Abstract
Objective
To evaluate pharmacologic and psychotherapy treatment among individuals with postpartum depression or anxiety living in urban and rural areas of the United States.
Methods
This is a retrospective cohort study of individuals diagnosed with postpartum depression or anxiety within 12 months of delivery from 2011 to 2020 in the MarketScan Commercial Research Database. Individuals with postpartum depression or anxiety were identified using ICD codes. Individuals previously receiving treatment were excluded. The exposure, urban residence, was determined by residence in a metropolitan area. The primary outcome was a composite of pharmacologic and psychotherapy treatment for depression or anxiety. Secondary outcomes included components of the primary outcome and time from delivery to first prescription fill. Multivariable modeling adjusted for clinically relevant confounders including history of depression or anxiety that was untreated. Sensitivity analysis included individuals with prior treatment.
Results
Of 181,185 individuals included, 159,001 (88%) had urban residence. Individuals with postpartum depression or anxiety residing in urban areas were less likely to receive treatment (aHR 0.90; 95% CI 0.88–0.92) than those residing in rural areas, driven by pharmacologic treatment (aHR 0.79; 95% CI 0.77–0.81). Time from delivery to first prescription fill was longer in urban (mean 134 days, 95% CI 133–134) compared to rural areas (mean 128 days, 95% CI 127–130). Primary results were similar in sensitivity analyses.
Conclusion
Urban residing individuals with postpartum depression or anxiety were less likely to receive pharmacologic or psychotherapy treatment than rural residing individuals in this insured US cohort. This finding was driven by differences in pharmacologic treatment among groups.
Keywords: postpartum anxiety, postpartum depression, rural, treatment, urban
1. INTRODUCTION
Rural America is disproportionately affected by chronic diseases, dichotomy of age ranges, and higher poverty rates [1]. Mental health care in these areas is limited by lack of availability, accessibility and acceptability [1]. A 2016 literature review found a higher prevalence of postpartum depression in rural regions of the United States as compared to urban regions, consistent with data from international studies finding higher rates of postpartum depression in individuals living in rural regions [2, 3]. Two recent studies, in the United States and the United Kingdom, found a higher risk of perinatal depression among rural individuals compared to urban but this did not persist after adjusting for measures of socioeconomic status and support [4, 5].
There are few data regarding perinatal mental health treatment in rural America despite the relative burden [2]. Cultural, social, and family demands, in addition to provider comfort with types of treatment in pregnant and breastfeeding individuals, are likely different in rural and urban areas, and may influence the type of treatment that an individual receives. A study among all adults by Ziller et al. found that rural areas have less office‐based mental health visits and more pharmacologic prescribing compared to urban areas based on the 2003–2004 Medical Expenditure Panel Surveys in the United States [6]. While perinatal and postpartum depression and anxiety can be treated through both psychotherapy and pharmacologic interventions, there are limited data on treatment rates of these interventions across the United States and how this may differ between urban and rural residing individuals.
The objective of our study was to compare the treatment of postpartum depression and anxiety among individuals residing in urban areas compared to those in rural areas. Specifically, we examined both pharmacologic and psychotherapy interventions in postpartum individuals with depression and anxiety. We hypothesized that treatment would differ between groups with individuals residing in urban areas receiving treatment more often than individuals residing in rural areas.
2. METHODS
We performed a retrospective cohort study using the IBM MarketScan database. MarketScan is a commercial insurance‐based claims database including data from all 50 states. It includes information on outpatient and inpatient services, pharmaceuticals, and de‐identified individual characteristics. This analysis was exempt from Institutional Review Board approval as all data were de‐identified.
All deliveries to females aged 13–50 years between 2011 and 2020 were identified using a validated algorithm of International Classification of Diseases Ninth Revision (ICD‐9) and Tenth Revision (ICD‐10) codes [7, 8]. The cohort was limited to pregnancies that were delivered between 20 and 42 weeks. Gestational age at time of delivery was generated from ICD‐9 and ICD‐10 codes of neonatal linked claims, which were linked in the MarketScan database starting in 2011, or in unlinked claims calculated from antepartum procedures or the maternal delivery code as a median national value [7, 9, 10–14]. Only the first eligible delivery for each individual was included. Individuals with a diagnosis of depression or anxiety in the 12 months after delivery were included in the analysis. Individuals were excluded if they did not have 12 months of pre‐delivery enrollment in MarketScan, received treatment for depression or anxiety in the 12 months prior to delivery, or had a diagnosis of bipolar disorder, schizoaffective disorder, schizophrenia, or psychosis in the 12 months prior to delivery [15, 16]. Mental health diagnoses were based on ICD‐9 and ICD‐10 codes (Supporting Information) [15, 16, 17, 18, 19].
The primary exposure was location of residence—urban or rural—based on metropolitan statistical area. Areas are designated as metropolitan statistical areas, micropolitan statistical areas, or neither by the United States Census Bureau and the Office of Management and Budget [20, 21]. Utilizing the National Center for Health Statistics Urban‐Rural Classification Scheme for Counties, in statistical analyses these designations are frequently dichotomized into urban (metropolitan statistical areas) or rural (non‐metropolitan statistical areas which include micropolitan statistical areas or non‐core areas) as it is consistent across all states [22, 23, 24, 25, 26]. The primary outcome was treatment of depression or anxiety in the 12 months after delivery. Treatment was defined as pharmacologic, psychotherapy, or both. Secondary outcomes included components of the primary outcome (pharmacologic or psychotherapy) and time from delivery to first fill of a prescription.
Medications were defined as psychotropic prescriptions identified from prior studies, assessed for clinical applicability (LG), and grouped into antidepressants and anxiolytics (Supporting Information) [18, 27, 28]. Medication prescription fills were identified by National Drug Codes (NDC) in MarketScan pharmacy claims for the specified antidepressant and anxiolytic medications. Psychotherapy visits were ascertained by insurance claims for outpatient services, revenue codes or procedure codes. Treatment—pharmacologic or psychotherapy—required both a diagnosis of depression or anxiety and the specified antidepressant, anxiolytic, or psychotherapy visit to exclude those obtaining treatment for other indications. We recorded the severity of mental health diagnosis in the 12 months following delivery, with severe disease defined as any of the following: inpatient admission for depression or anxiety, receiving mental health services within an inpatient setting, a diagnosis of psychosis, or attempted suicide or self‐harm based on ICD‐9 and ICD‐10 codes [15, 29, 30].
Demographic data included maternal age at delivery, gestational age at delivery, mode of delivery, depression or anxiety (in the 12 months prior to delivery), and region (Northeast, Midwest, South, West, other). Confounders such as substance use disorder were identified by ICD‐9 and ICD‐10 codes and recorded for alcohol, substance, and tobacco use in the 12 months prior to delivery and in the postpartum period [15, 16]. Comorbidities were identified using a validated maternal comorbidity index (MCI) [31] based on the 6 months prior to delivery. Severe maternal morbidity index (SMM) was assessed using the Centers for Disease Control and Prevention [32] criteria during delivery admission. Both indices were based on ICD codes (Supporting Information). Birth‐related factors included preterm birth (less than 36 weeks), congenital malformations, [33] stillbirth, and neonatal mortality.
To account for individuals who lost commercial insurance after delivery, censoring was used at the time they left the database or at 12 months after mental health diagnosis was recorded. Demographic and baseline clinical characteristics were compared between individuals living in urban and rural areas using Chi‐squared or Kruskal–Wallis tests for variables, as appropriate. The incidence of the primary and secondary outcomes was reported by rural and urban area of residence. To compare the probability of treatment based on urban or rural residence, a univariate Cox proportional hazards ratio was performed for primary and secondary outcomes. A Cox proportional hazards model was developed to determine factors associated with occurrence of treatment (medication, psychotherapy, or both) in the postpartum period in urban and rural areas; this included age, a history of depression or anxiety in the 12 months prior to delivery, substance use disorder, presence of any factor in the maternal comorbidity index, presence of any severe maternal morbidity, infant‐related factors, and severity of illness. A priori we planned to control for a history of depression or anxiety, and any factor in the Cox proportional hazards model that remained significant in univariable modeling. Adjusted hazard ratios were calculated with 95% confidence intervals. As a secondary outcome, we examined time (days) from delivery to first fill of a medication, and time (days) from diagnosis to first fill of a medication, between urban and rural individuals.
In a sensitivity analysis, the cohort was redefined to include individuals who received treatment for depression or anxiety in the 12 months prior to delivery. An additional sensitivity analysis was completed in a time to treatment analysis, redefining the cohort to include individuals who received treatment (medications or therapy) in the postpartum period but did not have a diagnosis code for depression or anxiety. This was done to improve accuracy of our claims‐based data and decrease the chance that we were missing individuals who were receiving treatment but did not have a diagnosis code. Lastly, we examined the incidence of a diagnosis of postpartum depression or anxiety in the baseline population between urban and rural individuals to ensure a difference in diagnosis between groups was not contributing to potential findings. We followed STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) reporting guidelines for observational studies [34].
3. RESULTS
A total of 181,185 individuals with postpartum depression or anxiety met inclusion/exclusion criteria for analysis (Figure 1). Most individuals lived in urban areas (88%) (Table 1). There were differences between urban and rural individuals in median maternal age at delivery (urban 31 years, rural 28 years; p < 0.001), gestational age at delivery (urban 38 weeks, rural 37 weeks; p < 0.001), and rates of preterm birth < 36 weeks (urban 33.3%, rural 36.2%; p < 0.001). Region of residence in the United States was statistically different between groups although the most common location for both urban and rural‐residing individuals was the South followed by the Midwest, West, and Northeast (Table 1). Individuals living in rural areas were more likely to use tobacco (antenatal urban 2.91%, rural 5.5% p < 0.001; postpartum urban 3.4%, rural 5.88% p < 0.001) and illicit substances than urban individuals (antenatal urban 1.03%, rural 1.33% p < 0.001; postpartum urban 1.33%, rural 1.74% p < 0.001). There was no statistically significant difference between groups in depression or anxiety prior to delivery (urban 53.2%, rural 52.8%; p = 0.322).
FIGURE 1.
Population cohort.
TABLE 1.
Baseline characteristics of the analytic population.
| Variable | Urban | Rural | All | p value |
|---|---|---|---|---|
| Demographics | ||||
| Population density | 159,001 (87.8%) | 22,184 (12.2%) | 181,185 (100%) | |
| Maternal age (years) | 31 (27–34) | 28 (24–32) | 30 (26–34) | p < .001 |
| Region | p < .001 | |||
| Northeast | 27,652 (17.4%) | 2540 (11.5%) | 30,192 (16.7%) | |
| Midwest | 35,403 (22.3%) | 6485 (29.2%) | 41,888 (23.1%) | |
| South | 58,702 (36.9%) | 10,602 (47.8%) | 69,304 (38.3%) | |
| West | 34,264 (21.6%) | 2554 (11.5%) | 36,818 (20.3%) | |
| Unknown | 2980 (1.87%) | 3 (0.01%) | 2983 (1.65%) | |
| Birth characteristics | ||||
| Gestational age (weeks) | 38 (35‐39) | 37 (34‐39) | 38 (35‐39) | p < .001 |
| Neonatal mortality a | 106 (0.1%) | 12 (0.08%) | 118 (0.09%) | |
| Stillbirth | 1759 (1.11%) | 255 (1.15%) | 2014 (1.11%) | p = .589 |
| Congenital malformations | 4967 (3.12%) | 624 (2.81%) | 5591 (3.09%) | p = .013 |
| Preterm delivery (< 36w) | 52,995 (33.3%) | 8034 (36.2%) | 61,029 (33.7%) | p < .001 |
| Comorbidities | ||||
| Severe Maternal Morbidity | p < .001 | |||
| 0 | 152,561 (96%) | 21,399 (96.5%) | 173,960 (96%) | |
| ≥ 1 | 6440 (4%) | 785 (3.5%) | 7225 (4%) | |
| Maternal Comorbidity Index | p = .624 | |||
| 0 | 96,549 (60.7%) | 13,432 (60.6%) | 109,981 (60.7%) | |
| ≥ 1 | 62,452 (39.3%) | 8752 (39.4%) | 71,204 (39.3%) | |
| Mental health history | ||||
| History depression or anxiety | 84547 (53.2%) | 11717 (52.8%) | 96,264 (53.1%) | p = .322 |
| Prior to delivery substance use | 1638 (1.03%) | 296 (1.33%) | 1934 (1.07%) | p < .001 |
| Postpartum substance use | 2115 (1.33%) | 385 (1.74%) | 2500 (1.38%) | p < .001 |
| Prior to delivery alcohol use | 304 (0.19%) | 39 (0.18%) | 343 (0.19%) | p = .681 |
| Postpartum alcohol use | 1281 (0.81%) | 144 (0.65%) | 1425 (0.79%) | p = .015 |
| Prior to delivery tobacco use | 4620 (2.91%) | 1221 (5.5%) | 5841 (3.22%) | p < .001 |
| Postpartum tobacco use | 5399 (3.4%) | 1304 (5.88%) | 6703 (3.7%) | p < .001 |
Neonatal mortality is only calculated for births prior to 2016; p value not calculated.
A smaller proportion of urban individuals received treatment with medications or psychotherapy for postpartum depression or anxiety compared to rural individuals (urban 55.1%, rural 59.4% p < 0.001; Table 2). This persisted in examining treatment with medication alone (urban 31.8%, rural 39.6% p < 0.001) and treatment with psychotherapy alone (urban 44.8%, rural 46% p < 0.001).
TABLE 2.
Treatment of postpartum depression or anxiety.
| Outcome | Urban (n = 159,001) | Rural (n = 22,184) | HR (95% CI) | aHR a (95% CI) |
|---|---|---|---|---|
| Treatment | 87,543 (55.1%) | 13,186 (59.4%) | 0.88 (0.86‐0.90) | 0.90 (0.88‐0.92) |
| Medication | 50,569 (31.8%) | 8789 (39.6%) | 0.76 (0.74‐0.78) | 0.79 (0.77‐0.81) |
| Therapy | 71,251 (44.8%) | 10,204 (46%) | 0.97 (0.95‐0.99) | 0.99 (0.97‐1.01) |
Abbreviations: aHR, adjusted hazard ratio; HR, hazard ratio.
Adjusted for SMM, MCI, age, substance, alcohol and tobacco use, preterm birth, congenital malformations, depression/anxiety in the year before delivery, and severe mental health diagnosis.
In unadjusted and adjusted models, individuals in urban areas with postpartum depression or anxiety were less likely to receive treatment (medication or psychotherapy) than individuals in rural areas (aHR 0.90 95% CI 0.88–0.92; Table 2). In examining the components of the composite, this appeared to be driven by a difference in medication treatment between groups (aHR 0.79; 95% CI 0.77–0.81). While urban individuals were also less likely to receive treatment with psychotherapy than rural individuals (HR 0.97 95% CI 0.95–0.99), this was no longer statistically significant after controlling for confounders (aHR 0.99 95% CI 0.97–1.01).
Adjustments included a priori history of depression or anxiety and factors that remained significant in modeling; full model for factors associated with the occurrence of treatment is listed in the Supporting Information. Individuals with substance use in the year prior to delivery, severe mental health, or any severe maternal morbidity were more likely to receive treatment. Individuals with a diagnosis of depression or anxiety in the year prior to delivery, preterm deliveries, congenital malformations, or any maternal comorbidity index were less likely to receive any treatment.
Time from delivery to first fill of a medication was longer in urban than rural individuals but was not clinically meaningful (Mean 134 days Urban, 128 days Rural, p < 0.001) (Figure 2). Time from diagnosis to first fill of a medication was also longer in urban than rural individuals but was again not clinically meaningful (Mean 46 days Urban, 41 days Rural, p < 0.001).
FIGURE 2.
Time until first medication fill from mental health diagnosis (MH dx) and from end of pregnancy (EOP). Green line, rural individuals; gray line, urban individuals. Vertical lines denote 95% CI.
In sensitivity analysis re‐defining the cohort to include individuals with treatment in the 12 months prior to delivery, individuals with postpartum depression or anxiety in urban areas were less likely to receive any treatment (aHR 0.97; 95% CI 0.96–0.99), including treatment with medications (aHR 0.86; 95% CI 0.84–0.87), but more likely to receive treatment with psychotherapy (aHR 1.08; 95% CI 1.06–1.10). In this sensitivity analysis, the differences appear to be driven by the rate of continuing treatment, where prior treatment with medications and/or therapy weighed heavily in the adjusted hazard ratios for treatment postpartum (prior medication treatment aHR 2.25, 95% CI 2.22–2.27; prior therapy treatment aHR 2.58, 95% CI 2.55–2.61).
The sensitivity analysis including those with treatment (medication or therapy) in the postpartum period regardless of diagnosis codes did not change our primary findings. Individuals living in urban areas were still less likely than individuals in rural areas to receive any treatment (medication or therapy) (aHR 0.84 95% CI 0.83‐0.85), treatment with medications (aHR 0.76 95% CI 0.74‐0.77), or treatment with therapy (0.95 95% CI 0.94‐0.97). Lastly, in examining the baseline population, a diagnosis of depression or anxiety in the 12 months postpartum was seen in clinically similar proportions of urban (11.23%) and rural (11.91%) individuals (p < 0.001).
4. DISCUSSION
Individuals with postpartum depression or anxiety in urban areas were less likely to receive new treatment with medications and/or psychotherapy, than those in rural areas. This finding was driven by differences in treatment with medication among groups. When we examined all treatment, rather than just new treatment, medications were still less likely to be prescribed to urban individuals than rural individuals in the postpartum period for depression or anxiety.
Our findings are similar to other studies in the adult population, pharmacologic prescribing occurred more often in rural areas than urban areas [6]. We thought this might be due to higher rates of postpartum depression in rural as compared to urban areas as suggested in other studies [2, 3]. However, a diagnosis of depression or anxiety in the 12 months postpartum was seen in clinically similar proportions of urban and rural individuals in the baseline population. Additionally, results were similar when we included all individuals who received treatment in the postpartum period. This accounted for the potential error of providers not entering a proper diagnosis code. Lastly, in our Cox regression analysis we a priori controlled for a prior diagnosis of depression or anxiety and still found that individuals in urban areas were less likely to receive treatment in the postpartum period than those in rural areas.
Consistent with our finding that new medication treatment of postpartum depression or anxiety is filled over four months after delivery, other studies have also found that recurrent or new onset depression often persists past the 2 or 6 week postpartum visit [35, 36]. Around 30% of patients diagnosed with postpartum depression still meet diagnostic criteria between 4 months and 3 years postpartum [35, 36].
While only a hypothesis, the difference in treatment seen in our study may be due to the need occurring outside the typical 6‐week postpartum period and easier access to a provider who will see a patient long term for mental health care, such as a family medicine or internal medicine primary care provider. Typically, a patient may only have contact with their delivering Obstetric provider through the 6‐week postpartum visit. In a rural location the delivering Obstetric provider is more likely to be a family medicine provider [37, 38, 39]. Such physicians may see patients for more than 6 weeks postpartum, as well as see them for their newborn's care, and have a longer relationship with them outside of obstetrics. As many rural individuals have limited access to birthing hospitals, they may be driving further distances to care and may also have limited access to in person pharmacotherapy options. Additionally, as rural patients are more likely to receive treatment for mental health care from primary care providers than behavioral health specialists, they may have immediate access to treatment options, instead of relying on a referral to a specialist [40, 41]. These providers should continue to be supported and included in discussions on treatment of postpartum depression and anxiety in the United States.
Primary care providers in rural areas may also feel more comfortable treating postpartum depression or anxiety than Obstetricians and Gynecologists. Obstetricians and Gynecologists feel it is their responsibility, but do not feel comfortable treating postpartum depression [42, 43]. Obstetricians and Gynecologists, who are more likely to deliver in urban areas [38], need better psychopharmacotherapy education during their training and through continued medical education [44]. Enhancing capabilities of Obstetrics and Gynecology providers to treat postpartum depression and anxiety may allow improved access, continuity of care, and potential space for more severe individuals to see Psychiatrists and Psychiatric Nurse Practitioners.
This study has several limitations. Use of administrative data risks under‐ascertainment and misclassification. There are likely individuals that had a diagnosis of postpartum depression or anxiety but did not have the ICD code entered. A lower baseline rate of postpartum depression or anxiety (11‐12%) was seen for both urban and rural populations compared to prior studies, this may reflect a lack of entry of diagnoses codes by providers. However, this would bias our results towards the null making it more likely that a difference exists between treatment of rural and urban individuals. Individuals may have been prescribed a medication but did not fill the prescription. Nonetheless, by using only medications that were filled by the patient our study captures real world data and is more clinically relevant to patient care. Psychotherapy can be accessed by patients outside of insurance systems and we were not able to capture this information. We were unable to measure socioeconomic status, race/ethnicity, education, social support, or relationship status. Our estimate of the metropolitan statistical area was based on the primary payor but may not coincide with the pregnant individual. Data were only included for those with commercial insurance reducing the generalizability of findings. Recent studies have examined a difference in pregnancy and postpartum insurance coverage between urban and rural residents; rural residents were more likely to be uninsured than urban residents [23]. Uninsured individuals, or intermittent coverage, is associated with less prenatal care and recommended postpartum care compared to those with continuous insurance [45]. Obtaining more granular data to examine these components is integral to future studies. There is also the possibility of a type 1 error in our study as a sample size was not calculated.
Our study has several strengths. We had a large and diverse population, representative of much of the United States. The IBM MarketScan database accurately reflects the general ratio of urban to rural individuals in the United States [46]. The breadth of information available to us allowed us to examine multiple different medications, diagnosis codes and control for many confounders. Additionally, we were able to tell if the prescription was filled which is more consistent with real world clinical scenarios. By excluding individuals who previously received treatment in our primary analysis we avoided misclassification. However, in sensitivity analysis including these individuals we found the same results for our primary outcome and medication treatment, reassuring us that our initial results were reliable.
5. CONCLUSION
Despite safety profiles and known efficacy of medications and psychotherapy for postpartum depression and anxiety, we identified differences in the treatment of these conditions based on rural or urban location in commercially insured individuals. Variation in access, engagement, and receipt of care may contribute to these differences. Our study results are likely underestimated given less than half of individuals with postpartum depression seek treatment [47]. Promotion of follow up for treatment and efforts to understand differences in treatment between urban and rural individuals during the postpartum period are warranted.
CONFLICT OF INTEREST STATEMENT
Unrelated to this work, the University of Utah receives funds for M.C.S time from Gilead Science Incorporated and KoKo Medical. Other authors declare no conflicts of interest.
Supporting information
Supporting Information
ACKNOWLEDGMENTS
We thank the University of Utah Obstetrics and Gynecology Department and the Department of Surgery and Surgical Population Analysis Research Core for data collection, management, and analysis. The research reported in this manuscript was funded in part by the Utah Stimulating Access to Research in Residency (StARR) under Award Number 1R38HL167282‐01.
An earlier version of the abstract as an ePoster was presented at the 2023 ACOG Annual Clinical Scientific Meeting in Baltimore, Maryland, from May 19 to 21.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
REFERENCES
- 1. Human, J. , and Wasem C.. 1991. “Rural Mental Health in America.” American Psychologist 46(3): 232–9. 10.1037//0003-066x.46.3.232. [DOI] [PubMed] [Google Scholar]
- 2. Mollard, E. , Hudson D. B., Ford A., and Pullen C.. 2016. “An Integrative Review of Postpartum Depression in Rural U.S. Communities.” Archives of Psychiatric Nursing 30(3): 418–24. 10.1016/j.apnu.2015.12.003. [DOI] [PubMed] [Google Scholar]
- 3. Villegas, L. , McKay K., Dennis C.‐L., and Ross L. E.. 2011. “Postpartum Depression among Rural Women From Developed and Developing Countries: A Systematic Review.” Journal of Rural Health 27(3): 278–88. 10.1111/j.1748-0361.2010.00339.x. [DOI] [PubMed] [Google Scholar]
- 4. Ginja, S. , Jackson K., Newham J. J., Henderson E. J., Smart D., and Lingam R.. 2020. “Rural‐Urban Differences in the Mental Health of Perinatal Women: A UK‐Based Cross‐Sectional Study.” BMC Pregnancy and Childbirth 20(1): 464. 10.1186/s12884-020-03132-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Nidey, N. , Tabb K. M., Carter K. D., Bao W., Strathearn L., Rohlman D. S., Wehby G., and Ryckman K.. 2020. “Rurality and Risk of Perinatal Depression among Women in the United States.” Journal of Rural Health 36(1): 9–16. 10.1111/jrh.12401. [DOI] [PubMed] [Google Scholar]
- 6. Ziller, E. C. , Anderson N. J., and Coburn A. F.. 2010. “Access to Rural Mental Health Services: Service Use and Out‐of‐Pocket Costs.” Journal of Rural Health 26(3): 214–24. 10.1111/j.1748-0361.2010.00291.x. [DOI] [PubMed] [Google Scholar]
- 7. Bruno, A. M. , Horns J. J., Allshouse A. A., Metz T. D., Debbink M. L., and Smid M. C.. 2023. “Association between Periviable Delivery and New Onset of or Exacerbation of Existing Mental Health Disorders.” Obstetrics and Gynecology 141(2): 395–402. 10.1097/AOG.0000000000005050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kuklina, E. V. , Whiteman M. K., Hillis S. D., Jamieson D. J., Meikle S. F., Posner S. F., and Marchbanks P. A.. 2008. “An Enhanced Method for Identifying Obstetric Deliveries: Implications for Estimating Maternal Morbidity.” Maternal and Child Health Journal 12(4): 469–77. 10.1007/s10995-007-0256-6. [DOI] [PubMed] [Google Scholar]
- 9. Ailes, E. C. , Simeone R. M., Dawson A. L., Petersen E. E., and Gilboa S. M.. 2016. “Using Insurance Claims Data to Identify and Estimate Critical Periods in Pregnancy: An Application to Antidepressants.” Birth Defects Research Part A: Clinical and Molecular Teratology 106(11): 927–34. 10.1002/bdra.23573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Hornbrook, M. C. , Whitlock E. P., Berg C. J., Callaghan W. M., Bachman D. J., Gold R.…Williams S. B.. 2007. “Development of an Algorithm to Identify Pregnancy Episodes in an Integrated Health Care Delivery System.” Health Services Research 42(2): 908–27. 10.1111/j.1475-6773.2006.00635.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Kharbanda, E. O. , Vazquez‐Benitez G., Shi W. X., Lipkind H., Naleway A., Molitor B., and Nordin J. D.. 2012. “Assessing the Safety of Influenza Immunization during Pregnancy: The Vaccine Safety Datalink.” American Journal of Obstetrics and Gynecology 207(3 Suppl): S47–51. 10.1016/j.ajog.2012.06.073. [DOI] [PubMed] [Google Scholar]
- 12. Margulis, A. V. , Setoguchi S., Mittleman M. A., Glynn R. J., Dormuth C. R., and Hernández‐Díaz S.. 2013. “Algorithms to Estimate the Beginning of Pregnancy in Administrative Databases.” Pharmacoepidemiology and Drug Safety 22(1): 16–24. 10.1002/pds.3284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Rosenbloom, J. I. , Stwalley D., Lindley K. J., Michael Nelson D., Olsen M. A., and Stout M. J.. 2020. “Latency of Preterm Hypertensive Disorders of Pregnancy and Subsequent Cardiovascular Complications.” Pregnancy Hypertension 21: 139–44. 10.1016/j.preghy.2020.05.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Yusuf, A. , Chia V., Xue F., Mikol D. D., Bollinger L., and Cangialose C.. 2018. “Use of Existing Electronic Health Care Databases to Evaluate Medication Safety in Pregnancy: Triptan Exposure in Pregnancy as a Case Study” Pharmacoepidemiology and Drug Safety 27(12): 1309–15. 10.1002/pds.4658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lewkowitz, A. K. , Rosenbloom J. I., Keller M., Lopez J. D., Macones G. A., Olsen M. A., and Cahill A. G.. 2019. “Association between Severe Maternal Morbidity and Psychiatric Illness Within 1 Year of Hospital Discharge After Delivery.” Obstetrics and Gynecology 134(4): 695–707. 10.1097/AOG.0000000000003434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. World Health Organization . 1992. The ICD‐10 Classification of Mental and Behavioural Disorders Clinical Descriptions and Diagnostic Guidelines. Geneva: World Health Organization. [Google Scholar]
- 17. Fiest, K. M. , Jette N., Quan H., St Germaine‐Smith C., Metcalfe A., Patten S. B., and Beck C. A.. 2014. “Systematic Review and Assessment of Validated Case Definitions for Depression in Administrative Data.” BMC Psychiatry 14(1): 289. 10.1186/s12888-014-0289-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Hanley, G. E. , and Mintzes B.. 2014. “Patterns of Psychotropic Medicine Use in Pregnancy in the United States from 2006 to 2011 Among Women with Private Insurance.” BMC Pregnancy and Childbirth 14(1): 242. 10.1186/1471-2393-14-242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Townsend, L. , Walkup J. T., Crystal S., and Olfson M.. 2012. “A Systematic Review of Validated Methods for Identifying Depression Using Administrative Data.” Pharmacoepidemiology and Drug Safety 21(Suppl 1): 163–73. 10.1002/pds.2310. [DOI] [PubMed] [Google Scholar]
- 20. United States Census Bureau . “Delineation Files.” Accessed June 20, 2023. https://www.census.gov/geographies/reference‐files/time‐series/demo/metro‐micro/delineation‐files.html.
- 21. United States Census Bureau . “Housing Patterns and Core‐Based Statistical Areas.” Accessed June 20, 2023. https://www.census.gov/topics/housing/housing‐patterns/about/core‐based‐statistical‐areas.html.
- 22. Administration, Health Resources and Services . Defining Rural Population. U.S. Department of Health & Human Services. https://www.hhs.gov/guidance/document/defining‐rural‐population#:~:text=A%20Metro%20area%20contains%20a,(MSA)%20are%20considered%20rural. [Google Scholar]
- 23. Admon, L. K. , Daw J. R., Interrante J. D., Ibrahim B. B., Millette M. J., and Kozhimannil K. B.. 2023. “Rural and Urban Differences in Insurance Coverage at Prepregnancy, Birth, and Postpartum.” Obstetrics & Gynecology, 14(3): 570–81. 10.1097/AOG.0000000000005081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. U.S. Census Bureau . Understanding and Using American Community Survey Data—Defining “Rural” Areas. https://www.census.gov/content/dam/Census/library/publications/2019/acs/ACS_rural_handbook_2019_ch01.pdf. [Google Scholar]
- 25. Kozhimannil, K. B. , Interrante J. D., Basile Ibrahim B., Chastain P., Millette M. J., Daw J., and Admon L. K.. 2022. “Racial/Ethnic Disparities in Postpartum Health Insurance Coverage Among Rural and Urban U.S. Residents.” Journal of Women's Health 31(10): 1397–402. 10.1089/jwh.2022.0169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Prevention, Centers for Disease Control and . “NCHS Urban‐Rural Classification Scheme for Counties.” Accessed June 20, 2023. https://www.cdc.gov/nchs/data_access/urban_rural.htm.
- 27. Hershenberg, R. , Gros D. F., and Brawman‐Mintzer O.. 2014. “Role of Atypical Antipsychotics in the Treatment of Generalized Anxiety Disorder.” CNS Drugs 28(6): 519–33. 10.1007/s40263-014-0162-6. [DOI] [PubMed] [Google Scholar]
- 28. Methodology, WHO Collaborating Centre for Drug Statistics . 2021. Guidelines for ATC Classification and DDD Assignment 2021. Oslo, Norway, 20.
- 29. Hedegaard, H. , and Johnson R. L.. 2019. “An Updated International Classification of Diseases, 10th Revision, Clinical Modification (ICD‐10‐CM) Surveillance Case Definition for Injury Hospitalizations.” National Health Statistics Reports (125): 1–8. https://www.ncbi.nlm.nih.gov/pubmed/31751206. [PubMed] [Google Scholar]
- 30. Hedegaard, H. , Schoenbaum M., Claassen C., Crosby A., Holland K., and Proescholdbell S.. 2018. “Issues in Developing a Surveillance Case Definition for Nonfatal Suicide Attempt and Intentional Self‐Harm Using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD‐10‐CM) Coded Data.” National Health Statistics Reports (108): 1–19. https://www.ncbi.nlm.nih.gov/pubmed/29616901. [PubMed] [Google Scholar]
- 31. Bateman, B. T. , Mhyre J. M., Hernandez‐Diaz S., Huybrechts K. F., Fischer M. A., Creanga A. A., Callaghan W. M., and Gagne J. J.. 2013. “Development of a Comorbidity Index For Use in Obstetric Patients.” Obstetrics and Gynecology 122(5): 957–65. 10.1097/AOG.0b013e3182a603bb. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Promotion, National Center for Chroic Disease Prevention and Health . 2021. “How Does CDC Identify Severe Maternal Morbidity?.” Accessed October 5th. https://www.cdc.gov/reproductivehealth/maternalinfanthealth/smm/severe‐morbidity‐ICD.htm.
- 33. Huybrechts, K. F. , Kulldorff M., Hernández‐Díaz S., Bateman B. T., Zhu Y., Mogun H., and Wang S. V.. 2021. “Active Surveillance of the Safety of Medications Used During Pregnancy.” American Journal of Epidemiology 190(6): 1159–68. 10.1093/aje/kwaa288. [DOI] [PubMed] [Google Scholar]
- 34. von Elm, E. , Altman D. G., Egger M., Pocock S. J., Gotzsche P. C., Vandenbroucke J. P., and Initiative S.. 2007. “The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies.” Annals of Internal Medicine 147(8): 573–7. 10.7326/0003-4819-147-8-200710160-00010. [DOI] [PubMed] [Google Scholar]
- 35. Vliegen, N. , Casalin S., and Luyten P.. 2014. “The Course of Postpartum Depression: A Review of Longitudinal Studies.” Harvard Review of Psychiatry 22(1): 1–22. 10.1097/HRP.0000000000000013. [DOI] [PubMed] [Google Scholar]
- 36. Wisner, K. L. , Perel J. M., Peindl K. S., and Hanusa B. H.. 2004. “Timing of Depression Recurrence in the First Year after Birth.” Journal of Affective Disorders 78(3): 249–52. 10.1016/S0165-0327(02)00305-1. [DOI] [PubMed] [Google Scholar]
- 37. Deutchman, M. , Macaluso F., Bray E., Evans D., Boulger J., Quinn K., James K. A., et al. 2022. “The Impact of Family Physicians in Rural Maternity Care.” Birth 49(2): 220–32. 10.1111/birt.12591. [DOI] [PubMed] [Google Scholar]
- 38. Rayburn, W. F. , Klagholz J. C., Murray‐Krezan C., Dowell L. E., and Strunk A. L.. 2012. “Distribution of American Congress of Obstetricians and Gynecologists Fellows and Junior Fellows in Practice in the United States.” Obstetrics and Gynecology 119(5): 1017–22. 10.1097/AOG.0b013e31824cfe50. [DOI] [PubMed] [Google Scholar]
- 39. Young, R. A. 2017. “Maternity Care Services Provided by Family Physicians in Rural Hospitals” Journal of the American Board of Family Medicine 30(1): 71–77. 10.3122/jabfm.2017.01.160072 [DOI] [PubMed] [Google Scholar]
- 40. Killeen, D. , Jetty A., Peterson L. E., Bazemore A., and Jabbarpour Y.. 2022. “The Association of Practice Type and the Scope of Care of Family Physicians.” Journal of the American Board of Family Medicine. 10.3122/jabfm.2022.220172R1. [DOI] [PubMed] [Google Scholar]
- 41. Muench, U. , Jura M., Thomas C. P., Perloff J., and Spetz J.. 2022. “Rural‐Urban Prescribing Patterns by Primary Care and Behavioral Health Providers in Older Adults with Serious Mental Illness.” BMC Health Services Research 22(1): 1440. 10.1186/s12913-022-08813-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. LaRocco‐Cockburn, A. , Melville J., Bell M., and Katon W.. 2003. “Depression Screening Attitudes and Practices among Obstetrician‐Gynecologists.” Obstetrics and Gynecology 101(5 Pt 1): 892–8. 10.1016/s0029-7844(03)00171-6. [DOI] [PubMed] [Google Scholar]
- 43. Leiferman, J. A. , Dauber S. E., Heisler K., and Paulson J. F.. 2008. “Primary Care Physicians' Beliefs and Practices toward Maternal Depression.” Journal of Women's Health 17(7): 1143–50. 10.1089/jwh.2007.0543. [DOI] [PubMed] [Google Scholar]
- 44. Hutner, L. A. , Yeaton‐Massey A., Toscano M., Coulehan J., Hage B., Gopalan P., Miller E. S., et al. 2021. “Cultivating Mental Health Education in Obstetrics and Gynecology: A Call to Action.” American Journal of Obstetrics & Gynecology MFM 3(6): 100459. 10.1016/j.ajogmf.2021.100459. [DOI] [PubMed] [Google Scholar]
- 45. Admon, L. K. , Daw J. R., Winkelman T. N. A., Kozhimannil K. B., Zivin K., Heisler M., and Dalton V. K.. 2021. “Insurance Coverage and Perinatal Health Care Use Among Low‐Income Women in the US, 2015‐2017.” JAMA Network Open 4(1): e2034549. 10.1001/jamanetworkopen.2020.34549. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Dobis E. A., Krumel T. P. Jr., Cromartie J., Conley K. L., Sanders A., and Ortiz R.. 2021. Rural America at a Glance 2021 Edition. U.S. Department of Agriculture (Economic Research Service; ). https://www.ers.usda.gov/publications/pub-details?pubid=102575 [Google Scholar]
- 47. Azale, T. , Fekadu A., and Hanlon C.. 2016. “Treatment Gap and Help‐Seeking for Postpartum Depression in a Rural African Setting.” BMC Psychiatry 16(1): 196. 10.1186/s12888-016-0892-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.