Skip Navigation


Journal of Public Health Advance Access originally published online on February 13, 2008
Journal of Public Health 2008 30(2):178-185; doi:10.1093/pubmed/fdn005
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
30/2/178    most recent
fdn005v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Joffe, M.
Right arrow Articles by Palmer, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Joffe, M.
Right arrow Articles by Palmer, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2008, Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved

Use of time to pregnancy in environmental epidemiology and surveillance



Mike Joffe, Emeritus Reader1

Shantini Paranjothy, Walport Clinical Lecturer2

Hilary Fielder, Director of Screening Services3

Ronan Lyons, Professor4

Stephen Palmer, Professor2,5
1 Department of Epidemiology and Public Health, Imperial College, St Mary's Campus, Norfolk Place, London W2 1PF, UK
2 Department of Primary Care and Public Health, School of Medicine, Cardiff University, 5th Floor Neuadd Meirionnydd, Heath Park, Cardiff CF14 4XN, UK
3 Breast Test Wales, 18 Cathedral Road, Cardiff CF11 9LJ, UK
4 School of Medicine, Swansea University, Singleton Park, Swansea SA2 8PP, UK
5 Health Protection Agency, Local and Regional Services, Department of Primary Care and Public Health, Centre for Health Sciences Research, Cardiff University, 4th Floor Neuadd Meirionydd, Heath Park, Cardiff CF14 4XN, UK

Address correspondence to Mike Joffe. E-mail: m.joffe{at}imperial.ac.uk

Background Potential sources of environmental pollution, such as incinerators or landfill sites, can adversely affect reproduction and/or development. Time to pregnancy (TTP) is a validated measure of biological fertility that can be studied with relatively small populations.

Methods Pregnant local residents living within 3 km of a landfill site (‘exposed’ group, n = 200) or elsewhere in the Rhondda valleys (‘unexposed’ group, n = 400) were interviewed by health visitors or midwives. The response rate was 83%.

Results No difference was found in the TTP distributions between the exposed and unexposed groups. Relationships of TTP with covariates were consistent with the literature.

Conclusions In a context of public and scientific concern about possible reproductive toxicity, an interview study of TTP was highly acceptable to local women. A large enough sample to generate stable TTP distributions was readily achieved.

Keywords: environmental epidemiology, fertility, fecundity, landfill, reproduction, time to pregnancy


   Introduction
TOP
 Introduction
Methods
 Results
 Discussion
 Conclusion
 Acknowledgments
 References
 
Human reproduction and embryonic development can be adversely affected by exposure to environmental agents,1,2 and because of this, clusters of miscarriages or congenital anomalies are often attributed by the public to recognisable potential sources of pollution, such as incinerators or landfill sites. In such cases, typically, there are too few adverse events to allow causal hypotheses to be tested rigorously. End points that are more frequent allow small populations to be investigated. In this paper, the use of ‘time to pregnancy’ is explored as an end point that may have utility for investigating population environmental exposures.

Time to pregnancy (TTP) is an outcome measure of biological fertility, defined as the number of months that a couple takes to conceive, given unprotected intercourse. TTP is a functional measure of couple fertility,3 and its use is complementary to more mechanistic research on specific biological pathways and medical conditions in both sexes, and to studies using biomarkers of semen quality. As well as being a sensitive measure of biological fertility in the true sense, TTP is also affected by embryotoxic agents that cause early loss before recognition of pregnancy.3 TTP is well suited to surveillance, and has been suggested as an useful overall indication of a reproductive hazard when no specific hypothesis can be formulated.4

Whilst prospective TTP studies beginning in women before they become pregnant are feasible5 and have methodological advantages in the study of aetiological factors,5–8 they are often not practicable as they are expensive to conduct. They also tend to suffer from methodological problems such as the absence of a sampling frame and low and selective participation rates, leading to bias.9 However, retrospective TTP studies are easy to conduct, require a relatively short questionnaire, and have been found to be acceptable in a wide variety of cultures.3

Validation studies have demonstrated that women's self-reports give an accurate representation of the true TTP distribution,10–13 even with duration of recall up to 20 y.13 Men can also provide apparently valid information, generating the same distributions and analytical results as women drawn from the same population,14,15 albeit with more digit preference and item non-response (typically 10–15% compared with 5–10% for women). TTP is a property of the couple, so the covariates for both partners are required. As the outcome is the number of months that it took to conceive and TTP values are constrained to be integers, the discrete-time analogue of the Cox model is recommended for analysis.9,16 Alternatively, regression models for ordinal responses such as the continuation-ratio model can be used.17

The methodology of retrospective TTP-based research has evolved gradually since the mid-1980s, and the potential biases and other problems are now well understood. They can be effectively controlled if the studies are properly designed, conducted and analysed9,18 (see Discussion). Time trends14 and international differences19,20 have been studied, and risk factors for delay in TTP identified.15,21–23 Whilst there are many occupational studies of TTP, e.g.24,25 its potential for use in environmental epidemiology has not hitherto been realised. It is proposed that the following applications be considered:

  1. defined population studies in areas close to sources of emissions of agents that are perceived as possibly harmful;
  2. surveys representative of the general population that are sufficiently large to allow spatial heterogeneity to be detected;
  3. non-spatial studies that relate TTP to environmental agent exposure.
Data should be obtained from special surveys with a well-defined source population. Clinical samples (drawn from fertility clinics, for example) are unsuitable, due to strong selection and self-selection biases, and studies based on volunteer samples are also difficult to interpret and should be avoided. Two types of sampling frame are suitable for retrospective environmental studies of TTP:
  1. Pregnancy based studies: A sample of women who are pregnant or have recently had a baby are asked about TTP relating to that pregnancy. The advantage is that it is easy to define and contact the women and, as they are asked about a current or recent pregnancy, recall bias is minimal. Disadvantages include that sterile couples are excluded and subfecund couples are underrepresented.
  2. Cross-sectional studies: Couples are selected randomly from the general population, and asked about TTP either for all previous pregnancies or for just one, e.g. the first; it is also possible to enquire about unsuccessful attempts and sterility (see below). For many of them, the pregnany(ies)/attempt(s) may have been some years ago, which can be an advantage in certain circumstances, for example with a landfill site that has been operating for years, when pre-operational observations are valuable. As the focus is on previous exposures, this design approximates to a retrospective cohort study. The disadvantage is long recall with possible recall bias, especially for variables relating to exposure and for covariates (as also occurs e.g. in case control studies of cancer).
In its literal sense, the term ‘time to pregnancy’ refers only to the period of unprotected intercourse that leads to conception, but TTP studies should also collect information on contraceptive failures (see Discussion). In addition, it is possible also to include unsuccessful attempts (infertile phases and long-term infertility) if the study design is not pregnancy based. This is preferable, as stronger inferences can then be made: selection bias from including only fertile couples is avoided, and regression estimates are not conditional on achieved conception as they are with pregnancy based studies.26 The length of the unsuccessful attempt generally has to be at least six months for the couple to remember these attempts and provide valid information about them. Use of a cross-sectional sample rather than pregnancies may also allow study of a larger number of conceptions, by extending the number of years of observation.

We report here a study of TTP in the neighbourhood of the Nant-y-Gwyddon landfill site in the Rhondda valleys in South Wales, which can be regarded as a pilot study for the use of this approach in environmental epidemiology. Studies of residents had shown that women living nearby were at increased risk of giving birth to a child with a congenital anomaly. Rates were also increased for the years leading up to the opening of the site,27 and this raised concerns about the other local environmental exposures that could possibly explain the increased rate of congenital abnormalities in this area, requiring further investigation. However, as only up to four babies with congenital anomalies are born each year to families living within 3 km of the site, the statistical power was lacking to explore increases in risk over a short period of time. A follow-up study was therefore undertaken to explore this issue further, using TTP as an outcome measure. The aim of this study was to determine whether the distribution of TTP varied between residents living close to the landfill site and those further away.


   Methods
TOP
 Introduction
 Methods
Results
 Discussion
 Conclusion
 Acknowledgments
 References
 
Following approval by the Bro Taf Local Research Ethics Committee, women were identified from the local midwifery records. They were eligible for the study if they had achieved a pregnancy of at least 24 weeks and were resident in the Rhondda valleys. The first 600 consecutive pregnant women after July 1999 who satisfied these criteria comprised the study group, including 200 living in the five wards within 3 km of the landfill site (‘exposed’ group) and 400 in the other 10 wards, as experience has shown that approximately 200 pregnancies with TTP values are required to generate a stable distribution.3

A questionnaire, based on that used in the Asclepios project24,25 was administered by health visitors or midwives in face-to-face interviews. It documented demographic details, previous contraceptive method, TTP and self-reported environmental exposures. Reported covariates since stopping contraception included smoking habits of both partners, mother's alcohol consumption and use of medication (prescribed and over-the-counter), recreational drugs, use of hair colorants and exposure to solvents in paint. Personal educational level and current employment were also recorded, as was a ward-based deprivation score (Townsend Index of Material Deprivation). The questionnaire was pre-piloted to assess its acceptability. All interviewers were trained to administer the questionnaire.

Statistical analysis was carried out using STATA version 8.2. Relative risks and 95% confidence intervals (CI) were calculated. Differences in proportions were tested using Chi squared test for heterogeneity. A continuation-ratio model was fitted17 to estimate the odds ratio of becoming pregnant within the next month at a given time, comparing the exposed and unexposed groups.


   Results
TOP
 Introduction
 Methods
 Results
Discussion
 Conclusion
 Acknowledgments
 References
 
Questionnaires were completed from interviews with 499 out of 600 women identified as giving birth to a child from the midwifery register between July 1999 and June 2000, a response rate of 83%. Of those not interviewed, there was no response from 77 women despite three attempts by the health visitor to contact them; nine women did not achieve a pregnancy of 24 weeks; eight women refused to participate; three women had moved out of the area with no forwarding address; two were not interviewed because of illness; and two were found not to be residents of Rhondda. The electoral ward of residence for 481 of the 499 respondents were able to be identified, 326 in unexposed and 155 in exposed wards. Of the 499 pregnancies, 362 (73%) were eligible for a TTP value as they did not result from a contraceptive failure, and reported TTP values were available for 329 (91%) of these.

The mean ages of the 155 women living in the exposed wards (26.3 y) and their partners (28.0 y) were similar to the mean of the 326 women and partners in the unexposed wards (26.8 and 28.3 y, respectively) (Table 1). A higher proportion of women in the exposed group (45%) compared with the unexposed group (12%) were within the lowest quintile of the distribution of Townsend score for material deprivation.


View this table:
[in this window]
[in a new window]

  		Table 1 Demographic characteristics of women in exposed and unexposed groups

 
Women in the exposed wards were less likely to have used birth control prior to pregnancy, and were more likely to have planned the pregnancy, and to have taken prescribed medicines (Table 2). There was no significant difference in the proportion of women taking fertility treatment, folic acid supplements or other medicines, or tablets or sprays bought from a pharmacy. Medication for asthma was more commonly reported in the exposed wards (RR 4.21, 95% CI 1.29–13.76) although the numbers were small (eight cases and four controls). The results were similar when use of bronchodilator and steroid inhaler were analysed separately.


View this table:
[in this window]
[in a new window]

  		Table 2 Characteristics of women in exposed and unexposed groups

 
The cumulative TTP distributions (Figure 1) were very similar for residents of exposed and unexposed areas. There was no difference in fertility for those in the exposed group compared to those in the unexposed group [OR 1.02 (95% CI 0.78, 1.33)] (Table 3). Women younger than 25 y were more fertile. Fertility was significantly increased among women educated beyond the age of 16 y. It was also increased among women who had previously used contraception, as expected.9 Following adjustment for all maternal and paternal characteristics, there was still no significant difference in fertility between the exposed and unexposed groups. However, education level was more strongly associated with increased fertility (Table 3).


Figure 1
View larger version (10K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1 Cumulative distribution of time to pregnancy in exposed and unexposed wards.

 


View this table:
[in this window]
[in a new window]

  		Table 3 Odds ratio from the continuation-ratio model showing the relationship between fertility (TTP) and soci-demographic and lifestyle variables

 
Although there was a lower proportion of accidental pregnancies in the exposed group (16.4%) compared with (27.4%), this difference was not significant following adjustment for all maternal and paternal characteristics [adjusted OR 0.48 (95% CI 0.20, 1.15)]. Nevertheless it was considered important to explore whether this could have affected the findings, because accidental pregnancies are ineligible for a TTP value and are treated as missing values in the main analysis. A sensitivity analysis was therefore conducted in which pregnancies reported as accidental were included in the TTP analysis with an assigned value of zero, which would over-compensate any bias present, thereby providing upper and lower bounds for the magnitude of any association.9 The findings for the exposed/unexposed comparison were unchanged [adjusted OR 1.20 (95% CI 0.83, 1.73) compared with 1.17 (0.81, 1.68)]. The other findings were similarly robust, including the association with educational level, apart from ‘using birth control’ and ‘planned pregnancy’ which reversed in direction as expected (data available on request).


   Discussion
TOP
 Introduction
 Methods
 Results
 Discussion
Conclusion
 Acknowledgments
 References
 
Main finding of this study
The problems of small area studies to characterise risk precisely around a point source hazard are well recognised, and include a small study population, low levels of exposure that are difficult to define, and the possibility of migration reducing the stability of populations. Agencies responsible for protecting the public health need to use the most sensitive measures available to investigate adverse outcomes in communities.

This study showed no difference in TTP between residents of the Rhondda who lived close to the Nant-y-Gwiddon landfill site and those who lived some distance away. Although the power of the study was limited, it was sufficient to show the expected trends with age, previous contraception and educational level,15,21 although the confidence intervals included 1.0 after adjustment in the cases of previous contraception and age. The lack of association with living near to the landfill in 1999 provides some reassurance to couples living in the area. A more comprehensive study of risks to mothers living around landfill sites in Wales has shown that increased relative risks of congenital malformations after sites opened had declined to unity in 1998–2000.28

What is already known on this topic
The TTP methodology is well established in descriptive studies, e.g. of national differences19,20 and of trends,14,29 and has also been widely used to study lifestyle behaviours21–23 and occupational exposures.24,25 However, to our knowledge it has not previously used in the context of environmental studies.

Evidence on the reproductive health effects of landfill sites is still rather limited.1,2,27 This is the first study of biological fertility in the neighbourhood of such a site.

What this study adds
Our experience shows that a pregnancy-based TTP study is readily acceptable in the context of local concern about possible toxic effects from a local landfill site. It is likely that this would be true of other such environmental sources of chemical exposure that could be regarded as potentially toxic, such as an incinerator or an industrial site. This accords with previous experience on the acceptability of TTP studies,3 but contrasts with the typical situation with factory studies, where experience has not always been so positive.

The negative findings in this study cannot necessarily be extrapolated to other landfill sites, due to the heterogeneity in chemical exposures from such sites. All we can say is that this particular site appears not to affect the biological processes that are measured by TTP, i.e. those leading to conception and to the successful maintenance of early pregnancy.

Limitations of this study
Potential biases need to be considered. The response rate of 83% suggests that response bias is unlikely to be a serious problem when this is the case.

As with any study that finds no association, we have to consider whether this could have been due to methodological shortcomings. The relevant issues are measurement error and statistical power. There is some measurement error associated with retrospective self-reports of TTP,12 but in this study, the lack of a precise exposure measurement is likely to have been far more important, because distance from a point source is a somewhat crude means of assessing exposure to chemical agents.30,31 In addition, the size of the study, with 117 in the exposed and 200 in the unexposed group, means that the statistical power was highly limited. The negative findings therefore need to be treated with some caution.

This study was carried out using a pregnancy-based sample. As previously stated, it is preferable to employ a cross-sectional population sample, and for the questionnaire to include items on infertile phases (unsuccessful attempts). The use of a pregnancy-based method here means that we cannot exclude the possibility that exposures from the landfill site can lead to severe infertility, manifest as inability to conceive, but without affecting the more minor degree of subfertility that TTP is able to measure; however, this appears to be biologically implausible. In the UK context, it would be feasible to carry out a study of this kind with a sampling frame based on age-sex registers in primary care. Such a cross-sectional survey design would allow a TTP study to be combined with other topics if desired.

The age range chosen would depend on which years are relevant to the exposure of concern. It would be possible to enquire about all pregnancies/attempts per couple, which is statistically efficient, or just one (e.g. the first). Only data on the first are unbiased, as fertile couples may tend to have a larger number of children, and adverse reproductive experience may affect behaviour (‘reproductive compensation’). A further potential problem with including all pregnancies/attempts is that it is necessary to use frailty (mixed) models,32,33 and there are currently no standard programs for this in the context of survival analysis.

Other potential biases include planning bias, behaviour change bias and truncation bias.9,34 Planning bias could result if one exposure group has a higher degree of risk-taking and/or a lower degree of conscious planning, so that they are more prone to contraceptive failures, because TTP is not defined for accidental pregnancies. The standard way to handle this is to obtain the same covariate information on pregnancies that resulted from a contraceptive failure as for those eligible for TTP values. This makes possible a parallel logistic regression analysis of this outcome, with the same exposure variable and covariates as the main model, and a sensitivity analysis for the main analysis with the accidental pregnancies assigned a TTP value of zero.9 These additional analyses proved to be robust in the present study: the lack of association of TTP with exposure was confirmed, as was the strong association of TTP with maternal educational level.

All time-varying covariates, such as smoking habit, are asked in relation to the ‘starting time’, when unprotected intercourse began, otherwise if women tended to quit smoking after several months of trying to conceive, the relatively long TTP value would be associated with the non-smoking value, which could lead to smoking being wrongly classified as protective.9,34 Time variables, including age, are also analysed at the starting time. For example, if a woman starts trying to conceive at 35 y and only succeeds after 5 y, the value of 35 is associated with her TTP value rather than 40. This makes everybody equal at the starting time.

It is not necessary to ask for a great deal of detailed information. For example, frequency of intercourse is not generally included in TTP surveys, as this would be likely to discourage participation. Similarly, information on illnesses, shift work and travel during the TTP interval is omitted. The assumption is made that they are equally distributed across the exposure groups and/or that their effect is small at the group level.


   Conclusion
TOP
 Introduction
 Methods
 Results
 Discussion
 Conclusion
Acknowledgments
 References
 
Recommendations on the conduct of retrospective TTP studies are listed in Box 1. It is important to recognise that this type of study is unusually prone to bias if appropriate steps are not taken. To reduce the risk of bias, information on accidental pregnancies, and preferably also on unsuccessful attempts, should be regarded as an integral part of data collection and analysis. More detailed discussion of the methodological issues can be found in Joffe et al.,9 which has a suggested core questionnaire.


Box 1. Recommendations
  • External generalisability requires that the study sample is representative of the underlying population.
  • It is preferable to use a study design including periods of unsuccessful attempts, not only the couples who achieve a pregnancy.
  • Questionnaire design is important:
    • the criterion for eligibility of a TTP value is that no contraceptive method was being used;
    • TTP should be measured in months, ungrouped, and should include current pregnancies;
    • infertile phases of at least six months’ duration should be included, so that more complete inferences can be made about the whole population, and also to help avoid medical intervention bias and differential persistence;
    • ongoing attempts should be included, to ensure that the least fertile couples are not excluded and to help avoid truncation bias;
    • covariates should be measured not only for TTP but also for infertile phases and contraceptive failures, so that consistent analyses can be carried out for all three outcomes;
    • time-varying covariates should be measured at the starting time both for TTP and infertile phases, to reduce behaviour change bias.

  • Discrete-time survival analysis can be carried out using grouped-continuous models (for example, a grouped Cox model) or continuation ratio models. The two methods tend to produce qualitatively similar results, but continuation ratio models can be computationally easier to implement.
    • sensitivity to the choice of censoring time should be explored.

  • The analysis should be confined to births, excluding non-birth outcomes such as miscarriages, to avoid bias due to differential recall.
  • A parallel logistic regression analysis should be carried out using contraceptive failures as the outcome, to check whether the most fertile couples in certain exposure groups are being systematically excluded due to a higher propensity to have accidental pregnancies.
  • A sensitivity analysis should be carried out that includes contraceptive failures, assigning them a TTP value of zero, to check for sensitivity to planning bias; this will over-compensate any bias that is present.

 


   Acknowledgments
TOP
 Introduction
 Methods
 Results
 Discussion
 Conclusion
 Acknowledgments
References
 
The authors would like to thank Mrs Caroline Jones and Dr Sharon Hillier for their assistance with this study. The authors would also like to thank the health visitors and midwives in the Rhondda who administered the questionnaire; Mr Nathan Lester, National Public Health Service for Wales, for ascribing the ward codes and Townsend deprivation indices, and Mrs Bridie Sylvester and Mrs Andrea James for secretarial support.

This project received funding from Bro Taf Health Authority. They approved the study design but made no amendments to the study design proposed, and had no involvement in the collection, analysis and interpretation of data, in the writing of the report or in the decision to submit the paper for publication.

The authors declare that they have no competing interests.


   References
TOP
 Introduction
 Methods
 Results
 Discussion
 Conclusion
 Acknowledgments
 References
 

  1. Geschwind SA, Stolwijk JA, Bracken M, et al. Risk of congenital malformations associated with proximity to hazardous waste sites. Am J Epidemiol (1992) 135:1197–206.[Abstract/Free Full Text]
  2. Dolk H, Vrijheid M, Armstrong B, et al. Risk of congenital malformations near hazardous-waste landfill sites in Europe: the EUROHAZCON study. Lancet (1998) 352:423–27.[CrossRef][Web of Science][Medline]
  3. Joffe M. Time to pregnancy: a measure of reproductive function in either sex. Occup Environ Med (1997) 54:289–95.[Abstract/Free Full Text]
  4. Rowland A, Baird DD, Weinberg CR, et al. Reduced fertility among women employed as dental assistants exposed to high levels of nitrous oxide. New Engl J Med (1992) 327:993–97.[Abstract]
  5. Buck GM, Lynch CD, Stanford JB, et al. Prospective pregnancy study designs for assessing reproductive and developmental toxicants. Environ Health Perspect (2004) 112:79–86.[Web of Science][Medline]
  6. Chapin RE, Buck GM. Our once-in-a-lifetime opportunity. Environ Health Perspect (2004) 112:67–68.[Web of Science][Medline]
  7. Tingen C, Stanford JB, Dunson DB. Methodologic and statistical approaches to studying human fertility and environmental exposure. Environ Health Perspect (2004) 112:87–93.[Web of Science][Medline]
  8. Rockett JC, Buck GM, Lynch CD, et al. The value of home-based collection of biospecimens in reproductive epidemiology. Environ Health Perspect (2004) 112:94–104.[Web of Science][Medline]
  9. Joffe M, Key J, Best N, et al. Studying time to pregnancy by use of a retrospective design. Am J Epidemiol (2005) 162:115–24.[Abstract/Free Full Text]
  10. Baird DD, Weinberg CR, Rowland AS. Reporting errors in time-to-pregnancy data collected with a short questionnaire. Am J Epidemiol (1991) 133:1282–90.[Abstract/Free Full Text]
  11. Zielhuis GA, Hulscher MEJL, Florack EIM. Validity and reliability of a questionnaire on fecundability. Int J Epidemiol (1992) 21:1151–56.[Abstract/Free Full Text]
  12. Joffe M, Villard L, Li Z, et al. Long-term recall of time-to-pregnancy. Fertil Steril (1993) 60:99–104.[Web of Science][Medline]
  13. Joffe M, Villard L, Li Z, et al. A time to pregnancy questionnaire designed for long term recall: validity in Oxford, England. J Epidemiol Community Health (1995) 49:314–19.[Abstract/Free Full Text]
  14. Joffe M. Time trends in biological fertility in Britain. Lancet (2000) 355:1961–65.[CrossRef][Web of Science][Medline]
  15. Joffe M, Li Z. Male and female factors in fertility. Am J Epidemiol (1994) 140:921–29.[Abstract/Free Full Text]
  16. Fahrmeir L, Tutz G. Multivariate Statistical Modelling based on Generalized Linear Models (2001) 2nd edition. New York: Springer Verlag.
  17. Feinberg B. Analysis of Cross-Classified Data (1980) 2nd edition. Cambridge: Massachusetts Institute of Technology Press.
  18. Bonde JP, Joffe M, Sallmén M, et al. Validity issues relating to time-to-pregnancy studies of fertility. Epidemiology (2006) 17:347–49.[CrossRef][Web of Science][Medline]
  19. Karmaus W, Juul S. European infertility and subfecundity group: infertility and subfecundity in population-based samples from Denmark, Germany, Italy, Poland and Spain. Eur J Public Health (1999) 9:229–35.[Abstract/Free Full Text]
  20. Joffe M. Decreased fertility in Britain compared with Finland. Lancet (1996) 347:1519–22.[CrossRef][Web of Science][Medline]
  21. Bolumar F, Olsen J, Boldsen J. The European study group of infertility and subfecundity. Smoking reduces fecundity: a European multicenter study on infertility and subfecundity. Am J Epidemiol (1996) 143:578–87.[Abstract/Free Full Text]
  22. Bolumar F, Olsen J, Rebagliato M, et al. The European study group of infertility and subfecundity. Caffeine intake and delayed conception: a European multicenter study on infertility and subfecundity. Am J Epidemiol (1997) 145:324–34.[Abstract/Free Full Text]
  23. Olsen J, Bolumar F, Boldsen J, et al. The European study group of infertility and subfecundity. Does moderate alcohol intake reduce fecundability? A European multicenter study on infertility and subfecundity. Alcohol Clin Exp Res (1997) 21:206–12.[CrossRef][Web of Science][Medline]
  24. Kolstad HA, Bisanti L, Roeleveld N, et al, Asclepios. Time to pregnancy among male workers of the reinforced plastics industry in Denmark, Italy and The Netherlands. Scand J Work Env Health (2000) 26:353–58.
  25. Joffe M, Bisanti L, Apostoli P, et al, Asclepios. Time to pregnancy and occupational lead exposure. Occup Env Med (2003) 60:752–58.[Abstract/Free Full Text]
  26. Olsen J, Juul S, Basso O. Measuring time to pregnancy: methodological issues to consider. Hum Reprod (1998) 13:1751–53.[Free Full Text]
  27. Fielder HMP, Poon-King CM, Palmer SRP, et al. Health impact assessment of residents living near the Nant-y-Gwyddon landfill site. BMJ (2000) 320:19–22.[Abstract/Free Full Text]
  28. Palmer SR, Dunstan FD, Fielder H, et al. Risk of congenital anomalies after the opening of landfill sites. Environ Health Perspect (2005) 113(10):1362–5.[Web of Science][Medline]
  29. Scheike TH, Carstensen L, Keiding N, et al. Time trends in human fertility: a population based study from Sweden. Epidemiology. [E-pub ahead of print].
  30. Huang Y-L, Batterman S. Residence location as a measure of environmental exposure: a review of air pollution epidemiology studies. J Exp Anal Environ Epidemiol (2000) 10:66–85.[CrossRef][Web of Science][Medline]
  31. Kibble A, Harrison R. Point sources of air pollution. Occup Med (2005) 55:425–31.[Abstract]
  32. Scheike T, Jensen TK. A discrete survival model with random effects: an application to time to pregnancy. Biometrics (1997) 53:318–29.[CrossRef][Web of Science][Medline]
  33. Scheike TH, Petersen JH, Martinussen T. Retrospective ascertainment of recurrent events: an application to time to pregnancy. J Am Stat Assoc (1999) 94:713–25.[CrossRef][Web of Science]
  34. Weinberg CR, Baird DD, Wilcox AJ. Sources of bias in studies of time to pregnancy. Stat Med (1994) 13:671–81.[Web of Science][Medline]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
30/2/178    most recent
fdn005v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Joffe, M.
Right arrow Articles by Palmer, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Joffe, M.
Right arrow Articles by Palmer, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?