2. Materials and Methods
The investigation was planned as a prospective cohort study conducted to determine the daily probability of conception among healthy subjects. The research protocol was reviewed and approved by the Institutional Review Boards of Fondazione Lanza (Padua, Italy) and Georgetown University (Washington D.C., U.S.A.). The study was co-ordinated from the Department of Statistical Sciences of the University of Padua (Padua, Italy).
From 1992 through 1996, 782 women were recruited with the collaboration of seven European centres (Milan, Verona, Lugano, Düsseldorf, Paris, London and Brussels) providing services on fertility awareness and natural family planning. The entry criteria for the subjects were: women experienced in use of a Natural Family Planning method; married or in a stable relationship; between 18th and 40th birthday at admission; having at least had one menses after cessation of breastfeeding or after delivery; not currently taking hormonal medication or drugs affecting fertility. Neither partner could be permanently infertile and both had to be free from any illness that might cause sub-fertility, e. g., endocrine disorders. It was also required that couples did not have the habit of mixing incidences of unprotected and protected intercourse. Women were excluded if any one of the previous criteria was not fulfilled.
Data from an additional 99 subjects were also included retrospectively in view of their relevance to the aims of the study. These data came from a prospective investigation carried out in Auckland, New Zealand, in 1979-85 into the relationship between the interval from intercourse to fertilisation and the sex of the baby conceived. In this study recruitment was made from couples of proven fertility who were contemplating a further pregnancy. For the purpose of timing intercourse, these couples were instructed on how to recognise the fertile period of the menstrual cycle and anticipate ovulation from changes in cervical mucus. The woman partner also recorded her basal body temperature each day. The study design restricted the couples to only one act of intercourse during the fertile phase of the cycle [France et al 1984, France et al 1992]. This requirement, not respected in a few instances, was the probable cause of subjects frequently dropping out of the study if they had not achieved a pregnancy after 3-4 cycles of trying. The resulting short observational period of sexually active non-conception cycles is a plausible source of positive bias in the estimate of the level of daily fecundability in the present study. Therefore, while the Auckland data is of significant value to other aspects of the study, only results from the seven European centres have been used in determining daily probabilities of conception.
A description of the centres, with the names of the local principal investigators, is given in [Note 1].
2.2 Data Collection
In each centre the local principal investigator instructed selected natural family planning teachers about the purpose and the requirements of the study. After completing the instruction phase, the teachers screened and selected the subjects for admission into the study. A woman satisfying all the inclusion criteria was enrolled only after having given written informed consent. In order to ensure complete subject anonymity and confidentiality, each subject was assigned a study number and only the teacher maintained a personal relationship with the subject. The mutual trust established in this relationship was essential to maintaining the collection of quality reliable data of a sensitive personal nature, which encompassed sexual behaviour.
All the charts were periodically sent to the Department of Statistics at the University of Padua, where uniform evaluation for all cases of the recorded basal body temperature (BBT), taken on awakening in the morning before engaging in any activity, was conducted. Coding of mucus typology, in accordance with agreed common rules, was done in the local centres.
2.3 Study Factors
At entry into the study, the following information was collected: the month and year of birth of the woman and of her partner; the number of previous pregnancies, if any; the date of her last delivery (or miscarriage) and of the end of breastfeeding, if relevant; the date of last oral contraceptive pill taken, if any. Subsequently, after the collection of data had begun, it was decided to add the date of marriage for married couples and the sex of any baby conceived and born during the period of the study. This latter information is available for a large proportion of subjects.
In each menstrual cycle the woman was asked to record on a chart the days of her period and of any disturbance such as illness, broken sleep. She was asked to also record her basal body temperature on the chart for as many days as necessary to determine a clear post-ovulatory rise. She was further asked to observe and chart her cervical mucus symptoms daily during the cycle, and to record every episode of coitus, with specification of whether it was unprotected or protected (barrier methods, withdrawal, ...). Cycles in which even a single act of protected intercourse or of simple genital contact occurred were excluded from the analysis. The reliability of the information recorded of acts of intercourse was checked by the teacher in discussion with subjects at the end of each cycle. The importance of continuing to keep the record chart when subjects were trying to conceive a pregnancy was emphasised.
Charts were regularly collected by the teacher concerned. Following review at the local centre and scoring of the cervical mucus symptoms according to the common rules agreed for the study (Table 1), the charts were sent to the co-ordinating investigators in Padua for processing and entry into the data base [Note 2].
A menstrual cycle was characteristically defined as the interval in days from the beginning of one period of vaginal bleeding until the commencement of the next, where day 1 was the first day of fresh red bleeding, excluding any preceding days with spotting.
The "three over six rule" was used to determine the BBT shift, defined as follows: the first time in the cycle that three temperatures were recorded all of which were above the level of the immediately preceding six daily temperature recordings. Such a rule has been shown to perform well in predicting the start of the infertile period following ovulation [Marshall 1968]. Exceptions to the rule were permitted: a) if there was one "spike" temperature among the six at the lower level (a spike temperature was defined as a temperature which was 0.2 centigrades or more above both its immediate neighbouring temperatures); b) or, in a cycle in which the impact of illness or other disturbances could be discounted, if there were at least six lower temperatures recorded before the upward shift. In analyses in which the BBT rise was used as a conventional indicator for timing ovulation, the last day of lower temperatures was designated as day 0, the "BBT reference day", to which all preceding and following days were scaled according to their distance by integer numbers.
The cervical mucus peak day was defined as the last day with best quality mucus, in a specific cycle of the woman, by sensation or appearance, known retrospectively. This peak day was taken as "Mucus reference day" and identified as day 0.
A conception was assumed in the presence of a pregnancy going on at 60 days from the onset of the last menses or when before that term a miscarriage was clinically detected.
2.5 Statistical Analysis
All the following statistical analyses, performed in the Department of Statistical Sciences, at the University of Padua, were limited to cycles in which ovulation occurred, or at least appeared to occur, and BBT reference day and/or mucus reference day was identified.
We first chose the window of potential fertility to be the series of days relative to the identified day of ovulation such that a cycle without intercourse during these days never resulted in a pregnancy. Daily estimates of probability of conception (a simple division: day by day, number of pregnancies/number of acts of intercourse) were then calculated using cycles with only one intercourse during the putative window. Since the act responsible for conception was unknown in cycles with more than one act of intercourse in the fertile interval, a more sophisticated procedure was needed to estimate globally the daily fecundability in the general case with one or more than one act of intercourse in the window. For this purpose the Schwartz model [Schwartz, MacDonald, and Heuchel 1980] (see [2.5.1]), which is an extension of the one suggested by Barrett and Marshall [Barrett and Marshall 1969], was used. For each cycle, the probability of no conception is the probability the cycle is not viable plus the probability the cycle is viable and none of the intercourse acts result in successful fertilisation and survival to detection.
Inference was based on the likelihood: (i) parameter estimates were obtained by maximum likelihood, (ii) confidence intervals were then computed for each parameter of interest using the profile log-likelihood [Clayton and Hills 1993] and (iii) likelihood ratio tests were used to assess the significance of selected covariates.
Descriptive analysis was performed using SAS (see http://www.sas.com). R (http://www.r-project.org) was used to fit the Schwartz et al. model to the data. Functions and scripts are available upon request from the authors.
2.5.1 The Schwartz Model [Schwartz, MacDonald, and Heuchel 1980]
For each cycle, the observed outcome (conception/non conception) can be modelled as a Bernoulli random variable with parameter (the probability of success, i.e., the fecundability) that depends on the number and timing of the intercourse events.
Schwartz et al. [Schwartz, MacDonald, and Heuchel 1980] write fecundability as the product of three probabilities:
where P0 = pr (that a fertilizable ovule is produced)
Pv = pr (that the conceptus stays alive for at least six weeks | fertilized ovule)
Daily Fecundability: First Results from a New Data Base
Bernardo Colombo, Guido Masarotto
© 2000 Max-Planck-Gesellschaft ISSN 1435-9871