 (13) 6.3 (11) 12.2 (20) 3.8 (6) 59.4

Education:High 33.6 (54) 5.6 (9) 5.1 (8) 2.9 (5) 11.0 (18) 4.2 (7) 62.5Medium 21.4 (47) 6.9 (10) 6.9 (10) 5.0 (8) 12.3 (18) 4.6 (7) 67.0Basic 29.7 (46) 7.1 (11) 5.3 (8) 5.2 (8) 12.7 (20) 3.9 (6) 64.0

Disposable income:Highest 37.6 (54) 7.0 (10) 5.3 (8) 3.6 (5) 11.7 (17) 4.4 (6) 69.52nd 32.8 (48) 7.4 (11) 6.2 (9) 5.2 (8) 12.3 (18) 4.5 (7) 68.43rd 28.9 (43) 6.8 (10) 6.2 (9) 6.1 (9) 14.1 (21) 4.3 (7) 66.44th 25.9 (43) 6.1 (10) 6.2 (10) 5.8 (10) 12.1 (20) 4.0 (7) 60.1Lowest 19.4 (38) 5.5 (11) 5.5 (11) 4.4 (9) 11.3 (22) 4.1 (8) 50.4

Total 30.2 (48) 6.7 (11) 5.5 (9) 4.9 (8) 11.9 (19) 4.2 (7) 63.5

In 1988, diagnoses in the register were codedby the International Classification of Diseases,9th revision (ICD-9). The categories of in-dications associated with hysterectomy in thestudy were benign neoplasms, especially my-omas (ICD-9: 218), other fibroids (ICD-9:219-221), other disorders of the female genitaltract including endometriosis (ICD-9: 617),genital prolapse (ICD-9: 618), disorders ofmenstruation (ICD-9: 626), genital infections(ICD-9: 614-616), and malign neoplasms(ICD-9: 179-183).The statistical analysis was based on mul-

tidimensional tabulations of the data on hys-terectomies and the population at risk inrelation to age, socioeconomic indicators, andhospital catchment areas. The methods for de-fining hospital catchment areas are describedin detail by Keskimaki et al.5 Rates of hys-terectomy and 95% confidence intervals (95%CI) were related to the size of the populationin each age group. Direct age adjustment was

performed by using all Finnish women as thestandard population. Differences in the in-cidence of hysterectomy in relation to socio-economic position were estimated throughPoisson regression models using the GLIMprogram package.22 Results were adjusted for

age and hospital catchment area because of theover threefold regional differences in hys-terectomy rates in Finland.

ResultsThe total 1988 rate for hysterectomy in Finlandwas 63.5 per 10000 women. The rate washighest for women with the highest family in-come. Relative socioeconomic differences werealso largest across income groups. Accordingto social class and education, hysterectomyrates were highest among lower white collarwomen, blue collar women, or those who hadreceived a medium education (figure 1).The most frequent indication for hys-

terectomy was myoma, which accounted fornearly half of all the operations (table 2). Therates for myomas were higher among womenwith high socioeconomic status according to allthree socioeconomic indicators - the differencewas even twofold across income groups. Hys-terectomies due to disorders of menstruationtended to be more common in lower socio-economic groups. Prolapse was more prevalentand endometriosis less prevalent among farm-ers compared with other groups. For malign

75

70 _

Ki

ca)E 650

o 600a 5Co

a) 554-1

cz

50

45

69

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jech.bmj.com

/J E

pidemiol C

omm

unity Health: first published as 10.1136/jech.51.1.67 on 1 F

ebruary 1997. Dow

nloaded from

Luoto, Keskimaki, Reunanen

Table 3 Hysterectomy rates per 10 000 women aged 35 or more for hysterectomy due to benign indications adjusted forage in relation to socioeconomic status and surgical approach, Finland 1988

Surgical approaches

Hysterectomy Hysterectomyand unilateral Subtotal and bilateral Vaginal Otheroophorectomy hysterectomy oophorectomy hysterectomy approaches Total

Socioeconomic status Rate (%) Rate (%) Rate (%) Rate (%/o) Rate (%) Rate

Social class:Upper white collar 27.9 (48) 11.3 (19) 12.9 (22) 5.6 (10) 0.5 (1) 58.2Lower white collar 30.8 (49) 12.7 (20) 13.5 (22) 5.0 (8) 0.3 (1) 62.3Blue collar 27.7 (45) 15.2 (24) 13.3 (21) 5.7 (9) 0.3 (1) 62.2Farmer 19.1 (34) 15.5 (28) 14.0 (25) 6.7 (12) 0.3 (1) 55.6

Education:High 28.9 (50) 10.1 (17) 13.8 (24) 4.8 (8) 0.6 (1) 58.2Medium 28.9 (46) 13.0 (21) 13.2 (21) 7.0 (11) 0.3 (1) 62.5Basic 27.7 (46) 14.1 (23) 12.9 (21) 5.1 (9) 0.3 (1) 60.1

Disposable income:Highest 31.7 (49) 13.6 (21) 14.3 (22) 5.2 (8) 0.3(-) 65.12nd 29.2 (46) 13.5 (21) 14.8 (23) 6.0 (9) 0.3 (1) 63.93rd 29.0 (47) 14.0 (23) 12.6 (20) 6.1 (10) 0.4 (1) 62.04th 25.2(45) 12.7(23) 11.7(21) 6.1 (11) 0.3(1) 56.2Lowest 18.8 (41) 10.7 (23) 10.9 (23) 5.4 (12) 0.6 (1) 46.2

Total 27.6 (47) 13.1 (22) 12.7 (21) 5.5 (9) 0.4 (1) 59.3

.- 1: |-.1:"1 I.!:

.'

Figure 2 Rates for hysterectomy due to myoma performed in patients in private hospitals and pay beds and patients inpublic hospitals in relation to social class, education, and disposable income.

indications, there were no systematic ratedifferences across socioeconomic groups.The average age at the time of hysterectomy

for all operations was 48.3 (95%CI 48.1, 48.5)but it differed between indications and surgicalmethods. The age at hysterectomy was sig-nificantly higher among women who had ahysterectomy due to prolapse compared withthose who had a hysterectomy for other in-dications. Correspondingly, vaginal hys-terectomies (mean age 62.0, 95% CI 61.1,62.8), commonly undertaken for uterine pro-lapse in Finland, were performed for sig-nificantly older women than subtotalhysterectomies (mean age 45.2, 95% CI 44.8,45.6) or simple hysterectomies (mean age 43.1,95% CI 42.9, 43.2).Women in the farmer group (mean age 55.3,

95% CI 54.3, 56.2), with a basic education(mean age 50.7, 95% CI 50.4, 51.0) or thoseamong the two lowest income quintiles (meanage 52.9, 95% CI 52.2, 53.6 in the lowest and51.6, 95% CI 51.0, 52.2 in the 4th quintile)were significantly older when they underwenthysterectomy than women in the other groups.However, this was mainly due to the age struc-ture of the population at risk in these groups.

There were no significant differences in themean ages at hysterectomy in indication groupsin relation to education or occupation, exceptfor farmers whose age at hysterectomy washigher than in the other groups for all in-dications. For all indications, women in thehighest income category underwent hys-terectomy at a younger age than women in theother income categories.

Forty seven per cent of hysterectomies wereperformed for benign indications with uni-lateral oophorectomy or no oophorectomy,21% were hysterectomies with bilateral oophor-ectomy, 9% were vaginal hysterectomies, and22% were subtotal hysterectomies (table 3).Age adjusted social class differences were lar-gest for hysterectomies with unilateral oophor-ectomy or no oophorectomy. With regard toincome, those in the highest quintile had thehighest age adjusted rate for hysterectomy forall surgical approaches except vaginal hys-terectomy. Educational differences in respectof surgical approaches were small and con-tradictory.The risk for hysterectomy increased linearly

in relation to relative income. Adjustment forage and hospital catchment area by Poisson

70

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jech.bmj.com

/J E

pidemiol C

omm

unity Health: first published as 10.1136/jech.51.1.67 on 1 F

ebruary 1997. Dow

nloaded from

Socioeconomic variations in hysterectomy

regression did not change these relative socio-economic differences. Nor did the subsequentadditions of education and social class to themodel modify the income gradient for hys-terectomy rates. However, adjustment for in-come increased the relative rates amongwomenwith low education or occupational status.The proportion ofhysterectomies performed

in private hospitals was small - less than 1%.On the other hand, 18% of the hysterectomieswere performed in senior physicians' private(pay bed) patients in public hospitals. Figure2 shows separately the rates for hysterectomiesdue to myoma performed for regular patientsand patients treated in pay beds or in privatehospitals. According to all socioeconomic in-dicators, the higher the socioeconomic status,the larger the proportion of operations per-formed as private procedures. The findingswere similar for other benign indications.

DiscussionAccording to our study, Finnish women witha high disposable family income were morelikely to undergo hysterectomy than those witha low income, even after age, hospital catch-ment area, education, and occupational statuswere adjusted for. However, women with highstatus according to social class or educationhad lower hysterectomy rates than lower whitecollar or blue collar women, and those withmedium or basic education.The positive correlation of operation rates

and income was evident regarding proceduresdue to myomas, which accounted for half ofthe indications for hysterectomies. Social classand educational gradients in myoma operationsalso favoured the best off women. For otherindications, the differences between socio-economic groups were less systematic. Therates for procedures undertaken for disordersofmenstruation were highest among blue collarwomen and women with basic education, butwith regard to income levels these procedureswere most frequent in those with an inter-mediate income. For malign indications, therewere no systematic socioeconomic gradients.With regard to endometriosis, the gradientswere more or less opposite for educational andincome groups, and for social classes no cleartrend existed.For surgical approaches, our study did not

reveal any systematic socioeconomic differ-ences which were unrelated to indications forhysterectomy. For example, women with afarming background had a high rate of vaginalprocedures. In Finland this is a common inter-vention for treating uterine prolapse, a frequentindication among farming women, who tendedto be significantly older than women from othersocioeconomic groups.Our findings on trends in hysterectomy rates

according to different, highly correlated socio-economic indicators were partly contradictory.In particular, our findings on the dependencyof hysterectomies on family income were op-posite to findings from many earlier studies(table 1). In evaluation of these ambiguitiesseveral factors should be considered.

The adequacy of socioeconomic indicatorsin terms of social stratification may be onereason for the contradictory results. The defin-ition for family income in the study makesallowances for spouse's income, number ofchildren, taxes, and welfare benefits. It maytherefore take some sources of bias in socio-economic classification into account moreeffectively than other indicators. If a womanwas employed, her social class was classifiedaccording to her own occupation. However, insome cases husbands' occupations may reflectthe living conditions and social status offamiliesmore adequately. The value of education as asocioeconomic indicator is strongly dependenton age; in the older age groups there are veryfew women with higher education.23

In comparing our results regarding incomeand hysterectomy with those in previous stud-ies, problems of incompatibility are met.Different definitions of income may be an ex-planation for differences in our results andthose from other countries. Unlike our study,which used the quintiles of disposable familyincome, earlier studies often related hys-terectomies to the absolute family income level(table 1).Most earlier studies have been either eco-

logical studies or population surveys. The eco-logical studies using geographical proxies forsocioeconomic status are susceptible to the"ecological fallacy", and without a rather largesample size, population surveys based on thehistory of hysterectomies do not allow the as-sessment of present socioeconomic differencesin surgery rates. Population surveys also haveproblems related to the time lag between hys-terectomy and reported income, education, oroccupational status.'5 Education may be morestable, but occupational status and family in-come may change considerably over 10-20years. Consequently, it is not, for instance,possible to compare in any useful way ourresults with those of the earlier Finnish popu-lation survey2 on the prevalence of hys-terectomy.The most conspicuous finding, which also

explains the high hysterectomy rates in well offwomen in our study and may partly contributeto the high national hysterectomy rate inFinland compared with rates in similar coun-tries,6 is the high rate for myoma operations inwomen of high social status.There are several potential explanations for

this. Our results suggest that private medicinemay create easier access to gynaecological pro-cedures for the better off. In practice, Finlandhas a two tier health care system, in which thebetter offoften consult gynaecologists privately,but those of low social status usually visit gen-eral practitioners in municipal health centres.Most private specialist services are provided bypart time private practitioners who have fulltime jobs in hospitals. Private specialists canrefer their patients to their own hospitals, eitheras pay bed (private) or regular patients. Ac-cording to one survey, two thirds of pay bedpatients were referred by the attending hospitaldoctors.24 This situation may create a financialincentive for gynaecologists who are simul-

71

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jech.bmj.com

/J E

pidemiol C

omm

unity Health: first published as 10.1136/jech.51.1.67 on 1 F

ebruary 1997. Dow

nloaded from

Luoto, Keskimnki, Reunanen

taneously practising privately and in publichospitals to recommend hysterectomy forwomen willing to be treated at a pay bed evenif a more conservative treatment is an option.It has been suggested that private medicinemay create socioeconomic disparities similar tothose seen in hysterectomy for other pro-cedures, such as cataract operations and pro-statectomies,25 and this may also explain thehigh rates for surgical treatment of varicoseveins in the well off and urban women inFinland.26 However, their better access doesnot necessarily explain the specific finding onhigh rates for myoma operations in women ofhigh social status.With regard to uterine diseases, we have

morbidity data only for the incidence of andmortality from malignant conditions of the fe-male genital tract.2728 However, for myomaswe have no data on morbidity in relation tosocioeconomic group. Nor do we have data onpatients' treatment preferences, which are, inaddition to morbidity, important for operationdecisions in discretionary surgery like myomaoperations. Thus, the effect of these factors onsocioeconomic differences cannot be evaluated.One potential explanation for the high num-

bers of myoma operations in the better offwomen is that private gynaecologists may bemore competent to diagnose uterine myomasthan GPs in municipal health centres. Womenof higher social status are also likely to receivegynaecological examinations more often andregularly than the worse off. Consequently, theprobability of myomas being diagnosed couldbe higher among women of higher status.On the other hand the worse off, who usually

consult GPs, may be admitted to hospital withmenstrual disorders, a frequent symptom ofmyomas. In many cases, this admission dia-gnosis is subsequently recorded as a treatmentdiagnosis in the discharge register. The com-bined analysis of hysterectomies due to myomaand menstrual disorder did reduce the so-cioecononomic rate differences across socialclasses and educational levels, but the cleargradient according to disposable income re-

mained.Finally, it is also possible that the high rates

of myoma operations in the well off are relatedto their fewer pregnancies29 or the higher use ofhormone replacement therapy for menopausesymptoms30 in these groups in Finland. Lowparity has been observed to be related to theprobability of hysterectomy due to myoma'3and to the sizes of uterine myomas.3' Thegrowth ofmyoma is considered to be dependenton oestrogen and probably on progestin as

well,32 and in a recent study33 the numberand size of myomas were observed to increaseduring hormone replacement therapy.Even if there are clear socioeconomic differ-

ences in hysterectomy rates and they are ob-viously partly explained by better access to

gynaecological services in the better off, it isdifficult to judge whether our findings displaysocioeconomic inequities. The socioeconomicdifferences were systematic only for myomaoperations, which do not have clear criteria fortreatment. Treatment decisions on myoma, as

well as on other benign uterine conditions, areoften based on subjective symptoms and patientpreferences, which are difficult to evaluate.Moreover, it is currently known that hys-terectomies may have adverse outcomes, suchas complications or long term cardiovasculardisorders.34 Thus, in terms of health outcomes,it is difficult to say who are favoured - thebetter off with high rates or the worse off withlow rates. More research is needed on decisionmaking in hysterectomy and on the outcomesof all surgical as well as non-surgical treatmentsof gynaecological conditions.

1 Coulter A, McPherson K, Vessey M. Do British womenundergo too many or too few hysterectomies. Soc Sci Med1 988;27:987-94.

2 Luoto R, Hemminki E, Topo P, Uutela A, Kangas I. Hys-terectomy among Finnish women-prevalence and wo-men's own opinions. Scand _j Soc Med 1992;20:209-12.

3 Luoto R, Kaprio J, Keskim5ki I, Pohjanlahti J-P, RutanenE-M. Incidence, causes and surgical methods for hys-terectomy in Finland, 1987-89. Int _J Epidemiol 1994;23:348-358.

4 McPherson K, Wennberg JE, Hovind OB, Clifford P. Small-area variations in the use of common surgical procedures:an international comparison of New England, Englandand Norway. N Engl 7Med 1982;307:1310-14.

5 Keskimaki I, Aro S, Teperi J. Regional variation in surgicalprocedure rates in Finland. Scand J Soc Med 1994;22:132-38.

6 Madsen M, Andersen TF, Roepstorff C, et al. Rates ofsurgeryin the Nordic countries. Variation between and within nations.Copenhagen: NOMESCO 1994:43.

7 Bunker J P, Brown B W. The physician-patient as an in-formed consumer of surgical services. N Engl3Med 1974;290:1051-55.

8 Koepsell T D, Weiss N S, Thompson D J, Martin D P.Prevalence of prior 34 hysterectomy in the Seattle-Tacomaarea. Am

_Public Health 1980;70:40-47.

9 Coulter A, McPherson K. Socioeconomic variations in theuse of common surgical operations. BMJ 1985;291:183-87.

10 Meilahn E N, Matthews K A, Egeland G, Kelsey S F.Characteristics of women with hysterectomy. Maturitas1989;11:319-29.

11 Schofield MJ, Bennett A, Redman S, Walters WAW, Sanson-Fisher R W. Self-reported long-term outcomes of hys-terectomy. Br .7 Obstet Gynaecol 1991;98:1129-36.

12 Gittelsohn A M, Halpern J, Sanchez R L. Income, race,and surgery in Maryland. Am I Public Health 1991;81:1435-41.

13 Vessey M P, Villard-Mackintosh L, McPherson K, CoulterA, Yeates D. The epidemiology of hysterectomy: findingsin a large cohort study. Br 7 Obstet Gynaecol 1992;99:402-7.

14 Santow G & Bracher M. Correlates of hysterectomy inAustralia. Soc Sci Med 1992;34:929-42.

15 Kjerulff K, Langenberg P, Guzinski G. The socioeconomiccorrelated of hysterectomies in the United States. Am 7Publ Health 1993;83:106-8.

16 Kuh D, Stirling S. Socioeconomic variation in admissionfor diseases of female genital system and breast in a

national cohort aged 15-43. BMJ 1995;311:840-43.17 Nikiforov 0. General hospital care in Finland in the 1960s and

1970s. (In Finnish.) Helsinki: Health Services Researchby the National Board of Health in Finland 32, 1984.

18 Keskimaki I, Aro S. Accuracy of data on diagnosis, pro-cedures and accidents in the Finnish hospital dischargeregister. Int 7 Health Sciences 1991;2:15-21.

19 Central Statistical Office of Finland. Classification of socio-economic groups. Tilastokeskus: Hand books 17, 1989.

20 Organisation for economic co-operation and development.The OECD list of social indicators. Paris: OECD, 1982.

21 Toimenpidenimikkeistio 1983. (Codes for surgical procedures)Helsinki: Finnish Hospital League, 1983.

22 Payne CD ed. The GLIM system, release 3.77, manual.Oxford: The Numerical Algorithms Group, Royal Stat-istical Society, 1985.

23 Rejneveld SA, Gunning-Schepers LJ. Age, health and themeasurement of the socio-economic status of individuals.EurJ Public Health 1995;5:187-92.

24 Committee report on the development of ambulatory hospital care

andfreedom to choose the attending doctor (Hoitavan laakarinvalinnan ja poliklinikkatoiminnan kehittamistoimikunnanmietint6, in Finnish). Helsinki: Komiteamietint6 1982:19. Government Printing Centre, 1982.

25 Keskimaki I, Salinto M, Aro S. Private medicine and socio-economic differences in the rates of common surgicalprocedures in Finland. Health Policy 1996;36:245-59.

26 Sisto T, Reunanen A, Laurikka J, Impivaara 0, Heli6vaaraM, Knekt P et al. Prevalence and risk factors of varicoseveins in lower extremities: Mini-Finland Health Survey.EurISurg 1995;161:405-14.

27 Pukkala E. Cancer risk by social class and occupation. A

72

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jech.bmj.com

/J E

pidemiol C

omm

unity Health: first published as 10.1136/jech.51.1.67 on 1 F

ebruary 1997. Dow

nloaded from

Socioeconomic variations in hysterectomy

survey of 109,000 cancer cases among Finns of workingage. In: WahrendorfJ ed. Contributions to epidemiology andbiostatistics. Vol 7. Basel: Karger, 1995.

28 Valkonen T, Martelin T, Rimpela A, Notkola V, SavelaS. Socio-economic mortality differences in Finland 1981-90.Official Statistics of Finland, Population. Helsinki: Stat-istics Finland, 1993;1.

29 Nikander T. The women life course and the family formation.Helsink: Statistics Finland, 1992; 1. Official Statistics ofFinland, Population.

30 Topo P, Klaukka T, Hemminki E, Uutela A. Use ofhormonereplacement therapy in 1976-89 by 45-64 year old Finnishwomen. Epidemiol Community Health 1991:45:277-80.

31 Shikora SA, Niloff JM Bistrian BR, Forse RA, BlackburnGL. Relationship between obesity and uterine lei-omyomata. Nutrition 1991:7:251-55.

32 Rein MS, Barbieri RL, Friedman AJ. Progesterone: a criticalrole in the pathogenesis of uterine myomas. Am ObstetGynecol 1995:172:14-18.

33 Frigo P, Eppel W, Asseryanis E, et al. The effects ofhormonesubstitution in depot form on the uterus in a group of50 perimenopausal women - a vaginosonographic study.Matunitas 1995:21:221-25.

34 Luoto R, Kaprio J, Reunanen A, Rutanen E-M. Cardio-vascular morbidity in relation to ovarian function afterhysterectomy. Obstet Gynecol 1995;85:515-22.

73

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jech.bmj.com

/J E

pidemiol C

omm

unity Health: first published as 10.1136/jech.51.1.67 on 1 F

ebruary 1997. Dow

nloaded from