Abstract
Introduction
Identifying healthcare utilization and costs associated with active and passive smoking during pregnancy could help improve health management strategies.
Aims and Methods
Data are from the Newborn Epigenetics STudy (NEST), a birth cohort enrolled from 2005 to 2011 in Durham and adjacent counties in North Carolina, United States. Participants included those for whom prenatal serum samples were assayed and for whom administrative data were obtainable (N = 1045). Zero-inflated poisson regression models were used to assess associations between cotinine, adjusted for covariates (eg, race and ethnicity, age at delivery, cohabitation status, and education), and health care utilization outcomes. Generalized linear regression models were used to estimate average total charges. Simulation models were conducted to determine the economic benefits of reducing secondhand smoke and smoking during pregnancy.
Results
Increasing levels of cotinine were positively associated with parent’s number of emergency department (ED) visits (coefficient[b] = 0.0012, standard error [SE] = 0.0002; p < .001), the number of ICU hours (b = 0.0079, SE = 0.0025; p = .002), time spent in the ICU (b = 0.0238, SE = 0.0020, p < .001), and the number of OP visits (b = 0.0003, SE = 0.0001; p < .001). For infants, higher cotinine levels were associated with higher number of ED (b = 0.0012, SE = 0.0004; p = .005), ICU (b = 0.0050, SE = 0.001; p < .001), and OP (b = 0.0006, SE = 0.0002; p < .001) visits and longer time spent in the ED (b = 0.0025, SE = 0.0003; p < .001), ICU (b = 0.0005, SE = 0.0001; p < .001), and IP (b = 0.0020, SE = 0.0002; p < .001). Simulation results showed that a 5% reduction in smoking would correspond to a potential median cost savings of $150 533 from ED visits of parents and infants.
Conclusions
Our findings highlight the importance of smoke exposure cessation during pregnancy to reduce health care utilization and costs for both parents and infants.
Implications
This study reinforces the importance of reducing smoking and secondhand smoke exposure during pregnancy. Focusing on expanding cessation services to this group could help reduce morbidities observed within this population. Furthermore, there is the potential for healthcare cost savings to healthcare systems, especially for those with high delivery numbers. These cost savings are represented by potential reductions in ED, OP, and ICU hours and visits.
Introduction
Although the prevalence of smoking and secondhand smoke (SHS) exposure has declined, the rate has not declined equally for all groups, especially for vulnerable populations who continue to experience harm related to smoking and SHS.1,2 As such, pregnant persons and children are particularly vulnerable to smoke exposure. Of the 600 000 deaths attributed to SHS globally in 2004, 47% of deaths occurred in women, 28% in children, and 26% in men.3 Maternal smoking and SHS exposure during pregnancy increases the risk for placenta previa and abruption, bleeding, ectopic pregnancy, and hypertension.4–9 Smoke exposure during pregnancy is associated with preterm birth, low birth weight, perinatal/infant mortality, and childhood cancer.10–13 Furthermore, active smoking and SHS exposure during pregnancy are associated with alterations in epigenetic signatures associated with the pathogenesis of adult cancers.14,15
Despite the known consequences of smoking during pregnancy, 23% of pregnant persons report smoking in the 3 months prior to their pregnancy, 1%–21% smoke during their pregnancy in the United States (depending on the state),16 and 17% smoke during the postpartum months.17,18 These estimates do not account for smoke exposure from domestic partners, relatives, and other environmental sources. Furthermore, estimates of smoke exposure often rely on self-reporting, which could lead to underestimation.19,20 In both our studies and others, it has been shown that about 25% of pregnant persons who report nonsmoking have cotinine levels suggesting otherwise and that equal numbers of nonsmoking pregnant persons unreliably report being exposed to secondhand smoke.19,21,22 These data highlight the importance of using biomarkers to estimate smoke exposure, especially among pregnant persons where there are social pressures to underreport smoke exposure.20,23
Associated medical costs among those who smoke during pregnancy or among children exposed to SHS are non-trivial. Smoking during pregnancy has been estimated to add $724 to the average neonatal cost of a birth,24 which contributes to about 9% of the total direct medical costs in the first year of life,25 and results in greater utilization of health care services in early childhood.26 Nationally, SHS exposure is estimated to contribute annually to more than $5 billion in direct medical costs and $4.7 billion in lost productivity costs.27 Despite increasing knowledge about health costs associated with prenatal smoke exposure, few studies have estimated health care utilization and costs among pregnant persons and their newborn infants who are exposed to varying levels of tobacco smoke. Furthermore, few studies have used cotinine biomarkers to more accurately assess a range of tobacco smoke exposure that includes both SHS exposure and smoking during pregnancy. This study aimed to clarify this issue by estimating the associated healthcare utilization, as well as costs, associated with both prenatal SHS exposure and smoking during pregnancy, for pregnant persons and their infants during the first year of life. We examined the association between pregnant persons’ serum-cotinine and health care utilization, and health care costs of parents and their infants from the Newborn Epigenetics STudy (NEST), a birth cohort recruited from a Mid-Atlantic health care system.
Materials and Methods
Data and Study Sample
The study sample includes pregnant persons enrolled in NEST at Duke University Hospitals in Durham and its adjacent counties in North Carolina, United States. Detailed NEST recruitment strategies and information are described in other studies.28,29 Briefly, NEST is a prospective pre-birth cohort study designed to examine the associations between prenatal exposures, epigenetics, and childhood neurodevelopmental and health outcomes. Participants (N = 2595) were enrolled between 2005 and 2011. To be eligible, participants had to be at least 18 years of age, English or Spanish speaking, planning to use Duke or Durham Regional Hospital for their delivery and willing to provide a prenatal blood sample. At enrollment, participants completed baseline surveys, and blood samples were collected. Participants also provided consent for the research team to access their medical records and the medical records of their future infants for up to one year after the birth. For this study, we excluded participants whose serum samples were not assayed for prenatal cotinine (n = 1546), who did not deliver live babies (n = 87), and who were missing data on the date of delivery (n = 3). The retained sample was 1045 parents and their infants. The sample for analysis only included those participants with singleton births, since all participants with twin births were not assayed for prenatal cotinine.
Measures
Medical Records Data
Health encounter data were extracted from Duke University Health System data warehouse using the Duke Enterprise Data Unified Content Explorer. Duke Enterprise Data Unified Content Explorer is a research data extraction system that allows flexible, self-service applications for data extraction by clinical researchers and QI professionals across the Duke University Health System.30,31 All medical encounters during 1 year prior to delivery and 1 year following delivery were extracted for each participant.
Encounter data were classified into emergency department (ED) visits, inpatient stays (IP), or outpatient visits (OP). We defined the utilization period as 2 years for the pregnant person with one year prior to delivery and the first year after delivery. For infants, the utilization period was defined as the first year of life. For each subject, we calculated these utilization measures during the utilization period: the number of ED visits, the number of IP visits, and the number of OP visits. Subjects who did not have any of the corresponding visits had a count of 0 visits. The ED length of stay (ED_LOS) was calculated as the total number of hours spent in the ED. The length of stay was defined as the total number of days spent in the hospital and calculated as the sum of days from the time of admission to the time of discharge, and the total number of hours spent in the ICU (ICU_LOS). The total number of hours or days were set to 0 for subjects who did not have any of the corresponding visits. We also created indicator variables to identify if the participant had any ED, IP, OP visits, or was ever transferred to ICU.
To estimate the costs, total charges incurred during the utilization period were summed separately for pregnant persons and infants. Using the U.S. Consumer Price Index for all urban consumers, we adjusted to 2012 values to account for inflation.32 Then we converted the charges into costs using the cost-to-charge ratio for the state of North Carolina for the year 2012, which was 0.389, from the U.S. Department of Labor data on Medical Fee Schedule.33
Serum Cotinine
Blood samples were collected at enrollment during the 16th week of gestation (SD = 10.2). The samples were analyzed for serum cotinine using a liquid chromatography-mass spectrometry method as described previously.34–36 This method offered high-sensitivity and high-accuracy measurements of cotinine concentration for achieving a detection limit of 0.05 ng/mL. Cotinine was included both as a continuous variable and as a categorical variable in the analyses. Using recommendations from prior research, cotinine was categorized into three levels: (1) < 1 ng/mL (persons who never/formerly smoked with no SHS exposure), (2) 1–3 ng/mL (persons who never/formerly smoked exposed to SHS), and (3) > 3 ng/mL (persons who smoke).37
Covariates
The following covariates were included, which were obtained from a survey administered at enrollment: Race and ethnicity (Hispanic, non-Hispanic Black, non-Hispanic White, and non-Hispanic other), age at delivery, cohabitation status (married or living with a partner vs. single), education (at least some college vs. less than college), and parity (0, 1, or 2+).
Statistical Analysis
Estimating Utilization and Charges
Chi-square or ANOVA tests were used to compare individual-level characteristics of the pregnant person between different levels of cotinine. We estimated two sets of models, one for the pregnant person, and one for the infant. We used Poisson regression models to estimate the counts of the number of visits (ED, IP) and the number of hours or minutes spent in the hospital (LOS) or the ED (ED_LOS) as a function of prenatal cotinine serum levels adjusted for covariates. To account for outcomes with excess zeros, we used zero-inflated Poisson (ZIP) regression models. ZIP regression estimates the probability of having a zero count and adjusts for that probability when estimating the count. We used a logit inflation model to predict having no visits and being a subject with zero counts.
Missing charges for subjects with no utilization (8 pregnant persons, 49 infants) were assumed to be $0.00. Multiple imputations with 100 imputations were used to impute missing charges for those who had utilization during the study (256 pregnant persons, 334 infants) using regression models that predicted the charges by the subject characteristics and health utilization (numbers of ED, IP, OP ICU visits, and their length of stays). Generalized linear regression models with gamma distribution and log links were used to estimate the average total charges. An amount of $0.10 was added to the charges in order not to drop subjects with zero charges from the models with log links. Two-part models were also used to estimate the average charges for infants; the first part of the model estimated the probability of having $0.00 charges, and the second model estimated the average charges conditional on having >$0.00 charges. The marginal effects of cotinine (change per unit change of cotinine) on the charges are estimated as the slope of the non-linear cotinine-charges curve at a specific cotinine level, and was estimated at the average cotinine level of the sample. The predicted charges, marginal effects, and plots were estimated using the STATA margins and marginsplot commands.
All models included the pregnant person’s cotinine level as a continuous variable or as a categorical variable, and adjusted for covariates. Statistical significance was determined at p < .05.
Estimating Potential Savings of Reducing Smoking Exposure
Using the data and predictions from the regression models, we performed simulation studies to estimate the potential healthcare cost savings associated with smoking cessation interventions. In the first simulation, we assumed persons with active smoking who quit smoking, but were potentially still exposed to SHS. We randomly selected J smoking participants, where J represents the number of smoking participants that quit smoking and J = 25 to 225 with 25 increments. We calculated the potential savings for each subject as the difference in predicted total costs as a person who smokes for that subject and the predicted cost if that subject quit smoking. Then we calculated the sum of these savings for all subjects in that sample as the potential savings for the potential number of persons quitting smoking. We repeated this for 1000 iterations, and estimated the minimum, median, and maximum potential savings of those iterations. Similarly, in the second simulation, we estimated the total potential savings for reducing the number of those exposed to SHS. The potential savings were estimated by calculating the difference between the predicted cost as exposed to SHS and the predicted costs if not exposed to SHS.
Results
Descriptive Statistics
Table 1 provides descriptive statistics by the plasma cotinine level categories representing no smoke exposure, SHS exposure, and active smoking. Pregnant persons with cotinine levels indicating SHS exposure or active smoking were more likely to be younger at delivery. Persons who actively smoked were more likely to have previous deliveries. Pregnant persons who were Black and those who were less educated were more likely to be actively smoking or be exposed to SHS (p < .001). Those who had less than adequate gestational weight gain were more likely to be persons who smoked (p < .001).
Table 1.
Descriptive Table by Categories of Prenatal Blood Cotinine Levels
| Variable | Categories | ALL (n = 1045) | Non-exposed < 1 ng/mL (n = 643) |
SHS 1-3 ng/mL (n = 170) |
Persons who actively smoke > 3 ng/mL (n = 232) |
p-value |
|---|---|---|---|---|---|---|
| Parity | 0 | 382 (36.56%) | 261 (40.59%) | 58 (34.12%) | 63 (27.16%) | <.001 |
| 1 | 348 (33.30%) | 210 (32.66%) | 55 (32.35%) | 83 (35.78%) | ||
| ≥2 | 283 (27.08%) | 155 (24.11%) | 45 (26.47%) | 83 (35.78%) | ||
| Missing | 32 (3.06%) | 17 (2.64%) | 12 (7.06%) | 3 (1.29%) | ||
| Marital Status | Married/ living with partner | 345 (33.01%) | 171 (26.59%) | 63 (37.06%) | 111 (47.84%) | <.001 |
| Never married/ divorced/separated | 613 (58.66%) | 420 (65.32%) | 92 (54.12%) | 101 (43.53%) | ||
| Missing | 87 (8.33%) | 52 (8.09%) | 15 (8.82%) | 20 (8.62%) | ||
| Race/Ethnicity | Black | 529 (50.62%) | 278 (43.23%) | 101 (59.41%) | 150 (64.66%) | <.001 |
| Hispanic | 80 (7.66%) | 57 (8.86%) | 15 (8.82%) | 8 (3.45%) | ||
| Other | 51 (4.88%) | 37 (5.75%) | 6 (3.53%) | 8 (3.45%) | ||
| White | 385 (36.84%) | 271 (42.15%) | 48 (28.24%) | 66 (28.45%) | ||
| Education | High school degree or less | 359 (34.35%) | 158 (24.57%) | 66 (38.82%) | 135 (58.19%) | <.001 |
| Some college | 602 (57.61%) | 435 (67.65%) | 89 (52.35%) | 78 (33.62%) | ||
| Missing | 84 (8.04%) | 50 (7.78%) | 15 (8.82%) | 19 (8.19%) |
SHS = secondhand smoke.
Table 2 presents the health utilization of pregnant persons and infants. Pregnant persons exposed to SHS or who actively smoke were more likely to have had an ED visit and to be admitted to an ICU than pregnant persons not exposed to smoke. Pregnant persons exposed to SHS or who actively smoke had a higher number of ED visits, and spent longer hours at the ED and in the ICU. Infants exposed to pregnant persons who smoked or who were exposed to SHS during pregnancy were also more likely to have an ED visit, have a higher number of ED visits, and spent more hours in the ED department; they also were twice as likely to be admitted to the ICU.
Table 2.
Observed Health Utilization Outcomes of Pregnant Persons Within 1 Year Before and 1 Year After Delivery by Prenatal Blood Cotinine Levels and Children in First Year of Age by Pregnant Person’s Blood Cotinine Levels
| Variable | ALL (n = 1045) | Persons who never/formerly smoke <1 ng/mL (n = 643) |
SHS 1–3 ng/mL (n = 170) |
Persons who actively smoke >3 ng/mL (n = 232) |
||
|---|---|---|---|---|---|---|
| #Pregnant persons with no health records (%) |
8 (0.77%) | 5 (0.78%) | 2 (1.18%) | .62 | 1 (0.43%) | .58 |
| mean(std) | p-value | p-value | ||||
| #ED visits | 2.51 (3.86) | 1.63 (2.72) | 2.77 (3.35) | <.001 | 4.77 (5.57) | <.001 |
| Hours in ED | 5.00 (13.18) | 3.04 (10.35) | 6.08 (15.35) | .02 | 9.64 (16.78) | <.001 |
| #IP stays | 0.65 (0.55) | 0.67 (0.51) | 0.56 (0.59) | .02 | 0.67 (0.59) | .95 |
| Days IP | 1.84 (2.06) | 1.86 (2.04) | 1.60 (1.77) | .10 | 1.96 (2.30) | .57 |
| Hours in ICU | 0.10 (1.95) | 0.00 (0.00) | 0.00 (0.00) | NA | 0.45 (4.12) | .10 |
| #ICU stays | 0.01 (0.08) | 0.00 (0.04) | 0.00 (0.00) | .32 | 0.02 (0.15) | .04 |
| #OP visits | 25.39 (15.24) | 26.26 (15.14) | 24.46 (16.00) | .19 | 23.67 (14.84) | .024 |
| N(%) | ||||||
| #Had ED visit | 638 (61.05%) | 327 (50.86%) | 117 (68.82%) | <.001 | 194 (83.62%) | <.001 |
| #Had IP visit | 644 (61.63%) | 416 (64.70%) | 86 (50.59%) | <.001 | 142 (61.21%) | .31 |
| #Admitted to ICU | 6 (0.57%) | 1 (0.16%) | 0 (0.00%) | .61 | 5 (2.16%) | .002 |
| #Had OP visit | 1037 (99.23%) | 638 (99.22%) | 168 (98.82%) | .62 | 231 (99.57%) | .58 |
| #Children with no health records (%) | 49 (4.69%) | 28 (4.35%) | 14 (8.24%) |
0.0421
|
7 (3.02%) | 0.3729 |
| mean(std) | p-value | p-value | ||||
| #ED visits | 0.65 (1.30) | 0.46 (1.02) | 0.72 (1.28) | .02 | 1.14 (1.77) | <.001 |
| Hours in ED | 1.48 (4.39) | 1.06 (3.49) | 1.72 (3.73) | .04 | 2.45 (6.40) | <.001 |
| #IP stays | 1.03 (0.18) | 1.02 (0.17) | 1.03 (0.21) | .43 | 1.04 (0.20) | .13 |
| Days IP | 3.91 (9.92) | 3.49 (9.48) | 5.41 (14.14) | .11 | 4.02 (7.12) | .39 |
| Hours in ICU | 10.44 (109.84) | 6.73 (91.12) | 24.33 (192.20) | .27 | 10.96 (71.42) | .48 |
| #ICU stays | 0.05 (0.25) | 0.03 (0.19) | 0.07 (0.26) | .10 | 0.09 (0.37) | .03 |
| #OP visits | 8.18 (7.34) | 8.05 (7.34) | 7.83 (6.92) | .74 | 8.76 (7.62) | .22 |
| N(%) | ||||||
| #Had ED visit | 327 (31.29%) | 154 (23.95%) | 58 (34.12%) | .002 | 115 (49.57%) | <.001 |
| #Had IP visit | 992 (94.93%) | 612 (95.18%) | 155 (91.18%) | .81 | 225 (96.98%) | .29 |
| #Admitted to ICU | 46 (4.40%) | 20 (3.11%) | 11 (6.47%) | .03 | 15 (6.47%) | .03 |
| #Had OP visit | 767 (73.40%) | 467 (72.63%) | 118 (69.41%) | .94 | 182 (78.45%) | .13 |
p-values: t-test for differences in mean of continuous outcomes, and overall chi-square tests for categorical binary outcomes compared to persons who never/formerly smoke. Bold and italicized values indicate statistical significance. SHS = secondhand smoke; ED = emergency department; IP = inpatient; ICU = intensive care unit; OP = outpatient.
Multivariable Regression Results
The ZIP regression or Poisson regression models, adjusting for the covariates and including cotinine level as a continuous quantitative predictor and health care utilization as the outcome are shown in Table 3. The zero component of the ZIP models reflects the likelihood of having a zero number of outcome visits (eg, no ED visits), while the count component estimates the increase in the number of visits by the unit increase of pregnant person’s prenatal cotinine.
Table 3.
Pregnant Persons and Infants Utilization Outcome Regression Models
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| # ED | Hours in ED | # IP hospitalizations | LOS (days) |
# ICU visits! | # ICU hours! | # OP! | |
| b(se) | b(se) | b(se) | b(se) | b(se) | b(se) | b(se) | |
| Pregnant person | |||||||
| Count component | |||||||
| Prenatal cotinine | 0.0012*** | 0.0003 | −0.0012 | 0.0006 | 0.0079** | 0.0238*** | 0.0003** |
| (0.0002) | (0.0002) | (0.0008) | (0.0005) | (0.0025) | (0.0020) | (0.0001) | |
| Zero component | |||||||
| prenatal cotinine | −0.0143** | −0.0079*** | NA | 0.0030* | NA | NA | NA |
| (0.0051) | (0.0020) | (0.0012) | |||||
| Infants | |||||||
| Count component | |||||||
| Prenatal cotinine | 0.0012** | 0.0025*** | 0.0002 | 0.0020*** | 0.0050*** | 0.0005*** | 0.0006*** |
| (0.0004) | (0.0003) | (0.0005) | (0.0002) | (0.0010) | (0.0001) | (0.0002) | |
| Zero component | |||||||
| Prenatal cotinine | −0.0053* | −0.0026* | −0.0002 | 0.0043 | −0.0222 | −0.0074*** | −0.0004 |
| (0.0026) | (0.0012) | (143.7083) | (562.0955) | (0.0298) | (0.0018) | (0.0015) | |
Results from Zero-inflated or Poisson models adjusted for: age at delivery, parity, education, race, and marital status. Standard error in parentheses; *p < .05, **p < .01, ***p < .001;! Poisson models. b = coefficient estimate; se = standard error; ED = emergency department; IP = inpatient; ICU = intensive care unit; OP = outpatient; b = coefficient estimate; se = standard error.
For pregnant persons, the results of the zero component show lower cotinine levels are associated with a higher likelihood of not visiting the ED (b = −0.0143, p = .005) and no time spent in ED (b = −0.0079, p < .001). The count parts of the models show that increasing levels of cotinine were all positively associated with the number of ED visits (b = 0.0012, p < .001), number of ICU visits (b = 0.0079, p = .002), amount of time spent in the ICU (b = 0.0238, p < .001), and number of OP visits (b = 0.0003, p = .002). The adjusted predicted utilization with 95% confidence intervals estimated from those models are visually presented in Figure S1. The figure shows an increase in the number and hours spent in ED and ICU as cotinine increases. For example, the average number of hours in ED for subjects with 50 ng/L serum cotinine compared to those at 100 ng/L were 5.55 (SE = 0.21) and 6.31(SE = 0.33), respectively. The full ZIP models for pregnant persons are presented in Table S1. Figure S1 displays the Pregnant persons’ predicted utilization outcomes with 95% confidence intervals at the different levels of the pregnant person’s prenatal plasma cotinine levels.
Similarly, for infants, lower cotinine levels were associated with having no ED visits (b = -0.0053, P = .04), no time spent in ED (b = -0.0026, P = .03), and no time spent in the ICU (b = -0.0074, P<.001). Higher cotinine levels were associated with higher number of ED (b = 0.0012, P = .005), ICU (b = 0.0050, P<.001), and OP (b = 0.0006, P<.001) visits and longer time spent in the ED (b = 0.0025, P<.001), ICU (b = 0.0005, P<.001), and IP (b = 0.0020, P<.001). The full ZIP models for infants are presented inTable S2. Figure S2 displays the infants’ predicted utilization outcomes with 95% confidence intervals at the different levels of the pregnant person’s plasma cotinine levels.
Table 4 and Figure S3 present the results of the charges’ analysis. For pregnant persons, an increase in cotinine levels is associated with a significant increase in the total charges: $45.58 (standard error [se] = $15.80, P = .004) per unit increase in plasma cotinine levels. Infants’ charges were not statistically associated with cotinine.
Table 4.
GLM Model Coefficients to Estimate Total Charges
| Pregnant Person | Infant | |
|---|---|---|
| b! (se) | b! (se) | |
| Prenatal cotinine | 0.0033** | 0.0000 |
| (0.0012) | (0.0074) | |
| Squared prenatal cotinine | −0.0000 | 0.0000 |
| (0.0000) | (0.0000) | |
| Age at delivery | 0.0119 | −0.0258 |
| (0.0074) | (0.0336) | |
| Parity (0 parity is the reference) | ref | ref |
| 1 parity | 0.1179 | −0.4031 |
| (0.0965) | (0.3902) | |
| 2 + parity | 0.1502 | −0.3126 |
| (0.1088) | (0.4431) | |
| Missing parity | −2.6652*** | |
| (0.4987) | ||
| Black | 0.1320 | 0.1319 |
| (0.1019) | (0.3905) | |
| Hispanic | −0.0672 | 0.8488 |
| (0.1643) | (0.6221) | |
| Other races | −0.2003 | −1.8796** |
| (0.1890) | (0.7088) | |
| White | ref | ref |
| High school degree or less | ref | ref |
| Some college | 0.1396 | 0.6065 |
| (0.0956) | (0.3638) | |
| Missing Education | 0.1082 | −0.1140 |
| (0.3247) | (1.2574) | |
| Living with Partner | ref | ref |
| Living alone | −0.1299 | −0.2678 |
| (0.0917) | (0.3680) | |
| Missing Marital Status | −0.0402 | −0.3593 |
| (0.3154) | (1.2050) | |
| Constant | 9.0361*** | 10.2148*** |
| (0.2263) | (1.0012) | |
| N | 783 | 664 |
GLM = generalized linear model;! = coefficients of GLM with Gamma distribution and log link function. Living alone = single, widowed, or divorced; Living with partner = married or living with a partner. *p < .05, **p < .01, ***p < .001; Constant: model intercept. B = coefficient estimate; se = standard error.
Estimating Potential Savings Results
The results of the simulation study showed that a 5% reduction in the percentage of those who smoke (from the current 22.5% to 17.5%) could result in median savings of $150 533. Figure S4 presents the minimum, median, and maximum potential cost savings by the number of those who quit smoking. To put this in perspective, a successful prenatal smoking cessation program that reduces the rate by 50% (113 in 1000 pregnancies) has the potential to result in a median savings of $317 021 per 1000 pregnancies.
The potential savings of reducing exposure to SHS were much smaller. Figure S5 presents the minimum, median, and maximum potential cost savings by the number of SHS who would be unexposed. The simulation study showed that a 5% reduction in the percentage of subjects who are currently exposed to SHS (from 16% to 11%) could result in median savings of $3792. The distribution of log charges for pregnant persons and infants are presented in Figure S6.
Discussion
By leveraging biological samples from a birth cohort and hospital administrative data, this study documents the association between prenatal smoke exposure and health care utilization and healthcare costs among both infants and pregnant persons. Pregnant persons with elevated prenatal cotinine concentration had more ED, ICU, and OP visits and spent a longer time in the ICU. Additionally, elevations in cotinine concentration during pregnancy were associated with higher healthcare costs for pregnant persons during the perinatal phase. The results were similar for infants with respect to healthcare utilization; however, this did not translate into significantly greater healthcare costs.
The results demonstrated correlations between various confounding risk factors and the outcomes under investigation. Notably, age, race, parity, marital status, and educational attainment exhibited associations with some outcomes at varying significance levels with different and sometimes contradicting directions. However, it is imperative to recognize that these factors were included in the models as potential confounders to the effects of plasma cotinine on the outcomes. Consequently, the observed correlations may be biased, as the models did not adjust for confounders related to the association between these risk factors and the outcomes (Table 2 fallacy).38 Therefore, interpretations of these correlations may be misleading,39 highlighting the necessity for further research to validate these findings.
The findings in this current study support findings from other studies showing greater use of health care services and costs among children exposed to pregnant persons smoking during pregnancy.40–42 Whereas other studies have relied on self-report, this study is the first to take advantage of pregnant persons’ blood samples to assay cotinine biomarkers to assess low levels of smoke exposure and active smoking. Thus, the findings herein add to existing literature by using an objective measure of smoke exposure.
We also used the healthcare utilization data to simulate the cost savings of reducing active smoking and SHS exposure in pregnant persons. Reducing the number of persons who actively smoke by 5% could reduce costs by $150 533 for every 1000 births. According to the most recent US Preventive Services Task Force review of tobacco cessation interventions among pregnant persons, the average quit rate is 16.4% for any behavioral intervention, whereas any control group results in quit rates of 12.2%.43 Combined with data showing most infants are born in hospitals with more than 2000 births per year,44 considerable year-to-year cost savings can be achieved for hospitals that implement basic smoking cessation programs for pregnant persons. Current research shows that psychosocial interventions (eg, counseling, health education, social support) and nicotine replacement therapy in late pregnancy may be effective, though research is ongoing in this area.45,46 Reducing SHS exposure could be more challenging as this requires working with others who smoke around the pregnant person or attempting community-based efforts. The cost savings may be considerably lower, but it is important to the health of the pregnant person and infant to also address reducing SHS exposure.
The results should be interpreted within the context of limitations. Since this study was conducted within one healthcare system, there could be missing data if pregnant persons used multiple healthcare systems. Additionally, since these data were obtained from only areas in North Carolina, the results from this study may not fully generalize to other populations in other areas of the United States or even the state of North Carolina. Although the study’s sample is geographically limited, the meticulousness in methodology, demonstrated through the utilization of serum cotinine levels to gauge smoking exposure and the integration of electronic health records data to extract health utilization and cost, bolsters the validity of the correlational findings presented herein. Although we used an objective measure of smoke exposure, the half-life of cotinine in plasma is short (14–15 hours) and even shorter in pregnant individuals, thus it may not represent true exposure.47,48 Although the study did not find an association between high care costs and smoke exposure among infants, this association has been found in prior research.26 Differences in methodology (eg, self-report vs. biomarkers) could partially explain the discrepancies.
In sum, the findings highlight the importance of reducing smoking and smoke exposure during pregnancy. Expanding cessation services to this vulnerable group could help drive down morbidities observed within this population. In addition, there could be considerable healthcare cost savings to healthcare systems, especially for those with high delivery numbers.
Supplementary material
Supplementary material is available at Nicotine and Tobacco Research online.
Acknowledgments
Drs. Fuemmeler, Dahman, and Wheeler had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis.
Contributor Information
Bernard F Fuemmeler, Department of Family Medicine and Population Health, Virginia Commonwealth University, Richmond, VA, USA; Massey Cancer Center, Virginia Commonwealth University, Richmond, VA, USA.
Bassam Dahman, Department of Health Behavior and Policy, Virginia Commonwealth University, Richmond, VA, USA.
Trevin E Glasgow, Department of Public Health Sciences, University of Virginia, Charlottesville, VA, USA.
D Jeremy Barsell, Department of Family Medicine and Population Health, Virginia Commonwealth University, Richmond, VA, USA.
Jason A Oliver, TSET Health Promotion Research Center, Stephenson Cancer Center, Oklahoma City, OK, USA; Department of Psychiatry & Behavioral Sciences, Oklahoma State University Center for Health Sciences, Tulsa, OK, USA.
Junfeng Zhang, Environmental Science & Policy Division, Duke Global Health Institute and Nicholas School of the Environment, Durham, NC, USA.
Cathrine Hoyo, Department of Biological Sciences, Center for Human Health and the Environment, North Carolina State University, Raleigh, NC, USA.
Susan K Murphy, Department of Obstetrics and Gynecology, Duke University Medical Center, Durham, NC, USA.
F Joseph McClernon, Department of Psychiatry and Behavioral Sciences, Duke University, Durham, NC, USA.
David C Wheeler, Department of Biostatistics, Virginia Commonwealth University, Richmond, VA, USA.
Funding
This research was supported by the National Institute of Environmental Health Sciences (R01ES016772, R21ES014947, P01ES022831, and R24ES028531), the National Cancer Institute (R01CA239595), the US Environmental Protection Agency (RD-83543701), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD084487), the National Institute of Diabetes and Digestive and Kidney Diseases (R01DK085173), and the Duke Cancer Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders. Further, USEPA does not endorse the purchase of any commercial products or services mentioned in the publication. In addition, this research was supported in part by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences.
Declaration of Interests
The authors have no relevant financial or nonfinancial interests to disclose.
Author Contributions
Bernard Fuemmeler (Conceptualization [lead], Investigation [lead], Methodology [lead], Supervision [lead], Writing—original draft [lead], Writing—review & editing [lead]), Bassam Dahman (Formal analysis [lead], Writing—original draft [equal], Writing—review & editing [equal]), Trevin E. Glasgow (Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Jeremy Barsell (Data curation [equal], Formal analysis [equal], Writing—original draft [equal], Writing—review & editing [equal]), Jason Oliver (Investigation [equal], Supervision [equal], Writing—review & editing [equal]), Junfeng (Jim) Zhang (Formal analysis [equal], Writing—review & editing [equal]), Cathrine Hoyo (Conceptualization [equal], Investigation [equal], Methodology [equal], Writing—review & editing [equal]), Susan Murphy (Conceptualization [equal], Writing—review & editing [equal]), F Joseph McClernon (Conceptualization [equal], Writing—review & editing [equal]), and David Wheeler (Formal analysis [equal], Writing—review & editing [equal]).
Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
References
- 1. Massetti GM, Thomas CC, Ragan KR.. Disparities in the context of opportunities for cancer prevention in early life. Pediatrics. 2016;138(suppl 1):S65–S77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Homa DM, Neff LJ, King BA, et al. ; Centers for Disease Control and Prevention (CDC). Vital signs: disparities in nonsmokers’ exposure to secondhand smoke--United States, 1999-2012. MMWR Morb Mortal Wkly Rep. 2015;64(4):103–108. [PMC free article] [PubMed] [Google Scholar]
- 3. Oberg M, Jaakkola MS, Woodward A, Peruga A, Pruss-Ustun A.. Worldwide burden of disease from exposure to second-hand smoke: a retrospective analysis of data from 192 countries. Lancet. 2011;377(9760):139–146. doi: 10.1016/S0140-6736(10)61388-8 [DOI] [PubMed] [Google Scholar]
- 4. Ananth CV, Savitz DA, BowesWA, Jr, Luther ER.. Influence of hypertensive disorders and cigarette smoking on placental abruption and uterine bleeding during pregnancy. Br J Obstet Gynaecol. 1997;104(5):572–578. [DOI] [PubMed] [Google Scholar]
- 5. Cnattingius S. The epidemiology of smoking during pregnancy: smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine Tob Res. 2004;6(6):125–140. [DOI] [PubMed] [Google Scholar]
- 6. Castles A, Adams EK, Melvin CL, Kelsch C, Boulton ML.. Effects of smoking during pregnancy. Five meta-analyses. Am J Prev Med. 1999;16(3):208–215. [DOI] [PubMed] [Google Scholar]
- 7. Roelands J, Jamison MG, Lyerly AD, James AH.. Consequences of smoking during pregnancy on maternal health. J Womens Health (Larchmt). 2009;18(6):867–872. doi: 10.1089/jwh.2008.1024 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Rogers JM. Tobacco and pregnancy. Reprod Toxicol. 2009;28(2):152–160. [DOI] [PubMed] [Google Scholar]
- 9. Shiverick KT, Salafia C.. Cigarette smoking and pregnancy I: ovarian, uterine and placental effects. Placenta. 1999;20(4):265–272. [DOI] [PubMed] [Google Scholar]
- 10. Weitzman M, Gortmaker S, Sobol A.. Maternal smoking and behavior problems of children. Pediatrics. 1992;90(3):342–349. [PubMed] [Google Scholar]
- 11. Ferrante G, Antona R, Malizia V, et al. Smoke exposure as a risk factor for asthma in childhood: a review of current evidence. Allergy Asthma Proc. 2014;35(6):454–461. doi: 10.2500/aap.2014.35.3789 [DOI] [PubMed] [Google Scholar]
- 12. Thacher JD, Gruzieva O, Pershagen G, et al. Parental smoking and development of allergic sensitization from birth to adolescence. Allergy. 2016;71(2):239–248. doi: 10.1111/all.12792 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Nelson L, Valle J, King G, et al. Estimating the proportion of childhood cancer cases and costs attributable to the environment in California. Am J Public Health. 2017;107(5):756–762. doi: 10.2105/AJPH.2017.303690 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Joubert BR, Felix JF, Yousefi P, et al. DNA methylation in newborns and maternal smoking in pregnancy: genome-wide consortium meta-analysis. Am J Hum Genet. 2016;98(4):680–696. doi: 10.1016/j.ajhg.2016.02.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Fuemmeler BF, Dozmorov MG, Do EK, et al. DNA methylation in babies born to nonsmoking mothers exposed to secondhand smoke during pregnancy: an epigenome-wide association study. Environ Health Perspect. 2021;129(5):57010. doi: 10.1289/EHP8099 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. America’s Health Rankings Analysis of America’s Health Rankings Composite Measure. United Health Foundation AmericasHealthRankings.org; 2023. [Google Scholar]
- 17. D’Angelo D, Williams L, Morrow B, et al. ; Centers for Disease Control and Prevention (CDC). Preconception and interconception health status of women who recently gave birth to a live-born infant--Pregnancy Risk Assessment Monitoring System (PRAMS), United States, 26 reporting areas, 2004. MMWR Surveill Summ. 2007;56(10):1–35. [PubMed] [Google Scholar]
- 18. Tong VT, Jones JR, Dietz PM, D’Angelo D, Bombard JM; Centers for Disease Control and Prevention (CDC). Trends in smoking before, during, and after pregnancy - Pregnancy Risk Assessment Monitoring System (PRAMS), United States, 31 sites, 2000-2005. MMWR Surveill Summ. 2009;58(4):1–29. [PubMed] [Google Scholar]
- 19. Shipton D, Tappin DM, Vadiveloo T, et al. Reliability of self reported smoking status by pregnant women for estimating smoking prevalence: a retrospective, cross sectional study. BMJ. 2009;339:b4347. doi: 10.1136/bmj.b4347 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Dietz PM, Homa D, England LJ, et al. Estimates of nondisclosure of cigarette smoking among pregnant and nonpregnant women of reproductive age in the United States. Am J Epidemiol. 2011;173(3):355–359. doi: 10.1093/aje/kwq381 [DOI] [PubMed] [Google Scholar]
- 21. Wheeler DC, Boyle J, Jeremy Barsell D, et al. Tobacco retail outlets, neighborhood deprivation and the risk of prenatal smoke exposure. Nicotine Tob Res. 2022;24(12):2003–2010. doi: 10.1093/ntr/ntac164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Schechter JC, Fuemmeler BF, Hoyo C, et al. Impact of smoking ban on passive smoke exposure in pregnant non-smokers in the southeastern United States. Int J Environ Res Public Health. 2018;15(1):83. doi: 10.3390/ijerph15010083 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Florescu A, Ferrence R, Einarson T, et al. Methods for quantification of exposure to cigarette smoking and environmental tobacco smoke: focus on developmental toxicology. Ther Drug Monit. 2009;31(1):14–30. doi: 10.1097/FTD.0b013e3181957a3b [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Adams EK, Miller VP, Ernst C, et al. Neonatal health care costs related to smoking during pregnancy. Health Econ. 2002;11(3):193–206. [DOI] [PubMed] [Google Scholar]
- 25. Leung GM, Ho LM, Lam TH.. The economic burden of environmental tobacco smoke in the first year of life. Arch Dis Child. 2003;88(9):767–771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Batscheider A, Zakrzewska S, Heinrich J, et al. ; GINIplus and LISAplus study groups. Exposure to second-hand smoke and direct healthcare costs in children – results from two German birth cohorts, GINIplus and LISAplus. BMC Health Serv Res. 2012;12:344. doi: 10.1186/1472-6963-12-344 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Bonnie R. J, Stratton K, Wallace RB.. Ending the Tobacco Problem: Blueprint for the Nation. Washington, DC: Institute of Medicine; 2007. [Google Scholar]
- 28. Hoyo C, Murtha AP, Schildkraut JM, et al. Methylation variation at IGF2 differentially methylated regions and maternal folic acid use before and during pregnancy. Epigenetics. 2011;6(7):928–936. doi: 10.4161/epi.6.7.16263 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Liu Y, Murphy SK, Murtha AP, et al. Depression in pregnancy, infant birth weight and DNA methylation of imprint regulatory elements. Epigenetics. 2012;7(7):735–746. doi: 10.4161/epi.20734 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Horvath MM, Winfield S, Evans S, et al. The DEDUCE guided query tool: providing simplified access to clinical data for research and quality improvement. J Biomed Inform. 2011;44(2):266–276. doi: 10.1016/j.jbi.2010.11.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Horvath MM, Rusincovitch SA, Brinson S, et al. Modular design, application architecture, and usage of a self-service model for enterprise data delivery: The Duke Enterprise Data Unified Content Explorer (DEDUCE). J Biomed Inform. 2014;52:231–242. doi: 10.1016/j.jbi.2014.07.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Dunn A, Grosse SD, Zuvekas SH.. Adjusting health expenditures for inflation: A review of measures for health services research in the United States. Health Serv Res. 2018;53(1):175–196. doi: 10.1111/1475-6773.12612 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. U.S. Department of Labor - (OWCP) - Medical Fee Schedule | U.S. Department of Labor. U.S. Department of Labor. Accessed December 5, 2022. https://www.dol.gov/agencies/owcp/regs/feeschedule/fee/fee12/fs12_ccr [Google Scholar]
- 34. Bernert JT, Jacob P, Holiday DB, et al. Interlaboratory comparability of serum cotinine measurements at smoker and nonsmoker concentration levels: a round-robin study. Nicotine Tob Res. 2009;11(12):1458–1466. doi: 10.1093/ntr/ntp161 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Dempsey DA, Meyers MJ, Oh SS, et al. Determination of tobacco smoke exposure by plasma cotinine levels in infants and children attending urban public hospital clinics. Arch Pediatr Adolesc Med. 2012;166(9):851–856. doi: 10.1001/archpediatrics.2012.170 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Jacob P, Yu L, Duan M, et al. Determination of the nicotine metabolites cotinine and trans-3’-hydroxycotinine in biologic fluids of smokers and non-smokers using liquid chromatography-tandem mass spectrometry: biomarkers for tobacco smoke exposure and for phenotyping cytochrome P450 2A6 activity. J Chromatogr B Analyt Technol Biomed Life Sci. 2011;879(3-4):267–276. doi: 10.1016/j.jchromb.2010.12.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Benowitz NL, Bernert JT, Caraballo RS, Holiday DB, Wang J.. Optimal serum cotinine levels for distinguishing cigarette smokers and nonsmokers within different racial/ethnic groups in the United States between 1999 and 2004. Am J Epidemiol. 2009;169(2):236–248. doi: 10.1093/aje/kwn301 [DOI] [PubMed] [Google Scholar]
- 38. Westreich D, Greenland S.. The table 2 fallacy: presenting and interpreting confounder and modifier coefficients. Am J Epidemiol. 2013;177(4):292–298. doi: 10.1093/aje/kws412 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. VanderWeele TJ, Staudt N.. Causal diagrams for empirical legal research: a methodology for identifying causation, avoiding bias and interpreting results. Law Probab Risk. 2011;10(4):329–354. doi: 10.1093/lpr/mgr019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Narvestad H, Vestergaard CH, Rytter D, Bech BH.. Maternal smoking during pregnancy and offspring utilisation of health care services: A population-based cohort study. Paediatr Perinat Epidemiol. 2019;33(5):384–393. doi: 10.1111/ppe.12577 [DOI] [PubMed] [Google Scholar]
- 41. Adams EK, Melvin CL, Raskind-Hood C, Joski PJ, Galactionova E.. Infant delivery costs related to maternal smoking: an update. Nicotine Tob Res. 2011;13(8):627–637. doi: 10.1093/ntr/ntr042 [DOI] [PubMed] [Google Scholar]
- 42. Vaz LR, Jones MJ, Szatkowski L, et al. Estimating the health-care costs of children born to pregnant smokers in England: cohort study using primary and secondary health-care data. Addiction. 2018;113(7):1305–1316. doi: 10.1111/add.14183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Patnode CD, Henderson JT, Coppola EL, et al. Interventions for tobacco cessation in adults, including pregnant persons: updated evidence report and systematic review for the us preventive services task force. JAMA. 2021;325(3):280–298. doi: 10.1001/jama.2020.23541 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Handley SC, Passarella M, Herrick HM, et al. Birth volume and geographic distribution of US hospitals with obstetric services from 2010 to 2018. JAMA Netw Open. 2021;4(10):e2125373. doi: 10.1001/jamanetworkopen.2021.25373 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Chamberlain C, O’Mara-Eves A, Porter J, et al. Psychosocial interventions for supporting women to stop smoking in pregnancy. Cochrane Database Syst Rev. 2017;2(2):CD001055. doi: 10.1002/14651858.CD001055.pub5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Claire R, Chamberlain C, Davey MA, et al. Pharmacological interventions for promoting smoking cessation during pregnancy. Cochrane Database Syst Rev. 2020;2020(3):CD010078. doi: 10.1002/14651858.CD010078.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Benowitz NL, Hukkanen J, Jacob P.. Nicotine chemistry, metabolism, kinetics and biomarkers. Handb Exp Pharmacol. 2009;192:29–60. doi: 10.1007/978-3-540-69248-5_2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Dempsey D, Jacob P, Benowitz NL.. Accelerated metabolism of nicotine and cotinine in pregnant smokers. J Pharmacol Exp Ther. 2002;301(2):594–598. doi: 10.1124/jpet.301.2.594 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.