Journal of Critical Care 29 (2014) 199–203
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
Health insurance status and outcomes of critically ill obstetric patients: A prospective cohort study in Argentina☆ Daniela N. Vasquez a, b, c,⁎, Andrea V. Das Neves a, c, Vanina B. Aphalo b, Cecilia I. Loudet a, Javier Roberti c, Federico Cicora c, Matias Casanova b, Hector S. Canales a, Alfredo D. Intile b, Jose L. Scapellato b, Pablo M. Desmery b, Elisa Estenssoro a a b c
Hospital Interzonal General de Agudos Gral. San Martín, La Plata, Buenos Aires, Argentina Sanatorio Anchorena, City of Buenos Aires, Argentina Fundación para la Investigación y Asistencia de la Enfermedad Renal, La Plata, Buenos Aires, Argentina
a r t i c l e
i n f o
a b s t r a c t
Keywords: Critical care Insurance Pregnancy Pregnancy outcome Private sector Public sector
Purpose: In Argentina, uninsured patients receive public health care, and the insured receive private health care. Our aim was to compare different outcomes between critically ill obstetric patients from both sectors. Methods: This is a prospective cohort, including pregnant/postpartum patients requiring admission to 1 intensive care unit in the public sector (uninsured) and 1 in the private (insured) from January 1, 2008, to September 30, 2011. Results: A total of 151 patients were included in the study. In uninsured (n = 63) vs insured (n = 88) patients, Acute Physiology and Chronic Evaluation II (APACHE II) and Sequential Organ Failure Assessment scores were 11 ± 6.5 vs 8 ± 4 and 3 (2-7) vs 1 (0-2), respectively, and 84% vs 100% received prenatal care (P = .001 for all). Multiple organ dysfunction syndrome (MODS) was present in 32 (54%) uninsured vs 9 (10%) insured patients (P = .001), and acute respiratory distress syndrome developed in 18 (30.5%) of 59 vs 2(2%) of 88 (P = .001). Neonatal survival was 80% vs 96% (P = .003). Variables independently associated with the development of MODS were APACHE II (odds ratio, 1.30 [1.13-1.49]), referral from another hospital (odds ratio, 11.43 [1.86-70.20]), lack of health insurance (odds ratio 6.75 [2.17-20.09]), and shock (odds ratio 4.82 [1.54-15.06]). Three patients died, all uninsured. Conclusions: Uninsured critically ill obstetric patients (public sector) were more severely ill on admission and experienced worse outcomes than insured patients (private sector). Variables independently associated with MODS were APACHE II, shock, referral from another hospital, and lack of insurance. © 2014 Elsevier Inc. All rights reserved.
1. Introduction
higher income and educated people are insured and seek care in the private sector, whereas low-income and less-educated are uninsured and receive health care in the public sector [3]. Two US studies reported that critically ill uninsured patients received fewer critical care services and probably experienced higher hospital mortality than insured patients [5,6]. This suggests that lack of insurance may be an independent risk factor for death among critically ill patients [5,6]. A retrospective study of obstetric patients from Indonesia showed higher mortality and more severe morbidity in the public health sector compared with the private sector, but only 5% of patients in that study were critically ill [7]. In Australia, another population-based retrospective study revealed a higher level of obstetric intervention in the private sector [8]. Munnur et al [9] compared populations of critically ill obstetric patients from India (public hospital) and the United States (private hospital) and found that the former were more severely ill on admission, had higher maternal and fetal mortality, received lower levels of prenatal care, and experienced a grater delay in reaching hospital than US patients did.
Each year, nearly 300 000 maternal deaths occur, almost exclusively in developing countries, characterizing maternal mortality as a health indicator of social inequality [1]. In Argentina, where maternal and perinatal mortality are higher than in developed countries, 70% of these deaths occur in hospitals, mainly in intensive care units (ICUs), with significant differences across provinces and socioeconomic groups [2]. The Argentine health system is composed of public and private sectors, the latter consisting of work-related health insurance and by individually paid health insurance [3]. Almost 64% of the population has some type of paid health insurance [4]. Generally,
☆ Conflict of interest statement: We declare that we have no conflicts of interest. ⁎ Corresponding author. Intensive Care Unit, Hospital Interzonal General de Agudos, Gral. San Martín, St 1 and 70, La Plata, 1900, Buenos Aires, Argentina. Intensive Care Unit, Anchorena Clinic, 1872 Dr Tomas M. de Anchorena St (C1425ELP), City of Buenos Aires, Argentina. Tel./fax: +54 2214733200. E-mail address: [email protected] (D.N. Vasquez). 0883-9441/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2013.11.010
200
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
The objective of this study was to explore associations between health insurance status and severity of illness in critically ill obstetric patients. As maternal mortality has fallen in the last decades, other outcome indicators were proposed for assessing maternal care. In this case, we chose the development of multiple-organ dysfunction as a severity indicator and insurance status, presence of comorbidity, referral from other centers, and prenatal care, among others, as potential predictors. Our hypothesis was that uninsured critically ill obstetric patients were more severely ill than insured patients.
a level of significance of P ≤ .05 on the Wald test and/or confounding effects (variation coefficient ≥20%). The model was calibrated with the Hosmer-Lemeshow goodness-of-fit test to evaluate the discrepancy between observed and expected MODS values. Model discrimination was evaluated with the area under the receiver operating characteristics curve (AUROC). STATA 11.1 and SPSS 15 (SPSS, Inc, Chicago, Ill) were used for the analysis.
2. Materials and methods
The sample size of 57 patients in each group was calculated considering a severity of illness measure, namely, the percentage of patients requiring MV in the public health sector vs the private health sector, at an α = .05 and β = .8. This was based on a previous study [20], where 41% of patients required MV in the public sector, and a pilot study, where 15% of patients required MV in the private sector.
This was a prospective cohort study, which included all pregnant/ postpartum (b 42 days) patients requiring admission to an ICU in the public health sector (uninsured) vs an ICU in the private sector (insured) in Argentina between January 1, 2008 and September 30, 2011. The only exclusion criterion for entering the study was an unwillingness to participate, but none refused consent. The private clinic, located in the city of Buenos Aires, is a 186-bed center where 2000 children are born per year. The public hospital is a universityaffiliated 449-bed center in the city of La Plata in the province of Buenos Aires, where 3000 infants are delivered annually. Both hospitals are referral centers and offer the same standard health care. Both intensive care units (ICUs) are medical-surgical units managed by intensivists, with 14 beds in the public hospital and 12 in the private hospital, the nurse-to-patient ratio being 1:2 and 1:2-3, respectively. The training and experience of the nursing staff in both sectors are similar. Demographic data, comorbidity (Charlson score) [10], admission diagnosis, obstetric (occurring only in pregnant/postpartum patients) vs nonobstetric causes of admission (also occurring in nonpregnant patients) [11], obstetric history including ante/postpartum admission, gestational age, minimal vs standard prenatal care (at least 1 vs 5 visits for a term pregnancy) [12], parity, history of induced or spontaneous abortions, and type of delivery were recorded. Also registered were length of stay (LOS) in the ICU and in the hospital; severity of illness scores during the first 24 hours in the ICU, using the worst values for each parameter (Acute Physiology and Chronic Evaluation II [APACHE II] [13] and Sequential Organ Failure Assessment [SOFA] [14]); level of intervention in the ICU (mechanical ventilation [MV], days on MV, Therapeutic Intervention Scoring System 28 [TISS28] [15], central lines, or Swan-Ganz catheter requirement); and complications in the ICU such as acute respiratory distress syndrome (ARDS) [16], septic [17] or hypovolemic shock [18], and multipleorgan dysfunction syndrome (MODS) (dysfunction of ≥ 2 organs according to the SOFA score) [17]. Intensive care unit and hospital maternal mortality [19] and fetal-neonatal losses were recorded. Induced abortions were separately analyzed from fetal losses. Severe neonatal morbidity was also documented. 2.1. Statistical analysis Categorical variables are shown as numbers (percentages), and continuous variables, as mean ± SD or median (interquartile range [IQR]), according to their distribution. Continuous normally and nonnormally distributed variables were compared with Student t test and the Wilcoxon test, respectively, and categorical variables were compared with the χ 2 or Fisher test. Multiple comparisons between categorical variables were performed using the multiple χ 2 test with a Bonferroni correction. P ≤ .05 was considered significant. A multivariable analysis adjusting for potential confounders was performed to evaluate the relationship between predictor and outcome variables. A multiple logistic regression model was built using MODS as the dependent variable. Variables included in the model were those related to MODS in the univariate analysis with P ≤ .20. The multivariate model was built manually, including variables with
2.2. Sample size calculation
2.3. Ethical considerations This study was approved by the ethics committee of each center and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Informed written consent was requested from each patient or caregiver/relative, in the event the patient was unable to provide consent upon admission to the ICU. Once able to consent for themselves, those patients consented voluntarily. 3. Results Over the study period, 151 obstetric patients were admitted to both ICUs: 63 in the public sector (all uninsured) and 88 in the private sector (all insured). These admissions represented 7% of all ICU admissions (n = 884) in the public and 3.4% in the private sector (n = 2601; P b .000). During the same period, 10 260 deliveries were performed in the public sector and 6970 in the private, totaling an ICU admission rate of 614/100 000 (0.61%) and 1262/100 000 (1.26%), respectively (P b .000). Uninsured patients were younger, illustrated more comorbidity, were more severely ill on admission, and stayed longer in the ICU than insured patients (Table 1). Three uninsured patients died, 2 during the first 24 hours after admission to the ICU. Causes of admission for these Table 1 Characteristics and outcomes of uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
n Age (y) Charlson 0 (without comorbidity) APACHE II APACHE risk (%) SOFA24 Location before admission to ICU Operating room Ward Emergency Other hospital Antepartum admission ICU-LOS (d) Hospital-LOS (d) ICU mortality Hospital mortality
Uninsured (public)
Insured (private)
63 (42) 26 ± 7.5 51 (81) 11 ± 6.5 12 (5-20) 3 (2-7)
88 (58) 33 ± 6 83 (94) 8±4 6 (2.5-13) 1 (0-2)
21 (33) 17 (27) 14 (22) 11 (18)a,b 14 (22) 4 (2-9) 11 (6-21) 2 (3.2) 3 (4.8)
45 (51) 19 (22) 22 (25) 2 (2) 23 (26) 3 (2-4) 6 (4-8) 0 (0) 0 (0)
P .000 .029 .000 .000 .000 .004
.55 .008 .000 .17 .057
Data are shown as mean ± SD, median (IQR), or n (%). SOFA24 indicates Sequential Organ Failure Assessment (during the first 24 hours of admission). a Post hoc analysis to specifically evaluate where the difference was: other hospital vs operating room, P = .006. b Post hoc analysis to specifically evaluate where the difference was: other hospital vs emergency department, P = .048.
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
patients were subarachnoid hemorrhage, severe pulmonary hypertension, and chorioamnionitis. What is more, their fetuses and neonates did not survive. The 2 groups had similar obstetric history, with the same percentage of women reporting a previously induced abortion. Prenatal care was less frequent among uninsured patients (Table 2). Admission diagnoses were classified into 2 broad categories: obstetric (occurring only in pregnant/postpartum patients) and nonobstetric (also occurring in the general population). In this regard, obstetric causes of admission were more frequent in the private sector (insured patients) (n = 73; 83%) than in the public sector (uninsured patients) (n = 42; 67%); P = .021. The main causes of admission in both groups were hypertensive disease of pregnancy and hemorrhage, whereas septic abortion was reported only among uninsured patients (Table 3). The number of interventions and complications in the ICU was higher in uninsured patients from the public sector (Table 4). The incidence of renal dysfunction and failure differed according to the chosen criteria. With a creatinine value of 0.8 mg/dL as the upper cut-off point for renal dysfunction in pregnant patients [21], its incidence was 24.5% (37 patients), including 25 uninsured patients (40%) and 12 insured (14%) (P = .000). However, when the SOFA score was used (creatinine upper cut-off level of 1.2 mg/dL), these numbers decreased to 9 (14.5%) and 5 (6%), respectively (P = .09). Only 2 patients, both uninsured, showed renal failure according to the SOFA score (with creatinine ≥3.5 mg/dL) and 1 required dialysis. Risk factors for severe morbidity were evaluated with multivariate analysis, using MODS as the outcome measure (Table 5). The independent risk factors for MODS were lack of insurance, referral from another hospital, APACHE II score, and development of shock (model discrimination AUROC, 0.91 [0.85-0.96]). Fetal-neonatal survival was 45 (80%) of 56 among uninsured vs 78 (96%) of 81 among insured patients (P = .003). Severe morbidity in neonates did not differ in either group (Table 6). 4. Discussion Uninsured critically ill obstetric patients (public health sector) were more severely ill on admission and had a higher rate of severe morbidity than insured patients (private health sector). Fetalneonatal loses were higher among uninsured patients, and although not statistically significant, maternal mortality was limited to the same group. Uninsured critically ill obstetric patients were younger, showed a higher prevalence of comorbidity and higher severity scores on admission, and were more frequently referred from other hospitals
Table 2 Obstetric history of uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients Total
Uninsured (public)
Insured (private)
P
32 ± 8 112/122 (92) 92/97 (95) 2 (1-4) 50/122 (41) 17/118 (14.5) 31/119 (26)
31 ± 8 38/48 (84) 25/29 (86) 2 (1-4) 19/43 (44) 9/40 (22.5) 7/41 (17)
33 ± 8 74/74 (100) 67/68 (98.5) 2 (1-3) 31/79 (39) 8/78 (10) 24/78 (31)
.07 .000 .027 .15 .59 .07 .10
Patients with known route of delivery Cesarean 111/140 (78) Vaginal delivery 10/140 (7)) Discharged from ICU 14/140 (10) pregnant
44/53 (83) 4/53 (7.5) 5/53 (9.5)
67/87 (77) 6/87 (7) 9/87 (10)
.39 1 1
Gestational agea Prenatal careb Standard prenatal carec Parity Previous abortions Induced abortions Spontaneous abortions
Data are presented as mean ± SD, median (IQR), or n (%). a Gestational age on admission to ICU or at the end of the current pregnancy (weeks). b At least 1 prenatal visit. c At least 5 prenatal visits in a term pregnancy [12].
201
Table 3 Causes of admission to ICU for uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
Obstetric Hypertensive disease of pregnancy Severe preeclampsia HELLP Eclampsia Eclampsia + HELLP Gestational hypertension Abortion Obstetric hemorrhage Postoperative period after cesarean ± hysterectomy for placenta accreta (scheduled) Ectopic pregnancy Puerperal sepsis Intrauterine fetal death Nonobstetric Sepsis Severe pneumonia Peritonitis Upper urinary tract infection Neurologic Subarachnoid hemorrhage Seizures (no eclampsia) Postoperative period after cesarean for underlying diseases Respiratory failure Others
Uninsured (public)
Insured (private)
P
42 (67) 21 (33) 1 7 10 3 0 5 (8)a 7 (11) 6 (9.5)
73 (83) 42 (48) 23 5 11 1 2 1 (1)b 24 (27) 1 (1)
.021 .07
0 1 (1.5) 2 (3) 21 (33) 12 (19) 10 2 0 2 (3) 2 0 3 (5)
4 (4.5) 0 1 (1) 15 (17) 6 (7) 3 2 1 2 (2) 0 2 1 (1)
.14
0 4 (6)
.08 .016 .02
.022
NS
3 (3.5) 3 (3.5)
Data are shown as n (%). a All were septic abortion. b Spontaneous abortion.
compared with insured patients. These findings were similar to the Indian patients in the study of Munnur et al [9] and to the Indonesian public sector patients in the study of Adisasmita et al [7]. The cost of transport to health centers, staff attitudes toward women with lower incomes, and poor communication skills may be some of the barriers to access proper and timely care for uninsured patients, resulting in greater illness severity on admission [22]. Once patients were admitted to the ICU, uninsured critically ill obstetric patients from the public sector showed a higher incidence of severe morbidity, such as MODS or ARDS, leading to a higher level of intervention, compared with insured patients from the private sector. In contrast, the level of intervention for obstetric patients from the private sector was higher compared with those patients from the public sector in a population-based study in Australia [8]. Other
Table 4 Level of intervention and complications for uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
n TISS28 MV MV (d) Pulmonary artery catheter Shock Type of shock Hypovolemic Septic Hypovolemic and septic Other ARDS MODS Central line
Uninsured (public)
Insured (private)
63 25 ± 9 27 (43) 7 (2-13) 2 (3.6) 22 (37)
88 20 ± 9 12 (14) 1 (0.5-4) 3 (3) 21 (24)
10 (45) 8 (36) 3 (14) 1 (5) 18/59 (30.5) 32 (54) 39 (70)
19 (90)a 1 (5) 0 (0) 1 (5) 2/88 (2) 9 (10) 23 (26)
P .001 .000 .003 1 .092 .036
.000 .000 .000
Data are shown as mean ± SD, median (IQR), or n (%). a Post hoc analysis to specifically evaluate where the difference was: hypovolemic vs septic shock, P = .042.
202
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
Table 5 Multivariate analysis for variables associated with MODS among critically ill obstetric patients Variable
Odds ratio (95% confidence interval)
P
Lack of insurance Referral from another hospital APACHE II Shock
6.75 (2.17-20.09) 11.43 (1.86-70.20) 1.30 (1.13-1.49) 4.82 (1.54-15.06)
.001 .007 .000 .007
Variables were adjusted by age, prenatal care, appropriate prenatal care, nonobstetric cause of admission, obstetric cause of admission, presence of comorbidity, hospital LOS, TISS28, use of mechanical ventilation, and ARDS (model discrimination AUROC, 0.91 [0.85-0.96]).
studies have shown a lower level of intervention and higher mortality among uninsured than among insured critically ill patients [5,6,23]. Although the Indian patients in the study of Munnur et al [9] were more severely ill than US patients, they received fewer invasive procedures and showed higher mortality. The same was true for patients from the public sector in the study of Adisasmita et al [7]. Hospital and prenatal care are free in Argentina, so affordability was not a barrier to receiving required procedures once the patient was admitted to hospital or the ICU. In this study, 92% of the 151 patients received some type of prenatal care, indicating a clear improvement over a previous retrospective study from Argentina [20], possibly a result of recent national health programs to diminish maternal mortality. However, uninsured patients still received less frequent prenatal care than insured, similar to Indian patients in the study of Munnur et al [20]. In Argentina, abortion is illegal except in specific cases outlined by law, such as when the mother's life is in jeopardy or when pregnancy is the result of rape in a mentally incapable female [24]. Nevertheless, the incidence of induced abortion did not differ significantly in either health sector. What was significant was the fact that there were no cases of septic abortion in the insured patients. This situation might indicate that although abortion is undertaken by women from both groups, it takes place in worse conditions for low-income, lesseducated, and uninsured women from the public sector, and therefore, complications are more frequent. Similar results have been reported in an Indonesian study, where induced abortion is also illegal and septic abortion was more common in patients admitted to hospitals in the public health sector [7]. Causes of admission to the ICU were classified as obstetric and nonobstetric. In both health sectors, at least two thirds of patients were admitted for obstetric reasons, consistent with most reports [25]. However, nonobstetric causes of admission were more frequent among uninsured (public health sector) than insured (private health sector) patients (33% vs 17%, respectively), which might be associated with a higher prevalence of comorbidity [26] and greater illness severity [11,27]. The most common obstetric causes of admission were hypertensive disease of pregnancy and obstetric hemorrhage, as has been reported throughout the world [25].
Table 6 Characteristics, level of intervention, and outcomes of fetuses/neonates from uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
Fetal/neonatal survival Hospital-LOS (d) Neonatal ICU Appropriate birth weight for gestational age MV CPAP RDS
Uninsured (public)
Insured (private)
P
45/56 14 21/38 34/37
78/81 (96) 9 (5-23.5) 38/66 (58) 54/64 (84)
.003 .138 .82 .36
11/38 (29) 15/37 (40.5) 18/78 (23)
.05 .78 .05
(80) (6-30) (55) (92)
11/20 (55) 8/18 (44) 19/45 (42)
Data are presented as n (%) and median (IQR). CPAP indicates continuous positive airway pressure; RDS, respiratory distress syndrome.
In terms of severe morbidity, ARDS was more frequent among uninsured patients. Shock developed equally in both sectors, but septic shock was more common in the public sector and hypovolemic shock in the private sector, the latter associated with a higher incidence of obstetric hemorrhage. The lower incidence of hypovolemic shock in the public sector was predictable, as the public hospital is a referral center for placenta accreta management. Renal dysfunction was more common in uninsured patients as well as among Indian patients in the study of Munnur et al [7]. This is the first study indicating a higher incidence of renal dysfunction in pregnant/ postpartum patients using creatinine cut-off levels adjusted to pregnancy, instead of the commonly used scores for general ICU population, which clearly underestimate this condition. The development of MODS is considered a measure of severe morbidity, and its risk factors were investigated. Variables independently associated with the development of MODS were APACHE II score, shock, referral from another hospital, and lack of health insurance. Higher APACHE II score and the development of shock as risk factors for MODS have well-known biological plausibility. On the other hand, the association between referral from other centers and MODS may be related to delay in obtaining proper care [28]. Lack of health insurance was an independent risk factor for multiple-organ dysfunction. Although this might be related, in part, to the high proportion of referred patients among the uninsured, it does not seem to be the only explanation. Lack of prenatal care, a key factor for reducing maternal mortality, comorbidity, and MODS, was not associated in the multivariate model. Our hypothesis was that the high prenatal coverage (~ 90%) precluded finding an association given our sample size. As Zeeman [28] has addressed, maternal outcomes might be adversely affected by inadequate use of prenatal care and a delay more than 24 hours between onset of illness and ICU admission. The latter may be related to a delay in the decision to seek care, in accessing services and in the provision of appropriate care once the health facility is reached [31]. In this study, the higher incidence of MODS among uninsured patients does not seem to be related to differential management in each sector, as the 2 hospitals offered similar services. Moreover, the level of intervention was even higher among uninsured patients. These patients were sicker upon admission to the ICU, probably due to a delay in deciding to seek care, which may be related to sociocultural factors, and to accessibility problems, such as distance and transport cost or availability. One factor possibly affecting the decision to seek care could be maternal education, which is inversely related to maternal mortality, even after adjusting for marital status, economic status, institutional capacity, or health care investment [29,30]. A higher level of education improves a woman's ability to perceive signs and symptoms in her own body and to seek care and inquire about personal health. Moreover, education may help women overcome communication barriers, gain self-confidence, and feel empowered to make correct decisions. Although gestational age did not differ between groups, fetalneonatal loss was higher among uninsured patients. This may be related to the higher prevalence of comorbidity in this population as well as the greater severity of their condition upon admission and the higher incidence of severe morbidity. However, severe morbidity and interventions among surviving neonates were similar in both groups. Thus, we might hypothesize that the severity of maternal illness upon admission determines fetal-neonatal mortality but might not influence the progression of the disease in those surviving. Several limitations of this study need to be noted. This was an observational study and, as such, prone to bias and confounding. We performed a multivariate analysis to monitor for possible confounders. The study only represents women with access to a health care facility and who were admitted to an ICU; however, most deliveries and maternal deaths occur within the health care system in
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
Argentina [2]. This research was carried out in private and public centers representing both high- and low-income populations, respectively. Finally, other potentially important variables, such as maternal education or other barriers that marginalized women experience when accessing care, were not measured. 5. Conclusions In summary, uninsured critically ill obstetric patients were more severely ill on admission and displayed higher levels of severe morbidity than insured patients. Similarly, fetal-neonatal losses were more frequent among uninsured patients. Risk factors for developing MODS were APACHE II score, shock, referral from another hospital, and lack of insurance. However, some level of prenatal care, a measurement that has been shown to have great impact on reducing maternal mortality, was received by nearly 90% of patients. Therefore, other preventive measures, such as improving social conditions, particularly regarding maternal education and women's empowerment as well as free universal access to safe legal abortions, should be implemented to reduce severe maternal morbidity. Acknowledgments This study was supported by the Argentinean National Ministry of Health and the Argentinean Society of Critical Care. The funding sources had no role in study design, data collection, data analysis, data interpretation, writing of the report, and decision to submit the manuscript for publication. The English in this document has been checked by at least 2 professional editors, both native speakers of English. For a certificate, please see: http://www.textcheck.com/certificate/RJhRYm. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jcrc.2013.11.010.
References [1] Lozano R, Wang H, Foreman KJ, et al. Progress towards Millennium Development Goals 4 and 5 on maternal and child mortality: an updated systematic analysis. Lancet 2011;378:1139–65. [2] Argentinean Ministry of Health - Pan American Health Organization. Basic Indicators: Argentina 2010; 2011 [http://wwwmsalgovar and http://wwwopsar]. [3] Pan American Organization of Health. Health in the Americas. Argentina: Scientific and Technical Publication No 622; 2007 39–61. [4] INDEC. Argentinian National Institute of Statistics and Censuses. http//www. censo2010.indec.gov.ar/resultadosdefinitivos_totalpais.asp.2010; 2010. [5] Fowler RA, Noyahr LA, Thornton JD, et al. An official American Thoracic Society systematic review: the association between health insurance status and access, care delivery, and outcomes for patients who are critically ill. Am J Respir Crit Care Med 2011;181:1003–11.
203
[6] Lyon SM, Benson NM, Cooke CR, et al. The effect of insurance status on mortality and procedural use in critically ill patients. Am J Respir Crit Care Med 2011;184: 809–15. [7] Adisasmita A, Deviany PE, Nandiaty F, et al. Obstetric near miss and deaths in public and private hospitals in Indonesia. BMC Pregnancy Childbirth 2008;8:10. [8] Roberts CL, Tracy S, Peat B. Rates for obstetric intervention among private and public patients in Australia: population based descriptive study. BMJ 2000;321: 137–41. [9] Munnur U, Karnad DR, Bandi VD, et al. Critically ill obstetric patients in an American and an Indian public hospital: comparison of case-mix, organ dysfunction, intensive care requirements, and outcomes. Intensive Care Med 2005;31: 1087–94. [10] Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [11] Karnad DR, Lapsia V, Krishnan A, et al. Prognostic factors in obstetric patients admitted to an Indian intensive care unit. Crit Care Med 2004;32:1294–9. [12] Schwarcz RUA, Lomuto C, Martinez I, Galimberti D, Garcia OM, Etcheverry ME, Queiruga M. Prenatal care. Practice guidelines for the preconception and prenatal care. Argentina: National Ministry of Health; 2001 1–39 [Available at: http:// www.msal.gov.ar]. [13] Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818–29. [14] Vincent JL, de Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on “sepsis-related problems” of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793–800. [15] Malstam J, Lind L. Therapeutic intervention scoring system (TISS)—a method for measuring workload and calculating costs in the ICU. Acta Anaesthesiol Scand 1992;36:758–63. [16] Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818–24. [17] Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101:1644–55. [18] Oyelese Y, Scorza WE, Mastrolia R, et al. Postpartum hemorrhage. Obstet Gynecol Clin North Am 2007;34:421–41. [19] WHO. International Statistical Classification of Diseases and Related Health Problems, 10th revision. Geneve, Switzerland: World Health Organization; 1992. [20] Vasquez DN, Estenssoro E, Canales HS, et al. Clinical characteristics and outcomes of obstetric patients requiring ICU admission. Chest 2007;131:718–24. [21] Yeomans ER, Gilstrap III LC. Physiologic changes in pregnancy and their impact on critical care. Crit Care Med 2005;33:S256–8. [22] Silal SP, Penn-Kekana L, Harris B, et al. Exploring inequalities in access to and use of maternal health services in South Africa. BMC Health Serv Res 2012;12:120. [23] Danis M, Linde-Zwirble WT, Astor A, et al. How does lack of insurance affect use of intensive care? A population-based study. Crit Care Med 2006;34:2043–8. [24] Codigo penal de la nación Argentina. http://www.codigopenalonline.com.ar/ codigo_penal_argentino_delitos_contra_las_personas.html. [25] Pollock W, Rose L, Dennis CL. Pregnant and postpartum admissions to the intensive care unit: a systematic review. Intensive Care Med 2010;36:1465–74. [26] Lombaard H, Pattinson RC. Underlying medical conditions. Best Pract Res Clin Obstet Gynaecol 2008;22:847–64. [27] Mhyre JM, Bateman BT, Leffert LR. Influence of patient comorbidities on the risk of near-miss maternal morbidity or mortality. Anesthesiology 2011;115:963–72. [28] Zeeman GG. Obstetric critical care: a blueprint for improved outcomes. Crit Care Med 2006;34:S208–14. [29] Filippi V, Ronsmans C, Campbell OM, et al. Maternal health in poor countries: the broader context and a call for action. Lancet 2006;368:1535–41. [30] Karlsen S, Say L, Souza JP, et al. The relationship between maternal education and mortality among women giving birth in health care institutions: analysis of the cross sectional WHO Global Survey on Maternal and Perinatal Health. BMC Public Health 2011;11:606. [31] Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Soc Sci Med 1994;38:1091–110.
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
Health insurance status and outcomes of critically ill obstetric patients: A prospective cohort study in Argentina☆ Daniela N. Vasquez a, b, c,⁎, Andrea V. Das Neves a, c, Vanina B. Aphalo b, Cecilia I. Loudet a, Javier Roberti c, Federico Cicora c, Matias Casanova b, Hector S. Canales a, Alfredo D. Intile b, Jose L. Scapellato b, Pablo M. Desmery b, Elisa Estenssoro a a b c
Hospital Interzonal General de Agudos Gral. San Martín, La Plata, Buenos Aires, Argentina Sanatorio Anchorena, City of Buenos Aires, Argentina Fundación para la Investigación y Asistencia de la Enfermedad Renal, La Plata, Buenos Aires, Argentina
a r t i c l e
i n f o
a b s t r a c t
Keywords: Critical care Insurance Pregnancy Pregnancy outcome Private sector Public sector
Purpose: In Argentina, uninsured patients receive public health care, and the insured receive private health care. Our aim was to compare different outcomes between critically ill obstetric patients from both sectors. Methods: This is a prospective cohort, including pregnant/postpartum patients requiring admission to 1 intensive care unit in the public sector (uninsured) and 1 in the private (insured) from January 1, 2008, to September 30, 2011. Results: A total of 151 patients were included in the study. In uninsured (n = 63) vs insured (n = 88) patients, Acute Physiology and Chronic Evaluation II (APACHE II) and Sequential Organ Failure Assessment scores were 11 ± 6.5 vs 8 ± 4 and 3 (2-7) vs 1 (0-2), respectively, and 84% vs 100% received prenatal care (P = .001 for all). Multiple organ dysfunction syndrome (MODS) was present in 32 (54%) uninsured vs 9 (10%) insured patients (P = .001), and acute respiratory distress syndrome developed in 18 (30.5%) of 59 vs 2(2%) of 88 (P = .001). Neonatal survival was 80% vs 96% (P = .003). Variables independently associated with the development of MODS were APACHE II (odds ratio, 1.30 [1.13-1.49]), referral from another hospital (odds ratio, 11.43 [1.86-70.20]), lack of health insurance (odds ratio 6.75 [2.17-20.09]), and shock (odds ratio 4.82 [1.54-15.06]). Three patients died, all uninsured. Conclusions: Uninsured critically ill obstetric patients (public sector) were more severely ill on admission and experienced worse outcomes than insured patients (private sector). Variables independently associated with MODS were APACHE II, shock, referral from another hospital, and lack of insurance. © 2014 Elsevier Inc. All rights reserved.
1. Introduction
higher income and educated people are insured and seek care in the private sector, whereas low-income and less-educated are uninsured and receive health care in the public sector [3]. Two US studies reported that critically ill uninsured patients received fewer critical care services and probably experienced higher hospital mortality than insured patients [5,6]. This suggests that lack of insurance may be an independent risk factor for death among critically ill patients [5,6]. A retrospective study of obstetric patients from Indonesia showed higher mortality and more severe morbidity in the public health sector compared with the private sector, but only 5% of patients in that study were critically ill [7]. In Australia, another population-based retrospective study revealed a higher level of obstetric intervention in the private sector [8]. Munnur et al [9] compared populations of critically ill obstetric patients from India (public hospital) and the United States (private hospital) and found that the former were more severely ill on admission, had higher maternal and fetal mortality, received lower levels of prenatal care, and experienced a grater delay in reaching hospital than US patients did.
Each year, nearly 300 000 maternal deaths occur, almost exclusively in developing countries, characterizing maternal mortality as a health indicator of social inequality [1]. In Argentina, where maternal and perinatal mortality are higher than in developed countries, 70% of these deaths occur in hospitals, mainly in intensive care units (ICUs), with significant differences across provinces and socioeconomic groups [2]. The Argentine health system is composed of public and private sectors, the latter consisting of work-related health insurance and by individually paid health insurance [3]. Almost 64% of the population has some type of paid health insurance [4]. Generally,
☆ Conflict of interest statement: We declare that we have no conflicts of interest. ⁎ Corresponding author. Intensive Care Unit, Hospital Interzonal General de Agudos, Gral. San Martín, St 1 and 70, La Plata, 1900, Buenos Aires, Argentina. Intensive Care Unit, Anchorena Clinic, 1872 Dr Tomas M. de Anchorena St (C1425ELP), City of Buenos Aires, Argentina. Tel./fax: +54 2214733200. E-mail address: [email protected] (D.N. Vasquez). 0883-9441/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2013.11.010
200
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
The objective of this study was to explore associations between health insurance status and severity of illness in critically ill obstetric patients. As maternal mortality has fallen in the last decades, other outcome indicators were proposed for assessing maternal care. In this case, we chose the development of multiple-organ dysfunction as a severity indicator and insurance status, presence of comorbidity, referral from other centers, and prenatal care, among others, as potential predictors. Our hypothesis was that uninsured critically ill obstetric patients were more severely ill than insured patients.
a level of significance of P ≤ .05 on the Wald test and/or confounding effects (variation coefficient ≥20%). The model was calibrated with the Hosmer-Lemeshow goodness-of-fit test to evaluate the discrepancy between observed and expected MODS values. Model discrimination was evaluated with the area under the receiver operating characteristics curve (AUROC). STATA 11.1 and SPSS 15 (SPSS, Inc, Chicago, Ill) were used for the analysis.
2. Materials and methods
The sample size of 57 patients in each group was calculated considering a severity of illness measure, namely, the percentage of patients requiring MV in the public health sector vs the private health sector, at an α = .05 and β = .8. This was based on a previous study [20], where 41% of patients required MV in the public sector, and a pilot study, where 15% of patients required MV in the private sector.
This was a prospective cohort study, which included all pregnant/ postpartum (b 42 days) patients requiring admission to an ICU in the public health sector (uninsured) vs an ICU in the private sector (insured) in Argentina between January 1, 2008 and September 30, 2011. The only exclusion criterion for entering the study was an unwillingness to participate, but none refused consent. The private clinic, located in the city of Buenos Aires, is a 186-bed center where 2000 children are born per year. The public hospital is a universityaffiliated 449-bed center in the city of La Plata in the province of Buenos Aires, where 3000 infants are delivered annually. Both hospitals are referral centers and offer the same standard health care. Both intensive care units (ICUs) are medical-surgical units managed by intensivists, with 14 beds in the public hospital and 12 in the private hospital, the nurse-to-patient ratio being 1:2 and 1:2-3, respectively. The training and experience of the nursing staff in both sectors are similar. Demographic data, comorbidity (Charlson score) [10], admission diagnosis, obstetric (occurring only in pregnant/postpartum patients) vs nonobstetric causes of admission (also occurring in nonpregnant patients) [11], obstetric history including ante/postpartum admission, gestational age, minimal vs standard prenatal care (at least 1 vs 5 visits for a term pregnancy) [12], parity, history of induced or spontaneous abortions, and type of delivery were recorded. Also registered were length of stay (LOS) in the ICU and in the hospital; severity of illness scores during the first 24 hours in the ICU, using the worst values for each parameter (Acute Physiology and Chronic Evaluation II [APACHE II] [13] and Sequential Organ Failure Assessment [SOFA] [14]); level of intervention in the ICU (mechanical ventilation [MV], days on MV, Therapeutic Intervention Scoring System 28 [TISS28] [15], central lines, or Swan-Ganz catheter requirement); and complications in the ICU such as acute respiratory distress syndrome (ARDS) [16], septic [17] or hypovolemic shock [18], and multipleorgan dysfunction syndrome (MODS) (dysfunction of ≥ 2 organs according to the SOFA score) [17]. Intensive care unit and hospital maternal mortality [19] and fetal-neonatal losses were recorded. Induced abortions were separately analyzed from fetal losses. Severe neonatal morbidity was also documented. 2.1. Statistical analysis Categorical variables are shown as numbers (percentages), and continuous variables, as mean ± SD or median (interquartile range [IQR]), according to their distribution. Continuous normally and nonnormally distributed variables were compared with Student t test and the Wilcoxon test, respectively, and categorical variables were compared with the χ 2 or Fisher test. Multiple comparisons between categorical variables were performed using the multiple χ 2 test with a Bonferroni correction. P ≤ .05 was considered significant. A multivariable analysis adjusting for potential confounders was performed to evaluate the relationship between predictor and outcome variables. A multiple logistic regression model was built using MODS as the dependent variable. Variables included in the model were those related to MODS in the univariate analysis with P ≤ .20. The multivariate model was built manually, including variables with
2.2. Sample size calculation
2.3. Ethical considerations This study was approved by the ethics committee of each center and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Informed written consent was requested from each patient or caregiver/relative, in the event the patient was unable to provide consent upon admission to the ICU. Once able to consent for themselves, those patients consented voluntarily. 3. Results Over the study period, 151 obstetric patients were admitted to both ICUs: 63 in the public sector (all uninsured) and 88 in the private sector (all insured). These admissions represented 7% of all ICU admissions (n = 884) in the public and 3.4% in the private sector (n = 2601; P b .000). During the same period, 10 260 deliveries were performed in the public sector and 6970 in the private, totaling an ICU admission rate of 614/100 000 (0.61%) and 1262/100 000 (1.26%), respectively (P b .000). Uninsured patients were younger, illustrated more comorbidity, were more severely ill on admission, and stayed longer in the ICU than insured patients (Table 1). Three uninsured patients died, 2 during the first 24 hours after admission to the ICU. Causes of admission for these Table 1 Characteristics and outcomes of uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
n Age (y) Charlson 0 (without comorbidity) APACHE II APACHE risk (%) SOFA24 Location before admission to ICU Operating room Ward Emergency Other hospital Antepartum admission ICU-LOS (d) Hospital-LOS (d) ICU mortality Hospital mortality
Uninsured (public)
Insured (private)
63 (42) 26 ± 7.5 51 (81) 11 ± 6.5 12 (5-20) 3 (2-7)
88 (58) 33 ± 6 83 (94) 8±4 6 (2.5-13) 1 (0-2)
21 (33) 17 (27) 14 (22) 11 (18)a,b 14 (22) 4 (2-9) 11 (6-21) 2 (3.2) 3 (4.8)
45 (51) 19 (22) 22 (25) 2 (2) 23 (26) 3 (2-4) 6 (4-8) 0 (0) 0 (0)
P .000 .029 .000 .000 .000 .004
.55 .008 .000 .17 .057
Data are shown as mean ± SD, median (IQR), or n (%). SOFA24 indicates Sequential Organ Failure Assessment (during the first 24 hours of admission). a Post hoc analysis to specifically evaluate where the difference was: other hospital vs operating room, P = .006. b Post hoc analysis to specifically evaluate where the difference was: other hospital vs emergency department, P = .048.
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
patients were subarachnoid hemorrhage, severe pulmonary hypertension, and chorioamnionitis. What is more, their fetuses and neonates did not survive. The 2 groups had similar obstetric history, with the same percentage of women reporting a previously induced abortion. Prenatal care was less frequent among uninsured patients (Table 2). Admission diagnoses were classified into 2 broad categories: obstetric (occurring only in pregnant/postpartum patients) and nonobstetric (also occurring in the general population). In this regard, obstetric causes of admission were more frequent in the private sector (insured patients) (n = 73; 83%) than in the public sector (uninsured patients) (n = 42; 67%); P = .021. The main causes of admission in both groups were hypertensive disease of pregnancy and hemorrhage, whereas septic abortion was reported only among uninsured patients (Table 3). The number of interventions and complications in the ICU was higher in uninsured patients from the public sector (Table 4). The incidence of renal dysfunction and failure differed according to the chosen criteria. With a creatinine value of 0.8 mg/dL as the upper cut-off point for renal dysfunction in pregnant patients [21], its incidence was 24.5% (37 patients), including 25 uninsured patients (40%) and 12 insured (14%) (P = .000). However, when the SOFA score was used (creatinine upper cut-off level of 1.2 mg/dL), these numbers decreased to 9 (14.5%) and 5 (6%), respectively (P = .09). Only 2 patients, both uninsured, showed renal failure according to the SOFA score (with creatinine ≥3.5 mg/dL) and 1 required dialysis. Risk factors for severe morbidity were evaluated with multivariate analysis, using MODS as the outcome measure (Table 5). The independent risk factors for MODS were lack of insurance, referral from another hospital, APACHE II score, and development of shock (model discrimination AUROC, 0.91 [0.85-0.96]). Fetal-neonatal survival was 45 (80%) of 56 among uninsured vs 78 (96%) of 81 among insured patients (P = .003). Severe morbidity in neonates did not differ in either group (Table 6). 4. Discussion Uninsured critically ill obstetric patients (public health sector) were more severely ill on admission and had a higher rate of severe morbidity than insured patients (private health sector). Fetalneonatal loses were higher among uninsured patients, and although not statistically significant, maternal mortality was limited to the same group. Uninsured critically ill obstetric patients were younger, showed a higher prevalence of comorbidity and higher severity scores on admission, and were more frequently referred from other hospitals
Table 2 Obstetric history of uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients Total
Uninsured (public)
Insured (private)
P
32 ± 8 112/122 (92) 92/97 (95) 2 (1-4) 50/122 (41) 17/118 (14.5) 31/119 (26)
31 ± 8 38/48 (84) 25/29 (86) 2 (1-4) 19/43 (44) 9/40 (22.5) 7/41 (17)
33 ± 8 74/74 (100) 67/68 (98.5) 2 (1-3) 31/79 (39) 8/78 (10) 24/78 (31)
.07 .000 .027 .15 .59 .07 .10
Patients with known route of delivery Cesarean 111/140 (78) Vaginal delivery 10/140 (7)) Discharged from ICU 14/140 (10) pregnant
44/53 (83) 4/53 (7.5) 5/53 (9.5)
67/87 (77) 6/87 (7) 9/87 (10)
.39 1 1
Gestational agea Prenatal careb Standard prenatal carec Parity Previous abortions Induced abortions Spontaneous abortions
Data are presented as mean ± SD, median (IQR), or n (%). a Gestational age on admission to ICU or at the end of the current pregnancy (weeks). b At least 1 prenatal visit. c At least 5 prenatal visits in a term pregnancy [12].
201
Table 3 Causes of admission to ICU for uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
Obstetric Hypertensive disease of pregnancy Severe preeclampsia HELLP Eclampsia Eclampsia + HELLP Gestational hypertension Abortion Obstetric hemorrhage Postoperative period after cesarean ± hysterectomy for placenta accreta (scheduled) Ectopic pregnancy Puerperal sepsis Intrauterine fetal death Nonobstetric Sepsis Severe pneumonia Peritonitis Upper urinary tract infection Neurologic Subarachnoid hemorrhage Seizures (no eclampsia) Postoperative period after cesarean for underlying diseases Respiratory failure Others
Uninsured (public)
Insured (private)
P
42 (67) 21 (33) 1 7 10 3 0 5 (8)a 7 (11) 6 (9.5)
73 (83) 42 (48) 23 5 11 1 2 1 (1)b 24 (27) 1 (1)
.021 .07
0 1 (1.5) 2 (3) 21 (33) 12 (19) 10 2 0 2 (3) 2 0 3 (5)
4 (4.5) 0 1 (1) 15 (17) 6 (7) 3 2 1 2 (2) 0 2 1 (1)
.14
0 4 (6)
.08 .016 .02
.022
NS
3 (3.5) 3 (3.5)
Data are shown as n (%). a All were septic abortion. b Spontaneous abortion.
compared with insured patients. These findings were similar to the Indian patients in the study of Munnur et al [9] and to the Indonesian public sector patients in the study of Adisasmita et al [7]. The cost of transport to health centers, staff attitudes toward women with lower incomes, and poor communication skills may be some of the barriers to access proper and timely care for uninsured patients, resulting in greater illness severity on admission [22]. Once patients were admitted to the ICU, uninsured critically ill obstetric patients from the public sector showed a higher incidence of severe morbidity, such as MODS or ARDS, leading to a higher level of intervention, compared with insured patients from the private sector. In contrast, the level of intervention for obstetric patients from the private sector was higher compared with those patients from the public sector in a population-based study in Australia [8]. Other
Table 4 Level of intervention and complications for uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
n TISS28 MV MV (d) Pulmonary artery catheter Shock Type of shock Hypovolemic Septic Hypovolemic and septic Other ARDS MODS Central line
Uninsured (public)
Insured (private)
63 25 ± 9 27 (43) 7 (2-13) 2 (3.6) 22 (37)
88 20 ± 9 12 (14) 1 (0.5-4) 3 (3) 21 (24)
10 (45) 8 (36) 3 (14) 1 (5) 18/59 (30.5) 32 (54) 39 (70)
19 (90)a 1 (5) 0 (0) 1 (5) 2/88 (2) 9 (10) 23 (26)
P .001 .000 .003 1 .092 .036
.000 .000 .000
Data are shown as mean ± SD, median (IQR), or n (%). a Post hoc analysis to specifically evaluate where the difference was: hypovolemic vs septic shock, P = .042.
202
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
Table 5 Multivariate analysis for variables associated with MODS among critically ill obstetric patients Variable
Odds ratio (95% confidence interval)
P
Lack of insurance Referral from another hospital APACHE II Shock
6.75 (2.17-20.09) 11.43 (1.86-70.20) 1.30 (1.13-1.49) 4.82 (1.54-15.06)
.001 .007 .000 .007
Variables were adjusted by age, prenatal care, appropriate prenatal care, nonobstetric cause of admission, obstetric cause of admission, presence of comorbidity, hospital LOS, TISS28, use of mechanical ventilation, and ARDS (model discrimination AUROC, 0.91 [0.85-0.96]).
studies have shown a lower level of intervention and higher mortality among uninsured than among insured critically ill patients [5,6,23]. Although the Indian patients in the study of Munnur et al [9] were more severely ill than US patients, they received fewer invasive procedures and showed higher mortality. The same was true for patients from the public sector in the study of Adisasmita et al [7]. Hospital and prenatal care are free in Argentina, so affordability was not a barrier to receiving required procedures once the patient was admitted to hospital or the ICU. In this study, 92% of the 151 patients received some type of prenatal care, indicating a clear improvement over a previous retrospective study from Argentina [20], possibly a result of recent national health programs to diminish maternal mortality. However, uninsured patients still received less frequent prenatal care than insured, similar to Indian patients in the study of Munnur et al [20]. In Argentina, abortion is illegal except in specific cases outlined by law, such as when the mother's life is in jeopardy or when pregnancy is the result of rape in a mentally incapable female [24]. Nevertheless, the incidence of induced abortion did not differ significantly in either health sector. What was significant was the fact that there were no cases of septic abortion in the insured patients. This situation might indicate that although abortion is undertaken by women from both groups, it takes place in worse conditions for low-income, lesseducated, and uninsured women from the public sector, and therefore, complications are more frequent. Similar results have been reported in an Indonesian study, where induced abortion is also illegal and septic abortion was more common in patients admitted to hospitals in the public health sector [7]. Causes of admission to the ICU were classified as obstetric and nonobstetric. In both health sectors, at least two thirds of patients were admitted for obstetric reasons, consistent with most reports [25]. However, nonobstetric causes of admission were more frequent among uninsured (public health sector) than insured (private health sector) patients (33% vs 17%, respectively), which might be associated with a higher prevalence of comorbidity [26] and greater illness severity [11,27]. The most common obstetric causes of admission were hypertensive disease of pregnancy and obstetric hemorrhage, as has been reported throughout the world [25].
Table 6 Characteristics, level of intervention, and outcomes of fetuses/neonates from uninsured (public health sector) vs insured (private health sector) critically ill obstetric patients
Fetal/neonatal survival Hospital-LOS (d) Neonatal ICU Appropriate birth weight for gestational age MV CPAP RDS
Uninsured (public)
Insured (private)
P
45/56 14 21/38 34/37
78/81 (96) 9 (5-23.5) 38/66 (58) 54/64 (84)
.003 .138 .82 .36
11/38 (29) 15/37 (40.5) 18/78 (23)
.05 .78 .05
(80) (6-30) (55) (92)
11/20 (55) 8/18 (44) 19/45 (42)
Data are presented as n (%) and median (IQR). CPAP indicates continuous positive airway pressure; RDS, respiratory distress syndrome.
In terms of severe morbidity, ARDS was more frequent among uninsured patients. Shock developed equally in both sectors, but septic shock was more common in the public sector and hypovolemic shock in the private sector, the latter associated with a higher incidence of obstetric hemorrhage. The lower incidence of hypovolemic shock in the public sector was predictable, as the public hospital is a referral center for placenta accreta management. Renal dysfunction was more common in uninsured patients as well as among Indian patients in the study of Munnur et al [7]. This is the first study indicating a higher incidence of renal dysfunction in pregnant/ postpartum patients using creatinine cut-off levels adjusted to pregnancy, instead of the commonly used scores for general ICU population, which clearly underestimate this condition. The development of MODS is considered a measure of severe morbidity, and its risk factors were investigated. Variables independently associated with the development of MODS were APACHE II score, shock, referral from another hospital, and lack of health insurance. Higher APACHE II score and the development of shock as risk factors for MODS have well-known biological plausibility. On the other hand, the association between referral from other centers and MODS may be related to delay in obtaining proper care [28]. Lack of health insurance was an independent risk factor for multiple-organ dysfunction. Although this might be related, in part, to the high proportion of referred patients among the uninsured, it does not seem to be the only explanation. Lack of prenatal care, a key factor for reducing maternal mortality, comorbidity, and MODS, was not associated in the multivariate model. Our hypothesis was that the high prenatal coverage (~ 90%) precluded finding an association given our sample size. As Zeeman [28] has addressed, maternal outcomes might be adversely affected by inadequate use of prenatal care and a delay more than 24 hours between onset of illness and ICU admission. The latter may be related to a delay in the decision to seek care, in accessing services and in the provision of appropriate care once the health facility is reached [31]. In this study, the higher incidence of MODS among uninsured patients does not seem to be related to differential management in each sector, as the 2 hospitals offered similar services. Moreover, the level of intervention was even higher among uninsured patients. These patients were sicker upon admission to the ICU, probably due to a delay in deciding to seek care, which may be related to sociocultural factors, and to accessibility problems, such as distance and transport cost or availability. One factor possibly affecting the decision to seek care could be maternal education, which is inversely related to maternal mortality, even after adjusting for marital status, economic status, institutional capacity, or health care investment [29,30]. A higher level of education improves a woman's ability to perceive signs and symptoms in her own body and to seek care and inquire about personal health. Moreover, education may help women overcome communication barriers, gain self-confidence, and feel empowered to make correct decisions. Although gestational age did not differ between groups, fetalneonatal loss was higher among uninsured patients. This may be related to the higher prevalence of comorbidity in this population as well as the greater severity of their condition upon admission and the higher incidence of severe morbidity. However, severe morbidity and interventions among surviving neonates were similar in both groups. Thus, we might hypothesize that the severity of maternal illness upon admission determines fetal-neonatal mortality but might not influence the progression of the disease in those surviving. Several limitations of this study need to be noted. This was an observational study and, as such, prone to bias and confounding. We performed a multivariate analysis to monitor for possible confounders. The study only represents women with access to a health care facility and who were admitted to an ICU; however, most deliveries and maternal deaths occur within the health care system in
D.N. Vasquez et al. / Journal of Critical Care 29 (2014) 199–203
Argentina [2]. This research was carried out in private and public centers representing both high- and low-income populations, respectively. Finally, other potentially important variables, such as maternal education or other barriers that marginalized women experience when accessing care, were not measured. 5. Conclusions In summary, uninsured critically ill obstetric patients were more severely ill on admission and displayed higher levels of severe morbidity than insured patients. Similarly, fetal-neonatal losses were more frequent among uninsured patients. Risk factors for developing MODS were APACHE II score, shock, referral from another hospital, and lack of insurance. However, some level of prenatal care, a measurement that has been shown to have great impact on reducing maternal mortality, was received by nearly 90% of patients. Therefore, other preventive measures, such as improving social conditions, particularly regarding maternal education and women's empowerment as well as free universal access to safe legal abortions, should be implemented to reduce severe maternal morbidity. Acknowledgments This study was supported by the Argentinean National Ministry of Health and the Argentinean Society of Critical Care. The funding sources had no role in study design, data collection, data analysis, data interpretation, writing of the report, and decision to submit the manuscript for publication. The English in this document has been checked by at least 2 professional editors, both native speakers of English. For a certificate, please see: http://www.textcheck.com/certificate/RJhRYm. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jcrc.2013.11.010.
References [1] Lozano R, Wang H, Foreman KJ, et al. Progress towards Millennium Development Goals 4 and 5 on maternal and child mortality: an updated systematic analysis. Lancet 2011;378:1139–65. [2] Argentinean Ministry of Health - Pan American Health Organization. Basic Indicators: Argentina 2010; 2011 [http://wwwmsalgovar and http://wwwopsar]. [3] Pan American Organization of Health. Health in the Americas. Argentina: Scientific and Technical Publication No 622; 2007 39–61. [4] INDEC. Argentinian National Institute of Statistics and Censuses. http//www. censo2010.indec.gov.ar/resultadosdefinitivos_totalpais.asp.2010; 2010. [5] Fowler RA, Noyahr LA, Thornton JD, et al. An official American Thoracic Society systematic review: the association between health insurance status and access, care delivery, and outcomes for patients who are critically ill. Am J Respir Crit Care Med 2011;181:1003–11.
203
[6] Lyon SM, Benson NM, Cooke CR, et al. The effect of insurance status on mortality and procedural use in critically ill patients. Am J Respir Crit Care Med 2011;184: 809–15. [7] Adisasmita A, Deviany PE, Nandiaty F, et al. Obstetric near miss and deaths in public and private hospitals in Indonesia. BMC Pregnancy Childbirth 2008;8:10. [8] Roberts CL, Tracy S, Peat B. Rates for obstetric intervention among private and public patients in Australia: population based descriptive study. BMJ 2000;321: 137–41. [9] Munnur U, Karnad DR, Bandi VD, et al. Critically ill obstetric patients in an American and an Indian public hospital: comparison of case-mix, organ dysfunction, intensive care requirements, and outcomes. Intensive Care Med 2005;31: 1087–94. [10] Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [11] Karnad DR, Lapsia V, Krishnan A, et al. Prognostic factors in obstetric patients admitted to an Indian intensive care unit. Crit Care Med 2004;32:1294–9. [12] Schwarcz RUA, Lomuto C, Martinez I, Galimberti D, Garcia OM, Etcheverry ME, Queiruga M. Prenatal care. Practice guidelines for the preconception and prenatal care. Argentina: National Ministry of Health; 2001 1–39 [Available at: http:// www.msal.gov.ar]. [13] Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818–29. [14] Vincent JL, de Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on “sepsis-related problems” of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793–800. [15] Malstam J, Lind L. Therapeutic intervention scoring system (TISS)—a method for measuring workload and calculating costs in the ICU. Acta Anaesthesiol Scand 1992;36:758–63. [16] Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818–24. [17] Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101:1644–55. [18] Oyelese Y, Scorza WE, Mastrolia R, et al. Postpartum hemorrhage. Obstet Gynecol Clin North Am 2007;34:421–41. [19] WHO. International Statistical Classification of Diseases and Related Health Problems, 10th revision. Geneve, Switzerland: World Health Organization; 1992. [20] Vasquez DN, Estenssoro E, Canales HS, et al. Clinical characteristics and outcomes of obstetric patients requiring ICU admission. Chest 2007;131:718–24. [21] Yeomans ER, Gilstrap III LC. Physiologic changes in pregnancy and their impact on critical care. Crit Care Med 2005;33:S256–8. [22] Silal SP, Penn-Kekana L, Harris B, et al. Exploring inequalities in access to and use of maternal health services in South Africa. BMC Health Serv Res 2012;12:120. [23] Danis M, Linde-Zwirble WT, Astor A, et al. How does lack of insurance affect use of intensive care? A population-based study. Crit Care Med 2006;34:2043–8. [24] Codigo penal de la nación Argentina. http://www.codigopenalonline.com.ar/ codigo_penal_argentino_delitos_contra_las_personas.html. [25] Pollock W, Rose L, Dennis CL. Pregnant and postpartum admissions to the intensive care unit: a systematic review. Intensive Care Med 2010;36:1465–74. [26] Lombaard H, Pattinson RC. Underlying medical conditions. Best Pract Res Clin Obstet Gynaecol 2008;22:847–64. [27] Mhyre JM, Bateman BT, Leffert LR. Influence of patient comorbidities on the risk of near-miss maternal morbidity or mortality. Anesthesiology 2011;115:963–72. [28] Zeeman GG. Obstetric critical care: a blueprint for improved outcomes. Crit Care Med 2006;34:S208–14. [29] Filippi V, Ronsmans C, Campbell OM, et al. Maternal health in poor countries: the broader context and a call for action. Lancet 2006;368:1535–41. [30] Karlsen S, Say L, Souza JP, et al. The relationship between maternal education and mortality among women giving birth in health care institutions: analysis of the cross sectional WHO Global Survey on Maternal and Perinatal Health. BMC Public Health 2011;11:606. [31] Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Soc Sci Med 1994;38:1091–110.
