Lessons learned from the Philadelphia Collaborative Preterm Prevention Project: the prevalence of risk factors and program participation rates among women in the intervention group

Approximately 65-75% of infant mortality in the U.S. is related to preterm birth (PTB), and premature births account for almost half of all chronic and severe morbidity in young children [1],[2]. Every year more than 450,000 babies in the U.S. are born prematurely (defined as less than 37 completed weeks of gestation), accounting for 11.6 percent of all live birth deliveries. Consequently, PTB has been the subject of a considerable amount of epidemiological and medical research. Unfortunately, the causes are still not well-understood, and thus the path to reducing PTB remains somewhat elusive [2].

The strongest predictor of PTB is a previous PTB. Women with a previous PTB are between 2 to 5-fold greater risk for a second PTB, compared to women with a previous full term delivery [3]. Moreover, the earlier the gestational age of a PTB, the greater the risk for a repeat PTB. As many as 50% of women with a prior history of PTB occurring at or before 32 weeks gestational age (GA) have a subsequent PTB [3],[4].

Past efforts to reduce PTB have been largely unsuccessful, perhaps because such efforts target a single risk factor, and the corresponding services or interventions have been offered only during the course of pregnancy. What may be needed, therefore, are more comprehensive, long term interventions, beginning prior to conception and thus outside the context of traditional prenatal care [2],[5],[6].

In light of the extremely high rate of repeat preterm delivery, the most obvious target group for inter-conceptional risk reduction is women with a history of preterm birth. Unfortunately, little is known about this manifestly high risk group in terms of the degree of willingness or ability to participate in appropriate interventions or services in the postpartum or inter-conceptional period.

Presented here are the result of an analysis of data from the Philadelphia Collaborative Preterm Prevention Project (PCPPP), a large randomized control trial designed to identify and reduce risk factors for repeat preterm births among women who previously delivered moderately to severely premature (<35 weeks gestation) infants. PCPPP data were examined to determine both the prevalence of the major risk factors for PTB among enrollees, the frequency with which enrollees agreed to treatment, as well as the actual participation rates pertaining to each of the intervention arms of the study.

As Figure 1 shows, 6 percent of women had no identified risk factors, 14 percent had only one risk factor, 22 percent had two risk factors, 31.4 percent had three risk factors and 25.9 percent had 4 or more major risk factors for a PTB. Prevalence rates for each of the major risk factors assessed are presented in Table 3. Of the women in the study sample 57.3 percent were diagnosed with urogenital tract infection, and 58.6 percent with periodontal disease. About 39 percent of women in the intervention group were identified as cigarette smokers. Just over 60 percent of the women had CES-D scores of >16; of these 81 or 17.2 percent were determined to be clinically depressed. A total of 106 or 22.5 percent of enrollees were determined to be reading at low levels of literacy. The assessment of housing conditions revealed that 392 of the women, or 83.2 percent, were living in an unstable housing situation.

Intervention/treatment acceptance rates are presented in column 2 of Table 3. In all cases, the majority of women was receptive and thus agreed to the respective treatment protocols. Specifically, of those diagnosed with a urogenital tract infection, 90.2% stated they were willing to accept treatment and took the medication in the presence of PCPPP staff. Acceptance rates for periodontal disease, clinical depression, literacy and housing were 88.6%, 95.1%, 68.9%, and 91.1%, respectively. In comparison, among those who were identified as smokers and therefore eligible for treatment, only 57.3% stated they were willing to accept treatment.

Participation rates, as percentage of women eligible for treatment varied considerably across the intervention arms of the study (Table 3: column 3). Exactly 85.0 percent of women diagnosed with urogenital tract infection met the definition of minimal participation described earlier, compared to 48.7 percent for periodontal disease. Among identified smokers, 28.1 percent participated by receiving some form of intervention designed to help women reduce or quit smoking. For the women with depressive symptomatology and who were subsequently determined to be clinically depressed the corresponding participation rate was 74.1 percent. The rate of participation for women identified with low literacy levels was 40.2 percent. Finally, among those women determined to be living in an unstable housing conditions, 66.1 percent participated by receiving some form of housing assistance.

There were fewer instances where race/ethnicity and/or insurance status were significantly associated with rates of acceptance, with the general trend toward more willingness to accept treatment for minority and Medicaid or uninsured women. Specifically, Black (91.9%) and Hispanic (83.7%) women who were eligible for periodontal treatment were more likely to accept treatment than were White women (40.0%), while uninsured women or women on Medicaid were more likely than women with private insurance to accept the intervention for low literacy (74.4 vs. 45.0%), and more likely to accept help with housing problems (93.8 vs. 83.3%). Notably, as shown in column 3 of Table 4, there were no differences according to either race/ethnicity or health insurance status in terms of the rates of participation (among those who accepted treatment) for any of the treatment arms of the study, a finding which we discuss in more detail below.

The main purpose of the PCPPP trial was to determine if a set of systematic, coordinated, evidenced-based risk-reduction strategies implemented during the inter-conception period would reduce the risk of a subsequent PTB. We are unaware of any similar large scale RCT attempting to mitigate multiple known risk factors for repeat PTB, in a cohort of women in the immediate postpartum period following a premature delivery.

The findings presented here document the prevalence of the risk factors in the study cohort, the willingness of PCPPP enrollees to agree to receive treatment, and the extent to which women participated in the medical and other services offered as part of the PCPPP trial. As noted earlier, participation rates were calculated based on the lowest possible yet reasonable thresholds for classifying an enrollee as having `participated’. In addition, the PCPPP trial was designed to incorporate every reasonable effort to eliminate, or at least minimize all known barriers to accessing treatment or services.

Not surprisingly, given that the sample consisted of women who had just delivered prematurely, the prevalence of the risk factors in the study sample appear to be high when compared to available data for similar sociodemographic populations. Recent data available from the Centers for Disease Control (CDC) , for example, revealed that 24.7 percent of adults 30 to 34 years of age had mild to severe periodontal disease, and that rates increased markedly with age for both men and women [22]. The presence of periodontal disease in the CDC study was established according to the CDC/American Academy of Periodontology definitions, which are not directly comparable to ours. However, using a full-mouth examination and clinical confirmation by a periodontist we found that 58.5 percent of the PCPPP intervention group had moderate to severe periodontal disease. Pocket depth is generally considered an important marker for periodontal disease and all women identified as having periodontal disease in the PCPPP trial had at least one site with pocket depth ≥4 mm. By comparison, the detailed data from the above-mentioned CDC study showed that 29.6 percent of all adults 30–34 had at least one site with pocket depth >4 mm and that rates increased markedly with age. This is consistent with the conclusion that PCPPP women had a prevalence of periodontal disease considerably higher than that in the general population of women of similar age, especially in light of the fact that the average age of the PCPPP intervention group was only 25.6.

Data from the recent National Health Interview Survey reveal that about 21 percent of women 18–44 years age are current smokers [23], compared to almost 40 percent reported here. With respect to the prevalence of depression, reports tend to vary considerably depending on the screening instrument used, the cut-offs used to define depression and depression severity, or whether point or period prevalence rates are in question, among other things. Hence the precise `true’ prevalence of depression among postpartum women or women of childbearing age is unclear. A systematic review of the literature by Gavin and colleagues revealed that, approximately 19 percent of new mothers suffer from some form of depression at some point during the first three months following delivery, and as many as 7 percent have been diagnosed as having suffered from major depression [24]. In the PCPPP study a total of 471 women in the intervention group were screened for depressive symptomatology during the first 3 months postpartum; of those 237 or 29 percent were identified as being “possibly depressed” according to the Center for Epidemiological Studies of Depression Scale (CES-D). Subsequent evaluation of the “possibly depressed” group using the SCID revealed that 16 percent of the intervention group was clinically depressed, more than twice the equivalent period prevalence rate reported by Gavin and colleagues.

Urogenital tract infections screened for in the PCPPP study intervention included the more commonly reported conditions of bacterial vaginosis, chlamydia, and gonorrhea. The prevalence of these conditions in the study sample was 49, 10, and 1.7 percent, respectively, compared to 19, 2.2, and 0.24 percent reported elsewhere for young adult women or women of childbearing age in general [25]-[27].

In addition to the prevalence of the risk factors targeted for intervention in the PCPPP trial we also documented the rate at which eligible women were willing to participate in the respective interventions. With the exception of the smoking intervention the overwhelming majority of women identified with other risk factors expressed the willingness to receive the appropriate treatment or services. The analysis of participation patterns revealed that eligible women availed themselves of the various PCPPP treatments and services in widely varying degrees, ranging from a high of 85 percent (for infection) to a low of 28 percent (for smoking). Approximately 2 out of every 5 women identified with low literacy levels eventually received any related services, and fewer than half of the women with confirmed periodontal disease ever had any follow-up treatments. Again, in each case, every reasonable effort was made to remove or mitigate what are the most often-cited `barriers’ to receiving health care and related services -- including full reimbursement for costs and personalized accommodations related to transportation to and from clinical sites, free childcare, flexible and convenient clinical hours, and even home visits when necessary.

Not surprisingly, in most cases, the prevalence of risk factors was significantly higher for minority as compared to white women, and for low income women, as measured by insurance status. Black women enrolled in the study, for example, were more than two times more likely than white women to have been diagnosed with a urogenital tract infection and more than two and one-half times more likely to have been diagnosed with periodontal disease. In addition, both Black and Hispanic women were significantly more likely than white women to have reported housing instability, and were also significantly more likely to have elevated rates of low literacy.

There was no evidence to suggest that minority and low income women were either more reluctant to accept or able to take advantage of medical treatments or services offered as part of the PCPPP trial. Where rates of acceptance differed to any degree they tended to be higher among minority as opposed to white women (in the case of periodontal disease and low literacy) or higher among women who were uninsured or on Medicaid, as compared to privately insured women (in the case of housing instability). Moreover, there were no significant differences in participation rates associated with either race/ethnicity or insurance status, pertaining to any of the treatments or services offered as a part of the PCPPP interventions. The absence of any meaningful relationship between these socioeconomic factors and participation rates for the PCPPP trial suggests that the strategies used to make treatments and services as accessible as possible efforts were fairly successful.

The findings presented here also highlight the possibility of widely varying participation rates pertaining to interventions or programs offering multiple health and health-related treatments and services to women in the pre- or inter-conception period. Undoubtedly, participation will vary according to the perception of the benefits to those who enroll. Many other factors, however, will likely influence participation rates across `types’ of service or treatments. Ensuring that high-risk women will be able or willing to avail themselves of multiple appropriate treatments or services during the pre- or inter-conceptional period is increasingly being recognized as necessary to reduce the risk of PTB, since the underlying cause may be the cumulative effect of exposures and conditions which predate pregnancy, and may not be detectable or adequately addressed within the gestational period [2],[6],[7],[28]. Unfortunately, the forces that motivate, enable, or inhibit women from accessing needed services outside the context of prenatal care, in the pre- or inter-conceptional period, are likely to be quite complex and are by no means fully understood. Successfully identifying the appropriate strategies for increasing access to and use of risk-reduction strategies and services may well require innovative and multi-level methodologies that match the complex nature of the problem.

In summary, further research in this area appears warranted. We believe that we systematically and thoroughly addressed the well known and widely cited barriers associated with women’s access to and use of health-related services. Efforts to maximize participation of PCPPP enrollees were quite extensive, conducted as part of a fully-funded, large-scale randomized control trial to reduce repeat preterm birth. Nonetheless, participation rates varied and in some cases were quite low, even when defined in fairly minimal terms. This raises not only an important question about what more needs to be done in order to successfully deliver multiple high-quality risk-reduction interventions and services to high risk women, but also the relative costs and benefits of doing so. The more the chances of successfully reducing preterm birth rates hinge and are placed on interventions targeting high risk women prior to conception, the more important an adequate understanding of how to successfully reach such women with appropriate services and treatment becomes.

DW prepared the manuscript, participated in analysis and interpretation. LM made substantial contributions to the data analysis, interpretation, manuscript preparation, JC is the Principal Investigator for the PCPPP and made substantial contributions toward analysis and interpretation and manuscript preparation. All authors read and approved the final manuscript.