Discussion Questions

• ■
  

  Why is this topic worthwhile?

  
  
• ■
  

  What are the advantages and disadvantages of the study design?

  
  
• ■
  

  What were the primary and secondary objectives of the study?

  
  
• ■
  

  Did the research assistant provide appropriate information to participants?

  
  
• ■
  

  Was the proper clinic employee surveyed?

  
  
• ■
  

  What is the clinical impact of this study?

Introduction

Preconception care is essential, according to the Centers for Disease Control and Prevention. The agency notes that significant problems can occur by the 11th or 12th week of pregnancy. It also points out that 11% of pregnant women smoke, and 10% drink alcohol. As for women with the potential to become pregnant, 69% do not take folic acid supplements, 31% are obese, 4% have preexisting medical conditions, and 3% take teratogenic medications. Obviously, then, women should be seen soon after a pregnancy is discovered—if not before. Yet, Nettleman and colleagues found that a considerable number of practices delay that crucial first appointment.

Anthony Shanks, MD, and George A. Macones, MD, MSCE, Associate Editor

Shanks: The study presented by Dr Nettleman and colleagues addresses a very thought-provoking subject: the office-based component of delayed entry into prenatal care. Is this a worthwhile topic to study? Why?

Groden: Theoretically, there should be no office-based contribution to delays in obtaining prenatal care; in reality, this may not be the case. The first weeks of pregnancy are some of the most crucial since organogenesis takes place, and this period presents a unique opportunity for intervention. However, late presentation to the obstetrician’s office often causes this opportunity to be missed. Whether late presentation is the fault of the patient or of the office has never been systematically studied until now, and it is an important topic to examine because the results could affect obstetric practices across the country—from community-based practices to large academic institutions. If the office is a factor in postponed prenatal care, its practices can be altered to eliminate the delay, thereby optimizing early prenatal care for patients.

The secondary goal was to assess the advice given to pregnant women when they were scheduling the appointments. It is also useful to examine this aspect of early prenatal care, as there are many vital interventions that can have a lasting effect on the fetus. Patients who take daily medications or who engage in potentially harmful behaviors may not recognize that they need to make changes without medical advice. Additionally, interventions like folic acid and prenatal vitamins are most effective in the early weeks of the first trimester. The earliest chance the practitioner has to offer these interventions might be the phone call the patient makes to set up her first prenatal appointment. Overall, this study offers an important evaluation of early intervention points, and as such, it is very worthwhile.

Shanks: We think that in an ideal world, there should be no barriers to medical care—especially for pregnant women. This topic is intriguing because it explores how much culpability falls on providers.

Study Design

Shanks: What kind of study design is this? What are the advantages/disadvantages of this type of study?

Reed: A cross-sectional design was used. This is a useful format when describing variables and their distribution patterns. The researchers evidently selected this design because they were sampling obstetrical offices in a single state over a short period; it allows for rapid and inexpensive data collection and can be advantageous when associations are examined. In addition, a cross-sectional study can be included as the first step in a cohort study without adding extra cost.

Unlike a cohort study, cross-sectional studies only measure prevalence—they lack the longitudinal time dimension. Other weaknesses of this design include difficulty establishing causal relationships from observational data; and, they are impractical if used to study rare outcomes.

Shanks: Cross-sectional studies are observational studies that provide a “snapshot” at a particular point in time. As mentioned, a major strength of cross-sectional studies over cohort studies or clinical trials is that they are fast and inexpensive. There is no loss to follow-up in this design because you obtain all the information at once. As you mentioned, this type of study design can provide information about prevalence—the proportion who have a disease or condition at a particular point in time. A weakness of cross-sectional studies is the difficulty in establishing causal relationships from data obtained.

Shanks: What were the primary and secondary objectives of this study? Are there any alternative study designs that answer these questions?

Carbone: The primary objective of this study was to assess the office-based component of delay in obtaining prenatal care. Appointment availability was assessed and the ability of the patient to negotiate an earlier appointment date was examined. The secondary objective was to assess the advice given to a pregnant woman when she calls to schedule her first prenatal appointment. To accomplish this objective, the questions asked by the office were evaluated as was the advice given to the patient.

An alternative study design would be to perform a patient survey during pregnancy. This would be a good way to assess the patients’ overall satisfaction with the advice they were given and the time frame in which they obtained their first prenatal visit. With this approach, the authors could also see if the delay in establishing prenatal care had any impact on the pregnancy outcomes. What’s more, the investigators could examine the differences between federally-funded clinics versus private obstetrical-gynecological practices versus maternal-fetal medicine offices.

Shanks: Another option would be a case-control study. Pregnancies could be evaluated for outcomes of interest and then retrospectively evaluated to determine when their prenatal care began. While case-control studies are useful for evaluating rare outcomes, they are unable to estimate incidence or prevalence in a population. Another weakness of case-control studies is the susceptibility to bias in both the cases and controls, though this can be limited in the study design. Ultimately, this type of study design is useful for generating hypotheses—something that the authors were able to do in their study.

Shanks: What do you think about the study’s inclusion/exclusion criteria? Were the analytic statistical methods appropriate for the study design?

Goetzinger: To be included in the study, physicians’ offices had to be located in a single state and have phone numbers that were listed in a commercial directory service. Offices were excluded from the study if they were the only listed obstetrical office in the county, had an invalid phone number listing, could not be reached after multiple attempts, were not accepting new patients, did not accept self-referrals, did not accept the insurance product, or did not have appointment scheduling available. All of these criteria are reasonable and make the study feasible, although they may introduce some bias into the study. For example, counties with only 1 listed obstetrical office phone number were excluded to maintain anonymity. While anonymity is an important aspect to the study, these particular counties may be the very ones in which access to early prenatal care might be limited.

Exactly how many counties were excluded based on this factor or on the county’s population density was not noted. The remaining clinics were excluded based on logistical factors that made them impossible to study (eg, researchers were unable to get through on the phone line or the offices were not accepting new appointments). These same logistical factors might also mimic some of the barriers faced by patients when trying to schedule their first prenatal appointment. Readers need to consider these inclusion/exclusion criteria in order to accurately identify the study population. This will allow them to determine whether the study results are generalizable to their own patient population.

The statistical methods were also appropriate for this study design. χ ² tests were used to compare categorical data and Student t tests were used to compare continuous variables. Finally, linear regression was used to evaluate for an association between gestational age at appointment time and clinic density. Results were reported as an r ² value (the coefficient of determination). As a reminder, r ² is a test of “goodness of fit” of the model and shows how closely X and Y are related. The r ² values can fall between 0 and 1, with 0 representing no correlation and 1 representing perfect correlation. In this study, linear regression was used to evaluate for an association between gestational age at appointment time and clinic density, yielding an r ² value of 0.05.

Shanks: Excellent. Did the research assistant (RA) provide appropriate information? Can you envision different responses if the RA provided different information?

Groden: The RA gave the date of her last menstrual period and the results of her urine pregnancy test. When prompted, she revealed she had Blue Cross insurance, was married, 29 years old, a gravida 1, and without health problems. These were all suitable responses. It would be interesting to perform a similar study but supply different answers; factors like insurance or the presence of serious medical problems might have altered the responses. If a patient had chronic medical problems, would the patient have been seen earlier? Would she have been referred to a maternal-fetal medicine specialist? If the patient reported she did not have insurance, would she have gotten an appointment? Would she have been offered the same appointment as the insured caller or a later one? It is quite possible that the results of this study would have differed if the caller had given different information.