Predictive Factors for Loss to Postpartum Follow-Up Among Low Income HIV-Infected Women in Texas

Abstract

Guidelines for HIV primary care include visits every 3 months (up to 6 months in those with stable HIV). During pregnancy, women with HIV commonly attend once weekly to once monthly visits; however, after delivery, many are lost to follow-up. Our goal was to assess the frequency of loss to primary care follow-up postpartum and to identify predictors of loss to care. A retrospective chart review of HIV-infected women in a Houston prenatal program was done. Optimal care was defined as one visit to HIV primary care providers (PCPs) every 6 months within the first year after delivery, and loss to follow-up as no visits within the first postpartum year. Multivariate logistic regression analysis was used to identify factors associated with loss to follow-up. Charts (n=213) were analyzed for follow-up with PCPs. The loss to follow-up rate was 39% in the first postpartum year. Associated factors were younger age, black race, late entry to prenatal care, and no plans for contraception. Predictors of loss to primary care after pregnancy can be used to identify specific subpopulations of pregnant women at highest risk for falling out of care.

Introduction

Given that women aged 25–44 years account for the majority of new HIV cases, many are in their reproductive years and end up seeking medical care only during pregnancy.¹ In a study that reviewed records from 15 U.S. health departments, 64.3% of HIV-positive women attended at least one prenatal visit prior to delivery. As expected, these patients had better outcomes than those with no prenatal care. When compared to the 26.8% who did not receive prenatal care, the vertical transmission rate to the neonate in the prenatal care group was 1.6% versus 8.5% in patients with no prenatal care.² Few studies in the literature have looked specifically at postpartum follow-up rates in HIV-infected women in the United States. In a U.S. study conducted in Jackson, Mississippi, postpartum follow-up rates with HIV providers were as low as 37%; loss to follow-up was associated with late entry into prenatal care.³ However, among the populations studied, 48% in the Mississippi study resided in a rural area and 89% were black, making this population different from the Harris County population (the catchment area for this study), which is primarily urban and has a larger Hispanic component A retrospective review of 25 women in Switzerland between 2004 and 2012 suggested a transient decrease in adherence to medication at delivery (from 97% to 93%) with resumption of adherence during the first 6 months postpartum; however, the subjects were part of a program focused on adherence enhancement.⁴

The purpose of this study was to identify predictors of loss to care. HIV-infected pregnant women, when compared to nonpregnant women, are younger in age and have less advanced HIV with higher CD4 counts.⁵ This population is the ideal group in which to establish long-term care, as these patients are likely to have the maximal survival benefits. Maximal survival benefits are expected in patients that present to care prior to developing WHO stage 3 or 4 HIV infection and are able to follow up with an HIV provider indefinitely. From previous studies, it appears that women more likely to be lost to follow-up were younger in age, of black race, and had a history of late entry to prenatal care.⁶ Current HIV guidelines recommend asymptomatic, clinically stable patients be monitored every 3–6 months. Numerous studies have shown that morbidity and mortality increase in persons that do not follow up routinely.⁷ Also, many measures of retention in care have been defined.⁸ For the purpose of this study, we defined three groups—similar to those used by Rana et al.³ in the one other study of factors associated with loss to postpartum care in the United States: optimal follow-up (minimum of one clinical visit to PCP every 6 months within the first year after delivery), suboptimal follow-up (one visit to PCP within 1 year after delivery), and lost to care (no PCP visits postpartum).

Materials and Methods

HIV-infected women in the Harris Health System (HHS) clinics in Houston, Texas receive prenatal care at the Northwest Health Center (NWHC) and the Lyndon B. Johnson obstetrical clinics. HHS provides prenatal care for 50–70 HIV-infected women annually. At the first prenatal visit, a client meets with a clinical nurse and case manager to facilitate patient education and ensure that she has appropriate transportation and funding for medications during pregnancy. After 6 weeks postpartum, women must transfer to a primary care provider, which is most often an infectious disease specialist. Those who had an HIV primary care provider (PCP) prior to pregnancy are scheduled to return to their original PCP; those newly diagnosed during the pregnancy are linked to a PCP whom they have not met before.

After obtaining approval from the Baylor College of Medicine Institutional Review Board and the HHS Office of Research, a retrospective chart review was performed with the extraction of data from HHS electronic medical records, which contained outpatient and inpatient progress notes, labs, and delivery records. We included all HIV- infected women from the HHS clinics who received prenatal care from mid-2006 to 2011 and did not transfer out of the county system prior to delivery. A total of 256 prenatal charts were reviewed; 16 were excluded based on transfer during prior to delivery or up to 1 year postpartum. The remaining 240 charts represented 213 women, as some women had repeat pregnancies. Repeat pregnancies were excluded for the analysis. A total of 213 charts were analyzed for HIV PCP follow-up. Assigning each chart a number ensured anonymity.

The HIV viral load assay had a detection limit in 2006–2009 of <400; from 2010 to 2011 the detection limit was <48. Data on demographics, socioeconomic factors, obstetrical information, and HIV parameters were extracted from the electronic charts into a previously formulated data collection form. Variables in the data collection form were: Age, gravity and parity, race, estimated date of delivery (EDD), gestational age at initial OB visit, mode of delivery, preterm delivery, desire for contraception as elicited at prenatal visits, HIV status of partner, disclosure of status to partner, history of substance abuse, primary language, insurance, psychiatric diagnoses, age at HIV diagnosis, last CD4 count and viral load prior to pregnancy, CD4 and viral load at initial OB visit, CD4 count and viral load at delivery, antiretroviral therapy prior to pregnancy, postpartum depression and Edinburgh score when available. The HIV status of partner, disclosure to partner, substance abuse history, and evidence of postpartum depression were self-reported by patients.

Data were collected with the following end points: up to 6 months postpartum with HIV PCP, and 7–12 months postpartum with HIV PCP. For the purposes of this analysis, suboptimal follow-up was defined as at least one clinical visit with an HIV PCP in the 12-month postpartum period, while optimal follow-up was defined as at least one clinical visit in each 6 month period during 1 year postpartum.

The study groups were compared via Fisher exact test for grouped data, Kruskal-Wallis nonparametric test for continuous data. Logistic regression was used for multivariate analysis of factors associated with loss to follow-up. All variables with p less than or equal to 0.10 were included in the analysis. p Values of less than 0.05 were considered statistically significant. All analysis was done in SAS.

Results

A total of 213 charts were analyzed for HIV PCP follow-up. The median age group of subjects was 20–29, racial distribution was: 62% black, 34% Hispanic, and 4% white. Of the previously diagnosed HIV-positive women, 31% were on ART at initial OB visit. Twenty-nine percent of patients self-reported a substance abuse history, 19% had a psychiatric diagnosis, 66% had Medicaid, and 32% had state-funded Children's Health Insurance Program (CHIP). Twenty-seven percent of partners were HIV-positive, 38% HIV-negative, and 36% unknown. Seventy-seven percent of women had disclosed status to their partner. Characteristics of the study group are reported in Table 1.

Table 1.

Characteristics of Patients

Variable	N (%)
Black	131/213 (62)
Hispanic	73/213 (34)
White	9/213 (4)
Age
<20	18/213 (8)
20–29	119/213 (56)
>30	76/213 (36)
Parity
Primiparity (parity≤1)	66/213 (31)
Multiparity (parity>1)	147/213 (69)
Preterm delivery	35/202 (17)
Mode of delivery
Vaginal	141/210 (67)
Cesarean	69/210 (33)
Complications in pregnancy^a	87/212 (41)
Desired contraception
Bilateral tubal ligation	53/213 (25)
Implanon	7/213 (3)
Intrauterine device	25/213 (12)
Depot medroxyprogesterone acetate	46/213 (22)
Oral contraceptives	14/213 (7)
Condoms	60/213 (28)
None	8/213 (4)
History of substance abuse	61/211 (29)
HIV status of partner
Positive	57/213 (27)
Negative	80/213 (38)
Unknown	76/213 (36)
Disclosure to partner	164/213 (77)
Psychiatric diagnoses	40/213 (19)
Insurance
Medicaid	140/212 (66)
CHIP	67/212 (32)
Other	5/212 (2)
Language
English	150/213 (70)
Spanish	58/213 (27)
Other	5/213 (2)
New diagnosis in pregnancy	68/213 (32)
ART prior to pregnancy	66/213 (31)
Evidence of postpartum depression	26/155 (17)

Complications included but were not limited to hypertension, diabetes, anemia, short cervix.

As shown in Table 2, the variables which were significantly associated with no follow up were younger age, higher gestational age at first OB visit, younger age at HIV diagnosis, black race, lower initial viral load, no complications during pregnancy, no plans for contraceptive use, and not taking ARV therapy prior to pregnancy. These variables that showed significance were then included in the multivariate analysis (Table 3).

Table 2.

Factors Associated with Follow-Up with HIV Provider

Variable	No follow-up N=83	Suboptimal N=54	Optimal N=76	p Value
Age (years)
Mean (SD)	26.1 (5.5)	27.2 (6.8)	28.6 (5.7)	0.04
Gestational age at first OB visit (weeks)
Mean (SD)	19.4 (8.7)	16.6 (8.7)	14.6 (7.0)	0.002
Gestational age at first HIV OB visit (weeks)^a
Mean (SD)	20.6 (9.0)	17.2 (9.2)	15.9 (7.9)	0.003
Gestational age at delivery (weeks)
Mean (SD)	37.0 (5.5) N=81	36.4 (6.5)	36.3 (6.4)	0.71
Age at HIV dx (years)
Mean (SD)	22.0 (7.1)	23.0 (7.6)	24.7 (7.0) N=75	0.08
Last prior to pregnancy CD4
Mean (SD)	853 (1300) N=35	702 (1221) N=25	428 (210) N=42	0.11
Last prior to pregnancy VL, N	34	25	40
<400, %	35	52	45
[401–1000], %	6	8	3
>1000, %	59	40	53	0.53
CD4 count at initial OB visit
Mean (SD)	566 (789)	497 (567)	551 (735)	0.66
VL at initial OB visit, N	80	53	76
≤400, %	25	21	34
[400–1000], %	6	4	5
>1000, %	69	75	61	0.45
CD4 delivery
Mean (SD)	618 (561)N=77	631 (857)	552 (466)	0.55
VL at delivery, N	78	53	76
≤400, %	77	77	79
[400–1000], %	9	8	7
>1000, %	14	15	14	0.99
Gravidity
Mean (SD)	2.9 (1.6)	3.0 (1.9)	3.2 (1.8)	0.43
Parity
Mean (SD)	2.3 (1.4)	2.2 (1.2)	2.5 (1.4)	0.70
Race, N	83	54	76
Hispanic, %	25	31	46
White, %	4	7	3
Black, %	71	61	51	0.05
Preterm, N	79	51	72
No, %	86	78	82
Yes, %	14	22	18	0.51
Gestational age (weeks), N	81	54	76
<20, %	2	6	5
[20–25), %	0	0	0
[25–28), %	0	0	1
[28–34), %	7	2	5
[34–37), %	6	19	11
≥37, %	84	74	78	0.26
Mode of delivery, N	81	53	76
Vaginal, %	70	62	67
Cesarean, %	30	38	33	0.61
Complications during pregnancy, N	83	53	76
No, %	67	58	50
Yes, %	33	42	50	0.09
New diagnosis in pregnancy, N	83	54	76
No, %	69	70	66
Yes, %	31	30	34	0.85
Desired contraception, N	83	54	76
DMPA, %	18	13	32
All other, %	75	83	68
None, %	7	4	0	0.01
HIV status of partner, N	83	54	76
Positive, %	27	30	25
Negative, %	34	33	45
Unknown, %	40	37	30	0.57
Disclosure of status to partner, N	83	54	76
Yes, %	71	80	82
No, %	23	17	9
Unknown, %	6	4	9	0.15
History of substance abuse, N	83	53	75
No, %	64	72	79
Yes, %	36	28	21	0.12
Primary language, N	83	54	76
English, %	78	70	62
Spanish, %	19	24	38
Other, %	2	6	0	0.02
Insurance, N	83	53	76
Medicaid, %	71	66	61
CHIP, %	27	30	38
Other, %	2	4	1	0.51
Psychiatric disorder, N	83	54	76
No, %	77	81	86
Yes, %	23	19	14	0.40
ART prior to pregnancy, N	83	54	76
No, %	76	67	63
Yes, %	24	33	37	0.19

Women diagnosed with HIV for the first time during their pregnancy had at least one visit to another obstetrical provider before being referred for HIV OB care.

Table 3.

Odds Ratios for Loss to Follow-Up with HIV Provider

	Odds ratio estimates
Effect	Point estimate	95% Wald confidence limits
Age	0.943	0.896	0.992
GA at first OB visit	1.069	1.031	1.109
Black	2.155	1.151	4.035
No contraceptives	7.889	1.371	45.398

For the logistic regression analysis, the patients were divided into two groups for comparison (Table 3). The suboptimal and optimal follow-up cohorts were grouped together and compared with the no show cohort. The odds ratio for the no show group was higher among younger women, those with late entry to HIV OB prenatal care, black race, and with no plans for contraceptive use.

Discussion

Pregnant women with HIV are often more diligent about attending prenatal visits than routine primary care visits when they are not pregnant. During pregnancy, HIV-infected women tend to be highly motivated to take ARVs in order to prevent HIV transmission to their infants. Given the adherence to prenatal care visits seen with most patients, pregnancy is an ideal time to engage women in care of their HIV.

Several retrospective studies have looked at postpartum follow-up rates in the U.S. In the general population, many women will present to care for pregnancy and delivery; however, the loss to follow-up rate after delivery tends to be as high as 23–54%. A study from the University of California San Francisco followed postpartum follow-up rates in patients with gestational diabetes. The population in this study consisted of an ethnically and socioeconomically diverse group. The rate of postpartum follow-up to primary care providers for diabetes screening was 33.4%.⁹

Because of the high rates of HIV/AIDS in Sub-Saharan Africa, many of the HIV follow-up studies are outside of the United States. One study from University of Capetown looked at the loss to follow-up rates in HIV-positive postpartum women compared to HIV-positive nonpregnant women. The 3-year rate of loss to follow-up in postpartum women was 32% compared to 13% in the nonpregnant women. In this study, loss to follow-up was defined as women on antiretroviral therapies (ART) who had not attended the clinic for 3 months.⁵ Kreitchmann et al. describe a postpartum decrease in adherence to antiretroviral therapy also occurring in Latin America (Brazil, Peru, and Argentina); in their population the authors identified predictors of low adherence: smoking, alcohol use, and older age.¹⁰

With the development of antiretroviral therapies, there have been tremendous improvements in survival for HIV-infected patients; HIV has become a manageable chronic disease. For maximal survival benefits, patients need to enter care before advanced stage immunosuppression and remain in care indefinitely. However, national data have shown as many as half of patients who know they are infected do not receive routine care. The outcomes of poor adherence to therapy and inadequate follow-up to care were highlighted in a large HIV/AIDS study conducted on a Veterans Affairs population in Houston that evaluated the consequences of lack of follow-up in the first year after starting antiretroviral therapy (ART). Patients who did not attend clinical visits regularly were less likely to receive ART consistently, more likely to develop resistant HIV strains, and less able to achieve viral suppression. The mortality rates were higher (21%) in patients with lower retention in care rates versus 15% in the group that had optimal follow-up.¹¹

A woman who is pregnant tends to feel a strong responsibility to her unborn child, providing extra motivation for care not present in the nonpregnant adult. After delivery, staying in care may be partially dependent on how well the concept of “self care” has been integrated into her perception of being infected with HIV. Investigation of barriers to care among predominantly African American nonpregnant women of color in the southern United States has identified stigma, financial challenges (lack of private insurance), depression, and physical symptoms as commonly cited issues. Women who scored higher on scales of self-determination, autonomy, and relatedness were less likely to report multiple barriers to care.¹² Quinlivan et al. identified that a loss of autonomy and perception of powerlessness often is associated with a new diagnosis of HIV, and that linkage to care requires a feeling of support at the time of testing (knowledge of the importance of care is not enough). Retention in care requires reestablishment of autonomy and self-pride.¹³ In pregnant women, this concept might translate into emotional support at the time of testing and during prenatal care; seeing their personal success in achieving an undetectable viral load prior to delivery and receiving support for the duration of prenatal care offer women a chance to regain personal pride.

We were surprised by the findings of a preterm birth rate in the 17% range (14% among those with no follow-up with PCP, 22% among those with suboptimal follow-up, and 18% among those with optimal follow-up) (Table 2). Although the differences among groups are not statistically significant, the rate in all groups initially seemed high. However, a review of the preterm birth rate in Texas revealed that in 2011, 13% of all births were preterm; among African Americans the rate was 17% and among Hispanics 12.5%, suggesting that the preterm birth rate in this study is consistent with the local baseline and is, most likely, not a result of HIV infection.¹⁴ Also, a subanalysis of the preterm births revealed that 5/35 were a result of inductions secondary to pregnancy-associated hypertension. Our findings suggest that retention in care remains a significant problem in our minority Southern HIV-infected population of postpartum women. Within in the 12-month postpartum period, 83 patients, or 39%, were lost to follow-up to HIV providers, while only 36% had optimal HIV follow-up. The most significant factors associated with loss to follow-up postpartum were young age, late entry to prenatal care, black race, and no plans for contraceptive use. Patients who are younger appear to be more likely to put off receiving care, perhaps because they think they have a lot of time until the disease progresses. Previous studies have shown similar characteristics in patients more likely to be lost to follow-up.³ One characteristic that was specific to no follow-up was younger age of HIV diagnosis. The trend of younger and healthier patients being lost to follow-up is consistent with the national trend as younger women are less likely to visit PCPs routinely. The finding of no contraceptive use is a new predictor not previously reported in the literature.

Return to clinic for the 6-week routine postpartum visit with the obstetrical clinician was also documented in the data collection; however, it was not included in this report. One trend we noted was that patients were more likely to return for the 6-week postpartum visit than to the HIV provider. Possible explanations for why women may have returned to the obstetrical provider include: (1) their desire for contraception such as IUD insertion or implant placement, (2) since the 6-week postpartum visit is with a provider they have met before, they feel more comfortable returning, while the visit with the HIV provider may be with someone they have not met before, and (3) many postpartum women have had a recent diagnosis in the pregnancy and may need extra time to process it; the diagnosis is a major life-altering event, which may act as a barrier to care.⁵ In addition, this loss to follow-up postpartum may be a transient effect due to having a newborn at home; many of these women may return to care once their children are of school age and they no longer need childcare in order to come to their physician appointments.

Proposed interventions to increase postpartum follow-up for HIV care include (1) having pregnant patients visit the HIV PCP during the pregnancy just so they are able to develop a relationship with the PCP and may be more likely to return after their deliveries, and (2) creating a support group for the patients and having social workers, obstetrical providers, and HIV providers (PCPs) speak to the women to explain their options for follow-up once they have delivered.

Strengths of the study include the large number of HIV-positive pregnant women who receive care at the HHS. As a result of this analysis, we now have specific predictors of women at risk for loss to care whom we can effectively target by offering interventions such as extra counseling, support groups, and help with transportation.

A limitation of this study was occasional CD4 counts and viral loads that were missing, as some patients did not follow-up frequently enough to obtain labs at the appropriate intervals. Another limitation of the study is that patients may choose to get care for their HIV outside HHS. Since we used the HHS database, we cannot capture information of patients that received their HIV care elsewhere in Houston. However, HHS is the major provider of HIV care to the indigent HIV population of Houston. Also, our study addresses women in a Southern minority population and may not be applicable to other populations.

Our study does not address the consequences of suboptimal adherence and its clinical meaning. Someone who is suboptimally adherent to follow-up may not be taking medications regularly. Prescription refills are often done at the discretion of the provider (sometimes refilled without a clinical visit); also the patients' ability to pay for medications often changes. Many of these patients need the AIDS Drug Assistance Program to pay for their medications postpartum, which has logistical requirements that need to be met. One may argue that the definition of optimal follow-up and suboptimal follow-up may not be appropriate in this population, as even in suboptimal follow-up, patients may not be adequately adherent to HIV medications. Lack of retention to care in the HIV population is increasingly recognized and described.¹⁵ Interventions to improve retention to care are being designed and evaluated.

The findings of this study suggest that we should stress adequate follow-up in care to the population of HIV-infected women who are younger, black, present late for prenatal care, and who express little interest in postpartum contraception. Taking into consideration the limited number of studies of follow-up in HIV-positive women and even fewer in postpartum women in the U.S., this a subject in which further research is needed. Future studies should investigate interventions aimed at the women that are at highest risk to be lost to primary care after delivery.

Footnotes

Acknowledgments

We would like to acknowledge Drs. Maya Sato and Kosuke Yasukawa for their assistance with data review.

Author Disclosure Statement

No competing financial interests exist.

References

CDC. (2013, April 24). HIV among Women. Retrieved May 26, 2013 from http://www.cdc.gov.

Whitmore

, Taylor

, Espinoza

, et al. Correlates of mother-to-child transmission of HIV in the United States and Puerto Rico. Pediatrics, 2012; 129; e74.

Rana

, Gillani

, Flanigan

, et al. Follow-up care among HIV-infected pregnant women in Mississippi. J Women's Health, 2010; 19:1863–1867.

Gertsch

, Michel

, Locatelli

, et al. Adherence to antiretroviral treatment decreases during postpartum compared to pregnancy: A longitudinal electronic monitoring study. AIDS Patient Care STDs, 2013; 27:208–210.

Kaplan

, Orrell

, Zwane

, et al. Loss to follow-up and mortality amongst pregnant women referred to a community clinic for antiretroviral treatment. AIDS, 2008; 22:1679–1681.

Martin

, Navaratne

, Khan

, et al. Pregnant women with HIV infection can expect healthy survival: Three-year follow-Up. J Acq Immune Def Synd, 2006; 143:186–192.

Aberg

, Gallant

, Anderson

, et al. Primary care guidelines for the management of persons infected with human immunodeficiency virus: Recommendations of the HIV Medicine Association of the Infectious Diseases Society of America. Clin Infect Dis, 2004; 39:609–629.

Mugavero

, Wesfall

, Zinski

, et al. Measuring retention in care: The elusive gold standard. J Acq Immune Def Synd, 2012; 61:574–580.

Stasenko

, Liddell

, Cheng

, et al. Patient counseling increases postpartum follow-up in women with gestational diabetes mellitus. Am J Obstet Gynecol, 2011; 204:522.e1–e6.

10.

Kreitchmann

, Harris

, Kakehasi

, et al. for the NISDI Lilac Study Team. Antiretroviral adherence during pregnancy and postpartum in Latin America. AIDS Patient Care STDs, 2012; 26:486–495.

11.

Giordano

, Gifford

, White

Jr , et al. Retention in care: A challenge to survival with HIV infection. Clin Infect Dis, 2007; 44:1493–1499.

12.

Toth

, Messer

, Quinlivan

. Barriers to HIV care for women of color living in the Southeastern US are associated with physical symptoms, social environment, and self-determination. AIDS Patient Care STDs, 2013; 27:613–620.

13.

Quinlivan

, Messer

, Adimora

, et al. Experiences with HIV testing, entry, and engagement in care by HIV-infected women of color, and the need for autonomy, competency, and relatedness. AIDS Patient Care STDs, 2013; 27:408–415.

14.

Texas Department of State Health Statistics. Healthy Texas Babies Infant and Maternal Health Data. http://www.dshs.state.tx.us/HTBExpertPanelHomework2010.pdfpp.12-16 (updated July 2012).

15.

Gardner

, McLees

, Steiner

, et al. The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clin Infect Dis, 2011; 52:793–800.