Opportunity costs of postpartum care: A national survey of U.S. providers’ priorities and practice

Background: In the wake of new guidelines for improved postpartum care, we examined postpartum care providers’ views to identify aspects of care that may be inefficient, as well as tradeoffs that may be being made between different types of critical care, given a provider’s limited time with each patient. Methods: We surveyed 600 randomly-sampled U.S. providers about postpartum care timing, the priority of specific aspects of care, and the frequency with which they are provided. Results: The survey response rate was 43% across medical specialties. More than 40% of providers reported a single 6-week visit (or more infrequent care) is optimal, regardless of new guidelines. Certain types of postpartum care critical to minimizing maternal mortality risk (e.g., depression screening), were highly valued and routinely performed. Other aspects of care (e.g., pelvic exam), were identified as systematic inefficiencies. Screening for intimate partner violence and addiction were performed less often than similarly-valued care. Certain types of care (e.g. transitioning to parenthood), were identified as crucial but were not currently addressed by national postpartum care practice guidelines. Approximately 25% of respondents regarded distance care as a feasible alternative to much of the postpartum care currently provided in-person. Conclusions: Critical opportunity costs in postpartum care delivery were identified that could have implications for maternal mortality risk. Clear guidelines on the highest value care for each visit, and which practitioner should provide that care, could diminish the burden faced by postpartum care providers, improving universal delivery of needed postpartum care. Complementary distance care (telemedicine or evidence-based apps) represents a novel and potentially equitable approach to implementing new care recommendations.

The traditional practice of a single postpartum appointment 6 weeks after birth has lately been called into question. The American College of Obstetricians and Gynecologists (ACOG) published new guidelines in 2018 specifically recommending postpartum care "become an ongoing process, rather than a single encounter, with services and support tailored to each woman's individual needs." 1 The guidelines were, in part, published in response to the increasing U.S. maternal mortality rate, 2 which stand in contrast to global maternal mortality patterns. 3 Maternal mortality review committees report that over 60% of U.S. maternal deaths may have been prevented with more timely diagnoses and effective treatment for postpartum onset conditions, as well as improved patient knowledge of warning signs. 4 Furthermore, research has demonstrated the majority of maternal deaths occur in the 42 days following birth, 5 with approximately 25% occurring after women are discharged from the hospital following pregnancies and deliveries that appear uncomplicated. Pregnancy-related deaths are highest among non-Hispanic black women, women at older ages, 6 and those with public health insurance. 7 Thus, traditional practices may result in a lack of care during a critical time, particularly for those women at baseline higher risk of maternal mortality.
Yet the ways in which postpartum care should be prioritized and structured for each individual woman remains an open question. Existing guidelines for postpartum appointment care provision vary significantly in their scope and detail. 8 This could make it particularly challenging for U.S. providers to know how fellow care providers with different clinical specialties are meeting the postpartum needs of shared patients or how to prioritize their own postpartum care provision, particularly under a mandate of earlier, more frequent, and more personalized care. Specifically, providers will need to determine what aspects of care to deliver at each postpartum care visit, how many visits are appropriate, and how to tailor that care to the individual patient's needs at each of those visits. To date, however, there is limited evidence even on postpartum care provider priorities overall.
This study seeks to examine the priorities and actual practices for postpartum care among a national sample of postpartum healthcare providers in the context of practically implementing new care recommendations. We analyze the tradeoffs that providers may make with limited appointment time and competing priorities for care, exploring digital complements to routine postpartum care delivery.

Methods
The mailing addresses of 6,000 active obstetrician-gynecologist (OB-GYN) physicians, Professional mailing addresses were verified with a Google search. Paper surveys were mailed in March 2018-May 2018. Each mailing contained a paper survey and prepaid return envelope, a cover letter with a $10 bill for participation, and a QR code for optional online survey completion. Surveys were formatted following Dillman's Total Design Method to improve validity and response rate. 25 Formatting included hand writing the mailed envelopes, using individual stamps rather than prepaid envelopes, personalizing the cover 6 letters with the clinician's title, and providing multiple options for completing the survey (online and paper). Three weeks after the initial contact was sent, a short letter was mailed reminding clinicians to complete the survey. Participants whose mail was returned to sender or who returned an uncompleted survey were not included in the follow-up mailing.

Statistical analysis
All paper survey responses were double-entered. All demographic variables and questionnaire responses were summarized by medical specialty (OB-GYN, Family Medicine, and Nurse Midwives) using descriptive statistics (mean and standard deviation for continuous measures; count and percent for categorical measures). We compared differences between practitioner groups using t-tests, ANOVA, and Chi-squared tests. All open-ended responses were coded by two independent coders for thematic content and a Cohen's kappa was calculated to measure inter-rater reliability.
To quantify mismatches in care priorities and actual practice, the effect size (Cohen's d) was calculated for differences in ratings. To ensure that effects were not attributable to different use of the two scales (importance and frequency), both scales were normalized for analysis. Items for which frequency had a higher mean score than importance, or vice versa, were considered to be 'inefficiencies.' Items with comparable importance scores, but differences in frequency, were considered instances in which providers may potentially be forced to make 'tradeoffs' in care.
The statistical software package SPSS 25.0 (SPSS Inc., Chicago, IL) was used for all data analyses.

Consent to participate
This study was approved by the University of Pittsburgh Institutional Review Board in November 2017 under the protocol number 17100584. The need for consent was waived 7 for this study by the University of Pittsburgh Institutional Review Board.

Respondents
Of the 600 surveys sent to providers (200 per specialty subgroup), 50 were excluded from analysis (twenty-nine surveys were returned to sender as an undeliverable address and twenty-one recipients returned an uncompleted survey due to being retired or having a non-relevant specialty). A total of 20 surveys were returned with no explanation. These 20 surveys were categorized as refusals and were not excluded.
Based on eligible responses, the overall response was rate 43% (236/550). The majority -72% (170/236) -returned the survey by mail. Table 1 provides respondents' demographic data and their characterization of their patient population by provider specialty. Region was determined from the respondent's reported primary practice zip code. Priorities for care Table 2 shows the mean Likert-scale-rated priorities for each postpartum care item compared with mean reported frequency of practice of that item across provider types. It also shows the effect size of the difference in priority and practice (Cohen's d), which appropriately adjusts for non-normal distribution. Notes * Importance scale ranged from "1 = not at all" to "5 = extremely," with a midpoint of "3 = moderately." † Frequency scale ranged from "1 = never" to "5 = always," with a midpoint of "3 = Indicates item is performed more frequently than it is valued.
In terms of specific aspects of care, there was generally high correspondence between valued and performed care. For example, depression screening was an item that was both highly valued and frequently performed, as was birth-related and pregnancy-onset complications. There were, however, a few large inefficiencies in care: the pelvic exam, counseling regarding resumption of sexual activity, and intimate partner violence screening. The first two items were performed more frequently than the level at which they are valued. Intimate partner violence screening, on the other hand, was performed less often than would be expected considering its value.
There were also some consistent differences in care provision and valued care by provider type. Several items were valued and performed differently depending on provider type.  The mean appointment attendance rate was 75% for patients of providers who supported telemedicine approaches, and 81% for patients of providers who did not, (t(198) = 1.97, P = .05), illustrating that those with lower postpartum attendance rates, tend to report higher support for telemedicine. As one provider noted, "There is benefit to human touch 13 and contact. There are also many nonverbal cues that could be missed when not viewing the total person. However, telemedicine beats no visit at all."
These 'tradeoffs' in care under time constraints may be the result of professional norms, personal comfort level or experience, or heuristics for prioritizing clinically pressing care which is more concrete to address. The fact that our open-ended questions identified an additional category of priority care provision (transitioning to parenthood) that was not covered by existing guidelines, further highlights the constraints of providing all desirable aspects of care during a single appointment. For providers who feel that a single 6-week appointment is the most postpartum care needed may not have or take the opportunity to probe into these risks.
Provision of transportation and home-visits were popular suggested solutions to care access barriers. While transportation assistance has had mixed documented effectiveness on appointment attendance, [27][28][29] in-home visits are generally successful for assessing and reducing postpartum depression. 30 However, they are also expensive to scale.
Telemedicine may present a cost-efficient alternative for providing postpartum psychosocial care for those practices where telemedicine is feasible. While was notably absent from open-ended responses about overcoming care access barriers, when prompted to think about it, telemedicine was supported as a replacement for postpartum care by almost a quarter of respondents.
Many respondents reported caveats to telemedicine support in open-ended text responses.
While it was suggested that most aspects of a physical exam (e.g., pelvic exam or some contraceptive method provision) require in-person care, many of the highly valued psychosocial assessments could be performed virtually, either through telephone or videoconferencing. Therefore, guidelines that specify which types of postpartum care provision are best suited to an in-person versus a telemedicine visit could address some of these concerns. For example, an early telemedicine visit could be used to assess behavioral health and ask about physical symptoms. An in-person physical exam could then be scheduled for those women who report concerning physical symptoms or require further in-person counseling. Some providers voiced concern that telemedicine could negatively impact patient's comfort with disclosure or the provider's ability to detect subtle signs of psychosocial risk factors, such as depression. Given the time and effort required to coordinate video conferencing between patient and provider (as opposed to telephone contact), further investigation is warranted into both cost effectiveness and the relative rates of disclosure with telemedicine and other forms of distance care.
An important consideration for distance care is existing health inequities. Women with public insurance and African-American women are at considerably higher risk of severe postpartum maternal morbidity and mortality. Results showing that postpartum visit attendance is less common among women with public insurance were replicated here in providers' self-reports.
Given that internet access is also lower among this population, telemedicine may not be a feasible across-the-board solution to more frequent care provision. Currently, 94% of reproductively-aged women own a smartphone across sociodemographic groups 31 and early research suggests that African-American women may actually be more reliant on digital sources for accessing health information. 32 As such, evidence-based mobile apps, with provider-oversight, may be the most promising alternative for engaging and monitoring postpartum risk among a diverse population of peripartum women and could even help bypassing structural racism or care access inequities.
Our findings also suggest that care provision could and perhaps should be delineated and

Limitations
Our sample was small and, although respondents were randomly sampled from national lists, the generalizability of our findings is limited. While the consistency in both value and realized practice across specialties in our sample could reflect a consensus among practitioners, it is also possible that our sample is biased towards those with strong opinions regarding postpartum care practices; therefore, caution should be exercised in considering any differences by specialty.

Consent for publication
Not Applicable

Availability of data and materials
The datasets created during the current study is available from the corresponding author on reasonable request.