Impact of In-Hospital Paternity Outreach Efforts

     CHAPTER IV

     IMPACT OF IN-HOSPITAL PATERNITY OUTREACH EFFORTS

A.    Evaluation Methodology

     The Colorado paternity demonstration project involved four
hospitals in the Denver area. The in-hospital intervention was
initiated at slightly different timepoints at each hospital site.
We initiated at-birth orientations about paternity in University
Hospital in October 1992. At-birth orientations about paternity
were introduced in Mercy Hospital and St. Joseph Hospital in
September 1993. We initiated the at-birth paternity overture in
Denver General Hospital in June 1993.

     At all hospitals, the at-birth orientation involved the
routine presentation of oral and written information about
paternity to all unmarried mothers and fathers (when he was
available at the hospital). The orientation was made following
the baby's delivery but before the mother's discharge from the
hospital. It stressed the benefits of paternity for babies and
the rights and responsibilities associated with paternity for
parents. Unmarried parents were offered the opportunity to
acknowledge paternity on a voluntary basis and interested parents
were assisted with the acknowledgement process.  Paternity
acknowledgement was treated as a confidential process; voluntary
acknowledgements were not reported to the child support
enforcement agency.

     At all hospital sites, identical information was collected
about every woman exposed to the paternity orientation. Most of
the information could be readily extracted from the birth
certificate worksheet. This included standard demographic items
about unmarried mothers and their partners. In addition,
paternity outreach workers collected some limited information
that went beyond the birth certificate worksheet including
whether the baby's father had attended the birth and whether or
not one or both parents had signed the acknowledgement form.

     The assessment of the impact of the in-hospital paternity
outreach effort involved a comparison of pre- and
post-intervention paternity patterns at each hospital setting.
Our picture of paternity patterns prior to the demonstration
project was derived from automated records of births to unmarried
parents at the four participating hospitals in 1991. Across the
four hospitals, there were 4,260 births to unmarried parents in
1991. During this year, none of the hospitals presented
information about paternity and the voluntary acknowledgement
process to unmarried parents on a regular basis.

     Our information about paternity patterns following the
introduction of the at-birth overture came from the database
compiled by paternity workers affiliated with the demonstration
project. Workers completed data collection forms for each
unmarried parent exposed to the paternity overture following
delivery. During the demonstration project, at-birth orientations
about paternity were made to 3,902 unmarried parents. Thus, our
post-intervention database consists of 3,902 births across the
four participating hospitals during 1993-1994.  A copy of the
data collection form used to compile the post-intervention
database appears in Appendix C.

     Table IV-1 presents the numbers of cases in our comparisons
of pre- and post-intervention patterns across the four hospital
sites.


                           Table IV-1
         Numbers of Unmarried Births Prior to and During
         the Paternity Demonstration Project by Facility


Facility                Pre-intervention       Post-intervention*
                           (1991)                (1993-1994)


Denver General              1,697                      1,574
                           (39.8%)                     (40%)

Mercy                          87                         297
                            (2.0%)                      (7.6%)

St. Joseph                  1,290                          650
                           (30.3%)                      (16.7%)

University                  1,186                        1,381
                           (27.8%)                      (35.4%)

Total                       4,260                        3,902

*Only includes cases seen in the project, not all unmarried
births at these facilities.


B.   The Populations Served

          Can observed changes in voluntary paternity
acknowledgement levels following the initiation of the
demonstration project safely be attributed to the at-birth
paternity overture? Are any changes in paternity acknowledgement
levels due to changes in the demographic profile of the unmarried
populations served at the four participating facilities?

          To answer these questions, we compared selected
demographic characteristics of unmarried mothers who delivered at
each facility in 1991 with patterns for 1993-1994. The analysis
revealed that the types of populations served by each of the
hospitals remained relatively constant at the two timepoints.

          In two of the hospital facilities, Denver General and
University, unmarried mothers looked virtually the same both
prior to and following the initiation of the demonstration
project. Race, education, age, employment and prior numbers of
births for unmarried mothers were all quite similar at both pre-
and post-demonstration project timepoints. These patterns may
besummarized as follows:

o    In Denver General Hospital, the average age of unmarried
     mothers in 1991 was 22.3 as compared with 22.8 in 1993-1994.
     The proportion of unmarried mothers who were White was 12
     percent in 1991 as compared with 16 percent in 1993-1994.
     The proportion of mothers with less than a high school
     education was 66 percent in 1991 and 70 percent in
     1993-1994.

o    In University Hospital, the average age of unmarried mothers
     in 1991 was 22.7 as compared with 23.4 in 1993-1994. The
     proportion of unmarried mothers who were White was 49
     percent in both 1991 and 1993-1994. The proportion of
     mothers with less than a high school education was 45
     percent in 1991 versus 44 percent in 1993-1994.

     In two of the hospitals, Mercy and Saint Joseph, the
post-demonstration project population was more heavily White and
somewhat more apt to have had no prior births. While these
demographic characteristics are correlated with voluntary
paternity acknowledgement levels, they are offset by the fact
that mothers who delivered in 1993-1994 at these facilities had
somewhat lower education levels than their counterparts in 1991.
Education is also correlated with voluntary paternity
acknowledgement patterns with women with lower levels of
education tending to reject the paternity option.  There was no
difference in the average age of unmarried mothers who delivered
at these hospital facilities at the pre- and post-project
timepoints. Nor were there consistent differences in the tendency
to be employed during pregnancy. The following examples are
illustrative of patterns at Mercy and Saint Joseph Hospitals.

o    At Mercy Hospital, the proportion of unmarried mothers who
     were White rose from 24 percent in 1991 to 46 percent in
     1993-1994. During the same time period, the proportion of
     unmarried mothers with prior children dropped from 51
     percent to 32 percent. At the same time, the proportions
     employed during pregnancy remained approximately 35 percent
     during both timepoints and the proportion with less than a
     high school education was consistently 40 percent.

o    At Saint Joseph Hospital, the proportion of unmarried
     mothers who were White rose from 44 to 60 percent while the
     proportion with prior births dropped from 35 to 26 percent.
     Simultaneously, the proportion with less than a high school
     education rose from 25 to 32 percent.


       Taken together, it appears that while there were
somemodest differences in the unmarried populations served at
some of the facilities, none of the differences were very
substantial or consistent. As a result, we can rule out the
hypothesis that observed differences were due to changes in the
patient profile at the participating hospitals prior to and
following the initiation of the program.

C.   Other Possible Sources of Sample Bias

     Virtually all unmarried mothers were exposed to the
paternity outreach at University Hospital, since the birth
registration clerk assumed responsibility for presenting the
paternity orientation to unmarried mothers soon after the
inception of the demonstration project.  At that hospital, the
number of women exposed to the orientation during 1993-1994
essentially equalled the total number of unmarried births. There
was no sampling process or potential source of bias due to sample
factors.

     At the other three hospitals, however, the paternity
overture was only presented to a segment of the unmarried
population that delivered.  In general, presentations were made
to all women who delivered on days when a project paternity
worker was at the hospital.  Although there was no obvious bias
or selection process in the scheduling of presentations, there is
nevertheless a possibility that the segment exposed to the
orientation was unrepresentative of the total population of
unmarried patients.

     To determine whether this was indeed the case, we
comparedthe voluntary acknowledgement rate achieved in the
database of cases compiled for this demonstration project with
the universe of births occurring at each hospital facility during
the last quarter of 1993. The analysis revealed that
acknowledgement rates for the two groups of births at each
facility were virtually identical. This suggests that the samples
of mothers exposed to paternity orientations at the project
hospitals were essentially equivalent to the total unmarried
population of mothers served at each facility.

D.   Changes in Rates of Paternity Acknowledgement

     In each hospital setting, the process of systematically
presenting the paternity option to unmarried parents led to
dramatic increases in the voluntary acknowledgement rate. In
1991, prior to the project, voluntary acknowledgement rates at
the four participating hospitals ranged from 13 percent to 24
percent. In 1993-1994, both parents signed the paternity
acknowledgement form in 27 percent to 52 percent of the births
where they were given the option to do so. Figure IV-1 compares
the proportions who voluntarily acknowledged paternity at each
project facility in 1991 and in 1993-1994, following the
initiation of at-birth paternity presentations.

     Despite these gains in levels of voluntary paternity
acknowledgement following the introduction of the in-hospital
paternity effort, substantial proportions of unmarried parents
continued to disavow paternity and reject the option.  Indeed,
one-half to three quarters of the unmarried parents at each
participating facility refused to sign the voluntary
acknowledgement during 1993-1994 when presented with the
opportunity.

     Naturally, some proportion of parents who fail to
acknowledge paternity in the hospital immediately following
delivery will proceed to acknowledge on a voluntary basis at a
later date. Based upon the 1991 data for Colorado as a whole, an
additional 6 percent of unmarried parents acknowledge during the
three-year period following their children's birth. In addition,
approximately 12 to 14 percent obtain paternity establishments by
court order within a three-year time frame.

     At this point, we are unable to determine whether our
project hospitals will experience a comparable rise in voluntary
paternity acknowledgements over time.  On the one hand, it is
possible that the project's aggressive at-birth outreach efforts
will more effectively serve interested unmarried parents and this
will lessen the number of acknowledgements achieved at a later
date. On the other hand, some proportion of unmarried parents may
continue to acknowledge a few years following birth no matter
what the quality of the at-birth effort. Researchers who have
studied paternity acknowledgement patterns in Massachusetts have
failed to find conclusive evidence of trade-offs between at-birth
and post-birth acknowledgements. Although Massachusetts cities
where in-hospital acknowledgement rates are high generally have
low estimated post-birth acknowledgement rates, some areas with
average in-hospital rates still have substantial numbers of
post-birth acknowledgements (Williams et al, 1995).


E.   Summary

     The overall impact of the in-hospital paternity
demonstration project at the four participating facilities was
dramatic.  At Denver General and St. Joseph Hospitals, the
voluntary acknowledgement rate more than doubled.  At Mercy and
University Hospitals, the voluntary acknowledgement rate rose by
approximately 65 percent.  These increases could not be
attributed to changes in the unmarried populations served at each
facility which resembled one another at both pre- and
post-project timepoints.  Nor did changes in the voluntary
acknowledgement rate reflect sample factors or other sources of
bias.