Semiparametric modeling of grouped current duration data with preferential reporting

Trends in time-to-pregnancy in the USA: 2002 to 2017

STUDY QUESTION

Has there been there a temporal change in time-to-pregnancy (TTP) in the USA.

SUMMARY ANSWER

Overall, TTP was stable over time, but a longer TTP for women over 30 and parous women was identified.

WHAT IS KNOWN ALREADY

Fertility rates in the USA have declined over the past several years. Although these trends have been attributed to changing reproductive intentions, it is unclear whether declining fecundity (the biologic ability to reproduce measured by TTP in the current report) may also play a role. Indeed, trends based on declining sperm quality and higher utilisation of infertility treatment suggest fecundity may be falling.

STUDY DESIGN, SIZE, DURATION

This cross-sectional survey data from the National Survey of Family Growth was administered from 2002 to 2017. The surveys are based on nationally representative samples of reproductive-aged women in the USA. Interviews were conducted in person or through computer-assisted self-administration of sensitive questions.

PARTICIPANTS/MATERIALS, SETTING, METHODS

The study included women who self-reported time spent trying to become pregnant allowing utilisation of the current duration approach to estimate the total duration of pregnancy attempt (i.e. TTP). In all, 1202 participants were analysed over each study period. To estimate a TTP distribution overall and by parity, we used a piecewise constant proportional hazards model that accounts for digit preference. Accelerated-failure-time regression models, which were weighted to account for the sampling design, were used to estimate time ratios (TRs). Models were adjusted for age, BMI, race, education, relationship status, parity, pelvic inflammatory disease treatment and any reproductive problems.

MAIN RESULTS AND THE ROLE OF CHANCE

Of the participants analysed, the average age was 31.8 and BMI was 28.6, which was similar across the survey periods. Relationship status was the only demographic characteristic that changed over time. All other variables remained constant across the study periods. Overall, TRs comparing TTP between 2002 and 2017 increased slightly (TR: 1.02, 95% CI: 0.99, 1.04). When stratified by parity, parous women had a longer TTP over the later years of the study (TR: 1.04, 95% CI: 1.01, 1.06). TTP remained constant for nulliparous women. Similarly, TTP also increased over time for women over age thirty (TR: 1.02, 1.00, 1.05) but not for women under age thirty.

LIMITATIONS, REASONS FOR CAUTION

Small changes in data collection over time may have impacted the findings. We accounted for this in sensitivity analyses using imputed data. Overall, TRs were slightly attenuated using the imputed data, but represented similar patterns to the original data. Results for parous women and women over 30 remained consistent in the sensitivity analyses.

WIDER IMPLICATIONS OF THE FINDINGS

Consistent with reports of falling fertility rates and sperm counts, this study suggests parous and older couples in the USA may be taking longer to become pregnant. Although trends were suggestive of a small overall increase in TTP, particularly for parous women and women over age thirty, additional data are needed to attempt to understand these trends given the societal, economic and public health implications related to fecundity.

STUDY FUNDING/COMPETING INTEREST(S)

Funding was provided by National Institutes of Health grant R03HD097287 to A.C.M. There are no competing interests.

TRIAL REGISTRATION NUMBER

N/A.

Parametric modelling of prevalent cohort data with uncertainty in the measurement of the initial onset date

In prevalent cohort studies with follow-up, if disease duration is the focus, the date of onset must be obtained retrospectively. For some diseases, such as Alzheimer's disease, the very notion of a date of onset is unclear, and it can be assumed that the reported date of onset acts only as a proxy for the unknown true date of onset. When adjusting for onset dates reported with error, the features of left-truncation and potential right-censoring of the failure times must be modeled appropriately. Under the assumptions of a classical measurement error model for the onset times and an underlying parametric failure time model, we propose a maximum likelihood estimator for the failure time distribution parameters which requires only the observed backward recurrence times. Costly and time-consuming follow-up may therefore be avoided. We validate the maximum likelihood estimator on simulated datasets under varying parameter combinations and apply the proposed method to the Canadian Study of Health and Aging dataset.

Prevalent cohort studies and unobserved heterogeneity

Consider lifetimes originating at a series of calendar times [Formula: see text]. At a certain time [Formula: see text] a cross-sectional sample is taken, generating a sample of current durations (backward recurrence times) of survivors until [Formula: see text] and a prevalent cohort study consisting of survival times left-truncated at the current durations. A Lexis diagram is helpful in visualizing this situation. Survival analysis based on current durations and prevalent cohort studies is now well-established as long as all covariates are observed. The general problems with unobserved covariates have been well understood for ordinary prospective follow-up studies, with the good help of hazard rate models incorporating frailties: as for ordinary regression models, the added noise generates attenuation in the regression parameter estimates. For prevalent cohort studies this attenuation remains, but in addition one needs to take account of the differential selection of the survivors from initiation [Formula: see text] to cross-sectional sampling at [Formula: see text]. This paper intends to survey the recent development of these matters and the consequences for routine use of hazard rate models or accelerated failure time models in the many cases where unobserved heterogeneity may be an issue. The study was inspired by concrete problems in the study of time-to-pregnancy, and we present various simulation results inspired by this particular application.

Fecundity and Infertility Among Women with Disabilities in the United States

Background: Knowledge of fecundity and infertility in women with a disability (WWD) is limited. This study aims to compare the fecundity and infertility experiences of women with a self-identified disability, in domains of sensory (hearing and vision loss), cognitive (difficulty concentrating, remembering, and making decisions), and physical (difficulty walking or climbing stairs) disabilities, with those without a self-identified disability. Materials and Methods: Cross-sectional survey data from the National Survey of Family Growth (NSFG), 2011-2015, were analyzed. A final analytic sample of 383 women from the NSFG, aged 18-44 years, was included to study the fecundity rates of WWD and women without a disability, using the current duration approach. Results: Fecundity hazard ratios (FHRs) and the proportion of women experiencing infertility after 12 months of actively attempting pregnancy were reported for the comparison group and by each disability type. Women with a self-identified cognitive disability experienced significant decreases in fecundity (FHR = 0.56; 95% confidence interval [CI]: 0.30-0.88) when adjusting for age, education level, parity, living in a metropolitan area, and infertility services. Furthermore, the estimated proportion of infertile women without a disability was 0.38 (95% CI: 0.23-0.62) versus 0.51 (95% CI: 0.34-0.72) for women with a self-identified cognitive disability. Conclusions: Women with a self-identified cognitive disability experienced significant decreases in fecundity.

Estimating infertility prevalence in low-to-middle-income countries: an application of a current duration approach to Demographic and Health Survey data

STUDY QUESTION

Can infertility prevalence be estimated using a current duration (CD) approach when applied to nationally representative Demographic and Health Survey (DHS) data collected routinely in low- or middle-income countries?

SUMMARY ANSWER

Our analysis suggests that a CD approach applied to DHS data from Nigeria provides infertility prevalence estimates comparable to other smaller studies in the same region.

WHAT IS KNOWN ALREADY

Despite associations with serious negative health, social and economic outcomes, infertility in developing countries is a marginalized issue in sexual and reproductive health. Obtaining reliable, nationally representative prevalence estimates is critical to address the issue, but methodological and resource challenges have impeded this goal.

STUDY DESIGN, SIZE, DURATION

This cross-sectional study was based on standard information available in the DHS core questionnaire and data sets, which are collected routinely among participating low-to-middle-income countries. Our research question was examined among women participating in the 2013 Nigeria DHS (n = 38 948). Among women eligible for the study, 98% were interviewed.

PARTICIPANTS/MATERIALS, SETTING, METHODS

We applied a CD approach (i.e. current length of time-at-risk of pregnancy) to estimate time-to-pregnancy (TTP) and 12-month infertility prevalence among women ‘at risk’ of pregnancy at the time of interview (n = 7063). Women who were 18–44 years old, married or cohabitating, sexually active within the past 4 weeks and not currently using contraception (and had not been sterilized) were included in the analysis. Estimates were based on parametric survival methods using bootstrap methods (500 bootstrap replicates) to obtain 95% CIs.

MAIN RESULTS AND THE ROLE OF CHANCE

The estimated median TTP among couples at risk of pregnancy was 5.1 months (95% CI: 4.2–6.3). The estimated percentage of infertile couples was 31.1% (95% CI: 27.9–34.7%)—consistent with other smaller studies from Nigeria. Primary infertility (17.4%, 95% CI: 12.9–23.8%) was substantially lower than secondary infertility (34.1%, 95% CI: 30.3–39.3%) in this population. Overall estimates for TTP >24 or >36 months dropped to 17.7% (95% CI: 15.7–20%) and 11.5% (95% CI: 10.2–13%), respectively. Subgroup analyses showed that estimates varied by age, coital frequency and fertility intentions, while being in a polygynous relationship showed minimal impact.

LIMITATIONS, REASONS FOR CAUTION

The CD approach may be limited by assumptions on when exposure to risk of pregnancy began and methodologic assumptions required for estimation, which may be less accurate for particular subgroups or populations. Unrecognized pregnancies may have also biased our findings; however, we attempted to address this in our exclusion criteria. Limiting to married/cohabiting couples may have excluded women who are no longer in a relationship after being blamed for infertility. Although probably rare in this setting, we lack information on couples undergoing infertility treatment. Like other TTP measurement approaches, pregnancies resulting from contraceptive failure are not included, which may bias estimates.

WIDER IMPLICATIONS OF THE FINDINGS

Nationally representative estimates of TTP and infertility based on a clinical definition of 12 months have been limited within developing countries. This approach represents a pragmatic advance in our ability to measure and monitor infertility in the developing world, with potentially far-reaching implications for policies and programs intended to address reproductive health.

STUDY FUNDING/COMPETING INTERESTS

There are no competing interests and no financial support was provided for this study. Financial support for Open Access publication was provided by the World Health Organization.

Proximity and gravity: modeling heaped self-reports

Self-reported daily cigarette counts typically exhibit a preponderance of round numbers, a phenomenon known as heaping or digit preference. Heaping can be a substantial nuisance, as scientific interest lies in the distribution of the underlying true values rather than that of the heaped data. In principle, we can estimate parameters of the underlying distribution from heaped data if we know the conditional distribution of the heaped count given the true count, denoted the heaping mechanism (analogous to the missingness mechanism for missing data). In general, it is not possible to estimate the heaping mechanism robustly from heaped data only. A doubly-coded smoking cessation trial data set that includes daily cigarette count as both a conventional heaped retrospective recall measurement and a precise instantaneous measurement offers the rare opportunity to directly estimate the heaping mechanism. We propose a novel model that describes the conditional probability of the self-reported count as a function of its proximity to the truth and its intrinsic attractiveness, denoted its gravity. We apply variations of the model to the cigarette count data, illuminating the cognitive processes that influence self-reporting behaviors. The principal application of the model will be to enabling the correct analysis of heaped-only data sets. Copyright © 2017 John Wiley & Sons, Ltd.

SEX, LIES AND SELF-REPORTED COUNTS: BAYESIAN MIXTURE MODELS FOR HEAPING IN LONGITUDINAL COUNT DATA VIA BIRTH-DEATH PROCESSES

Surveys often ask respondents to report non-negative counts, but respondents may misremember or round to a nearby multiple of 5 or 10. This phenomenon is called heaping, and the error inherent in heaped self-reported numbers can bias estimation. Heaped data may be collected cross-sectionally or longitudinally and there may be covariates that complicate the inferential task. Heaping is a well-known issue in many survey settings, and inference for heaped data is an important statistical problem. We propose a novel reporting distribution whose underlying parameters are readily interpretable as rates of misremembering and rounding. The process accommodates a variety of heaping grids and allows for quasi-heaping to values nearly but not equal to heaping multiples. We present a Bayesian hierarchical model for longitudinal samples with covariates to infer both the unobserved true distribution of counts and the parameters that control the heaping process. Finally, we apply our methods to longitudinal self-reported counts of sex partners in a study of high-risk behavior in HIV-positive youth.

Female exposure to phenols and phthalates and time to pregnancy: the Maternal-Infant Research on Environmental Chemicals (MIREC) Study

OBJECTIVE

To assess the potential effect of bisphenol A (BPA), triclosan (TCS), and phthalates on women's fecundity, as measured by time to pregnancy (TTP).

DESIGN

Pregnancy-based retrospective TTP study.

SETTING

Not applicable.

PATIENT(S)

A total of 2,001 women during the first trimester of pregnancy recruited between 2008 and 2011 (the Maternal-Infant Research on Environmental Chemicals (MIREC) Study), with 1,742 women included in the BPA analysis, 1,699 in the TCS analysis, and 1,597 in the phthalates analysis.

INTERVENTION(S)

None.

MAIN OUTCOME MEASURE(S)

Fecundability odds ratios (FORs) estimated using the Cox model modified for discrete time data.

RESULT(S)

The BPA concentrations were not statistically significantly associated with diminished fecundity either in crude or adjusted models. Women in the highest quartile of TCS (>72 ng/mL) had evidence of decreased fecundity (FOR 0.84; 95% confidence interval, 0.72-0.97) compared with the three lower quartiles as the reference group. Exposure to phthalates was suggestive of a shorter TTP, as indicated by FORs greater than 1, although the 95% confidence interval always included 1.

CONCLUSION(S)

Elevated TCS exposure may be associated with diminished fecundity. BPA and phthalates showed no negative impact; on the contrary, some phthalates might be associated with a shorter time to pregnancy. A major limitation of the study was that only one measurement of exposure was available for each woman after conception. Further research is necessary to test these findings.