Interview setting and technique presumably alter efficacy of the postpartum interview. It is commonly believed that interviews are most effective when both parties are positioned at the same physical level.
To test the hypothesis that women's satisfaction and learning are improved when postpartum visits are conducted by pediatricians sitting in a chair rather than standing and are further improved when female pediatricians sit on the mother's bed.
Randomized double-blind study.
University hospital newborn nursery.
Seventy-five mothers on their first postpartum day. Patients' postpartum interviews were conducted by physicians assigned randomly to (1) sitting on the edge of the bed near the mother's feet, (2) sitting in a chair near the side of the bed, or (3) standing at the foot of the bed.
Main Outcome Measures
Within 60 minutes of completing the interview, an investigator (B.S.W.), blinded to physician position, asked each mother questions related to her satisfaction with the interview and the information that she was given. Differences among position groups were compared by means of 1-way analysis of variance or Kruskal-Wallis, χ2, or Fisher exact tests; P<.05 was considered statistically significant.
Potential confounding factors were similar among patients assigned to each of the 3 interview positions. The estimated duration of the interviews, the degree of satisfaction, and information retention were similar in each group.
Physicians need not make special efforts to conduct postpartum interviews in a seated position.
LABORATORY TESTS and diagnostic procedures are an important aspect of medical care. However, physician-patient interactions are arguably the fundamental basis of medical care. Efficacy of physician-patient interactions is critical not only in terms of patient satisfaction but also because effective interactions augment learning1 and may improve patients' ability to manage health-related activities.
Interview setting and technique presumably alter efficacy of the interaction. For example, interviews that encourage patients to ask questions are more satisfying and deal better with patients' concerns.2Immediacy is the principle that welcoming postures and close positioning between the communicators lead the addressee to infer that the communicator is more accepting.3 Greater immediacy facilitates and improves interviews because it leads to more effective encounters.4 One might assume, based on this theory, that sitting near patients would improve immediacy and, thus, satisfaction with interviews. This is consistent with the common belief that interviews are most effective when both parties are at the same physical level (ie, both sitting).1 Some female physicians further believe that women welcome closer interactions, such as sitting on the side of the hospital bed during interviews.
Optimal interactions are especially important when patients have a serious illness or are emotionally fragile. Postpartum women are often stressed, having frequently been denied adequate sleep during labor and delivery. Furthermore, roughly 20% are recovering from a cesarean delivery—a major operation that is often associated with considerable discomfort. Finally, the birth of a child is an emotionally charged experience that may limit the mother's ability to focus and assimilate new information. These factors put postpartum women at risk for ineffective interactions. We therefore tested the hypothesis that maternal satisfaction and learning are improved when postpartum visits are conducted by pediatricians sitting in a chair rather than standing and are further improved when female pediatricians sit on the mother's bedside.
With approval from the University of Louisville Human Studies Committee and written informed consent, we enrolled 75 mothers on their first postpartum day. The study was conducted in the newborn nursery at the University of Louisville Hospital. Mothers with infants in the intensive care nursery were ineligible to participate in the study.
Primiparous and multiparous women were enrolled. Exclusion criteria included inability to speak English fluently, serious postpartum complications, a history of serious learning disorders (eg, mental retardation), or indication on the patient's medical chart of major mental illness (eg, major depression or schizophrenia).
An investigator (B.S.W.) other than the interviewing pediatrician obtained informed consent. One of 2 female pediatricians (X.L.V. or M.K.) assigned to the newborn nursery then interviewed the patient. Physician position during the interview was assigned randomly to (1) sitting on the edge of the bed near the mother's feet (bed group), (2) sitting in a chair near the side of the bed (chair group), or (3) standing at the foot of the bed (standing group). Randomization was based on computer-generated codes maintained in sequentially numbered opaque envelopes; the envelopes were opened by the pediatrician just before each interview.
Interviews were scripted, with the scripts minimally modified as appropriate for individual mothers. The interviewer reassured mothers that their newborns were healthy and gave information to the new mothers detailing (1) the placement of the infant during sleep, (2) correct infant car seat use, (3) how to care for the umbilical cord, (4) how to define a fever and what to do for a fever, and (5) initial infant weight loss and feeding. Maternal questions were answered as usual. However, specific maternal-infant problems (eg, breastfeeding difficulties and newborn jaundice) that were not raised by the mother were explained during a separate visit unrelated to the study. Interviewers attempted to keep the duration of the interviews similar for each position (bed, chair, and standing).
An investigator (B.S.W.) blinded to physician position asked the mother a series of questions 15 to 60 minutes after the interview was completed.
The interviewing pediatricians timed each interview with a stopwatch. The number of questions asked by the patients was recorded.
The questionnaire given to the patient asked for the mother's perception of how long the pediatrician spent with her and her overall satisfaction with the encounter. Satisfaction was rated as excellent, very good, good, fair, or poor. We recorded potential confounding factors, including maternal age, ethnicity (white, African American, Asian, or Latino), parity, type of delivery (vaginal or cesarean), opioid use, and educational level (middle school, high school graduate, college graduate, or graduate degree). Pain was evaluated with the use of a 100-mm-long visual analog scale, with 0 mm indicating no pain and 100 mm indicating the worst imaginable pain. We also evaluated maternal retention of information presented during the interview. Learning was assessed by asking the mothers to answer the following questions:
What is the proper sleeping position for your baby (back, side, or front)?
Where should an infant car seat be located (front or rear seat)?
In which direction should your baby face (forward or backward)?
When can you expect the baby's umbilical cord to fall off (in weeks)?
What rectal temperature is considered a fever in a newborn?
When can you expect the baby to return to birth weight (in weeks)?
We assumed that the scripted interview would require about 5 minutes but would extend to close to 10 minutes with a typical number of maternal questions. We also assumed that patients would estimate the time elapsed during the standing interview reasonably accurately. We further assumed that sitting in a chair by the bedside would increase the mean ± SD time by 2 minutes, to 12 ± 2 minutes, and that sitting on the side of the bed would increase the time by an additional 2 minutes, to 14 ± 2 minutes. A sample-size estimate based on these assumptions indicated that 20 patients per group would provide 80% power for detecting a difference between the 3 groups at α = .05. We therefore enrolled 75 patients.
Our primary outcome measures were the mother's perception of how long the pediatrician spent with her, her overall satisfaction with the interview, the number of questions the patient asked, and retention of information presented during the interview. Secondary outcomes are given in Table 1.
When outcomes were normally distributed, differences in means among the 3 position groups were compared by means of a 1-way analysis of variance; Newman-Keuls posttests were used to compare individual groups. The nonparametric Kruskal-Wallis test was used for nonnormally distributed data. Categorical outcomes were analyzed by means of the χ2 test, except when small expected-frequency counts existed. In those cases, Fisher exact test was used. Data are presented as counts and percentages, means ± SDs, or medians and interquartile ranges. P<.05 was considered statistically significant.
One patient assigned to the chair group had recently been sedated and fell asleep in the middle of the interview. Another patient, assigned to the standing group, was inadvertently interviewed by a physician not involved in the study. These 2 patients were excluded, leaving 24 in the bed group, 27 in the chair group, and 22 in the standing group (Figure 1). None of the patients reported having had a prepartum consultation with a pediatrician regarding their upcoming newborn.
Flow diagram of study participants.
Patients in the 3 interview position groups were comparable with regard to amount of pain, actual duration of interview, type of delivery, ethnicity distribution, parity, and use of pain medication (Table 1). The mothers in the bed group were slightly older than those in the other 2 groups. All patients, except for 1 who had a graduate degree, had a middle or high school education.
There were no significant differences among the groups in the patients' estimated duration of their interviews, their overall satisfaction with their interviews, or the number of questions they asked during their interviews (Table 2). None of the interviews were rated as fair or poor by the patients. Our data provided 80% power for detecting a difference of 4 minutes in the estimated encounter duration. We also had 80% power to detect a 30% reduction (88% to 58%) in the proportion of patients who scored their satisfaction as excellent or very good vs good.
There were no differences between the groups in the number of correct responses to the questions asked 15 to 60 minutes after the interview (Table 3). Stratifying by parity did not alter the significance level of any of our results.
Postpartum interviews are usually the initial interaction between pediatricians and mothers. It is also when mothers are given considerable practical information about their newborns and important infant-safety information. However, postpartum women are often stressed and at high risk for poorly assimilating critically important information. We thus tested a simple, cost-free method of improving maternal satisfaction and learning during postnatal visits.
Patients in each designated position group overestimated the duration of the interviews by a factor of 2. Their estimates of the duration of the interviews conducted in each of the 3 randomly assigned positions were nonetheless similar. This result is inconsistent with our hypothesis and with the common clinical impression that the seated position augments the perceived duration of physician-patient encounters.5 However, our patients were of low socioeconomic status, with most achieving only a middle or high school education; it is possible that better-educated populations might be more sensitive to physician position during interviews. Similarly, physician position may prove more important in the context of a serious illness or when physicians deliver bad news.5
Most patients were satisfied with their postpartum pediatric interviews, with none rating the encounters as fair or poor. Satisfaction was comparable for each interview position. We thus failed to confirm our hypothesis that satisfaction would be improved when pediatricians sat in a chair at the bedside or sat at the foot of the patient's bed. It would nonetheless be a mistake to conclude that satisfaction among the 3 groups was comparable because the perceived duration was similar. In fact, there is considerable evidence that satisfaction correlates poorly with encounter duration.6
Patient satisfaction with medical encounters is related to physician courtesy and the amount of information provided.7 In contrast, satisfaction correlates poorly with nonverbal behaviors, such as eye contact, body positioning, and physical contact.7 Our results are consistent with this observation, as our encounters were scripted (and were thus similar in terms of courtesy and information content), with only body position differing among the groups.
We also failed to confirm our theory that information retention would be improved by an interview conducted in a seated position. Instead, information retention was good in all groups. The questions we asked represented the key information that new mothers should obtain from a postpartum interview. However, much of the information may already have been familiar to the mothers, regardless of parity. It thus remains possible that learning among the 3 groups would have differed more had the patients been presented with more complicated information.
There was considerable confusion in each of the 3 groups about the body temperature defining fever. Nearly a third of the patients responded that fever was identified by a rectal temperature of 104°F (40°C), rather than 100.4°F (38°C), as they had been instructed. At least in this population, the subtlety of the decimal point appeared to be confusing. Our data suggest that patients will be more likely to correctly identify hyperthermia if fever is defined with an integer value, such as 101°F (38.4°C).
The major limitation of our study was that the same 2 investigators conducted all interviews. Both were well aware of the study hypothesis; there was thus an unavoidable potential for bias. Yet, our inability to detect differences among the position groups suggests that investigator bias did not affect our study results. A second limitation is that this was a small study; therefore, the finding of no difference among the groups must be evaluated with caution. The patients in the bed group were slightly but significantly older than those in the other 2 position groups. However, the difference was only 3 to 4 years, which is of no clinical importance.
In summary, we failed to confirm our hypothesis that sitting near mothers improves satisfaction with postpartum pediatric interviews and increases the amount of information retained. Our results thus suggest that physicians need not make special efforts to conduct postpartum interviews in a seated position. Patients may be more likely to correctly identify hyperthermia if body temperature is defined with an integer value, such as 101°F, rather than a decimal value, such as 100.4°F.
Corresponding author: Daniel I. Sessler, MD, Outcomes Research Institute, 501 E Broadway, Suite 210, Louisville, KY 40202 (e-mail: Sessler@Louisville.edu).
Accepted for publication September 26, 2002.
This study was supported by grant GM 58273 from the National Institutes of Health, Bethesda, Md; the Joseph Drown Foundation, Los Angeles, Calif; and the Commonwealth of Kentucky Research Challenge Trust Fund, Louisville.
Mothers are given important information about their newborns during the postpartum visit. Interview setting and technique presumably alter efficacy of the interaction. For example, it is commonly believed that interviews are most effective when both parties are positioned at the same physical level (ie, both sitting), and some female physicians further believe that women welcome closer interactions, such as sitting on the side of the hospital bed during interviews.
We failed to confirm that sitting near the patient improves satisfaction with postpartum pediatric interviews or increases the amount of information retained. Our results suggest that physicians need not make special efforts to conduct postpartum interviews in a seated position.
Valdes XL, Kurbasic M, Whitfill BS, Sessler DI. Postpartum InterviewsFactors Affecting Patients' Learning and Satisfaction. Arch Pediatr Adolesc Med. 2003;157(4):327-330. doi:10.1001/archpedi.157.4.327