Rating the seriousness of maternal and child health outcomes linked with pregnancy weight gain

Abstract

Background:

Current pregnancy weight gain guidelines were developed based on implicit assumptions of a small group of experts about the relative seriousness of adverse health outcomes. Therefore, they will not necessarily reflect the values of women.

Objective:

To estimate the seriousness of 11 maternal and child health outcomes that have been consistently associated with pregnancy weight gain by engaging patients and health professionals.

Methods:

We collected data using an online panel approach with a modified-Delphi structure. We selected a purposeful sample of maternal and child health professionals (n=84) and women who were pregnant or recently postpartum (patients) (n=82) in the U.S. as panelists. We conducted 3 concurrent panels: professionals only, patients only, and patients and professionals. During a 3-round online modified-Delphi process, participants rated the seriousness of health outcomes (round 1), reviewed and discussed the initial results (round 2), and revised their original ratings (round 3). Panelists assigned seriousness ratings (0, [not serious] to 100 [most serious]) for infant death, stillbirth, preterm birth, gestational diabetes, preeclampsia, small for gestational age (SGA) birth, large for gestational age (LGA) birth, unplanned cesarean delivery, maternal obesity, childhood obesity, and maternal metabolic syndrome.

Results:

Each panel individually came to a consensus on all seriousness ratings. The final median seriousness ratings combined across all panels were highest for infant death (100), stillbirth (95), preterm birth (80), and preeclampsia (80). Obesity in children, metabolic syndrome in women, obesity in women, and gestational diabetes had median seriousness ratings ranging from 55 to 65. The lowest seriousness ratings were for SGA birth, LGA birth, and unplanned cesarean delivery (30 to 40).

Conclusion:

Professionals and women rate some adverse outcomes as being more serious than others. These ratings can be used to establish the range of pregnancy weight gain associated with the lowest risk of a broad range of maternal and child health outcomes.

Background

The Institute of Medicine gestational weight gain guidelines are used by health care providers worldwide to counsel women on how much weight they should gain in pregnancy ^{1, 2}. In 2009, the IOM recommendations were revised to consider adverse impacts not only on the infant, but also on the mother ¹. They recognized that in addition to preterm birth, infant death, small-for-gestational age birth, and child obesity, pregnancy weight gain was also associated with unplanned cesarean delivery and postpartum weight retention¹.

The 2009 IOM committee identified an important challenge in creating guidelines that minimize the risk of a number of different health outcomes related to pregnancy weight gain ¹. Establishing an optimal range of weight gain that considers multiple health outcomes simultaneously is challenging because women and their health care providers may view some outcomes as being more important or serious than others. For example, an infant death may be viewed as more serious than an unplanned cesarean delivery. In an effort to account for the relative severity of different health outcomes related to pregnancy weight gain, the IOM commissioned a quantitative risk trade-off analysis ¹. Unfortunately, health utility data were only available for 3 outcomes. With many other key outcomes left out, the relevance of the work was limited. Although perinatal scoring tools have been developed to combine different outcomes into a single composite and weight outcomes according to their severity ^3–5, their value is limited when studying gestational weight gain because they either do not consider both mother and child or include short- and long-term health outcomes.

In the absence of stakeholder input, guidelines will be created based on implicit assumptions of a small group of experts about the relative seriousness of adverse health outcomes. Therefore, they will not necessarily reflect the values of women ⁶. The present study was conducted in response to the IOM Committee’s call for researchers to fill this knowledge gap ¹. We sought to determine the seriousness of 11 health outcomes that have been consistently associated with pregnancy weight gain by engaging maternal and child health professionals and women who were pregnant or postpartum using an online modified-Delphi panel process.

Methods

The Delphi technique is a well-established method for exploring viewpoints of diverse stakeholder groups on a specific topic ⁷. In an iterative process, panelists score items, provide a rationale for their ratings, review other panelists’ responses, and revise their initial scores. The process is anonymous, minimizing the negative effects of group decision-making, such as “groupthink. ⁸ Delphi panels support the engagement of more diverse and representative panelists, reduce costs, increase time-efficiency, and anonymize discussions to reduce biases based on participant status or personality ^{9, 10}.

We collected data using an online panel approach with a modified-Delphi structure called ExpertLens™, which was created by researchers at the RAND Corporation ^11–18. This approach replaces traditional face-to-face meetings with anonymous moderated online discussion boards and facilitates and automates data collection and analysis. ExpertLens allows engagement of a large number of geographically-diverse panelists by providing them with an opportunity to anonymously share their perspectives and interact with other participants using their own computer at their convenience ¹⁹. It has been used in numerous studies to elicit opinions from diverse stakeholder groups including researchers, providers, policy makers, patients, and community members ^11–18.

Our study protocol has been described in detail elsewhere ²⁰. Our final methods were informed by a pilot study of the 3-round ExpertLens process ²⁰. While surveys aim to recruit a large, representative sample, the goal of an expert professional panel is to purposefully recruit the most knowledgeable individuals in the field to elicit their expertise, which in this context we considered experience with pregnancy either professionally or personally. We used social media and our professional networks to recruit a group of 90 maternal and child health professionals who have worked in the field at least 5 years. Content expertise was our only selection criterion for this panel.

We additionally used social media to recruit 90 women who were pregnant or no more than two years postpartum at enrollment. Panel membership on the patient’s panel was chosen using content expertise as a primary selection criterion, and racial/ethnic and geographic diversity as secondary selection criteria. To ensure we met these goals, we initially screened a sample of 365 women who responded to our advertisement, and then purposively selected all the non-white women and at least one woman from every US state. We then randomly selected among the remaining women to reach our recruitment target of 90 women. Such a purposive sampling approach is typical for stakeholder engagement panels ²¹.

From October 9th to November 26th, 2019, we conducted 3 concurrent panels: a panel of 60 professionals, a panel of 60 patients, and a mixed panel of 30 professionals and 30 patients. We conducted homogeneous panels and a mixed panel to evaluate differences in ratings within and between the stakeholder groups and the impact of exposing participants in a mixed panel to alternative views. Both patients and professionals were randomly assigned to either a homogeneous or a mixed panel and were informed of their assignment before the start of the study.

The purpose of the panels was to rate the seriousness of maternal and child health outcomes that have been consistently associated with pregnancy weight gain and are clearly operationalized or measured in most research studies. We selected 11 health outcomes based on a thorough review of the literature, focusing on the outcomes identified as relevant by the 2009 IOM committee ¹ and a recent Delphi consensus panel on nutrition in pregnancy ²². The outcomes included infant death, stillbirth, preterm birth, gestational diabetes, preeclampsia, small for gestational age (SGA) birth, large for gestational age (LGA) birth, unplanned cesarean delivery, obesity in children, obesity in women, and metabolic syndrome in women.

We developed background information for each outcome, including its definition and short- and long-term consequences (Appendix S1) ²⁰. We based this information primarily on UpToDate, a well-known evidence-based clinical resource. To make the information more accessible for patients, we relied on UpToDate’s “The Basics” text, which provides short overviews written with plain language principles.

Each panel completed a 3-round ExpertLens process. In Round 1, panelists reviewed background information for each health outcome (Appendix S1). We asked them to rate the seriousness of each outcome using a rating scale of 0 to 100, where 0 corresponded to not serious at all and 100 corresponded to the most serious. Panelists also provided open-ended rationales for their ratings. In Round 2, panelists were presented with a summary for each outcome which included their seriousness ratings relative to the panel’s medians and quartiles, the other panelists’ rationales, and a table summarizing the themes of these rationales. They then participated in an asynchronous, anonymous online discussion with other panel participants over the span of 14 days ¹⁰. In the mixed panel, the anonymous screen names indicated whether the participant was a professional or a patient. Participants were able to leave as many comments as they wanted and could respond directly to other participants’ comments while Round 2 was open. Finally, in Round 3, participants revised their Round 1 seriousness ratings based on group feedback and discussion and could explain any changes they made to their scores. Each round was open 1‒2 weeks. Participants were given a gift card for $165.

In contrast to some Delphi panels that require participants to reach consensus and use consensus as a criterion for determining the number of data collection rounds, we informed participants about the number of rounds before the start of the data collection ²³. In contrast to some Delphi panels that require participants to reach consensus and use consensus as a criterion for determining the number of data collection rounds, we informed participants about the number of rounds before the start of data collection. We did not aim to have panels reach consensus on the outcomes’ relative seriousness because we believe that the variation in scores represents the reality of the diverse experiences of women and care providers and should be reported in the literature.

Statistical analysis

We first calculated the degree of within-panel consensus of seriousness ratings using the RAND/UCLA Appropriateness Method’s (RAM) approach ²⁴. Because RAM’s approach was developed for 9-point scales and looks at the distribution of responses between the three groups categorized based on tertiles, we first divided the seriousness ratings into 3 groups: serious (68 to 100), moderate (34 to 67), and low (0 to 33). We then applied the RAM’s approach to determining agreement and the final group rating. We also examined differences in distributions of scores across panels, using a 10-point difference in the median scores as evidence of a meaningful difference.

We defined the final seriousness ratings based on panelists’ Round 3 responses. If panelists did not answer a given question in Round 3, we used their Round 1 seriousness scores ¹¹. We summarized final seriousness scores for each panel by calculating the median, interquartile range (25^th‒75^th percentile), 5^th and 95^th percentiles.

We also examined scores according to various participant characteristics descriptively. We did not use statistical tests to quantify differences in scores, as we anticipated relatively small numbers in some categories, leading to sparse data when cross-tabulating according to multiple participant characteristics.

Ethics approval

Ethics boards at the University of Pittsburgh and The RAND Corporation determined this study to be exempt from review.

Results

Of the 180 invited participants, 92% (84 professionals and 82 patients) participated in at least one panel round. Among the 166 who participated in at least one round, 78% (64 professionals and 66 patients) participated in all 3 rounds, 92% (76 professionals and 76 patients) participated in at least 2 rounds, and 8% (8 professionals and 6 patients) participated in only one round. Round 1, Round 2, and Round 3 participation rates were 88%, 77%, and 83%, respectively. A total of 115 panelists (64%) posted at least one comment in Round 2 (mean [standard deviation] number of comments was 8.1 [7.2], range 1‒36). Our final analytic sample included the 166 panelists who participated in Round 1 or Round 3. Of these, 86% participated in both rounds.

We achieved our goal of ensuring racial/ethnic and geographic diversity in panelists (Table 1). Approximately one-third of both groups were non-white. Hispanic participants made up 12% and 23% of the professionals and patients, respectively. Among both participant groups, all geographic regions were represented.

Table 1.

Characteristics of ExpertLens™ panel participants, n=166^a

Characteristics	Professionals (n=84) n (%)	Patients (n=82) n (%)
Race
White/European/Middle Eastern	54 (64)	51 (62)
Black/African American	13 (15)	10 (12)
Native American/Alaska Native	1 (1)	1 (1)
Native Hawaiian/Other Pacific Islander	0 (0)	0 (0)
Asian/Indian	8 (10)	7 (9)
Other race not listed	4 (5)	6 (7)
Two or more races	4 (5)	7 (9)
Hispanic ethnicity
Yes	10 (12)	19 (23)
No	74 (88)	63 (77)
Region of residence
West	17 (20)	12 (15)
Midwest	14 (17)	24 (29)
Northeast	20 (24)	13 (16)
South	33 (39)	33 (40)
Division of residence
Pacific	13 (15)	5 (6)
Mountain	4 (5)	7 (9)
West North Central	5 (6)	10 (12)
West South Central	8 (10)	8 (10)
East North Central	9 (11)	14 (17)
East South Central	2 (2)	6 (7)
Middle Atlantic	9 (11)	8 (10)
South Atlantic	23 (27)	19 (23)
New England	11 (13)	5 (6)
Self-identified gender
Male	12 (14)	0 (0)
Female	71 (86)	82 (100)
Other	0 (0)	0 (0)
Age (years)
<30	2 (2)	18 (22)
30–39	31 (37)	60 (73)
40–49	32 (38)	4 (5)
≥50	19 (23)	0 (0)
Education
Some high school	0 (0)	1 (1)
High school graduate or equivalent	0 (0)	8 (10)
Bachelors degree	1 (1)	21 (27)
Masters degree	7 (8)	26 (34)
Doctoral degree	76 (91)	21 (27)
Years worked in maternal and child health		n/a
5–9	23 (27)
10–14	28 (33)
≥15	33 (39)
Professional work setting
Academic	56 (67)	n/a
Hospital or doctor’s office	7 (8)
Government	10 (12)
Community health organization	11 (13)
Primary professional role
Researcher	64 (76)	n/a
Health care provider	15 (18)
Administrator or policy maker	2 (2)
Public health worker	3 (4)
Currently pregnant
Yes	n/a	22 (27)
No		60 (73)
Number of live births
0	n/a	5 (6)
1		28 (34)
2–3		42 (51)
≥4		7 (9)

^aIncludes participants who completed at least round 1 or round 3 of the study.

Most professionals identified as female, and were 40 years or older, and worked in maternal and child health for 15 or more years in primarily an academic setting. Approximately 76% were academics and 18% were health care providers. Patients tended to be 30 to 39 years old and with graduate level of education. Most participating patients were not currently pregnant and had 2 or 3 previous live births.

The analysis of individual panel ratings showed that each panel reached consensus on the seriousness of all health outcomes as determined by the RAM (results not shown). Comparing across panels, median scores were within 10 points for 8 of 11 outcomes (Figure 1, Table S1). There was a 15-point difference between panels for SGA birth, LGA birth, and metabolic syndrome. These differences were generally small, relative to the wide range in scores between outcomes within each panel (65- to 75-point ranges). Further, each panel ranked the outcomes similarly.

Figure 1.

Distribution of seriousness ratings of maternal and child health outcomes by panel type: professionals panel (red, n = 57), patients panel (green, n = 55), and mixed panel (orange, n = 27 professionals, n = 27 patients). The box represents the 25^th to 75^th percentiles and the line in the box is the median. The whiskers extend to 1.5 times the interquartile range of the lower and upper quartiles. The circles represent the outliers. SGA, small for gestational age birth; LGA, large for gestational age birth

Moreover, eight of the 11 outcomes were assigned the same seriousness decision by all 3 panels (Table S2). Infant death, stillbirth, preterm birth and preeclampsia were ranked as “serious” while childhood obesity, maternal obesity, and SGA birth were ranked as “moderately serious.” The patient panel and mixed panel scored unplanned cesarean delivery, LGA birth, and metabolic syndrome as less serious than the professionals panel; however, the differences in median scores between panels for these outcomes were relatively small, and the difference in seriousness ranking reflected primarily their proximity to the category cut-off.

Because of the similarities across the panel ratings, we decided to combine the data from all three panels to determine the overall seriousness ratings. The final median seriousness ratings combined across all panels were highest for infant death (100), stillbirth (95), preterm birth (80), and preeclampsia (80) (Figure 2, Table S1). Obesity in children, metabolic syndrome in women, obesity in women, and gestational diabetes had median seriousness ratings ranging from 55 to 65. The lowest final seriousness ratings were for SGA birth, LGA birth, and unplanned cesarean delivery (30 to 40). The interquartile range of scores (i.e., the absolute difference between the 25^th and 75^th percentiles) were 3.5 to 15 for the 4 most serious outcomes and were 17.5 to 30 for the remaining outcomes. The range between the highest and lowest score was wide for all outcomes except infant death and stillbirth.

Figure 2.

Distribution of final seriousness ratings of maternal and child health outcomes across the three panels, n = 166 professionals and patients. The box represents the 25^th to 75^th percentiles and the line in the box is the median. The whiskers extend to 1.5 times the interquartile range of the lower and upper quartiles. The circles represent the outliers. SGA, small for gestational age birth; LGA, large for gestational age birth.

Median scores differed moderately according to participant characteristics (typically within 10 points), particularly for less serious outcomes such an unplanned cesarean delivery and LGA birth (Table 2 and Table 3). The largest difference according to professionals’ characteristics was for unplanned cesarean delivery according to race-ethnicity (30-point difference). Among patients, the largest difference was for maternal obesity according to geographic region (25-point difference).

Table 2.

Professionals’ median ratings of the seriousness of maternal and child health outcomes according to select characteristics, n = 84

	Race-ethnicity			Region of residence				Panel type
	Non-Hispanic White	Non-Hispanic Black	Other	West	Midwest	Northeast	South	Professionals only	Mixed
Health Outcome	n=50	n=14	n=20	n=17	n=14	n=20	n=33	n=57	n=27
Infant death	100	99	100	100	100	100	100	100	100
Stillbirth	90	90	90	92	90	95	90	90	95
Preterm birth	75	83	80	80	80	70	80	80	80
Preeclampsia	80	80	80	75	73	80	80	80	78
Obesity in children	60	75	73	70	55	55	65	65	63
Metabolic syndrome in women	60	70	65	65	68	60	60	70	55
Obesity in women	58	65	65	60	68	50	60	60	55
Gestational diabetes	55	60	63	55	50	55	60	55	60
Small for gestational age birth	50	50	53	50	50	50	50	50	44
Unplanned cesarean delivery	30	39	50	40	23	30	40	35	30
Large for gestational age birth	30	35	49	33	33	33	40	35	35

Table 3.

Patients’ median ratings of the seriousness of maternal and child health outcomes according to select characteristics, n=82

	Race-ethnicity			Region of residence				Education		Panel type
	Non-Hispanic White	Non-Hispanic Black	Other	West	Mid-west	North-east	South	College or less	Graduate degree	Patients only	Mixed
Health outcome	n=41	n=10	n=31	n=12	n=24	n=13	n=33	n=30	n=47	n=55	n=27
Infant death	100	100	100	100	100	100	100	100	100	100	100
Stillbirth	99	100	100	100	99	99	100	100	100	100	100
Preterm birth	80	86	80	80	80	70	85	85	80	80	80
Preeclampsia	75	83	78	80	80	70	78	80	75	75	80
Obesity in children	60	68	70	50	68	60	65	60	65	65	60
Metabolic syndrome in women	60	58	60	55	55	60	60	50	60	60	55
Obesity in women	50	65	70	45	53	70	65	59	60	65	55
Gestational diabetes	50	60	50	45	50	50	50	50	50	50	60
Small for gestational age birth	36	45	35	30	40	30	40	40	35	35	40
Unplanned cesarean delivery	30	51	30	30	23	30	35	40	30	30	30
Large for gestational age birth	20	25	25	20	25	20	20	20	25	20	30

Comment

Principal findings

We found that this diverse group of 166 maternal and child health professionals and pregnant or postpartum women rated some adverse outcomes related to pregnancy weight gain as being more serious than others. In each of the 3 panels, participants came to a consensus on their seriousness decisions, although achieving consensus was not a goal of each panel. They scored infant mortality and stillbirth as the most serious, closely followed by preterm birth and preeclampsia. Child obesity, metabolic syndrome, obesity in women, and gestational diabetes were scored moderately serious. The least serious health outcomes were SGA birth, LGA birth, and cesarean delivery.

Strengths of the study

The size of our sample was in line with or larger than previous Delphi panels of perinatal health outcomes ^{22, 25, 26}. Like other Delphi panels, our sample was not representative. Three-quarters of professionals in our study were researchers, whose views on health outcomes may differ than health care providers. However, our successful recruitment of experts across a range of racial/ethnic groups and U.S. geographic regions provides some assurance that we captured diverse opinions. Further, our inclusion of pregnant and postpartum women as half the panel membership, including a panel consisting exclusively of patients, ensured patients’ views were taken into account in a meaningfully way and not overshadowed by the voices of health professionals. The survey included 11 common health outcomes consistently associated with pregnancy weight gain, which will allow investigators to account for the severity of a range of maternal and child health outcomes in subsequent research. Nevertheless, concerns over participant burden prevented us from including other outcomes that currently have an inconsistent relationship with pregnancy weight gain. Future work may consider the inclusion of breastfeeding, maternal mental health conditions, child cognition, subsequent fertility, and pregnancy loss or adverse pregnancy outcomes in the following pregnancy.

Limitations of the data

We recognize that there is no single “truth” when attempting to quantify the relative seriousness of different health outcomes, and that some individuals will view the same health outcome as being more serious than other individuals. Nevertheless, we argue that attempting to quantify this seriousness, even if imperfect, is better than the current approaches that do not reflect women’s and care providers’ values, such as studying outcomes in isolation or using a composite outcome that weights all components equally. We reported not only median scores, but also the range of scores for a given outcome, which will enable researchers to explore the impact of different weights through sensitivity analyses that use the highest and lowest values elicited from each panel. These findings will also inform policy-makers on the magnitude of variation in optimal ranges obtained from diverse opinions and account for the complex trade-off between low and high weight gain on maternal and child health.

Approximately 22% of participants did not complete all 3 rounds, but this is participation rate is substantially lower than many multi-round online Delphi panels ²⁷. Our use of a web-based platform may have reduced the quality of participant interaction due to information overload and difficulties following discussion threads ^{28, 29}. However, online discussion boards where participants are semi-anonymous often lead to more open and frank discussion. In-person panels where patients and clinicians interact with each other often suffer from unequal participation and perceptions that the input of patients is less important than that of clinicians ³⁰. Finally, we recruited patients using social media, which limited participation to women who were comfortable completing online surveys and had at least one device with internet access. However, given that most women of childbearing age are comfortable with internet-based platforms, selection bias is likely less of a concern than in studies of older individuals.

Interpretation

This study builds on previous Delphi committee-based efforts to advance consensus on outcomes related to nutrition in pregnancy. Rogozinska and colleagues²² conducted a Delphi survey of 26 researchers in 11 countries to develop a maternal, fetal, and neonatal composite outcome relevant to the evaluation of pregnancy diet and lifestyle studies. We extended this work by establishing the relative seriousness of the outcomes in their composite, as well as additional longer-term maternal and child outcomes. More broadly, our work serves as an example of how the ongoing The CoRe Outcomes in WomeN’s health (CROWN) initiative ³¹ could be extended to not only seek consensus on which outcomes should be collected and reported in a given research area, but also elicit stakeholder perspectives on their relative seriousness.

Two previous studies have rated the seriousness of maternal and child health outcomes related to pregnancy weight gain, but neither included patients’ input. The 2009 IOM committee ¹ calculated the loss in Quality-Adjusted Life Years (QALYs) associated with 3 outcomes: infant mortality, post-partum weight retention, and childhood obesity. Each case of infant mortality was associated with a loss of 80 QALYs, post-partum weight retention with losses of 3.1 and 4.2 QALYs in overweight and obese women, respectively, and childhood obesity with a loss of 9.8 QALYs. Our findings are broadly similar to these, in that they view childhood obesity as more serious than maternal obesity, and infant mortality as considerably more serious than either.

Oken and colleagues³² elicited weights for a composite outcome of 5 adverse events (preterm birth, SGA birth, LGA birth, child obesity, and post-partum weight retention) through a convenience sample of 12 Harvard researchers, as an extension to work relating pregnancy weight gain to a composite outcome. They identified LGA birth as the least serious outcome. They presented the weights for the remaining outcomes as the number of times more serious then LGA birth: child obesity x 2, post-partum weight retention x 3, SGA birth x 5, and preterm birth x 6. In our study, preterm birth was also identified as the most serious of these outcomes, but otherwise our findings differ in that SGA birth was deemed less serious than either child or maternal obesity, and child obesity was deemed more serious than maternal obesity. Our results may differ because our participant group included patients and was larger and more diverse.

A small number of other perinatal scoring tools that assign scores or seriousness ratings to different outcomes have also been developed, such as the MAIN index ³³ for mild-moderate perinatal morbidity and the Adverse Outcome Index ³ for evaluating the safety of labour and delivery. However, we could not use these tools because they include few of the outcomes deemed relevant for pregnancy weight gain.

We ³⁴, and others ^{35, 36}, have previously described analytic strategies for weighing adverse events according to their seriousness. Using these methods, the seriousness ratings we present could be applied to future epidemiologic studies seeking to establish the range of pregnancy weight gain associated with lowest risks of adverse maternal and child health outcomes. We intentionally included the median, range, and interquartile range of seriousness scores for all outcomes so that researchers can account for variability of opinions when using our seriousness scores in future studies.

Conclusions

We view the seriousness scores elicited by this study as a first step towards creating pregnancy weight gain guidelines that account for the relative seriousness of different health outcomes in an explicit, quantitative manner. These seriousness ratings can be used to conduct epidemiologic studies that seek to establish the range of pregnancy weight gain associated with the lowest risk of a broad range of maternal and child health outcomes that vary in their severity. Such an approach is essential for informing policy addressing the complex trade-off between low and high pregnancy weight gain on maternal and child health. We hope that our work stimulates other researchers to collect quantitative data on stakeholders’ views of the seriousness of short- and long-term maternal and child health outcomes.

Social media quote.

Professionals and women rate some adverse outcomes as being more serious than others. These ratings can be used to establish the range of pregnancy weight gain related to the lowest risk of a broad range of maternal and child health outcomes.

Synopsis.

Study Question

How seriously do patients and health professionals rate 11 maternal and child health outcomes that have been consistently associated with pregnancy weight gain?

What’s already known

Current pregnancy weight gain guidelines were developed based on implicit assumptions of a small group of experts about the relative seriousness of adverse health outcomes. Therefore, they will not necessarily reflect the values of women.

What this study adds

Professionals and women rate some adverse outcomes as being more serious than others. These ratings can be used to establish the range of pregnancy weight gain associated with the lowest risk of a broad range of maternal and child health outcomes.