Content uploaded by Joseph B Stanford
Author content
All content in this area was uploaded by Joseph B Stanford on Feb 13, 2019
Content may be subject to copyright.
Pregnancy Probabilities During Use of the Creighton
Model Fertility Care System
Margaret P. Howard, CNFPE; Joseph B. Stanford, MD, MSPH
Objective: To evaluate pregnancy probabilities during
use of the Creighton Model Fertility Care System (CrMS).
Design: Couples who began use of the CrMS were en-
tered into this observational cohort study. Follow-up in-
cluded detailed reviews of use of the CrMS. Pregnancy
probabilities were calculated with both net and gross life-
table analysis through 18 months.
Setting: A natural family planning service delivery pro-
gram based at an urban hospital in Houston, Tex.
Subjects: A group of 701 couples who received instruc-
tion in the CrMS were entered into the study. Most couples
(93%) were engaged or married. Most women were white
(83%), between the ages of 20 and 34 years (88%), and
college graduates (58%).
Main Outcome Measure: Pregnancies were classi-
fied based on a detailed evaluation involving the preg-
nant woman (usually with her partner).
Results: At 12 months, the following net pregnancy prob-
abilities were found per 100 couples: method-related preg-
nancies, 0.14; pregnancies caused by user and/or teacher
error, 2.72; pregnancies caused by achieving-related be-
havior (genital contact during a time known to be fer-
tile), 12.84; unresolved pregnancies, 1.43; and total preg-
nancies, 17.12. Pregnancy probabilities were similar when
stratified by the following reproductive categories: un-
complicated regular cycles, long cycles, discontinuing oral
contraceptives, breastfeeding, and other.
Conclusions: Pregnancy probabilities of the CrMS com-
pare favorably with those of other methods of family plan-
ning. Most pregnancies result from genital contact dur-
ing a known fertile time. Women need not have regular
cycles to use the CrMS successfully.
Arch Fam Med. 1999;8:391-402
PERIODIC ABSTINENCE to avoid
pregnancy has been in use
since “calendar rhythm” was
developed by Ogino and
Knaus1in the 1930s. Unfor-
tunately, the effectiveness of calendar
rhythm to avoid pregnancy is, under opti-
mal circumstances, only 85% to 94%.2In the
past few decades, methods of fertility aware-
ness have been developed, including the
ovulation method, which is based solely on
women’s observations of the vaginal dis-
charge of cervical mucus, and the sympto-
thermal method, which combines mucus
observations with the measurement of basal
body temperature and some calendar cal-
culations. These methods can be referred to
collectively as modern methods of natural
family planning (NFP). Because modern
methods of NFP are based on actual physi-
ologic signs of fertility and infertility rather
than calendar calculations, there are com-
pelling reasons to believe that modern NFP
can be more effective than calendar rhythm.3
Nevertheless, the effectiveness of mod-
ern methods of NFP to avoid pregnancy is
still a subject of controversy.4-6 Effective-
ness studies of the ovulation method have
yielded method-related pregnancy rates of
0 to 11.3 pregnancies per 100 woman-
years and total pregnancy rates of 0.4 to
39.7.7,8 For the symptothermal method,
method-related pregnancy rates have been
reported ranging from 0 to 13.1 and total
pregnancy rates ranging from 3.3 to 34.4.8,9
It is apparent that there is wide varia-
tion in pregnancy rates among NFP stud-
ies. Some factors contributing to this varia-
tion are common to all family planning
studies, including differences in the de-
mographics and potential for fertility
among the populations studied and dif-
ferences in time contributed by study sub-
jects.10 Also, while life-table analysis is a
more appropriate method for analyzing
pregnancies, most NFP studies have re-
ported Pearl rates. Life-table probabili-
ties are based on the proportion of sub-
This article is also available on our
Web site: www.ama-assn.org/family.
ORIGINAL CONTRIBUTION
From Fertility Care Services,
Covenant Health System,
Lubbock, Tex (Ms Howard);
and the Department of Family
and Preventive Medicine,
University of Utah, Salt Lake
City (Dr Stanford). Ms Howard
was formerly with the
Department of Natural Family
Planning, St Joseph’s Hospital,
Houston, Tex.
ARCH FAM MED/ VOL 8, SEP/OCT 1999
391
©1999 American Medical Association. All rights reserved.
jects who become pregnant over a specified period,
whereas Pearl rates are calculated based on the total num-
ber of cycles or months in a study.11 Pearl rates can vary
substantially based on the time each subject is in the study.
A detailed review of Pearl rates and life-table probabili-
ties designed for clinicians is available elsewhere.12
Other important factors that contribute to varia-
tion in results among NFP studies stem from issues that
are unique to NFP as opposed to other methods of fam-
ily planning. There are 3 areas of particular importance:
First, there is variability in the quality and standardiza-
tion of training for NFP teachers and for teaching pro-
cedures, causing variability in the type of instruction re-
ceived by NFP users (study subjects). The successful use
of NFP, more than other methods of family planning, re-
quires reliable instruction about proper use, including
appropriate observation of the symptoms and signs of fer-
tility, accurate interpretation of these signs, and sup-
port for questions and concerns. Second, the exclusion
by nearly all NFP studies of women who do not have regu-
lar menstrual cycles makes it impossible to generalize re-
sults beyond women who have regular menses. Finally,
NFP can be used either to avoid or to achieve preg-
nancy11,13; unlike other methods of family planning, the
choice to achieve a pregnancy does not presuppose dis-
continuation of the use of NFP. Hence, the evaluation
of the choices of NFP users is a centrally important al-
though controversial issue.6,14-20 In NFP studies, most preg-
nancies result from intercourse on a day that has been
identified as fertile by the method under study.6,21-23
The Creighton Model Fertility Care System (CrMS)
system was developed as a standardized version of the
METHODS
This study is based on clients served by the St Joseph Hos-
pital Natural Family Planning service program in Hous-
ton, Tex, from 1983 to 1989. During this time, the pre-
dominant type of NFP taught in this program was the CrMS.
All new users of the CrMS were considered eligible for this
study. Users were defined as clients who attended at least
1 follow-up session to learn the CrMS. Excluded were
couples with a history of infertility, women who were preg-
nant at the time they began use of the CrMS, and women
who were not genitally active at the time they began use of
the CrMS. No attempt was made to select for couples who
expressed strong motivation to avoid pregnancy for a time.
All subjects were entered into the study at the time
the woman began charting her cycle. Unlike some other
NFP studies,29-31 there was no preceding or separate learn-
ing phase for this study. Each couple contributed data un-
til they had completed 18 months of follow-up, the woman
became pregnant, or the couple left the study for some other
reason (as detailed below). At entry to the study, all women
were classified into one of the following reproductive sta-
tus categories: (1) uncomplicated regular cycles, defined
as 21 to 38 days32,33; (2) long cycles, defined as usually more
than 38 days; (3) age 40 years or older; (4) having discon-
tinued use of oral contraceptives within the past year; (5)
totally breastfeeding; (6) breastfeeding/weaning (defined
as having begun any type of supplemental feeding while
still breastfeeding); (7) within 3 menstrual cycles after birth
and not breastfeeding; or (8) within 1 cycle after an abor-
tion (induced or spontaneous). While these categories are
not necessarily mutually exclusive, for the purposes of this
study they were treated as such; each woman was placed
into the highest-numbered category that applied to her. For
example, a woman aged 41 years who stopped taking oral
contraceptives 11 months before study entry and was ex-
clusively breastfeeding a baby born 1 month before study
entry would have been listed as totally breastfeeding at the
beginning of this study.
Detailed description of the CrMS instruction has been
given elsewhere.25-28,34 After couples attended a group intro-
ductory session (occasionally the woman alone attended the
introductory session), they began immediately to chart the
woman’s daily observations of vaginal discharge. Accord-
ing to standard CrMS protocol, women were asked
to wipe across the vulva with flat white toilet tissue each time
they use the bathroom to check for any discharge and to re-
cord its characteristics (degree of stretch, color, and whether
it felt lubricative) according to a standardized recording sys-
tem. Women were taught to make external observations only
(ie, to not insert fingers into the vagina). Couples were in-
structed to abstain from genital contact for the first month
of observations, but not all couples followed this instruc-
tion. Two weeks after the introductory session, couples (or
sometimes women alone) received their first follow-up visit
by meeting individually with the Creighton Model teacher,
called an NFP practitioner. Over the following year, up to 8
follow-up visits took place. Generally these lasted 45 to 60
minutes each. During these visits, the NFP practitioner re-
viewed the couple’s understanding of the CrMS using a stan-
dardized list of assessment items, gave instruction tailored
to the individual circumstances, and reviewed all charting
completed by the woman/couple for accuracy, complete-
ness, and understanding of the times of fertility and infer-
tility. In general, the time of fertility begins at the first ap-
pearance of mucus discharge and ends at the end of the fourth
day after the peak day, defined in the CrMS as the last day
mucus is clear, stretches more than 1 inch, or feels lubrica-
tive. The average duration of mucus discharge is 5 to 6 days.
The CrMS also has special instructions to deal with breast-
feeding, other oligo-ovulatory states, and chronic vaginal dis-
charges. These are described in more detail elsewhere.25-27
Couples were explicitly asked at follow-up visits about
any use of withdrawal, use of barrier methods of contra-
ception, or contraceptive hormone use; those who used any
of these methods were excluded from the study. These cir-
cumstances were uncommon: fewer than 10 couples were
excluded for these reasons. We also reviewed any evi-
dence of pregnancy, as suggested by prolonged postovu-
latory phase of a menstrual cycle. Follow-up beyond 12
months was accomplished mainly by telephone contact.
Throughout instruction, couples were advised that the
CrMS can be used to achieve pregnancy as well as to avoid
pregnancy. It was emphasized that having genital contact
on a “day of fertility” has a high probability of resulting in
pregnancy and that this would be considered a free and re-
sponsible choice to potentially achieve pregnancy. Instruc-
tion was given during the initial CrMS session and rein-
forced during follow-up visits that “there is no such thing
as taking a chance with the ovulation method. You will know
whether you are fertile or infertile on any given day.”
ARCH FAM MED/ VOL 8, SEP/OCT 1999
392
©1999 American Medical Association. All rights reserved.
ovulation method originally described by Billings and Bill-
ings.24 The CrMS is characterized by rigorous profes-
sional training for teachers and standardized instruc-
tion procedures for clients learning the method, including
a uniform recording system for vaginal discharges.25-28 Fre-
quent follow-up and support is provided routinely dur-
ing the first year of use of the CrMS. This report evalu-
ates the pregnancy probabilities among couples served
by a large US center providing CrMS instruction.
RESULTS
The demographic characteristics of the study sample are
given in Table 1. In general, this was a well-educated, rela-
tively affluent group of couples. About three fourths of
women were of the Roman Catholic faith and most were
either married or engaged. Almost all women (96%) had
used the birth control pill at some point in their lives, and
almost one third (29.1%) had used it immediately prior to
beginning use of the CrMS. Slightly less than half of women
(46.3%) were classified as having uncomplicated regular
menstrual cycles (cycles lasting 21-38 days), without hav-
ing any of the noted additional factors affecting fertility.
A group of 701 clients contributed 6947.5 couple-
months to this study over 18 months of follow-up. The
net total pregnancy probabilities per 100 couples were
17.12 at 12 months and 21.26 at 18 months (Table 2).
The net non–pregnancy-related probabilities of leaving
the study were 27.39 at 12 months and 34.09 at 18 months
(Table 3). Thus, the all-cause net probabilities of leav-
ing this study (but not necessarily stopping use of the
CrMS) were 44.51 at 12 months and 55.35 at 18 months.
Importantly, the informed use of a fertile day for genital
contact by a couple was not considered to be evidence of
that couple necessarily “planning” or “intending” a preg-
nancy at a conscious level, but rather simply an acknowl-
edgment that the couple had chosen to engage in “achieving-
related behavior” (behavior that they knew to be likely to
cause pregnancy). Assessment of these dynamics of use of
the CrMS was accomplished by analyzing pregnancies in
a fashion consistent with the teaching and use of the CrMS.
Whenever a pregnancy occurred, a detailed evalua-
tion was done within the first 3 months of pregnancy. This
evaluation comprised a review of the circumstances sur-
rounding the pregnancy based on the daily CrMS record
from the woman/couple, the routine teaching documen-
tation of the CrMS practitioner, and a detailed interview
with the woman/couple, usually in person, but sometimes
by telephone. Based on this evaluation, all pregnancies were
classified into 1 of the following categories: method-
related pregnancies (occurring despite correct use of the
CrMS to avoid pregnancy), pregnancies related to error in
application of the CrMS by the woman/couple, pregnan-
cies related to error in teaching the CrMS on the part of
the teacher, pregnancies related to error on the part of both
the woman/couple and the teacher, pregnancies caused by
achieving-related behavior, or unresolved pregnancies (ie,
insufficient information was available to classify the preg-
nancy into 1 of the first 5 categories). All categories were
mutually exclusive (eg, if there was evidence of error by
both the teacher and the user, the pregnancy was classi-
fied as teacher and woman/couple error and not as woman/
couple error or teacher error). In all cases, the pregnancy
classification was discussed with the woman/couple in-
volved. Disagreement from the woman/couple regarding
the classification was rare, but if there was disagreement,
a second pregnancy evaluation was done by another indi-
vidual. Evaluations for all pregnancies in this study were
reviewed by one of us (M.P.H.).
We also grouped pregnancies together to estimate prob-
abilities for the following clinically relevant outcomes:
method-related pregnancy, avoiding-related pregnancy, and
extended-use pregnancy. Method-related pregnancy, some-
times referred to as method failure, denotes pregnancies
that occurred despite exactly correct use of a family plan-
ning method to avoid pregnancy, as objectively defined by
the investigators.12 We used the term avoiding-related preg-
nancy to denote pregnancies that occur despite correct use
of the CrMS to avoid pregnancy, as understood by the
couple. Thus, avoiding-related pregnancy included preg-
nancies that occurred because of errors (by the couple or
teacher) in addition to method-related pregnancies. We used
different assumptions to distribute unresolved pregnan-
cies among the other pregnancy categories to generate a
range of estimates for the probabilities of method-related
and avoiding-related pregnancy. For extended-use preg-
nancy,12 we included all pregnancies that occurred during
a given time after beginning use of the CrMS; we also made
further assumptions about pregnancies that may have oc-
curred beyond our knowledge among those who were
dropped from the study for reasons other than pregnancy.
This study involved review of the CrMS records as they
are routinely kept by CrMS practitioners (teachers). No ad-
ditional information was collected from couples, and all in-
formation was analyzed in a fashion that did not identify
individual couples; hence, informed consent of the sub-
jects was not sought for this study. The study protocol was
reviewed and approved as exempt by the University of Utah
Institutional Review Board for Human Subjects.
For each couple, the length of use of the CrMS prior to
leaving the study (because of either pregnancy or other rea-
sons) was calculated in ordinal months. The menstrual cycle
was not used as a unit of measurement for this study. Those
who were dropped from the study were then tallied by cat-
egory and by time contributed to the study, as described by
Hilgers,35 as adapted from Tietze and Lewit.36,37 From these
tabulations, net probabilities of leaving the study were cal-
culated for each category. A net life table gives the cumula-
tive probability at a specified follow-up time (such as 1 year)
of a subject exiting the study for each of a list of all possible
reasons. Net life-table probabilities have been widely used
to evaluate both contraceptives and NFP, but they are not
ideal for comparing different studies, because discontinua-
tion categories unrelated to pregnancy can influence preg-
nancy probabilities.11 Therefore, we also calculated gross life-
table probabilities.38 Gross life-table probabilities give an
estimate of the probability of pregnancy independent of any
other competing category of leaving the study and hence can
be used to compare pregnancy probabilities directly be-
tween different studies. However, gross probabilities have
not been commonly reported in family planning studies. We
reported our results in terms of probability of pregnancy rather
than failure, since the term pregnancy is more precise for
evaluating family planning methods.39
ARCH FAM MED/ VOL 8, SEP/OCT 1999
393
©1999 American Medical Association. All rights reserved.
Pregnancy probabilities by pregnancy classifica-
tion and by month of study are presented in Table 2. Most
pregnancies were caused by achieving-related behavior.
A small proportion of pregnancies were caused by user
error, teacher error, or a combination of user and teacher
errors (net, ,3.0 per 100 couples for all errors com-
bined). A few pregnancies were unresolved because of
insufficient data. There was only 1 method-related preg-
nancy. Table 2 presents both net and gross probabili-
ties. While the same trends are observable for each mea-
sure, the gross probabilities, as expected, are somewhat
higher.
Table 4 reports pregnancy probabilities at 12
months, stratified by the reproductive status of the woman
at entry into the study, using the same reproductive sta-
tus categories reported in Table 1. In this analysis, to-
tally breastfeeding and breastfeeding/weaning were com-
bined into the single category of breastfeeding. Because
of small numbers of couples in some strata, the follow-
ing categories were combined into a single category des-
ignated other: age 40 years or older, within 3 menstrual
cycles after birth, not breastfeeding, and within 1 cycle
after an abortion. The single method-related pregnancy
in the study occurred in a woman who had stopped tak-
ing oral contraceptives in the past year. Overall, preg-
nancy probabilities were remarkably similar between re-
productive categories. Pregnancy probabilities were
notably higher in the breastfeeding category. The low-
est total pregnancy probability at 12 months was in the
uncomplicated regular cycles category; this reflected a
lower probability of pregnancy because of achieving-
related behavior in this category.
Table 5 reports estimates of net probability for
method-related pregnancy (pregnancy that occurred
while the method was being used correctly), avoiding-
related pregnancy (incorporating errors in use of
the CrMS while it is still being used by the couple in a
way in which they expect to avoid pregnancy), and
extended-use pregnancy (incorporating all pregnancies
and estimates of pregnancies for women who left the
study for reasons not known to be related to pregnancy).
For method-related and avoiding-related pregnancies,
estimates are provided based on actual pregnancy prob-
abilities, with differing assumptions regarding the reso-
lution of unresolved pregnancies.35 For extended-use
pregnancies, a low estimate included pregnancies
known to have occurred in the study, and also assumed
that there was an immediate probability of pregnancy of
0.05 per 100 couples among those who were dropped
from the study for reasons not known to be related to
pregnancy. Similarly, the high estimate for extended-use
pregnancies assumes an immediate probability of preg-
nancy of 0.2 among those who were dropped from the
study for reasons not known to be related to pregnancy.
This percentage was chosen arbitrarily to exceed some-
what the overall 12-month pregnancy probability for
this study. At 12 months, across the assumptions used
for these estimates, the probability of method-related
pregnancy ranged from 0.16 to 1.57; the probability of
avoiding-related pregnancy ranged from 3.11 to 4.28,
and the probability of extended-use pregnancy ranged
from 18.47 to 22.51.
Table 1. Demographic and Reproductive Characteristics
of the Study Sample*
Clients, %
Age of women, y
15-19 2.6
20-24 25.7
25-29 40.3
30-34 21.9
35-39 7.5
$40 2.2
Average household income, $/y
,20 000 19.2
20 000-25 000 11.8
25 000-30 000 13.7
30 000-35 000 11.9
35 000-40 000 10.4
.40 000 33.1
Highest level of education
,High school graduate 2.2
High school graduate 12.6
Some college or vocational 26.7
College graduate 46.2
Graduate school 12.2
Ethnicity of women
White 83.2
African American 2.2
Hispanic American 11.9
Other 2.6
Religion of women
Roman Catholic 75.8
Protestant 13.3
Jewish 1.1
Other 9.0
None 1.2
Marital status
Single 5.6
Engaged 25.9
Married 66.9
Divorced or separated 1.6
Method of family planning used prior to study
Oral contraceptives 29.1
Intrauterine device 1.6
Barrier methods† 18.2
NFP methods‡ 13.8
Breastfeeding 1.4
Withdrawal 1.1
Other 1.9
None 32.7
Age at menarche, y
,10 4.4
11-12 40.3
13-14 44.1
$15 11.0
Reproductive status of woman§
Regular menstrual cycles (21-38 d) 46.3
Long cycles (usually .38 d) 4.9
Age $40 y 1.7
Discontinued oral contraceptive use within past year 29.7
Totally breastfeeding 4.4
Breastfeeding/weaning 10.6
Within 3 menstrual cycles after birth, not breastfeeding 1.6
Within 1 cycle after an abortion (induced or spontaneous) 0.7
*
Percentages may not add up to 100% because of rounding. NFP
indicates natural family planning.
†
Includes condom, spermicide, and/or diaphragm.
‡
Includes calendar rhythm, basal body temperature, symptothermal,
ovulation method, and self-devised natural methods.
§
Although these categories are not necessarily mutually exclusive, for the
purposes of this study they were treated as such. Each woman was placed
into the last category for which she qualified; each was classified in only
one category. They are listed in ascending order of priority. For example,
a woman aged 41 years who stopped taking oral contraceptives 11 months
before and was exclusively breastfeeding a baby would have been listed
in reproductive status category “totally breastfeeding” at the beginning
of this study.
ARCH FAM MED/ VOL 8, SEP/OCT 1999
394
©1999 American Medical Association. All rights reserved.
COMMENT
This study was designed to describe the real-world use
of the CrMS rather than to focus solely on avoiding preg-
nancy. Thus, this study differs both from traditional con-
traceptive efficacy studies and from many efficacy stud-
ies of NFP by including all pregnancies (including those
that may have been planned) in the classification of preg-
nancies and selection of subjects. This difference needs
to be understood to compare this study with other stud-
ies of family planning. However, the different approach
used in this study provides additional information that
is highly relevant to the clinical use of the CrMS. When
placed in appropriate context, the results of this study
can be compared with research on other methods of NFP
and contraception and can be used to advise couples about
the option of the CrMS for family planning. The results
of this study, in comparison with other studies, also raise
important questions for future research.
The results of this study are similar to those
found in previously published studies of the CrMS
that have used the same classifications of preg-
nancy.28,34,40 These classifications are an inherent part
of the way that the CrMS is taught and evaluated clini-
cally. In our experience, these pregnancy classifica-
tions are readily accepted by nearly all CrMS users.
The fact that each couple may choose during the study
to change their use of NFP (to avoid or to achieve
pregnancy)16 and that they may do so at any point
without necessarily announcing it in advance is recog-
nized and emphasized from the first CrMS session. We
believe that these pregnancy classifications accurately
reflect the decisions made by users of the CrMS as
they choose whether to have genital contact at any
given time.
A decision to have genital contact on a day of fer-
tility was not defined as planning a pregnancy, but it was
considered to be a decision to engage in achieving-
Table 2. Net and Gross Cumulative Probabilities of Pregnancy by Pregnancy Classification*
Ordinal
Month
Cumulative
Couple-Months,
No.
Pregnancy Probability per 100 Couples by Category of Pregnancy
Method-
Related
User
Error
Teacher
Error
User and
Teacher Errors
Achieving-Related
Behavior Unresolved Total†
Net
1 690.5 0.00 0.00 0.00 0.00 0.71 0.00 0.71
3 1934.0 0.00 0.29 0.29 0.43 3.85 0.14 4.99
6 3419.5 0.00 0.71 0.29 0.71 7.99 0.71 10.41
9 4639.5 0.00 1.14 0.29 0.86 10.56 1.28 14.12
12 5630.5 0.14 1.57 0.29 0.86 12.84 1.43 17.12
15 6414.0 0.14 1.57 0.29 0.86 15.83 1.43 20.11
18 6947.5 0.14 1.57 0.29 1.00 16.83 1.43 21.26
Gross
1 690.5 0.00 0.00 0.00 0.00 0.72 0.00 0.72
3 1934.0 0.00 0.32 0.31 0.48 4.15 0.17 5.37
6 3419.5 0.00 0.89 0.31 0.86 9.54 0.98 12.26
9 4639.5 0.00 1.64 0.31 1.12 13.43 1.92 17.67
12 5630.5 0.33 2.53 0.31 1.12 17.54 2.22 22.78
15 6414.0 0.33 2.53 0.31 1.12 23.85 2.22 28.69
18 6947.5 0.33 2.53 0.31 2.10 26.31 2.22 31.68
*
All pregnancy classifications are mutually exclusive (except total pregnancies).
†
Probabilities may not add to totals because of rounding.
Table 3. Net Cumulative Probability of Leaving the Study for Non–Pregnancy-Related Reasons by Reason for Leaving*
Ordinal
Month
Cumulative
Couple-Months,
No.
Net Probability per 100 Couples of Leaving the Study by Reason for Leaving†
To Use Other
Method of NFP
To Use Non-NFP
Contraceptive
Method
Moved out of Area,
Known to Still Be
Using the CrMS
Lost to
Follow-up Other Total
1 690.5 0.00 0.29 0.29 0.43 0.29 1.28
3 1934.0 0.29 2.28 2.43 3.00 1.85 9.84
6 3419.5 0.29 4.28 4.56 6.13 2.85 18.12
9 4639.5 0.43 4.71 5.85 8.84 2.85 22.68
12 5630.5 0.71 4.85 6.42 12.41 3.00 27.39
15 6414.0 0.71 5.14 6.56 14.84 3.14 30.39
18 6947.5 0.71 5.14 7.28 17.40 3.57 34.09
*
Probabilities may not add up to totals because of rounding. NFP indicates natural family planning.
†
These categories represent leaving the study but not necessarily discontinuing use of the Creighton Model Fertility Care Sysytem (CrMS).
ARCH FAM MED/ VOL 8, SEP/OCT 1999
395
©1999 American Medical Association. All rights reserved.
related behavior. A typical example of this is reflected in
the expression, “we weren’t trying to get pregnant for an-
other couple of months, but it didn’t matter that much,
so we thought we would take a chance.” On the other
hand, some of the pregnancies resulting from achieving-
related behavior clearly represented conscious planning
to get pregnant at that particular time. In this study, we
did not differentiate between these 2 circumstances be-
Table 4. Net and Gross Cumulative Pregnancy Probabilities per 100 Couples at 1 Year by Reproductive Category at Study Entry
and by Pregnancy Classification*
Reproductive Category
at Entry†
Cumulative
Couple-Months,
No.
Probability of Pregnancy at 1 Year by Category of Pregnancy‡
Method-
Related
User
Error
Teacher
Error
User and
Teacher Error
Achieving-Related
Behavior Unresolved Total‡
Net
Uncomplicated regular cycles 2616.5 0.00 1.55 0.31 0.62 10.25 1.24 13.98
Long cycles 294.0 0.00 0.00 0.00 0.00 14.71 2.94 17.65
Discontinued use of oral
contraceptives
1696.0 0.47 0.94 0.47 1.89 13.21 1.42 18.40
Breastfeeding§ 821.5 0.00 3.81 0.00 0.00 19.05 0.95 23.81
Other\202.5 0.00 0.00 0.00 0.00 14.29 3.57 17.86
All 5630.5 0.14 1.57 0.29 0.86 12.84 1.43 17.12
Gross
Uncomplicated regular cycles 2616.5 0.00 2.54 0.33 1.00 13.57 1.90 18.46
Long cycles 294.0 0.00 0.00 0.00 0.00 19.64 3.78 22.67
Discontinued use of oral
contraceptives
1696.0 1.08 1.47 0.51 2.16 18.56 2.07 24.33
Breastfeeding§ 821.5 0.00 6.18 0.00 0.00 25.59 2.11 31.66
Other\202.5 0.00 0.00 0.00 0.00 20.26 6.08 25.11
All 5630.5 0.33 2.53 0.31 1.12 17.54 2.22 22.78
*
All pregnancy classifications are mutually exclusive (except total pregnancies).
†
For further definition of reproductive categories, see Table 1 and the text.
‡
Probabilities may not add to totals because of rounding.
§
Includes total breast-feeding and breast-feeding/weaning.
\
Includes women who were aged 40 years or older, postpartum and not breastfeeding, or within 1 cycle after abortion (spontaneous or elective).
Table 5. Estimated Probabilities of Method-Related, Avoiding-Related, and Extended-Use Pregnancy
During Use of the Creighton Model Fertility Care System (CrMS)
Ordinal
Month
Estimated Net Probability per 100 Couples at 1 Year
Method-Related Pregnancy*Avoiding-Related Pregnancy† Extended-Use Pregnancy‡
High Estimate§ Low Estimate|High Estimate§ Low Estimate|High Estimate¶ Low Estimate#
1 0.00 0.00 0.00 0.00 0.97 0.78
3 0.14 0.00 1.14 1.03 6.90 5.47
6 0.71 0.00 2.43 1.84 13.95 11.30
9 1.28 0.00 3.57 2.51 18.57 15.24
12 1.57 0.15 4.28 3.11 22.51 18.47
15 1.57 0.15 4.28 3.07 26.11 21.61
18 1.57 0.15 4.42 3.21 27.96 22.93
*
The probability of method-related pregnancy is the probability of pregnancy if the method is used exactly according to instructions to avoid pregnancy. This is
comparable with “pregnancy during perfect use” in other studies.
†
The probability of avoiding-related pregnancy is the probability of pregnancy when the method is used consistently according to instructions to avoid
pregnancy, allowing for errors in use and/or teaching. Avoiding-related pregnancy differs from typical use pregnancy because it excludes pregnancies that result
from a couple deliberately having intercourse at a fertile time, without necessarily having made a conscious decision to achieve pregnancy. These estimates
include method-related, user error, teacher error, both user and teacher error, and, to a degree, unresolved pregnancies.
‡
The probability of extended-use pregnancy is the probability of pregnancy after beginning use of the method, regardless of whether the method continued to
be used within the period specified. Extended-use pregnancy differs from pregnancy during typical use in other family planning studies because it includes
pregnancies that occur among couples who stop using the family planning method to avoid pregnancy. These estimates include pregnancies from all categories in
this study, including those caused by achieving-related behavior and those that may have been planned. These estimates also include assumptions about
pregnancies that may have occurred among women who discontinued use of the CrMS for reasons not known to be related to pregnancy.
§
Includes all unresolved pregnancies.
\
Estimates are based on distributing unresolved pregnancies among the other pregnancy categories according to the same proportions as the remaining
pregnancies, as described by Hilgers.
35
¶
Includes an additional estimated 0.2 pregnancy rate attributed to subjects who left the study for reasons not known to be related to pregnancy. This includes
leaving to use another method of family planning, moving, or being lost to follow-up. For simplicity of calculation, pregnancies among these couples are assumed
to have occurred in the month that they left the study.
#
Includes an additional estimated 0.05 pregnancy rate attributed to subjects who left the study for reasons not known to be related to pregnancy.
ARCH FAM MED/ VOL 8, SEP/OCT 1999
396
©1999 American Medical Association. All rights reserved.
cause we wished to reflect the use dynamics of this ap-
proach in the clinical teaching of the CrMS. This con-
trasts with the approach used in most contraceptive ef-
ficacy studies, which requires study participants to declare
their intentions explicitly with regard to pregnancy at
given intervals (eg, the next cycle, the next month, or
the next year) and subsequently excludes from the study
couples who report explicit intentions to achieve preg-
nancy. In this context, we reemphasize that no attempt
was made in this study to select for couples who were
committed to avoiding pregnancy for any period. Thus,
the probabilities of achieving-related pregnancy (and to-
tal pregnancy probabilities) reported in this study in-
clude many pregnancies that most contraceptive effi-
cacy studies (and many NFP studies) would have excluded
on the basis that the couple had announced in advance
that they intended to achieve pregnancy. One previous
study of the CrMS evaluated all pregnancies (mostly
within the first trimester) and reported that of the total
pregnancies, couples reported that 56% were planned and
44% were unplanned.40 If we had followed the conven-
tion of other contraceptive research and excluded planned
pregnancies, the reported overall pregnancy probabili-
ties would have been lower. However, this would not have
completely reflected the use dynamics of the CrMS, which
can be used to achieve or avoid pregnancy.
In studies of other methods of NFP, pregnancies oc-
curring among couples who have knowingly engaged in
coitus on days defined as fertile by the NFP method un-
der study have been classified by various investigators
as pregnancies resulting from conscious departure from
the rules,21 informed-choice pregnancies,7,41 and pregnan-
cies resulting from risk-taking.42 These terms, although
conceptually distinct, may be comparable in practice with
our term of pregnancy caused by achieving-related be-
havior, as long as planned pregnancies are not ex-
cluded. Some NFP studies exclude planned pregnan-
cies, while others, like ours, do not. Most pregnancies
that occur in all NFP studies result from genital contact
during the fertile time, and variations in this category of
pregnancy are the main determinant of variations in to-
tal pregnancy rates or probabilities in NFP studies.7,43 The
dynamics of planning pregnancy are not well under-
stood.44 How intentions relate to sexual behavior among
NFP users (as well as users of other family planning meth-
ods) remains an area for further study.
In this study, we have separately identified preg-
nancies resulting from an error in application of the CrMS
by the couple, an error in teaching by the instructor, or
both. The common mechanism underlying these preg-
nancies is that they result from the couple having geni-
tal contact at a time that they believed to be infertile, but
that on review of their daily record was in fact defined
as fertile by the correct application of the CrMS. Ex-
amples of such errors include not remembering (or not
being taught) to observe for mucus discharge during the
light days of the menstrual flow and having genital con-
tact on a day of light flow in a cycle during which there
was an early mucus buildup with early ovulation, not re-
membering (or not being taught) to consider nonmen-
strual bleeding as a time of fertility, or misidentifying the
peak day (the estimated day of ovulation). These catego-
ries of pregnancy (which ideally would have a probabil-
ity of 0) constitute a key measure of the quality of teach-
ing that is critical to proper use of NFP. Thus, the clinical
significance of these categories of pregnancy is as a mea-
sure of the teaching process that is essential to optimal
use of NFP (in this case the CrMS). Other NFP studies
have reported teaching-related pregnancy probabilities
or rates (combining these 3 categories) ranging from 0
to 12.2 per 100 couples.7
Despite our best efforts, we were unable to obtain
sufficient information to classify a small number of preg-
nancies; these we reported as unresolved pregnancies.
These pregnancies can be assumed to have the same gen-
eral distribution as the pregnancies for which a classifi-
cation is available, or they can be assumed to consist en-
tirely of method-related pregnancies. This results in an
estimated range of probabilities for method-related and
avoiding-related pregnancies, as shown in Table 5.
The probabilities of avoiding-related pregnancy re-
ported in Table 5 represent the probability of pregnancy
that a couple can expect if they consistently use the CrMS
to avoid pregnancy; in other words, if they have genital con-
tact only on days they expect to be infertile according to
the CrMS instructions. The pregnancies that occur in this
circumstance are caused by an error that results in mis-
classifying a fertile time as infertile, either because of user
error or because of inaccuracy intrinsic to interpreting the
biological signs of fertility. We have further analyzed
whether the user error was completely or partially caused
by poor teaching or occurred in the setting of adequate
teaching. This probability has high clinical relevance for
users of NFP and is arguably a more relevant statistic to
quote to a couple as best possible effectiveness (ie, lowest
possible pregnancy probability) than the probability of
method-related pregnancy, since it accounts for human er-
ror in teaching or learning. However, these estimates of the
probability of avoiding-related pregnancy are not compa-
rable with probabilities of use-related pregnancy in con-
traceptive efficacy studies, because in contraceptive effi-
cacy studies (and some NFP studies), use-related pregnancy
is defined in a way that includes some pregnancies caused
by behaviors that the couple know are likely to result in
pregnancy. We did not use the traditional definition of use-
related pregnancy in this study because it does not coin-
cide with how the CrMS is taught.
In theory, it would be possible for a contraceptive
efficacy study to report a measure of the probability of
avoiding-related pregnancy similar to what we have used
in this study. For example, a study of a barrier method
could differentiate pregnancies resulting from the com-
plete lack of use of the barrier during a particular epi-
sode of sexual intercourse and pregnancies resulting from
improper use of a barrier because of misunderstanding
how to use it. To our knowledge, such a study does not
exist for contraceptive methods, despite suggestions of
some researchers that such distinctions would be clini-
cally meaningful.44
How then can one best compare the pregnancy prob-
abilities in this study with those of other family plan-
ning methods? Some might directly compare the total
pregnancy probability in our study with the total preg-
nancy rate or probability from studies of other NFP or
ARCH FAM MED/ VOL 8, SEP/OCT 1999
397
©1999 American Medical Association. All rights reserved.
contraceptive methods. For example, the probability of
pregnancy during the first year of typical use (or prob-
ability of use-related pregnancy) is reported for barrier
methods to be from 12% to 22%,45,46 for spermicides from
11% to 31%,17 and for oral contraceptives from 3% to
7%.45,46 However, in this type of comparison, the CrMS
net total pregnancy probability of 17.12 per 100 couples
is artificially high compared with pregnancy probabili-
ties of other methods because the CrMS probability in-
cludes pregnancies that the other methods would have
excluded as planned pregnancies.
Conceptually, the best direct comparison of the preg-
nancy probabilities of the CrMS in this study with other
methods of family planning would probably employ the
concept of extended-use pregnancy. The concept of ex-
tended-use pregnancy was developed to include compli-
ance and longer-term acceptability of family planning
methods as an essential component of their ultimate ef-
fect to avoid pregnancy. Extended-use pregnancies in-
clude all pregnancies that occur during the study pe-
riod, regardless of whether the method under study was
still in use to avoid pregnancy.10,12,38,47 For example, the
probability of an extended-use pregnancy over 1 year for
an oral contraceptive would take into account all preg-
nancies that occur within the year after beginning use of
the oral contraceptive, regardless of consistency of use,
including even those that occurred in women who dis-
continued use of the oral contraceptive altogether dur-
ing that year (and may or may not have resumed its use
during the same year). This would be comparable with
including pregnancies among NFP users who begin hav-
ing genital contact during the fertile time as well as among
those who abandon the use of NFP altogether. Probabili-
ties for extended-use pregnancy at 12 months by net life-
table analysis have been reported for the intrauterine de-
vice (range, 4.8-16.3 per 100 couples) and for oral
contraceptives (range, 8.4-39.5).48-50 Probabilities of ex-
tended-use pregnancy for reversible methods of contra-
ception are higher than the corresponding probabilities
of use-related pregnancy because some women discon-
tinue use of the contraceptive altogether. In this study,
we estimated a net probability of extended-use preg-
nancy of the CrMS at 12 months of 18.47 to 22.51 per
100 couples, based on varying assumptions of what per-
centage of persons who left the study for non–pregnancy-
related reasons ultimately ended up pregnant. How-
ever, these CrMS estimates include pregnancies that the
contraceptive studies would have excluded as having been
planned. Additionally, although the concept of extended-
use pregnancy has a well-respected history in the medi-
cal and family planning literature, it has fallen into dis-
use in recent years and has been reported only occasionally
in recent studies, although it is not always explicitly re-
ferred to as such.48 Despite this, we advocate its contin-
ued use because it is a highly relevant measure of preg-
nancy probability for a given method of family planning
from both clinical and demographic perspectives, par-
ticularly if it is defined in such a way as to include planned
pregnancies, as we have done in this study.
Taking these factors into consideration, it seems likely
that the pregnancy probabilities of the CrMS that would
be most comparable with the probabilities or rates of use-
related pregnancy reported for other methods of family
planning are between the estimates for avoiding-related
pregnancy and extended-use pregnancy given in Table 5.
As recommended in previous studies,12,51
we have reported gross and net life-table
probabilities for pregnancy. Since gross
probabilities adjust for subjects who left
the study for non–pregnancy-related rea-
sons (eg, moved or lost to follow-up), they are some-
what higher than net probabilities. Gross probabilities
are more comparable between studies, but few family plan-
ning studies have reported gross probabilities. While we
have emphasized net life-table probabilities in this re-
port because of their standard use, we have reported gross
probabilities for comparisons with other studies that may
also make these available. Gross probabilities from this
study cannot be directly compared with net probabili-
ties or Pearl rates from other studies.
The net life-table probabilities in this study are
based on a common denominator of all study subjects
using a method that was developed by leading contra-
ceptive researchers and is still widely used. However,
more recent literature has suggested that it is more accu-
rate to divide subjects or cycles of use into 2 groups—
perfect users (or perfect-use cycles) and imperfect users
(or imperfect-use cycles)—and to calculate separate
perfect-use and imperfect-use pregnancy probabilities
based on these separate denominators.45,51 Since the
CrMS is taught and used in a conceptual framework that
emphasizes that couples have the freedom, the responsi-
bility, and the capability to make choices about their
reproductive capacity at all times, we suggest that an
appropriate approach for the CrMS would be to evaluate
pregnancy probabilities for months when couples know-
ingly had genital contact during times of fertility (which
could be called an adjusted achieving-related pregnancy
probability), and pregnancy probabilities for months
when couples had genital contact only during times they
knew to be infertile (which could be called an adjusted
avoiding-related pregnancy probability). Unfortunately,
we do not have data on whether genital contact occurred
at the fertile time for every couple-month in our study
(we have such data only for months when pregnancy
occurred), and thus we cannot address this issue defini-
tively. However, the available evidence suggests that
couples who become pregnant during use of the CrMS
do so within very few cycles of genital contact during
the fertile time (76% within 1 cycle in 1 published
study).52 Therefore, the adjusted achieving-related preg-
nancy probability would likely be very high (very much
higher than the achieving-related pregnancy probability
reported in this study), and the adjusted avoiding-
related pregnancy probability would likely be only
slightly higher than the probabilities of avoiding-related
pregnancy reported in Table 5. This reasoning is consis-
tent with an analysis of a multinational study of the ovu-
lation method that found that genital contact only on
days of infertility within the ovulation method is highly
unlikely to result in pregnancy, whereas genital contact
on days of fertility is highly likely to result in preg-
ARCH FAM MED/ VOL 8, SEP/OCT 1999
398
©1999 American Medical Association. All rights reserved.
nancy.22 However, further research is needed to define
the exact probabilities for adjusted achieving-related and
adjusted avoiding-related pregnancy.
A related issue has to do with the first cycle of ob-
servation, during which couples are advised to abstain
from all genital contact to facilitate their learning of the
CrMS (to avoid confusion from seminal residue). Some
have suggested that this first cycle should be excluded
from analysis of pregnancy probabilities, since no expo-
sure to the possibility of pregnancy was to have oc-
curred. However, not all couples followed this instruc-
tion, and there were some pregnancies during the first
month of use. We have followed the approach of other
researchers studying the ovulation method by includ-
ing the first cycle in our analysis.22 Furthermore, alter-
nate analysis excluding the first cycle (and attributing
pregnancies to the second cycle) results in nearly iden-
tical 12-month life-table pregnancy probabilities (data not
shown).
It is possible that some couples with undiscovered
reduced fertility may have been included in this study,
which would lower the probability of pregnancy. Some
studies of NFP or contraception have dealt with this prob-
lem by including only couples of proven fertility (ie, those
who have previously had children).30 In this study, we
chose to maximize generalizability by including all couples
except those with a known history of infertility. It has
been suggested that couples who choose to use NFP or
who continue its use have, on average, lower fertility than
couples who use other methods. We know of no evi-
dence to support this suggestion. Couples who choose
to use NFP seem to be motivated by concerns about health,
by side effects from other methods, or by moral be-
liefs,53-55 reasons that are unlikely to be associated with
reduced fertility.
Most studies of NFP or contraceptive efficacy have
excluded women who have recently taken oral contra-
ceptives, who are breastfeeding, or who have other con-
ditions that might be associated with a temporary reduc-
tion in fertility. However, it is important to know how a
method performs among couples with these various cir-
cumstances. Therefore, this study included essentially all
new users of the CrMS. This allows for much wider gen-
eralizability of these results than has been possible with
previous studies of NFP, most of which have included
only women with regular, uncomplicated cycles. To pro-
vide relevant comparisons with such studies, we have re-
ported separate pregnancy probabilities for subjects with
uncomplicated regular cycles, as well as for long cycles,
discontinuing oral contraceptives, breastfeeding, and other
circumstances (Table 4). The total pregnancy probabil-
ity was lowest for couples with uncomplicated regular
cycles (net probability of 13.98 per 100 couples at 12 or-
dinal months for couples with uncomplicated regular
cycles vs 17.12 for the entire study). This does not mean
that the CrMS is intrinsically more effective for those with
uncomplicated regular cycles, because the probability of
method-related pregnancy is low for all of the groups,
and the differences are mainly caused by achieving-
related behavior. It seems that relatively new users of the
CrMS who have uncomplicated, regular cycles are less
likely to engage in achieving-related behavior (genital con-
tact on days known by the couple to be fertile). From
our clinical experience, we feel that this might be be-
cause new users of the CrMS rely to some extent on cal-
endar rhythm despite all the teaching of the CrMS to the
contrary. Couples who have irregular cycles and expe-
rience a time of fertility (as defined by mucus discharge
changes) that differs in calendar timing from what they
expect may be less likely to believe that this time is truly
fertile as indicated by the CrMS and thus may be more
likely to take a chance, at least initially. However, this
explanation is based on anecdotal reports and requires
further research.
While fertility usually returns rapidly after discon-
tinuing use of oral contraceptives,56 some cycle abnor-
malities exist in this circumstance that may make it more
difficult for a couple to begin use of NFP.57 These data
indicate that couples who discontinue use of oral con-
traceptives can use the CrMS effectively. Since a sub-
stantial proportion of those beginning use of the CrMS
are discontinuing use of oral contraceptives, this is an
important finding.
Similarly, a high proportion of couples who are be-
ginning use of the CrMS are breastfeeding, a situation for
which relatively few data exist on the effectiveness of NFP.
One notable exception is the lactational amenorrhea
method, a method of NFP based entirely on breastfeed-
ing that has been shown to have a pregnancy probabil-
ity of less than 2% in the first 6 months postpartum, as
long as breastfeeding is total without supplementation
and menses has not returned.58,59 However, most women
in the United States do not breastfeed in a way that meets
the criteria for the lactational amenorrhea method, and
pregnancy probabilities rise rapidly once weaning has been
initiated or menses has returned.60 Even couples who do
use the lactational amenorrhea method enter a period of
transition when other options for family planning are
needed.61 Unfortunately, we do not have data to report
what percentage of breastfeeding couple-months in this
study would have also met the criteria for the lacta-
tional amenorrhea method, so we cannot estimate an in-
dependent effect of the CrMS vs the intrinsic lower fer-
tility of the total breastfeeding state. In addition, we have
reliable information on breastfeeding status only at en-
try to the study; we do not know how long these women
continued to breastfeed, when they weaned, or whether
they were still breastfeeding at the time that they be-
came pregnant.
Since breastfeeding reduces fertility, the inclusion
of breastfeeding couples in this study theoretically could
have lowered overall pregnancy probabilities, similar to
the effect found in one previous study of the ovulation
method in postpartum women that found a lower over-
all pregnancy rate among women who were breastfeed-
ing.62 However, the opposite occurred: total pregnancy
probabilities were higher among the breastfeeding couples
(23.8 per 100 couples at 1 year for breastfeeders vs 17.1
for all couples). While counterintuitive, this finding does
have precedent: Labbok et al60 have reported higher rates
of pregnancy among women using the ovulation method
who were breastfeeding compared with women who were
not, and this occurred entirely after the criteria for the
lactational amenorrhea method were no longer met. They
ARCH FAM MED/ VOL 8, SEP/OCT 1999
399
©1999 American Medical Association. All rights reserved.
also presented data that suggested that this was not caused
by increased frequency of intercourse during the fertile
time. In contrast, our study indicated that the increased
total pregnancy probability during use of CrMS primar-
ily reflected a higher probability of pregnancy because
of achieving-related behavior. Overall, this study sug-
gests that the CrMS can be used effectively by breast-
feeding couples, but that breastfeeding couples may be
more likely than others to get pregnant as a result of geni-
tal contact on days known by the couple to be fertile. Fur-
ther research on the use of the CrMS in postpartum
couples (both breastfeeding and not) will be needed to
clarify the reasons for the higher probability of preg-
nancy among breastfeeding couples. It will be particu-
larly important to track the timing of supplementation,
weaning, and return of menses in relation to pregnancy
probabilities among women who are breastfeeding.
While the probability of leaving this study for rea-
sons other than pregnancy was 27.39 per 100 couples at
1 year, the probability of loss to follow-up was 12.4. Leav-
ing the study did not necessarily mean discontinuation of
use of the CrMS (Table 3). At 18 months, the probability
of loss to follow-up rose to 17.4, probably because the fol-
low-up mechanisms are less formalized after 1 year of use
of the CrMS. Probabilities or rates of loss to follow-up (gen-
erally at 1 year) in other studies have been reported from
0 to 59.5 for spermicides, from 0 to 43.8 for barrier meth-
ods, from 0.3 to 32.0 for oral contraceptives, and from 0.3
to 33.9 for other NFP methods.10
Women (and couples) are often not satisfied with
the generally available methods of family planning.63 There
is evidence that more women are potentially interested
in NFP methods than currently use them.64,65 Population-
based studies in Germany and the United States suggest
that up to 25% of women of reproductive potential may
be interested in using modern NFP methods to avoid preg-
nancy and up to 33% may be interested in using them to
achieve pregnancy. Most of this interest seems to be mo-
tivated by health concerns rather than religious con-
cerns.55,66 In light of this latent and largely unrecog-
nized demand, health care professionals need to be aware
of the availability and viability of modern methods of NFP.
We feel that the CrMS is particularly well suited to meet
this need because of its highly trained instructors and stan-
dardized teaching methods.3,28,34
Couples using the CrMS in this study were of rela-
tively high socioeconomic and educational status. How-
ever, there is no evidence that education is a prerequi-
site for the successful use of NFP. In a study of the Billings
Ovulation Method conducted by the World Health Or-
ganization,30 illiterate women in El Salvador learned to
understand their fertility cycles just as rapidly and eas-
ily as educated women in New Zealand and Ireland. Natu-
ral family planning has been used successfully in a vari-
ety of Third World countries.6,67,68 Furthermore, studies
of potential interest in NFP in developed countries have
shown no association with socioeconomic status or edu-
cation.55,66 We believe that the socioeconomic and edu-
cational status of couples in this study primarily reflects
knowledge of and access to the CrMS; thus, efforts to make
the CrMS available to a more diverse population are war-
ranted. We do emphasize that use of NFP requires the
cooperation of both partners and that NFP is not suit-
able for those who are not in monogamous relation-
ships.
Based on the results of this study, we suggest that
several points be emphasized in counseling couples about
the CrMS: First, if you use the CrMS perfectly to avoid
pregnancy, the probability of pregnancy within the first
year is less than 1%. However, accounting for errors
(which could be made by the users and/or the teacher),
the probability of pregnancy within the first year during
consistent use to avoid pregnancy is about 3% to 4%. Sec-
ond, those of normal fertility who have genital contact
on any day defined as fertile by the CrMS will very likely
become pregnant (one study suggests more than a 50%
probability of pregnancy within the first cycle).52 Third,
in our study, the probability that couples who started use
of the CrMS would be pregnant in 1 year was about 17%.
Most of those pregnancies were a result of the couple hav-
ing genital contact on a day they knew to be fertile, for
many reasons, including taking a chance or planning a
pregnancy. Fourth, women who have regular cycles, who
have irregular cycles, who are breastfeeding, or who are
discontinuing use of oral contraceptive pills can use the
CrMS successfully. Fifth, couples in which the woman
is breastfeeding may have a higher chance of pregnancy
with the CrMS if they have genital contact during a time
of fertility. Finally, successful use requires learning the
CrMS from a qualified Creighton Model instructor. Given
time constraints, this will not ordinarily be the physi-
cian or provider, but rather a trained instructor to whom
the provider can refer or who may provide instructional
services within the same office.
CONCLUSIONS
This study supports the potential effectiveness of the CrMS
in a broad spectrum of reproductive categories in a US
population of users. The strengths of this study in-
cluded the inclusion of couples in all reproductive situ-
ations, the comparison of pregnancy probabilities ac-
cording to reproductive status of the subject at entry, an
analytical approach that reflected how the CrMS was ac-
tually taught and used, reporting both gross and net prob-
abilities, results that provide information highly rel-
Contact Information
Creighton Model teachers (practitioners) are available in
most states and metropolitan areas. Some are associated
with health care institutions, such as hospitals or private
physician practices, some are associated with Catholic di-
ocesan offices, and some work independently. To locate a
Creighton Model teacher in your area, contact the Ameri-
can Academy of Natural Family Planning at 775-827-
2500, online at http://www.aanfp.org, or through the Pope
Paul VI Institute for the Study of Human Reproduction at
402-390-6600, online at http://www.popepaulvi.com.
Creighton Model teachers are trained in many centers in
the United States; a current listing of Creighton Model
teacher training programs can be obtained by contacting
the American Academy of Natural Family Planning.
ARCH FAM MED/ VOL 8, SEP/OCT 1999
400
©1999 American Medical Association. All rights reserved.
evant to the clinical use of the CrMS, and a reasonable
rate of loss to follow-up. The weaknesses of this study
included the lack of information about how many couples
who had genital contact during the fertile time did not
conceive, a lack of information about the timing of wean-
ing and the return of menses among the breastfeeding
subjects, and the inability to compare pregnancy prob-
abilities directly with those of many contraceptive stud-
ies because of the inclusion of pregnancies that would
have been excluded by those studies as planned preg-
nancies. Despite the weaknesses of this study, the re-
sults provide a strong basis for counseling couples about
this important option for family planning, as described
above. This research also suggests many pressing issues
for further inquiry, including the relationships among
achieving-related behavior (sexual behavior known to be
likely to cause pregnancy), taking chances, and plan-
ning pregnancy, both among NFP users and among us-
ers of other methods of family planning. Another
important area for further research is the relationships
among the infertility of lactational amenorrhea during
total breastfeeding, the return of fertility during wean-
ing and/or resumption of menses, and the use of the CrMS
in these circumstances.
Accepted for publication September 24, 1998.
We thank Charles E. Howard, MFA, CNFPP, for his
assistance with the collection and organization of data for
this study; Ken R. Smith, PhD, for his help with statistical
analysis; and the NFP practitioners for teaching the couples
who were subjects of this study.
Corresponding author: Joseph B. Stanford, MD, MSPH,
Department of Family and Preventive Medicine, Univer-
sity of Utah, 50 N Medical Dr, Salt Lake City, UT 84132
(e-mail: jstanford@dfpm.utah.edu).
REFERENCES
1. Knaus H. Periodical fertility and infertility in women.
Zentralbl Gynakol
. 1933;
57:1393-1409.
2. Dicker D, Wachsman Y, Feldberg D, Ashkenazi J, Yeshaya A, Goldman HA. Con-
traceptive diaphragm and the condom: a reevaluation and comparison of two
barrier methods with the rhythm method.
Contraception
. 1989;40:497-504.
3. Geerling JH. Natural family planning.
Am Fam Physician
. 1995;52:1749-1756.
4. Betts K. The Billings method of family planning: an assessment.
Stud Fam Plann
.
1984;15:253-266.
5. Benagiano G, Bastianelli C. Clinical trials of the ovulation method.
Suppl Int J
Gynecol Obstet
. 1989;1:91-98.
6. Ryder REJ. Natural family planning: effective birth control supported by the Catho-
lic Church.
BMJ
. 1993;307:723-726.
7. Klaus H. Natural family planning: a review.
Obstet Gynecol Surv
. 1982;37:128-150.
8. Kambic RT. Natural family planning use-effectiveness and continuation.
Am J
Obstet Gynecol
. 1991;165:2046-2048.
9. European Natural Family Planning Study Groups. Prospective European multi-
center study of natural family planning (1989-1992): interim results.
Adv Con-
tracept
. 1993;9:269-283.
10. Trussell J, Kost K. Contraceptive failure in the United States: a critical review of
the literature.
Stud Fam Plann
. 1987;18:237-283.
11. Lamprecht V, Trussell J. Natural family planning effectiveness: evaluating pub-
lished reports.
Adv Contracept
. 1997;13:155-165.
12. Trussell J, Hatcher RA, Cates W Jr, Stewart FH, Kost K. A guide to interpreting
contraceptive efficacy studies.
Obstet Gynecol
. 1990;76:558-567.
13. World Health Organization.
Fertility Awareness Methods: Report on a WHO Work-
shop
. New York, NY: United Nations Family Planning Association; 1987.
14. Billings JJ. Ovulation method of family planning [letter].
Lancet
. 1972;2:1193-1194.
15. Marshall J. Natural family planning [letter].
Lancet
. 1976;2:7987.
16. Klaus H, Fagan MU. Natural family planning: an analysis of change in procre-
ative intention.
J Am Med Womens Assoc
. 1982;37:231-241.
17. Woolley RJ. Contraception: a look forward, part I: new spermicides and natural
family planning.
J Am Board Fam Pract
. 1991;4:33-46.
18. Stanford JB. Contraception: natural family planning [letter].
J Am Board Fam Pract
.
1991;4:277-279.
19. Wilson MB. Natural family planning: cheap, effective, and free of side effects.
BMJ
. 1993;307:1357-1358.
20. Trussell J. Natural family planning: effective only if used perfectly [letter].
BMJ
.
1993;307:1003.
21. World Health Organization. A prospective multicentre trial of the ovulation method
of natural family planning, II: the effectiveness phase.
Fertil Steril
. 1981;36:591-
598.
22. Trussell J, Grummer-Strawn L. Further analysis of contraceptive failure of the
ovulation method.
Am J Obstet Gynecol
. 1991;165:2054-2059.
23. Weissmann MC, Foliaki L, Billings EL, Billings JJ. A trial of the ovulation method
of family planning in Tonga.
Lancet
. 1972;2:813-816.
24. Billings EL, Billings JJ, Catarinich M.
Billings Atlas of the Ovulation Method
. Mel-
bourne, Australia: Ovulation Method Research and Reference Centre of Austra-
lia; 1989.
25. Hilgers TW.
The Ovulation Method of Natural Family Planning
. Omaha, Neb: Pope
Paul VI Institute for the Study of Human Reproduction; 1983.
26. Hilgers TW, Daly KD, Hilgers SK, Prebil AM.
The Ovulation Method of Natural
Family Planning: A Standardized, Case Management Approach to Teaching: Book
One: Basic Teaching Skills
. Omaha, Neb: Creighton University Natural Family Plan-
ning Education and Research Center; 1982.
27. Hilgers TW, Prebil AM, Daly KD, Hilgers SK.
The Picture Dictionary of the Ovu-
lation Method and Other Assorted Teaching Aids
. Omaha, Neb: Creighton Uni-
versity Natural Family Planning Education and Research Center; 1982.
28. Fehring RJ, Lawrence D, Philpot C. Use effectiveness of the Creighton-Model Ovu-
lation Method of natural family planning.
J Obstet Gynecol Neonatal Nurs
. 1994;
23:303-309.
29. Medina JE, Cifuentes A, Abernathy JR, et al. Comparative evaluation of two meth-
ods of natural family planning in Colombia.
Am J Obstet Gynecol
. 1980;138:
1142-1147.
30. World Health Organization. A prospective multicentre trial of the ovulation method
of natural family planning, I: the teaching phase.
Fertil Steril
. 1981;36:152-158.
31. Wade ME, McCarthy P, Braunstein GD, et al. A randomized prospective study of
the use-effectiveness of two methods of natural family planning.
Am J Obstet
Gynecol
. 1981;141:368-376.
32. Treolar AE, Boynton RE, Behn BG, Brown BW. Variation of the human menstrual
cycle throughout reproductive life.
Int J Fertil Womens Med
. 1967;12:77-126.
33. Vollman RF.
The Menstrual Cycle
. Philadelphia, Pa: WB Saunders Co; 1977.
34. Hilgers TW, Stanford JB. Creighton-Model NaProEducation Technology for avoid-
ing pregnancy: use effectiveness.
J Reprod Med
. 1998;43:495-502.
35. Hilgers TW. The statistical evaluation of natural methods of family planning.
Int
Rev Nat Fam Plann
. 1984;8:226-264.
36. Tietze C, Lewit S. Statistical evaluation of contraceptive methods.
Clin Obstet Gy-
necol
. 1974;17:121-138.
37. Tietze C, Lewit S. Recommended procedures for the statistical evaluation of in-
trauterine contraception.
Stud Fam Plann
. 1973;4:35-42.
38. Trussell J, Menken J. Life table analysis of contraceptive failure. In: Hermalin AI,
Entwisle B, eds.
The Role of Surveys in the Analysis of Family Planning Pro-
grams
. Liege, Belgium: Ordina Editions; 1982:537-571.
39. Steiner M, Dominik R, Trussell J, Hertz-Picciotto I. Measuring contraceptive effec-
tiveness: a conceptual framework.
Obstet Gynecol
. 1996;88(suppl 3):24S-30S.
40. Doud J. Use-effectiveness of the Creighton Model of NFP.
Int Rev Nat Fam Plann
.
Spring 1985;9:54-72.
41. Brennan JJ, Klaus H. Terminology and core curricula in natural family planning.
Fertil Steril
. 1982;38:117-118.
42. Trussell J, Grummer-Strawn L. Contraceptive failure of the ovulation method of
periodic abstinence.
Fam Plann Perspect
. March-April 1990;22:65-75.
43. Guida M, Tommaselli G, Pellicano M, Palomba S, Nappi C. An overview of the
effectiveness of natural family planning.
Gynecol Endocrinol
. 1997;11:203-219.
44. Brown SS, Eisenberg L.
The Best Intentions: Unintended Pregnancy and the Well-
being of Children and Families
. Washington, DC: National Academy Press; 1995.
45. Trussell J, Hatcher RA, Cates W Jr, et al. Contraceptive failure in the United States:
an update.
Stud Fam Plann
. 1990;21:51-54.
46. Potter LS. How effective are contraceptives? the determination and measure-
ment of pregnancy rates.
Obstet Gynecol
. 1996;88(suppl 3):13S-23S.
47. Tietze C, Lewit S. Statistical evaluation of contraceptive methods: use-
effectiveness and extended use-effectiveness.
Demography
. 1968;5:931-940.
48. Polaneczky M, Slap G, Forke C, et al. The use of levonorgestrel implants (Nor-
plant) for contraception in adolescent mothers.
N Engl J Med
. 1994;331:1201-
1206.
ARCH FAM MED/ VOL 8, SEP/OCT 1999
401
©1999 American Medical Association. All rights reserved.
49. Tietze C, Lewit S. The IUD and the pill: extended use-effectiveness.
Fam Plann
Perspect
. April 1971;3:53-55.
50. Tietze C, Lewitt S. Use-effectiveness of oral and intrauterine contraception.
Fer-
til Steril
. 1971;22:508-513.
51. Trussell J. Methodological pitfalls in the analysis of contraceptive failure.
Stat
Med
. 1991;10:201-220.
52. Hilgers TW, Daly KD, Prebil AM, Hilgers SK. Cumulative pregnancy rates in pa-
tients with apparently normal fertility and fertility-focused intercourse.
J Reprod
Med
. 1992;37:864-866.
53. Von Fragstein M, Flynn A, Royston P. Analysis of a representative sample of natu-
ral family planning users in England and Wales, 1984-1985.
Int J Fertil
. 1988;
33(suppl):70-77.
54. Boys GA. Factors affecting client satisfaction in the instruction and usage of natu-
ral methods.
Int J Fertil
. 1988;33(suppl):59-64.
55. Freundl G, Frank P, Bauer S, Doring G. Demographic study on the family plan-
ning behaviour of the German population: the importance of natural methods.
Int J Fertil
. 1988;33(suppl):54-58.
56. Weisberg E. Fertility after discontinuation of oral contraceptives.
Clin Reprod Fer-
til
. 1982;1:261-272.
57. Stanford JB, Daly KD. Menstrual and mucus cycle characteristics in women dis-
continuing oral contraceptives [abstract].
Paediatr Perinat Epidemiol
. 1995;9:
A9.
58. Kennedy K, Rivera R, McNeilly A. Consensus statement on the use of breast-
feeding as a family planning method.
Contraception
. 1989;39:477-496.
59. Ramos R, Kennedy K, Visness C. Effectiveness of lacational amenorrhoea in pre-
vention of pregnancy in Manila, the Philippines: non-comparative prospective
trial.
BMJ
. 1996;313:909-912.
60. Labbok MH, Stallings RY, Shah F, et al. Ovulation method use during breast-
feeding: is there increased risk of unplanned pregnancy?
Am J Obstet Gynecol
.
1991;165:2031-2036.
61. Labbok MH, Howie PW. Overview and summary: the interface of breastfeeding,
natural family planning, and the lactational amenorrhea method.
Am J Obstet
Gynecol
. 1991;165:2013-2014.
62. Perez A, Labbok M, Barker D, Gray R. Use-effectiveness of the ovulation method
initiated during postpartum breastfeeding.
Contraception
. 1988;38:499-508.
63. Rosenfeld JA, Zahorik PM, Saint W, Murphy G. Women’s satisfaction with birth
control.
J Fam Pract
. 1993;36:169-173.
64. Stanford JB, Lemaire JC, Fox A. Interest in natural family planning among fe-
male family practice patients.
Fam Pract Res J
. 1994;14:237-249.
65. Klaus H, Fagan MU, Hoefler S, et al. Who has heard of NFP?
Int Rev Nat Fam
Plann
. Spring 1981;5:74-82.
66. Stanford JB, Lemaire JC, Thurman PB. Women’s interest in natural family plan-
ning.
J Fam Pract
. 1998;46:65-71.
67. Gray RH, Kambic RT, Lanctot CA, et al. Evaluation of natural family planning pro-
grammes in Liberia and Zambia.
J Biosoc Sci
. 1993;25:249-258.
68. Kambic RT, Gray RH, St Mart R, et al. Use-effectiveness among users of the symp-
tothermal method of family planning.
Int Fam Plann Perspect
. September 1991;
17:96-99.
Clinical Pearl
Probing Ulcers to Detect Osteomyelitis
If bone was palpable when probing pedal ulcers in diabetic patients,
osteomyelitis was likely. Bone was palpable by probe in 33 of 50 with osteomyelitis
and 4 of 26 without. (JAMA. 1995;273:721-723.)
ARCH FAM MED/ VOL 8, SEP/OCT 1999
402
©1999 American Medical Association. All rights reserved.
