1 School of Population Health, University of Queensland, Brisbane 4006, 2 Population Health & Clinical Sciences Division, Queensland Institute of Medical Research, Brisbane 4006 and 3 Queensland Fertility Group, Brisbane 4000, Australia
4 To whom correspondence should be addressed at: Centre for General Practice, School of Population Health, University of Queensland, Mayne Medical School, Herston Road, Herston, Brisbane, Queensland 4006, Australia. e-mail: marie-louise.dick{at}sph.uq.edu.au
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: definition/infertility/measurement/self-report/women
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Numerous measures of infertility have been used to research the aetiology, management and outcomes of infertility, with no single measure encompassing all dimensions. To date, limited attention has been devoted to systematically comparing different definitions of infertility. Marchbanks et al. (1989) used data from the Cancer and Steroid Hormone Study to examine the effect of five definitions of infertility (three definitions based on womens self-reported answers to direct questionnaire items, and two definitions based on computations from each respondents life calendars of reproductive and contraceptive events) on several research findings (Marchbanks et al., 1989
). They concluded that the definition of infertility impacts upon the prevalence of infertility, the age at fertility classification, and the probability of future conception. They stressed the importance of researchers being aware that the definition of infertility makes a difference.
The definition of infertility most commonly used in clinical practice is failure to conceive a pregnancy after 12 months of unprotected sexual intercourse (Healy et al., 1994). This definition has its basis in the finding that the majority (85%) of couples with no adverse factors are likely to achieve a pregnancy within 12 ovulatory cycles (Guttmacher, as cited in Forbes, 1995
). Arguably, the most accurate method of classifying couples as infertile according to this clinical definition is to use prospectively collected data regarding the time taken in efforts to conceive a pregnancy; however, the financial and logistic expenses involved in collecting such data prohibits its routine use in research studies. Joffe et al. (1995
), using data collected in the Oxford Family Planning Association contraceptive cohort study, validated womens retrospectively reported time-to-pregnancy (with the median recall time being 14 years) against time-to-pregnancy recorded annually by these same women. They report very good agreement between the two sources of information at the group level, but some misclassification was evident at the individual level, having implications for statistical power (Joffe et al., 1995
).
Self-reported difficulty conceiving is the measure of infertility used by patients when they first seek fertility advice from health practitioners. Both it and the clinical definition are used in epidemiological research, yet little is known about how they specifically relate to each other. In this study we aimed to compare these measures of infertility, and to explore the meaning and potential use of womens self-reported difficulties in conceiving as a measure of infertility. We also compared womens self-reported causes of infertility with data from their medical records.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Data collection
All women participated in a standard face-to-face interview in which responses to questions regarding various reproductive, demographic and environmental factors were sought. (The full questionnaire is available from the studys website: http://www.qimr.edu.au/research/labs/davidp/swh.) Three measures of infertility were derived from womens survey responses.
The self-reported difficulty in conceiving (DC) question
Question 39 in the Survey of Womens Health asked women Have you or your partner had any problems in conceiving a child and seen a doctor about this? (This question is hereafter referred to as the DC question.)
The calendar-derived time taken trying to conceive (calendar TTTC) measure
Two detailed calendars completed at the end of the survey provided additional information regarding womens fertility histories. The pregnancy and contraception calendar was completed by asking women to account, month by month, for details of their childbearing and contraceptive histories from the age of 15 to 50 years. Women were provided with a full list of contraceptive options from which to choose. The calendar specifically sought those time periods in womens reproductive lives when they were sexually active and using contraception; sexually active, not using contraception and trying to conceive; sexually active, not using contraception but not particularly trying to conceive; pregnant; and breastfeeding. If a woman was not sure when certain events happened, she was asked to estimate these details. The pregnancy and lactation calendar (completed for those women who conceived at least one pregnancy), recorded the date and outcome of each pregnancy, the occurrences and duration of full and partial breastfeeding, and the time it took for menses to return following each pregnancy. The average age of women participants in the original casecontrol study upon which this study was based was 55 years, and their average period of recall ranged from 10 to 40 years.
Each womans fertility status was categorized according to the accepted clinical definition of infertility (Healy et al., 1994), and information provided in the reproductive calendars, and a variable hereafter referred to as the calendar-derived time taken trying to conceive (calendar TTTC) was created. Women were classified as infertile if, at some stage during their reproductive lives, they were unable to conceive despite attempts for >12 consecutive months. Women who were not infertile, and who had never been pregnant, were classified as having untested fertility. The remaining women were classified as fertile. The distinction between primary and secondary infertility is not made in this paper.
Because of the decline in a womans natural fecundity with age, classification of fertility status was carried out by referring only to the information provided in womens calendars from the ages of 15 to 45 years. Women were considered not potentially fertile when they were using contraception, during the puerperium, and for the first 6 months of full-time breastfeeding if menses had not returned following birth (Lewis et al., 1991).
A protocol was devised to deal with missing data for 62 women for whom 12 months of the contraceptive calendar was not accounted. Relevant information provided elsewhere in the original questionnaire (such as reported age at first intercourse, duration of use of the oral contraceptive pill, and pregnancy and lactation calendar data) was reviewed, and uniform decisions regarding missing values were made (without knowledge of the womens underlying health status). Fertility status could not be determined for two women whose reproductive calendars contained no information.
The self-reported time taken trying to conceive (SR-TTTC) measure
Questions 37 and 38 in the Survey of Womens Health asked women for the maximum number of months required to conceive a pregnancy (if ever pregnant), or to try to conceive a pregnancy (if unsuccessful in their attempts). By combining womens responses to these two questions, and using the commonly accepted clinical definition of infertility (Healy et al., 1994), a variable known as self-reported time taken trying to conceive was created. Women were classified as infertile if they reported ever having attempted to conceive a pregnancy for >12 months; as having untested fertility if they had either never tried to conceive, or had only ever tried to conceive for
12 months; and as fertile if they were not already classified as infertile and had ever been pregnant.
Comparison with medical records
Those women who responded affirmatively to the difficulty in conceiving question (i.e. that they or their partner had had difficulty in conceiving for which they saw a doctor), were asked what they believed to be the cause of their fertility problem, and to provide the contact details of any doctor(s) they had consulted regarding this. With written consent, their doctors were approached to provide confirmatory details using a standard questionnaire. When a treating doctor was not nominated, or when the nominated doctor was deceased, retired or not contactable, the current medical practitioner was approached. When a womans medical records were available, her doctor was asked to indicate if any history of a difficulty in conceiving had been documented, and, if so, to supply details including the names of other doctors consulted. Telephone reminder calls were made to non-responding doctors 2 and 4 weeks after posting the questionnaires.
Data analysis
Prevalence rates of infertility were determined for each of the three measures of infertility. We calculated the sensitivity and specificity [with 95% exact binomial confidence intervals (CI)] of womens self-reported difficulty in conceiving, against fertility status determined using the classical definition of infertility applied to the calendar-derived TTTC and self-reported TTTC. We also compared the changes in sensitivity and specificity when the specified time period in the clinical definition was varied from 12 months. Women classified as having untested fertility were not included in these analyses. Self-reported reasons for the stated fertility problems were compared with the histories and causes of fertility problems documented in the womens medical records. Responses among women with and without ovarian cancer were analysed both separately and combined. As no significant differences were found, the analyses are not reported in the Results section. SAS statistical software (SAS Institute Inc., 1996) was used in the analyses.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Comparison of measures of infertility
Womens responses to the self-reported DC question were first compared with their calendar-derived fertility status using the clinical definition of infertility (Table I). After excluding the 134 women with untested fertility, only 66% of all the women who were considered infertile according to the clinical definition of infertility reported a difficulty in conceiving; in other words the DC question had a sensitivity of 66% (95% CI: 6071). Of the 1202 women who were considered fertile according to the clinical definition, only 62 (5%) reported a difficulty in conceiving, in other words the DC question was highly specific 95% (95% CI: 9396). Of the 260 women with tested fertility who responded affirmatively to the DC question, 198 (76%) recorded (in the calendar) at least one period of unsuccessfully trying to conceive for >12 consecutive months (i.e. the positive predictive value, or the proportion of respondents who test positive, who really have the condition being examined, was 76%) (Table I).
|
|
|
|
|
For 22 of the 41 women, a cause for the fertility problem was recorded in the medical records. Of these 22 women, 13 (59%) had self-reported causes that were concordant with the cause written in the medical records: male factor problems (n = 5); fallopian tube problems (n = 4); ovulation/hormonal problems (n = 2); endometriosis (n = 1); and idiopathic (n = 1). For six of the nine discordant pairs of reports, the clinically documented reason for the fertility problem was a uterine abnormality, whilst the women reported no medical reason found (n = 5) or a hormone imbalance (n = 1). Two womens medical practitioners reported the cause as idiopathic whilst these women had reported the cause as being due to a tubal abnormality or to internal injuries as a child respectively. One woman attributed her problem to a hectic lifestyle, whilst her medical records documented a male factor.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Prevalence estimates of infertility in the study population were 16% based on self-reported difficulty conceiving, and 20 and 23% based on the standard 12 month clinical definition applied to calendar-derived and self-reported times taken trying to conceive. All three estimates fall centrally in the broad 333% range of prevalences currently reported in the literature (Gunnell and Ewings, 1994; Schmidt and Münster, 1995
; Chandra and Stephen, 1998
; Karmaus and Juul, 1999
). Additionally, they fall within the 1626% range of lifetime prevalence estimates of infertility in a number of population-based studies. Malin et al. reported 16% of women answering affirmatively to a question Have you ever had difficulties in getting pregnant and/or having a child, whilst studies by Greenhall and Vessey, Gunnell and Ewings, and Schmidt, Münster and Helm (all using the 12 month clinical definition to define infertility) reported life-time prevalence estimates of infertility of 2426% (Greenhall and Vessey, 1990
; Gunnell and Ewings, 1994
; Schmidt et al., 1995
; Malin et al., 2001
).
Although the similarity in the estimates suggests prima facie that the self-reported difficulty in conception measure might be used interchangeably with the TTTC measures (and vice versa), a consideration of the sensitivity of the difficulty in conceiving question in relation to the two time-related measures based on the clinical definition (66 and 69%) implies otherwise. The self-reported difficulty in conceiving measure failed to identify approximately one-third of women who reported >12 consecutive months of unsuccessfully conceiving despite efforts. This sensitivity fell to 58% and rose to 78% when the time period in the clinical definition was varied from 6 to 36 months for the calendar-derived TTTC measure (Figure 1).
A closer exploration of what the self-reported difficulty in conceiving measure and the two time-related measures based on the clinical definition actually represent enables a more informed interpretation of the significance of this relatively low sensitivity. The clinical definition has its basis in the finding that the majority (85%) of couples with no adverse factors are likely to achieve a pregnancy within 12 ovulatory cycles (Guttmacher, as cited in Forbes, 1995). Although in practice, fertility represents a continuum, the 12 month time period without conceptions represents a reasonably pragmatic operational dividing line between more and less fertile women. In this study, the classification of a woman as infertile according to the clinical standard was used for two measures: self-reported TTTC, and calendar-derived TTTC. Given that previous studies have shown that menstrual and reproductive events can be recalled with reasonable accuracy (Bean et al., 1979
; Zielhuis et al., 1992
), and that the value of calendars to assist recall [for example by first recalling significant landmark events (Jobe and Mingay, 1989
)] is now supported by a strong body of theory and observation (Armstrong et al., 1992
), poor recall is unlikely to be a major source of error for the TTTC measures.
A comparison of the two time-related measures of infertility demonstrated reasonable similarity. When compared with the calendar-derived TTTC, self-reported TTTC demonstrated a sensitivity of 80% (95% CI: 7585) and a specificity of 97% (95% CI: 9598). Conversely, calendar-derived TTTC compared with self-reported TTTC gave a sensitivity of 86% (95% CI: 8190) and a specificity of 95% (95% CI: 9496).
The difficulty in conceiving question links two issues that are important from a public health planning perspective: whether women (or their partners) have ever had problems conceiving, and whether they have seen a doctor about it. It is a patient-centred measure; womens responses to the question will take into consideration their desire to have children, the point in time at which they consider a fertility problem exists, and factors influencing if and when they decide to seek medical advice. The question identified some women who had been trying to conceive for 12 months, but who were sufficiently concerned to seek medical advice regarding their fertility problem (the 62 false positives in Table I and the 61 false positives in Table II). On the other hand, the wording of the question precluded the detection of those women who were infertile according to the clinical definition, yet who did not seek medical attention (the 104 women with false negative responses in Table I and the 86 women with false negatives in Table II). That substantial proportions of infertile women do not seek medical advice has been previously documented in other population-based studies (Ducot et al., 1991
; Webb and Holman, 1992
; Gunnell and Ewings, 1994
; Schmidt et al., 1995
; Stephen and Chandra, 2000
). Furthermore, one of these studies demonstrated that 44% of the 62% of couples who sought medical advice actually did so prior to 6 months of attempts at conception (Ducot et al., 1991
).
The number of false negatives raises questions as to why women who have been unsuccessful in their efforts to conceive for more than a year have not sought medical advice. It points to a difference in perception of need by clinicians and patients, as clinicians generally recommend that couples experiencing infertility should not delay seeking medical advice beyond 12 months of efforts to conceive (and sooner if couples have known adverse risk factors) (Forbes, 1995; Morrison et al., 2001
). Additionally, given that indicators of social disadvantage are known to be major determinants of avoiding seeking help with infertility (Schmidt et al., 1995
; Wulff et al., 1997
; Chandra and Stephen, 1998
), and that a large international survey indicated poor public perception of infertility and its treatment (Adashi et al., 2000
), our findings support the public health imperative to better inform the general public about the utility and availability of infertility treatment options, their costs and success rates.
The DC question was highly specific (95%), and when the defining criteria of infertility for the clinical definition were altered, its specificity remained high despite fluctuations in the sensitivity (Figure 1). This suggests a potential role of the DC question to screen out women who do not have fertility problems, as it is unlikely to misclassify women without clinically defined infertility.
The DC question therefore has potential value for use in health services studies, and in studies aimed at quantifying the burden of perceived fertility problems in the community. However, its value is limited in aetiological studies where misclassification may be sufficient to obscure a true effect of infertility on disease.
Our attempts to validate womens self-reported fertility problems with medical records highlight the enormous difficulty in retrospectively documenting reproductive events. Medical records provided little information about the accuracy of most womens reports: of the 179 women for whom medical records were available, only 41 (23%) had their self-report of a fertility problem confirmed. There was complete concordance of self-reported infertility and the medical records maintained by those 18 doctors who were specifically consulted regarding the fertility problem. However, the low concordance rate (23/161 = 14%) with medical records completed by doctors not actually consulted regarding a fertility problem presumably represents the irrelevance of a womans fertility history in that context, leading to its never being reported or recorded. Our interpretation is that womens recall of fertility consultations is probably very good, but retrospectively confirming these reports with medical records is difficult. Our concordance rate of 59% for causes of infertility is similar to the 67% reported in another study (van Balen et al., 1996). Not surprisingly, those causes of infertility more readily diagnosed objectively by routine investigations were concordant, whilst less-objectively diagnosed conditions such as uterine factor were not.
In conclusion, there are many measures of infertility available for use in epidemiological studies, each defined differently, and each therefore potentially generating different associations. Researchers need to understand the definitional implications of infertility measures when choosing measures for a particular study. This study has gone some way to demonstrate the potential strengths and weaknesses of using womens self-reported difficulty in conceiving as a measure of infertility in epidemiological studies. It calls for improved education of the general public regarding the availability of infertility treatments and their potential risks and benefits.
![]() |
Acknowledgements |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Armstrong, B.K., White, E. and Saracci, R. (1992) Principles of Exposure Measurement in Epidemiology. Oxford University Press, Oxford.
Australian Electoral Commission (1999) Australian Electoral Commission Annual Report 19981999. Australian Electoral Commission, Canberra.
Bean, J.A., Leeper, J.D., Wallace, R.B., Sherman, B.M. and Jagger, H. (1979) Variations in the reporting of menstrual histories. Am. J. Epidemiol., 109, 181185.[Abstract]
Chandra, A. and Stephen, E.H. (1998) Impaired fecundity in the United States: 19821995. Fam. Plann. Perspect., 30, 3442.[ISI][Medline]
Ducot, B., Spira, A., Thonneau, P., Toulemon, L. and Leridon, H. (1991) Difficulties in conceiving. Discussion about methodology concerning the INED-INSERM survey carried out in France in 1988 on 3,181 women aged between 18 and 49. J. Gynecol. Obstet. Biol. Reprod. (Paris), 20, 643650.[Medline]
Forbes, K.L. (1995) Management of infertility in general practice. Mod. Med. Aust., 38, 1822.
Greenhall, E. and Vessey, M. (1990) The prevalence of subfertility: a review of the current confusion and a report of two new studies. Fertil. Steril., 54, 978983.[ISI][Medline]
Gunnell, D.J. and Ewings, P. (1994) Infertility prevalence, needs assessment and purchasing. J. Publ. Health Med., 16, 2935.[Abstract]
Healy, D.L., Trounson, A.O. and Andersen, A.N. (1994) Female infertility: causes and treatment. Lancet, 343, 15391544.[ISI][Medline]
Jobe, J.B. and Mingay, D.J. (1989) Cognitive research improves questionnaires. Am. J. Publ. Health, 79, 10531055.[ISI][Medline]
Joffe, M., Villard, L., Li, Z., Plowman, R. and Vessey, M. (1995) A time to pregnancy questionnaire designed for long term recall: validity in Oxford, England. J. Epidemiol. Community Health, 49, 314319.[Abstract]
Karmaus, W. and Juul, S. (1999) Infertility and subfecundity in population-based samples from Denmark, Germany, Italy, Poland and Spain. Eur. J. Publ. Health, 9, 229235.[Abstract]
Lewis, P.R., Brown, J.B., Renfree, M.B. and Short, R.V. (1991) The resumption of ovulation and menstruation in a well-nourished population of women breastfeeding for an extended period of time. Fertil. Steril., 55, 529536.[ISI][Medline]
Malin, M., Hemmink, E., Raikkonen, O., Sihvo, S. and Perala, M.L. (2001) What do women want? Womens experiences of infertility treatment. Soc. Sci. Med., 53, 123133.[CrossRef][ISI][Medline]
Marchbanks, P.A., Peterson, H.B., Rubin, G.L. and Wingo, P.A. (1989) Research on infertility: definition makes a difference. The Cancer and Steroid Hormone Study Group. Am. J. Epidemiol., 130, 259267.[Abstract]
Morrison, J., Carroll, L., Twaddle, S., Cameron, I., Grimshaw, J., Leyland, A., Baillie, H. and Watt, G. (2001) Pragmatic randomised controlled trial to evaluate guidelines for the management of infertility across the primary caresecondary care interface. Br. Med. J., 322, 12821284.
Purdie, D., Green, A., Bain, C., Siskind, V., Ward, B., Hacker, N., Quinn, M., Wright, G., Russell, P. and Susil, B. (1995) Reproductive and other factors and risk of epithelial ovarian cancer: an Australian case-control study. Int. J. Cancer, 62, 678684.[ISI][Medline]
SAS Institute Inc. (1996) Cary, NC, USA.
Schmidt, L. and Münster, K. (1995) Infertility, involuntary infecundity, and the seeking of medical advice in industrialized countries 19701992: a review of concepts, measurements and results. Hum. Reprod., 10, 14071418.[Abstract]
Schmidt, L., Münster, K. and Helm, P. (1995) Infertility and the seeking of infertility treatment in a representative population. Br. J. Obstet. Gynecol., 102, 978984.[ISI][Medline]
Stephen, E.H. and Chandra, A. (2000) Use of infertility services in the United States: 1995. Fam. Plann. Perspect., 32, 132137.[ISI][Medline]
Templeton, A., Fraser, C. and Thompson, B. (1990) The epidemiology of infertility in Aberdeen. Br. Med. J., 301, 148152.[ISI][Medline]
van Balen, F., Trimbos-Kemper, T. and Verdurmen, J. (1996) Perception of diagnosis and openness of patients about infertility. Patient Educ. Couns., 28, 247252.[CrossRef][ISI][Medline]
Webb, S. and Holman, D. (1992) A survey of infertility, surgical sterility and associated reproductive disability in Perth, Western Australia. Aust. J. Publ. Health, 16, 376381.[ISI]
Wulff, M., Hogberg, U. and Stenlund, H. (1997) Infertility in an industrial settinga population-based study from Northern Sweden. Acta Obstet. Gynecol. Scand., 76, 673679.[ISI][Medline]
Zielhuis, G.A., Hulscher, M.E. and Florack, E.I. (1992) Validity and reliability of a questionnaire on fecundability. Int. J. Epidemiol., 21, 11511156.[Abstract]
Submitted on September 23, 2002; resubmitted on July 15, 2003; accepted on September 10, 2003.