Sage Journals: Discover world-class research

Abstract

Background:

Veterans using Department of Veterans Affairs (VA) healthcare have a high burden of pre-pregnancy chronic disease that likely contributes to the observed high rate of pregnancy-related morbidity. Many common diseases frequently co-occur; understanding patterns of multimorbidity may inform the design and delivery of pre-pregnancy interventions to lower pregnancy morbidity risk.

Objective:

The current study sought to identify patterns of co-occurrence of pre-pregnancy chronic disease among Veterans.

Design:

We conducted a retrospective cohort study using VA administrative data.

Methods:

Our population included Veterans ages 18–45 with ⩾1 pregnancy outcome (ectopic, spontaneous abortion, stillbirth, and/or live birth) during fiscal years 2010–2019. Presence of common chronic diseases with implications for pregnancy was detected using encounter International Classification of Diseases, 9th and 10th Revision (ICD-9 and ICD-10) codes in the 2 years prior to pregnancy. Patients were grouped based on latent class models of diagnosis patterns; two to seven latent groups were examined for model fit and clinical interpretability.

Results:

We identified 56,853 pregnancies from 41,034 Veterans. More than half of pregnancies were complicated by an array of pre-pregnancy medical and mental health conditions that may negatively impact pregnancy health and contribute to adverse pregnancy outcomes. The most frequently occurring conditions included chronic pain (51.2% of pregnancies), depression (31.4%), anxiety (25.9%), and post-traumatic stress disorder (22.8%). A five-group model demonstrated the best balance between model fit and clinical interpretability. Groups included: “Pain and Mental Health” (28%), with high prevalence of chronic pain, depression, and anxiety; “Pain and Metabolic” (17%), high prevalence of chronic pain, obesity, and migraines; “Substance Use and Mental Health” (7%), high prevalence of alcohol use disorder, depression, and post-traumatic stress disorder; “Low Diagnosis” (43%), lower than average prevalence of diagnoses; and “High Complexity” (5%), high prevalence of conditions across multiple physiologic systems.

Conclusions:

We identified five distinct, clinically meaningful groups of Veterans based on co-occurring pre-pregnancy diseases. Tailoring interventions to these groups may address Veterans’ complex pre-pregnancy health risks effectively and efficiently.

Keywords

Veteran pre-pregnancy health chronic disease latent class analysis

Introduction

One in three women of reproductive age in the United States have at least one chronic medical condition such as diabetes, cardiovascular disease, and asthma.¹ Pre-existing chronic conditions are significant contributors to increasing U.S. rates of adverse pregnancy outcomes, including severe maternal morbidity and mortality.^2,3 Notably, 20% of women of reproductive age in the United States have two or more chronic conditions⁴; however, clinical treatment guidelines and disease management programs tend to focus on a single condition.^5,6 Moreover, the impact of multimorbidity on the health of pregnant and birthing people is poorly understood, though early research suggests that people with multimorbidity are at increased risk for severe maternal morbidity and mortality compared to those with one or no chronic conditions.^3,7

Identifying patterns of multimorbidity during the pre-pregnancy period and understanding their impact on pregnancy outcomes is particularly important for women Veterans. Women Veterans have a high burden of chronic medical and mental health conditions (e.g., hypertension, chronic pain, post-traumatic stress disorder)^8–10 and health risk behaviors (e.g., tobacco use, alcohol use),^11,12 with frequent comorbidity.^13,14 Likely related to this high burden of chronic disease, research also suggests that women Veterans who use Department of Veterans Affairs (VA) healthcare are at increased obstetric risk than women Veterans using outside care and the general population^15–17 with high rates both of pregnancy complications^16,18 and of adverse pregnancy outcomes.^8,17,19,20

Pre-pregnancy counseling and care for people with underlying chronic conditions, particularly those with multimorbidity, is critical for improving maternal and infant outcomes.²¹ Interventions that target a single chronic condition may be less effective at improving outcomes than those targeting a group or cluster of co-occurring conditions; furthermore, people with multimorbidity will often need several separate interventions that could create excessive treatment burden.²² Understanding patterns of multimorbidity may inform the design and delivery of patient-centered pre-pregnancy disease management programs to address modifiable risk factors, optimize Veterans’ overall health during and around pregnancy, and reduce their risk of adverse pregnancy outcomes. The objective of the current study was to identify clinically meaningful groupings based on patterns of pre-pregnancy chronic medical and mental health conditions. Using the empiric clustering technique latent class analysis (LCA), we sought to determine: (1) the most parsimonious number of patient classes that could describe multimorbidity patterns (2) the composition of chronic condition prevalence in each class; and (3) the demographic characteristics of each class, with the goal of revealing clinically meaningful groupings for potential tailored interventions.

Methods

We conducted a retrospective cohort study using national VA administrative data on Veterans ages 18–45 years who had one or more recorded pregnancy outcome (ectopic, spontaneous abortion, stillbirth, and/or live birth) in VA and/or used VA pregnancy care benefits for at least one pregnancy during fiscal years (FY) 2010–2019. These data include inpatient and outpatient billing data for prenatal, labor and delivery, and early postpartum care services provided either by VA or by VA-contracted community-care providers. Detailed methodology for identifying a cohort of pregnancies in VA and constructing relevant variables has been reported elsewhere.²³ Briefly, pregnancies were identified based on International Classification of Diseases, 9th and 10th Revision (ICD-9 and ICD-10) diagnosis and procedure, current procedural technology, or diagnosis-related group codes for inpatient, outpatient, and community care services (i.e., visits to non-VA providers that are paid for by VA). Guided by previously published algorithms^24,25 that we adapted for VA data, we (1) identified all claims that mapped to any code suggesting a pregnancy outcome, (2) distinguished unique pregnancy events, (3) established gestational ages for each pregnancy outcome, and (4) estimated the start date of pregnancy (last menstrual period (LMP)) by subtracting gestational age from the date of discharge or service.^24,25 This study was approved and determined to be exempt under category 4 by the VA Pittsburgh Healthcare System Institutional Review Board (IRBNet #1628889). We received a waiver of informed consent to complete these secondary analyses of de-identified administrative data.

To capture Veterans’ pre-pregnancy health risks, we identified common chronic medical and mental health conditions that can be impacted by pregnancy or that have been associated with increased risk for adverse health events during and after pregnancy.^8,26–28 We also identified two common pre-pregnancy health risk behaviors, alcohol use and tobacco use, that women are advised to cease during pregnancy but which they may engage in before knowing they are pregnant and which have been associated with increased risk for adverse pregnancy outcomes.^29–31 The complete list of pre-pregnancy health risks included in this study are provided in Supplemental Table 1. The presence or absence of each chronic condition or health risk behavior was operationalized with a dichotomous indicator (yes/no) for having one or more relevant ICD-9 or ICD-10 diagnosis code recorded in outpatient or inpatient encounters in the 2 years prior to the start of pregnancy.

Demographic characteristics of the sample were also extracted from the electronic medical record. Veteran age at pregnancy is dichotomized as above or below age 35, reflecting the clinical definition of “advanced maternal age” for people 35 and older. Self-reported race/ethnicity is defined by four categories: non-Hispanic Black (hereafter: Black), non-Hispanic white (hereafter: white), Hispanic, and non-Hispanic other (including American Indian/Alaska Native, Asian, and Native Hawaiian/Other Pacific Islander). Geographic location is described by U.S. census regions: Northeast, Midwest, South, or West. We describe rurality as large metro, small metro, micropolitan, or non-core rural. We also included VA enrollment priority group, a proxy for some types of Veteran vulnerability based on military service-connected disability rating, income, military service history, and other factors. Using previously described methods³² we collapsed the eight groups into five that are conceptually similar: high service-connected disability rating (groups 1 and 4), low service-connected disability rating (groups 2 and 3), low income without copayment (group 5), special circumstance (group 6), and copay (groups 7 and 8).^32,33 Finally, we categorized level of neighborhood disadvantage of the Veteran’s residence (using their most recent residential address relative to LMP) by linking the latitude and longitude of the Veteran’s address to matching census block groups in the Wisconsin Neighborhood Atlas.³⁴ This area deprivation index (ADI) is reported in quintiles, with the least disadvantaged neighborhoods in quintile 1 (ADI 1–29) and the most disadvantaged in quintile 5 (ADI 77–100).

We used descriptive statistics to assess sample characteristics and prevalence of each chronic medical and mental health condition in our study population. Comorbidity groups were modeled using LCA.³⁵ LCA is a probabilistic clustering technique that groups individuals into discrete classes based on patterns in the co-occurrence of input variables. Clustering techniques, which can identify otherwise unobservable subgroups in complex patient populations, previously have been used by healthcare systems in efforts to more effectively direct care.^36–39 In this analysis, we grouped individual pregnancies with the 26 chronic conditions listed in Supplemental Table 1 as inputs. We chose a priori to compare LCA models of two to seven classes, selecting a final model based on a combination of model fit statistics and clinical interpretability.^35,40 We evaluated model fit using including Akaike information criterion, Bayesian information criterion (BIC), adjusted BIC, and entropy R^2,35,41 as well as the mean predicted probability of class membership, which characterizes class separation (higher predicted probability represents better separation). We then compared the clinical utility of models, prioritizing those in which classes were: (1) internally cohesive, (2) clinically distinct from one another, and (3) large enough to be feasible for clinical intervention. To describe the characteristics of pregnancies by latent class, each pregnancy was assigned to the latent class with the highest model-based predicted probability. We present the observed prevalence of each characteristic by latent class. All analyses were conducted using SAS version 9.4.⁴² The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement⁴³; see Supplemental Table 2.

Results

We identified 56,853 pregnancies from 41,034 Veterans who used pregnancy-related care that VA either provided or paid for during FY 10–19. A five-group model demonstrated the best balance between model fit and clinical interpretability; model fit statistics are presented in Table 1. We found the five-group model to have the best model fit statistics while retaining clinically distinct profiles. We labeled groups based on the highest prevalence conditions within each, relative to the overall average (see Figure 1). Identified groups included: a “Pain and Mental Health” group (28%), with high prevalence of chronic pain, depression, anxiety, and post-traumatic stress disorder (PTSD); a “Pain and Metabolic” group (17%), with high prevalence of chronic pain, obesity, and migraines; a “Substance Use and Mental Health” group (7%), with high prevalence of alcohol use disorder, tobacco use, depression, and PTSD; a “Low Diagnosis” group (43%), with lower than average prevalence of diagnoses (more than half of pregnancies in this group had no chronic conditions); and a “High Complexity” group (5%), with high prevalence of conditions across multiple physiologic systems (pregnancies in this group had an average of seven unique conditions).

Table 1.

Model fit statistics.

Classes	AIC	BIC	CAIC	ABIC	Entropy R²	Mean predicted probability^a
2	37,145	37,619	37,672	37,451	0.68	0.908
3	33,178	33,894	33,974	33,639	0.57	0.807
4	30,529	31,486	31,593	31,146	0.58	0.762
5	29,201	30,400	30,534	29,975	0.58	0.742
6	28,407	29,847	30,008	29,336	0.58	0.711
7	28,061	29,744	29,932	29,146	0.56	0.680

Model fit is from global maximum reached within 100 random starts per model. AIC: Akaike information criterion; BIC: Bayesian information criterion; CAIC: Consistent Akaike information criterion; ABIC: Adjusted Bayesian information criterion.

Mean predicted probability for each patient’s best matching class, per model.

Figure 1.

Overall and within-class prevalence (%) of conditions for five-class model, n = 56,853.

Characteristics of the sample, in total and by latent class, are presented in Table 2. Most pregnancies occurred in Veterans under age 35, with only 21% pregnancies (n = 11,840) in Veterans ages 35 or older. Over half (53.4%) of pregnancies were to Veterans who identified as white, and another quarter (24.7%) were among Black Veterans; nearly half were to Veterans who lived in the South (47.7%) and to Veterans who lived in more disadvantaged neighborhoods (ADI quintiles 4 and 5, 39.6%). The most frequently occurring pre-pregnancy risks included chronic pain (51.2% of pregnancies), depression (31.4%), anxiety (25.9%), PTSD (22.8%), and tobacco use (18.3%).

Table 2.

Pregnancy characteristics by latent class groups (n = 56,853).

Characteristic	Total, n (%)	Groups^a (%)
		1: Low Diagnosis	2: SUD, mental health	3: Pain, mental health	4: Pain, metabolic	5: High complexity
		24,468 (43%)	3731 (7%)	16,191 (28%)	9502 (17%)	2961 (5%)
Age at LMP, mean (SD)	30.4 (4.9)	29.5 (4.9)	30.5 (4.6)	30.6 (4.6)	31.8 (4.8)	32.4 (4.8)
Age above “advanced maternal age” (%)	11,840 (20.8)	16.8	20.4	20.1	29.0	32.8
Race/ethnicity (%)
Non-Hispanic Black	14,047 (24.7)	22.4	24.3	23.2	32.2	28.3
Non-Hispanic White	30,332 (53.4)	53.4	58.5	55.7	47.6	52.2
Hispanic	6662 (11.7)	12.8	9.0	11.6	10.6	10.3
Non-Hispanic other	3197 (5.6)	6.4	5.0	5.3	4.9	4.7
Missing	2615 (4.6)	5.0	3.2	4.2	4.8	4.6
Census region (%)
Midwest	10,590 (18.6)	18.9	19.1	18.7	17.8	18.0
Northeast	5120 (9.0)	8.6	10.0	9.3	9.1	9.3
South	27,127 (47.7)	46.2	48.0	47.3	51.1	51.0
West	13,758 (24.2)	25.4	23.0	24.5	21.8	21.7
Missing	258 (0.5)	0.8	0.0	0.2	0.2	0.1
Rurality (UIC), %
Large metro	28,143 (49.5)	51.5	48.5	46.9	50.9	44.1
Small metro	21,049 (37.0)	35.2	37.4	39.2	36.7	40.5
Micropolitan	4662 (8.2)	7.7	8.7	8.6	7.9	10.0
Non-core rural	2702 (4.8)	4.6	5.2	5.1	4.2	5.2
Missing	297 (0.5)	0.9	0.2	0.2	0.3	0.2
ADI national rank, mean (SD)^b	52.3 (24.3)	51.1 (24.6)	55.5 (24.2)	52.3 (23.9)	53.1 (24.2)	55.8 (24.0)
ADI rank quintile (%)
1 (ranks 1–29, least distressed)	11,220 (19.7)	21.7	16.4	19.0	18.4	16.1
2 (30–44)	11,290 (19.9)	19.5	18.3	20.8	20.3	18.0
3 (45–58)	10,724 (18.9)	18.4	19.3	19.4	18.9	19.0
4 (59–76)	11,621 (20.4)	20.0	20.9	20.7	20.5	21.9
5 (77–100, most distressed)	10,912 (19.2)	18.1	23.3	18.7	20.0	23.6
Missing	1086 (1.9)	2.3	1.9	1.4	1.8	1.4
VA priority group, %
Disability level 1 (priority 1, 4)	21,792 (38.3)	18.1	56.7	56.7	40.0	76.5
Disability level 2 (priority 2, 3)	15,392 (27.1)	26.8	20.4	25.0	37.5	15.7
Low income (priority 5)	8915 (15.7)	20.5	16.3	11.2	13.7	6.1
Special circumstance (priority 6)	3189 (5.6)	8.9	3.4	3.5	3.2	0.7
Co-pay (priority 7, 8)	3020 (5.3)	7.8	2.6	3.0	5.2	1.0
Missing	4545 (8.0)	17.9	0.6	0.6	0.5	0.1
Number of chronic conditions, mean (SD)	2.4 (2.1)	0.5 (0.6)	4.7 (1.5)	3.4 (1.1)	2.8 (1.1)	7.0 (1.4)
Number of chronic conditions (%)
0	12,977 (22.8)	53.0	0	0	0	0
1	10,437 (18.4)	41.7	0	0	2.6	0
2	9621 (16.9)	5.3	5.3	24.1	44.4	0
3	8389 (14.8)	0	15.2	31.2	29.2	0
4	6542 (11.5)	0	26.2	25.3	15.4	0.1
5 or more	8887 (15.6)	0	53.3	19.4	8.4	99.9
Pregnancy outcome (%)
Live birth	43,298 (76.2)	80.4	73.7	73.7	73.1	67.9
Any stillbirth	411 (0.72)	0.5	0.7	0.8	1.0	1.1
SAB/ECT	13,144 (23.1)	19.1	25.7	25.6	26.0	31.0

SUD: substance use disorder; LMP: last menstrual period; UIC: urban influence code; ADI: area deprivation index; LB: live birth; SB: stillbirth; SAB: spontaneous abortion; ECT: ectopic pregnancy; VA: Veterans Health Administration; SD: standard deviation.

Group names were chosen based on the most prevalent conditions, but patients in these groups often have other comorbidities as well (see Figure 1). All p < 0.001, unadjusted.

Missing: n = 1086 (1.9%).

The high complexity group had the highest percentage of pregnancies (32.8%) occurring to Veterans 35 or older (see Table 1), while the low diagnosis group had the lowest (16.8%). The pain and metabolic group had the highest percent of pregnancies occurring among Veterans identified as non-Hispanic Black (32.2%) and those residing in the South (51.1%). The high complexity and substance use and mental health groups included the highest percentage of pregnancies among rural Veterans (5.2% in both groups) and among Veterans living in the most economically distressed areas (23.6% and 23.3%, respectively). Notably, pregnancies in the high complexity group had the highest prevalence of non-live birth outcomes, including stillbirth, spontaneous abortion, and ectopic pregnancy (32%), whereas the low diagnosis group had the lowest prevalence of non-live birth outcomes (19.1%).

Discussion

In this cohort of Veteran pregnancies for which VA either provided or paid for pregnancy-related care during FY 2010–2019, we identified five distinct latent groups based on co-occurrence patterns of pre-pregnancy chronic medical and mental health conditions. Nearly half of pregnancies in this sample had low overall rates of chronic disease; the remaining pregnancies, however, were complicated by an array of pre-pregnancy medical and mental health conditions that may negatively impact pregnancy health and contribute to adverse pregnancy outcomes.

A notable finding from this study was the significant contribution of pain diagnoses to each of the five latent groups we identified. In the general population, women are disproportionately affected by chronic pain disorders compared to men, and report greater pain severity and pain-related disability than males with the same pain modalities.⁴⁴ Within VA, provider-facing clinical recommendations suggest that women Veterans have higher rates of risk factors for chronic pain, including injuries sustained during military service, higher rates of depression and anxiety, and higher rates of sexual trauma, and that chronic pain is more common among women Veterans than among male Veterans.⁴⁵

The negative effects of chronic pain are widespread; pain can impact physical functioning, cognitive acuity, emotional wellbeing, interpersonal relationships, and overall quality of life.^46–49 Pain can also be accompanied by feelings of isolation, fear of dismissal by healthcare providers, and stigma associated with common treatments for pain disorders (e.g., opioids).^50–53 When pain co-occurs with other medical or mental health conditions, these consequences may be exacerbated. Pregnancy may further complicate pain diagnosis and treatment: for example, pregnancy may worsen certain types of pain (e.g., musculoskeletal) yet interfere with typical physical therapy recommendations, or pregnancy may offer a temporary respite from pain caused by another condition (e.g., rheumatoid arthritis) but require modification to typical medication regimens due to teratogenicity.⁵⁴ Our finding that chronic pain was the most prevalent pre-pregnancy comorbidity highlights the need for tailored pre-pregnancy counseling on the reciprocal effects of pain on pregnancy and pregnancy on pain as well as innovative treatment and intervention strategies during pregnancy that account for the intersection of pain with other common conditions.

VA’s existing primary care model, which relies on Patient-Aligned Care Teams and prioritizes Veterans’ whole health through care coordination, presents an ideal environment for developing and testing targeted pre-pregnancy care that addresses the complex needs of pregnant and birthing Veterans. Using LCA to identify multimorbidity clusters among pregnant Veterans offers a person-centered approach to identifying subgroups with shared characteristics who might benefit from common resources and interventions. An innovative contribution of this work is the potential to direct Veterans to healthcare services that address the intersection of multiple comorbid conditions rather than a more typical siloed approach. For example, a Veteran with comorbid substance use and mental health conditions might benefit from targeted and integrated behavioral health and substance use support during the perinatal period to optimize their health during and around pregnancy. Similarly, a treatment plan prioritizing increased physical activity and improving nutrition, which may be appropriate for a Veteran seeking to manage pre-pregnancy diabetes, may not be feasible or acceptable for a Veteran who is also managing chronic pain and who may have limited mobility. Understanding how both symptoms of and treatment for comorbid chronic conditions may be compounded is a critical first step toward developing new health promotion and chronic disease prevention strategies to provide tailored healthcare options to meet Veterans’ needs. Additional research to describe associations between identified patterns of multimorbidity and adverse maternal and pregnancy outcomes in this population, including severe maternal morbidity and early pregnancy loss, is ongoing and will provide further insight into targeting efforts to combinations of conditions that are most likely to impact pregnancy.

Strengths and limitations

This study is the first to explore potential clusters of disease among Veteran pregnancies and provides actionable information on patterns of comorbidity to improve Veteran-focused support and interventions in the pre-pregnancy and interpregnancy periods. A major strength of this work is our use of national VA administrative data over a multi-year timeframe, which enabled us to investigate diagnoses during a 2-year pre-pregnancy lookback period. This study also has limitations. Because our analyses were not hypothesis-driven, sample size was not identified a priori; all pregnancies meeting stated inclusion criteria were retained in the final sample. A methodological trade-off in this study is comparing the observed characteristics of pregnancies assigned to a latent class. An alternative method would be to incorporate auxiliary variables in the LCA model to estimate the odds of class membership by each characteristic.³⁵ This method would also incorporate the underlying uncertainty in the LCA model. However, due to concerns in the literature with model misspecification due to covariates, we chose to present observed prevalence.^55,56 In addition, because Veterans may use non-VA insurance as well as their VA coverage, it is possible that some chronic disease diagnoses are not reflected in our data. Furthermore, these data capture the presence of relevant chronic condition diagnoses in a Veteran’s health record but do not reflect the relative severity of any condition or whether a condition is currently being effectively managed. Finally, prior research suggests that Veterans who rely on VA healthcare may be higher risk than Veterans who use non-VA care¹⁹; findings based on these data may not be generalizable to Veterans not accessing VA or to a non-Veteran population.

Conclusion

In a national population of Veteran pregnancies during FY 2019–2019, LCA identified five groups based on patterns of comorbid pre-pregnancy chronic conditions that could be used to inform tailored interventions to improve pregnancy health and outcomes.

Supplemental Material

sj-docx-1-whe-10.1177_17455057261430209 – Supplemental material for Classifying Veterans’ pre-pregnancy health risks using latent class analysis

Supplemental material, sj-docx-1-whe-10.1177_17455057261430209 for Classifying Veterans’ pre-pregnancy health risks using latent class analysis by Deirdre A. Quinn, Franya Hutchins, Florentina E. Sileanu, Gregory T. Procario, Maria K. Mor, Ann-Marie Rosland, Jodie G. Katon, Lisa S. Callegari and Sonya Borrero in Women's Health

Footnotes

Acknowledgements

None.

Ethical considerations and consent to participate

This study was approved and determined to be exempt under category 4 by the VA Pittsburgh Healthcare System Institutional Review Board (IRBNet #1628889).

Consent to participate

We received a waiver of informed consent to complete these secondary analyses of de-identified administrative data.

Consent for publication

Not applicable.

Author contributions

Deirdre A. Quinn: Conceptualization; Investigation; Funding acquisition; Writing – original draft; Methodology; Writing – review & editing; Supervision; Project administration.

Franya Hutchins: Methodology; Writing – original draft; Writing – review & editing; Formal analysis.

Florentina E. Sileanu: Methodology; Writing – review & editing; Data curation; Project administration.

Gregory T. Procario: Writing – review & editing; Methodology; Data curation.

Maria K. Mor: Conceptualization; Methodology; Writing – review & editing; Supervision.

Ann-Marie Rosland: Conceptualization; Writing – review & editing; Methodology.

Jodie G. Katon: Conceptualization; Writing – review & editing.

Lisa S. Callegari: Conceptualization; Writing – review & editing.

Sonya Borrero: Conceptualization; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Dr. Quinn is supported by a VA Health Services Research and Development Career Development Award (20-224).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

Data from this study are not available to outside researchers given VA data privacy rules and IRB restrictions.

Supplemental material

Supplemental material for this article is available online.

Disclaimer

The views expressed herein are those of the authors and do not reflect those of the Department of Veterans Affairs or the U.S. Government.

References

Gawron

Sanders

Sward

, et al. Multi-morbidity and highly effective contraception in reproductive-age women in the US intermountain west: a retrospective cohort study. J Gen Intern Med 2020; 35(3): 637–642.

Trost

Beauregard

Chandra

, et al. Pregnancy-related deaths: data from Maternal Mortality Review Committees in 36 US States, 2017–2019. Centers for Disease Control and Prevention, US Department of Health and Human Services, 2022.

Brown

McKnight

Aker

Association between pre-pregnancy multimorbidity and adverse maternal outcomes: a systematic review. J Multimorb Comorbidity 2022; 12: 26335565221096584.

Gunja

Seervai

Zephyrin

, et al. Health and health care for women of reproductive age: how the United States compares with other high-income countries [cited 17 May 2025], https://www.commonwealthfund.org/publications/issue-briefs/2022/mar/health-and-health-care-women-reproductive-age (2022).

De La Rosa

Brady

Ibrahim

, et al. Co-occurrence of chronic pain and anxiety/depression symptoms in U.S. adults: prevalence, functional impacts, and opportunities. Pain 2024; 165(3): 666–673.

Fuhlbrigge

Lockey

RF.

Asthma and comorbid conditions. J Allergy Clin Immunol Pract 2021; 9(11): 3909–3910.

Beeson

Homer

CSE

Morgan

, et al. Multiple morbidities in pregnancy: time for research, innovation, and action. PLoS Med 2018; 15(9): e1002665.

Shaw

Asch

Kimerling

, et al. Posttraumatic stress disorder and risk of spontaneous preterm birth. Obstet Gynecol 2014; 124(6): 1111–1119.

Sheahan

Goldstein

Than

, et al. Women Veterans’ healthcare needs, utilization, and preferences in Veterans Affairs primary care settings. J Gen Intern Med 2022; 37(Suppl 3): 791–798.

10.

Breland

Frayne

Saecho

, et al. Sourcebook: women Veterans in the Veterans Health Administration. Volume 5: longitudinal trends in sociodemographics and utilization, including type, modality, and source of care. Veterans Health Administration, 2024.

11.

Wilson

Medenblik

Neal

, et al. Lifetime smoking patterns and preferences for smoking cessation among women Veterans receiving Veterans Health Administration Care. Qual Health Res 2019; 29(14): 2096–2107.

12.

Hoggatt

Lehavot

Krenek

, et al. Prevalence of substance misuse among US veterans in the general population. Am J Addict 2017; 26(4): 357–365.

13.

Lehavot

Hoerster

Nelson

, et al. Health indicators for military, veteran, and civilian women. Am J Prev Med 2012; 42(5): 473–480.

14.

Creech

Pulverman

Crawford

, et al. Clinical complexity in women Veterans: a systematic review of the recent evidence on mental health and physical health comorbidities. Behav Med 2021; 47(1): 69–87.

15.

Katon

Washington

Cordasco

, et al. Prenatal care for women Veterans who use Department of Veterans Affairs Health Care. Womens Health Issues 2015; 25(4): 377–381.

16.

Shaw

Joyce

Schmitt

, et al. Selection of higher risk pregnancies into Veterans Health Administration Programs: discoveries from linked Department of Veterans Affairs and California Birth Data. Health Serv Res 2018; 53(Suppl 3): 5260–5284.

17.

Quinn

Sileanu

Mor

, et al. Describing adverse pregnancy events and pregnancy-associated death among Veterans. J Womens Health (Larchmt) 2025; 34(2): 166–175.

18.

Katon

Mattocks

Zephyrin

, et al. Gestational diabetes and hypertensive disorders of pregnancy among women Veterans deployed in service of operations in Afghanistan and Iraq. J Womens Health (Larchmt) 2014; 23(10): 792–800.

19.

Hugin

Shaw

JG.

Obstetric outcomes in US. Veterans: emerging knowledge, considerations, and gaps. Semin Reprod Med 2019; 37(1): 17–23.

20.

Katon

Cypel

Raza

, et al. Deployment and adverse pregnancy outcomes: primary findings and methodological considerations. Matern Child Health J 2017; 21(2): 376–386.

21.

Lassi

Imam

Dean

, et al. Preconception care: Screening and management of chronic disease and promoting psychological health. Reprod Health 2014; 11(Suppl): S5.

22.

Salisbury

Man

Bower

, et al. Management of multimorbidity using a patient-centred care model: a pragmatic cluster-randomised trial of the 3D approach. Lancet 2018; 392(10141): 41–50.

23.

Quinn

Mor

Sileanu

, et al. Measuring female Veterans’ prepregnancy wellness using Department of Veterans Affairs’ Health Record Data. Obstet Gynecol 2021; 37(3): 471–480.

24.

Devine

West

Andrews

, et al. The identification of pregnancies within the general practice research database. Pharmacoepidemiol Drug Saf 2010; 19: 45–50.

25.

Ailes

Simeone

Dawson

, et al. Using insurance claims data to identify and estimate critical periods in pregnancy: an application to antidepressants. Birth Defects Res A Clin Mol Teratol 2016; 106(11): 927–934.

26.

Admon

Winkelman

Moniz

, et al. Chronic disease prevalence among mothers delivering in the United States, 2004–2013. Obstet Gynecol 2017; 129(5): S152–S153.

27.

Admon

Winkelman

TNA

Heisler

, et al. Obstetric outcomes and delivery-related health care utilization and costs among pregnant women with multiple chronic conditions. Prev Chronic Dis 2018; 15(4): 170397.

28.

Rocca

Boyd

Grossardt

, et al. Prevalence of multimorbidity in a geographically defined American population. Mayo Clin Proc 2014; 89(10): 1336–1349.

29.

Ethen

Ramadhani

Scheuerle

, et al. Alcohol consumption by women before and during pregnancy. Matern Child Health J 2009; 13(2): 274–285.

30.

Asp

Bergman

Lager

, et al. Alcohol exposure prior to pregnancy—does hazardous consumption affect placenta- and inflammatory-mediated pregnancy outcomes? A Swedish population-based cohort study. Acta Obstet Gynecol Scand 2022; 101(12): 1386–1394.

31.

Liu

Sun

, et al. Maternal cigarette smoking before and during pregnancy and the risk of preterm birth: a dose–response analysis of 25 million mother–infant pairs. PLoS Med 2020; 17(8): e1003158.

32.

Silva

Tarlov

French

, et al. Veterans Affairs Health System enrollment and health care utilization after the affordable care act: initial insights. Mil Med 2016; 181(5): 469–475.

33.

U.S. Department of Veterans Affairs. Enrollment Priority Groups. U.S. Department of Veterans Affairs [cited 1 March 2024], https://www.va.gov/health-care/eligibility/priority-groups/ (2024).

34.

Kind

AJH

Buckingham

. Making neighborhood-disadvantage metrics accessible—the neighborhood atlas. N Engl J Med 2018; 378(26): 2456–248.

35.

Collins

Lanza

ST.

Latent class and latent transition analysis: with applications in the social, behavioral, and health sciences. John Wiley & Sons, Inc., 2010.

36.

Yan

Kwan

Tan

, et al. A systematic review of the clinical application of data-driven population segmentation analysis. BMC Med Res Methodol 2018; 18(1): 121.

37.

Vuik

Mayer

Darzi

Patient segmentation analysis offers significant benefits for integrated care and support. Health Aff (Millwood) 2016; 35(5): 769–775.

38.

Tamer

Karam

Roderick

, et al. Know thy patient: a novel approach and method for patient segmentation and clustering using machine learning to develop holistic, patient-centered programs and treatment plans. NEJM Catal Innov Care Deliv. Epub ahead of print 23 August 2022. DOI: 10.1056/CAT.22.0084.

39.

Arnold

Thorpe

Mains-Mason

, et al. Empiric segmentation of high-risk patients: a structured literature review. Am J Manag Care 2021; 28(2): e69–e77.

40.

Sinha

Calfee

Delucchi

KL.

Practitioner’s guide to latent class analysis: methodological considerations and common pitfalls. Crit Care Med 2021; 49(1): e63–e79.

41.

Weller

Bowen

Faubert

SJ.

Latent class analysis: a guide to best practice. J Black Psychol 2020; 46(4): 287–311.

42.

SAS Institute Inc. SAS software. SAS Institute, Inc., 2013.

43.

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies. Ann Int Med 2007; 147(8): 573–577.

44.

Fillingim

King

Ribeiro-Dasilva

, et al. Sex, gender, and pain: a review of recent clinical and experimental findings. J Pain 2009; 10(5): 447–485.

45.

Department of Veterans Affairs. Pain management, Opioid safety, and PDMP (PMOP), https://www.va.gov/PAINMANAGEMENT/Providers/PainWomenVeterans.asp (2025, accessed 24 September 2025).

46.

Ray-Griffith

Morrison

Stowe

ZN.

Chronic pain prevalence and exposures during pregnancy. Pain Res Manag 2019; 2019: 1–7.

47.

Yang

Grol-Prokopczyk

. Chronic pain and friendship among middle-aged and older U.S. adults. J Gerontol Ser B 2021; 76(10): 2131–2142.

48.

Birkinshaw

Friedrich

Cole

, et al. Antidepressants for pain management in adults with chronic pain: a network meta-analysis. Cochrane Database Syst Rev 2023; 2023(5): CD014682.

49.

Khera

Rangasamy

Cognition and pain: a review. Front Psychol 2021; 12: 673962.

50.

Newton

Southall

Raphael

, et al. A narrative review of the impact of disbelief in chronic pain. Pain Manag Nurs 2013; 14(3): 161–171.

51.

Bean

Dryland

Rashid

, et al. The determinants and effects of chronic pain stigma: a mixed methods study and the development of a model. J Pain 2022; 23(10): 1749–1764.

52.

Bannon

Greenberg

Mace

, et al. The role of social isolation in physical and emotional outcomes among patients with chronic pain. Gen Hosp Psychiatry 2021; 69: 50–54.

53.

Ashton-James

Anderson

Mackey

, et al. Beyond pain, distress, and disability: the importance of social outcomes in pain management research and practice. Pain 2022; 163(3): e426–e431.

54.

Ray-Griffith

Wendel

Stowe

, et al. Chronic pain during pregnancy: a review of the literature. Int J Womens Health 2018; 10: 153–164.

55.

Nylund-Gibson

Masyn

KE.

Covariates and mixture modeling: results of a simulation study exploring the impact of misspecified effects on class enumeration. Struct Equ Model Multidiscip J 2016; 23(6): 782–797.

56.

Wurpts

Geiser

Is adding more indicators to a latent class analysis beneficial or detrimental? Results of a Monte-Carlo study. Front Psychol 2014; 5: 920.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB