Context. Hospital stays for newborns and their mothers after uncomplicated vaginal delivery have decreased from an average of 4 days in 1970 to 1.1 days in 1995. Despite the lack of population-based research on the quality-of-care implications of this trend, federal legislation passed in 1996 mandated coverage for 48-hour hospital stays after uncomplicated vaginal delivery.
Objective. To assess the impact of very early discharge (defined as discharge on the day of birth) on the risk of infant readmission during the neonatal period in a California healthy newborn population.
Design. Retrospective cohort study, based on a linked dataset consisting of the birth certificate, newborn, and maternal hospitalization record, and linked infant readmission records for all healthy, vaginally delivered, and routinely discharged California newborns from 1992 to 1995.
Outcome Measures. Very early discharge and infant readmission during the first 28 days of life.
Results. The percentage of infants discharged very early or early (after a 1-night stay) increased from 71% in 1992 to 85% in 1995. The percentage of infants discharged very early increased from 5.0% in 1992 to 5.7% in 1993 and 7.0% in 1994, then decreased to 6.7% in 1995. Characteristics that have been previously associated with suboptimal pregnancy outcomes were found to decrease the likelihood of very early discharge, eg, maternal complications, primiparity, and Hispanic, African American, South East Asian, or other Asian race/ethnicity.
The rate of readmission in the neonatal period initially decreased from 27.6 infants per 1000 in 1992 to 25.67 infants per 1000 in 1994, then increased to 30.2 infants per 1000 in 1995. For infants discharged early, no statistically significant increase in the risk of readmission was observed, compared with infants discharged after a 2+-night stay. The adjusted odds ratio (OR) for readmission was statistically significantly higher for infants who were discharged very early, compared with infants discharged early (OR: 1.27), first order births (OR: 1.21), infants born to mothers who experienced complications (OR: 1.11), infants with Medicaid insurance (OR: 1.23), and infants born to mothers who received adequate plus prenatal care (OR: 1.15). The risk was statistically significantly lower for female infants (OR: 0.75).
The proportion of infants rehospitalized for dehydration and low-risk infections over the 4 study years combined was statistically significantly higher in infants discharged very early (4.37‰ and 10.30‰, respectively), compared with infants discharged early (3.59‰ and 8.16‰, respectively) or after a 2+-night stay (2.91‰ and 7.95‰, respectively). The proportion of infants rehospitalized for dehydration increased statistically significantly from 2.89‰ in 1992 to 4.52‰ in 1995.
Conclusions. One-night stays with adequate antenatal and postnatal care outside the hospital do not increase the risk of readmission for healthy, vaginally delivered infants born in California. However, the decision to discharge infants on the day of birth should be applied conservatively because of the increased risk of infant readmission associated with very early discharge.
- DRG =
- diagnosis-related group •
- AAP =
- American Academy of Pediatrics •
- OR =
- odds ratio •
- CI =
- confidence interval
Hospital stays for newborns and mothers after uncomplicated vaginal delivery have decreased steadily since birth in the hospital setting became common.1 In the 1950s, the recommended postpartum hospital stay was 6 days.2 In the United States, hospital stays after vaginal deliveries gradually decreased from an average of 4 days in the 1970s to ∼2 days in 1992.1 In 1995, the mean length of stay for healthy, vaginally delivered, and routinely discharged infants in California was 1.1 days.
Reducing the number of days in the hospital after birth has several advantages: 1) the risk of iatrogenic infections is decreased; 2) mother and infant recover in a familiar home environment; and 3) hospital costs of caring for the infant and mother are reduced. However, concerns have also been expressed about potential disadvantages of early discharge: 1) breastfeeding is not established until the third or later postpartum day; 2) a number of conditions do not manifest themselves until 2 or more days after the delivery; 3) time is reduced for inhospital teaching on breastfeeding, infant care, women's health needs, and family planning and for psychosocial care.3–5 Although there are many published studies on the early discharge of infants, several reviews of this literature have pointed out a number of methodological difficulties, such as self-selected populations, small sample size, lack of controls, limitation of outcomes to death and readmission, and lack of knowledge on postdischarge follow-up care.2 ,5 Braveman et al5 concluded that “the currently available literature provides little scientific evidence to guide discharge planning for most apparently well newborns and their mothers.” Several large population studies have been conducted. Lee et al6 analyzed a population of 920 554 healthy infants with birth weights above 2500 g born from 1987 to 1994 in Ontario, Canada. They found that shorter hospital stays were associated with an increased number of readmissions during the first 14 days of life and that the severity of illness of these rehospitalized infants was increased. However, this study did not control for sociodemographic or maternal characteristics, such as maternal educational attainment or prenatal care. Liu et al7 assessed the safety of newborn early discharge using linked data for the State of Washington (n = 29 034). After adjusting for birth year, gestational age, maternal race/ethnicity, payer, any complication of pregnancy, sex, and parity, they found that discharging healthy infants within 30 hours of birth increased the risk of readmission within 28 days by 22%.
The goal of our study was to analyze readmission and discharge patterns in California before the enactment of the Newborns' and Mothers' Health Protection Act of 1996, (federal legislation that requires insurance companies to cover maternal and infant hospital stays for at least 48 hours after an uncomplicated vaginal delivery and for at least 96 hours after a cesarean section delivery8) using a methodology that would overcome many of the shortcomings described previously. By studying an entire cohort of healthy, California newborns delivered vaginally between 1992 and 1995 (n = 1 214 545), the study population was sufficiently large to allow us: 1) to examine the relationship among sociodemographic, medical, and maternal payer characteristics and the likelihood of very early (defined as discharge on the same day), early (defined as discharge after a 1-night stay), or later discharge (defined as discharge after a stay of 2 or more nights); and 2) to determine the risk-adjusted likelihood of readmission after early discharge. Because our study was not limited to specific “eligible for early discharge” populations, we were able to make a population-based statement about the effect of discharge practices on the likelihood of readmission. Finally, we used a definition of early discharge that is easily reproducible.
This study was based on a linked database of birth certificate, newborn and maternal hospitalization records, and any infant readmissions within the first 28 days of life for births occurring in California acute care hospitals during 1992–1995. Deterministic and probabilistic methodologies were used to obtain the linkages.9 ,10 Of the births in California acute care hospitals, 97.9% were linked successfully to the newborn and maternal hospital discharge record. Of the readmissions within 28 days of life, 97.7% were linked successfully to the birth hospitalization.
The study population consisted of all single live births vaginally delivered in California acute care hospitals reporting to the Office of Statewide Health Planning and Development in 1992, 1993, 1994, or the first 11 months of 1995 for whom the vital statistics birth record, the infant hospital discharge record, and the maternal hospital discharge record could be linked. Furthermore, we reduced the study population to include only diagnosis-related group (DRG) 391, normal newborns, who were discharged routinely to home.
For all 4 years combined, there were >2.1 million single live births in California for which the vital statistics birth record, infant hospital discharge record, and maternal hospital discharge record were linked. Of these infants, 1.2 million (57% of all births) met the study criteria of being vaginally delivered, routinely discharged, and assigned to DRG 391.
The American Academy of Pediatrics (AAP) defines early discharge and very early discharge, respectively, as stays of 48 and 24 hours or less, after uncomplicated vaginal delivery.11 Because discharge time was not included on the hospital discharge records of the Office of Statewide Health Planning and Development, we were unable to define length of stay in terms of hours. Instead, we used the following definitions. Very early discharge was defined as discharge occurring on the same day as the infant was born. Early discharge was defined as discharge after a 1-night stay in the hospital after birth. Other discharge was defined as discharge after 2 or more nights in the hospital after birth. Note that these criteria conservatively resemble the criteria set forth by the AAP. The group termed very early discharge is a subset of those infants discharged within 24 hours of delivery. The group termed early discharge is a subset of those infants discharged within 48 hours of delivery. We did not impose an upper limit on the length of stay for those infants who were discharged after 2 or more nights, because we used the criteria described previously to obtain a group of healthy infants only.
Simple descriptive statistics were used to compare: 1) the population of infants discharged very early, discharged early, and discharged after 2 or more nights in the hospital after birth; and 2) the population of infants readmitted within the first 28 days of life and those not readmitted. The Cochran-Armitage trend test for binomial proportions was used to test the null hypothesis of a linear trend in percentages and rates.12 Logistic regression analysis was used to obtain risk-adjusted odds ratios (ORs) for the variables related to discharge timing, controlling for maternal and infant demographic characteristics.
The following maternal and infant sociodemographic variables were used in the regression analysis:
Maternal complication(s) of pregnancy, labor, or delivery (none/any).FNa
Infant gender (male/female).
Maternal age (<18, 18–19, 20 to <30, and 30 years of age or older).
Marital status (single/married).b
Race/ethnicity (non-Hispanic white, Hispanic, black, Native American, South East Asian, other Asian, and other race).
Insurance type covering infant's hospital stay (MediCal, Private Insurance, health maintenance organization/preferred provide organization, and self-pay).
Parity after index birth (primiparous/multiparous).
Birth weight (<2500 g, 2500 to <4000 g, and ≥4000 g).
Prenatal care according to Kotelchuck13(inadequate, intermediate, adequate, and adequate plus).
Place of birth of mother (foreign born/US born).
In the logistic regression model, records with missing values for any of the predictor variables (4.8% of all records) were excluded from the analysis.
We used reclassifications of up to 25 International Classification of Diseases, Ninth Revision-Clinical Modificationcodes included in the infant discharge record to establish conditions contributing to a readmission, and we analyzed these conditions by discharge timing (unpublished data). To compare differences in the prevalence of diagnoses by discharge timing and differences in rates or percentages, we used Fisher's exact test.
Trends in Early Discharge
Table 1 displays trends in discharge timing for the study population in California from 1992 to 1995. In 1992, 4.9% of all vaginally delivered, healthy, and routinely discharged infants were discharged very early (on the day of birth). This percentage increased to 5.7% in 1993 and peaked at 7.0% in 1994. The percentage of infants discharged early (after a 1-night stay) increased each year throughout the study, from 66.1% in 1992 to 78.2% in 1995. Over this 4-year period, the percentage of infants who were discharged after at least 2 nights in the hospital decreased from 28.9% to 15.1%. All trends were statistically significant (P < .001).
Discharge Timing by Maternal and Infant Sociodemographic Characteristics
Table 2 displays descriptive statistics and adjusted ORs for discharge timing by maternal and infant sociodemographic characteristics for all study years combined. Crude ORs are not shown but were similar to the adjusted ORs displayed in the table. The factors most strongly associated with an increased likelihood of very early versus later discharge were self-pay (OR: 2.71) and managed care insurance (OR: 1.78). The factors most strongly associated with a decreased likelihood of very early versus later discharge were the presence of a maternal complication (OR: .41) and MediCal insurance (OR: .631). When comparing early versus later discharges, we found, for the most part, the same effects; however, their strength was reduced.
The statistics shown in Table 2 were calculated by combining all 4 years. We also examined the relationship between discharge timing and these characteristics for each year separately. With the exception of payer source, we observed a remarkable temporal stability in the association between the sociodemographic variables and discharge timing. For infants insured by managed care plans, the adjusted ORs for early discharge, compared with later discharge decreased from 1992 to 1993 from 1.32 (lower 95% confidence limit: 1.28; upper 95% confidence limit: 1.35) to 1.19 (1.16,1.23), then increased for each subsequent year to 1.31 (1.25,1.38). The adjusted ORs of very early discharge compared with later discharge steadily increased over the 4 years from 1.19 (1.13,1.26) in 1992 to 1.81 (1.66,1.98) in 1995. At the same time, the percentage of infants enrolled in a managed care plan increased from 31.1% in 1992 to 43.2% in 1995 (32.5% in 1993; 34.5% in 1994). For self-pay infants, the adjusted OR for early compared with later discharge decreased from 1.95 (1.86,2.04) in 1992 to 1.04 (.96,1.12) in 1995. The adjusted OR for very early compared with later discharge decreased from 3.57 (3.30,3.86) in 1992 to 1.64 (1.45,1.85) in 1995.
Trend in Readmission Rates
During the study, readmission rates initially decreased from 27.6 infants per 1000 births in 1992 to 25.7 infants in 1994 (Table 3). However, in 1995, the readmission rate increased to 30.2 infants per 1000. The trend was statistically significant using the Cochran-Armitage test for trends in binary proportions.
Readmissions by Maternal and Infant Sociodemographic Characteristics
For each of the years, 1992–1995, we evaluated the risk of readmission in the neonatal period associated with the timing of discharge and maternal and infant sociodemographic characteristics. We show the results for 1995 in Table 4 and point out differences to the other study years below.
For each year studied, we found the same set of factors to affect the likelihood of readmission in the neonatal period. The risk factor most strongly associated with reducing an infant's likelihood of readmission was self-pay insurance status (OR: .74). The morbidity gender differential translated into readmission rates as well; the risk of readmission for female infants was 75% of the risk for male infants.
The factor most strongly associated with increasing an infant's risk of readmission was infant birth weight. The risk of readmission for low birth weight infants was 53% higher than that of normal birth weight infants. Compared with non-Hispanic white race, other Asian race increased the likelihood of readmission by 38%. Compared with later discharge, very early discharge increased the probability of readmission by 30%. Medicaid insurance status (OR: 1.23), primiparity (OR: 1.21), and the presence of maternal complications (OR: 1.11) were also associated with an increased risk of readmission. No change in readmission risk was found for infants discharged early, compared with infants discharged later, or for newborns enrolled in a managed care plan.
The level of prenatal care received by the mother was also found to affect the risk of readmission. Although infants born to mothers receiving either inadequate or intermediate levels of prenatal care showed a decreased risk of readmission (95%), infants born to mothers who received adequate plus prenatal care had a 15% higher chance of being readmitted.
Finally, we observed an important trend in the adjusted OR of readmission for very early discharges, compared with later discharges; it declined from 1.55 (95% confidence interval [CI]: 1.41,1.70) in 1992, to 1.49 (CI: 1.35,1.64) in 1993, to 1.46 (CI: 1.32,1.61) in 1994, to 1.30 (CI: 1.18, 1.43) in 1995. This suggests that very early discharge becomes less of a risk for hospital readmission over time. (The details of time trends for all variables that are reported inTable 4 are available upon request.)
Prevalence of Diagnoses Leading to Readmission
Table 5 displays the diagnostic breakdown for readmission rates for all infants and by discharge timing. Jaundice was the condition most often diagnosed as a cause contributing to readmission (12.2‰). Although there was no difference in the prevalence of jaundice between very early and later discharges (13.08‰ vs 12.35‰), infants discharged very early were more likely to be rehospitalized with jaundice as a contributing diagnosis than infants discharged early (13.08‰ vs 12.14‰; Fisher's exact test,P ≤ .01). Dehydration was a factor in 3.49‰ of readmissions. The chance of being rehospitalized with a diagnosis of dehydration was statistically significantly higher for infants discharged very early (4.37‰), compared with infants discharged early (3.59‰) or later (2.91‰; P ≤ .01).
Congenital lower bowel obstructions, left-sided obstructive cardiac defects, right-sided obstructive cardiac defects, low-risk infections, and other causes of readmission were statistically significantly more likely to be contributing factors in readmissions for infants discharged very early (.52‰, .49‰, .44‰, 10.30‰, and 11.27‰, respectively), compared with infants discharged early (.21‰, .23‰, .18‰, 8.16‰, and 6.32‰, respectively) or infants discharged later (.20‰, .20‰, .13‰, 7.95‰, and 6.10‰, respectively).FNc
No statistically significant differences by discharge timing were found in the prevalence of moderate-risk infections—defined as infections with a mortality rate between 5% and 15%—or high-risk infection—defined as infections with a mortality rate of >15%.
We found a remarkable trend in the prevalence of dehydration among rehospitalized infants over the 4-year period (Table 6). In 1992, 2.99‰ of infants were rehospitalized with a contributing diagnosis of dehydration, 3.0‰ in 1993, 3.6‰ in 1994, and 4.5‰ in 1995. This trend was statistically significant (P ≤ .01) for all infants, as well as for infants who were discharged early or later.
California experienced a significant increase in the percentage of apparently healthy infants discharged very early or early in its healthy newborn population between 1992 and 1995 (Table 1). Although 29% of newborns were discharged after 2 or more nights in 1992, this percentage decreased to 15% in 1995.
An analysis of the sociodemographic factors associated with very early and early discharge indicated that those factors that were previously associated with a higher likelihood of adverse birth outcomes were found to decrease the likelihood of a same-day discharge. This result reflects positively on current discharge practices.
One of the most important factors contributing to the sharp rise in very early and early discharges was the increased presence of managed care in California in 1995. The percentage of infants enrolled in a managed care plan increased from 31% in 1992 to 43% in 1995. At the same time, our study found that enrollment in a managed care plan after controlling for other sociodemographic characteristics increased the likelihood of very early or early discharge (adjusted OR: 1.7 and 1.5, respectively). Furthermore, this likelihood of a very early discharge for newborns enrolled in a managed care plan increased each year throughout the study (1992, adjusted OR: 1.19; 1995, adjusted OR: 1.81).
Did the increase in early and very early discharge have negative implications for newborn morbidity in our cohort of healthy infants? We evaluated readmissions in the neonatal period to assess infant well-being. After a downward trend in the readmission rate during the first 3 years (1992, 27.6‰; 1993, 26.6‰; and 1994, 25.7‰), the overall rate of readmission jumped between 1994 and 1995 (30.2‰).
In a logistic regression analysis adjusting for sociodemographic factors, we found that for each study year, very early discharge was associated with an increased likelihood of readmission, compared with later discharge (adjusted OR: 1.30), whereas early discharge did not lead to an increased likelihood of readmission. The magnitude of the effect of very early discharge timing on the risk of readmission decreased from 1992 (adjusted OR: 1.55) to 1995 (adjusted OR: 1.30). The reduction in the OR of readmission for the very early discharge group may indicate increased experience among physicians in selecting infants eligible for very early discharge. In contrast, to some extent the inflated likelihood of readmission might be a result of selection-bias: a doctor is more likely to readmit a newborn if he/she was discharged very early. This notion is supported by the finding that mothers who received adequate plus prenatal care were also more likely to have their child readmitted. A diminishing effect of this selection bias over time might have contributed to a reduction in the OR.
In addition to discharge timing, other factors found to be associated with an increased risk of readmission included low birth weight (OR: 1.53), other Asian race/ethnicity (OR: 1.38), Medicaid insurance enrollment (OR: 1.23), and primiparity (OR: 1.21). The existence of low birth weight infants in a cohort of newborns classified as apparently healthy was surprising. We compared the percentage of low birth weight infants in our study to the percentage found in the Washington study undertaken by Liu et al7 that had also restricted the study population to apparently healthy and vaginally delivered infants. The percentage of low birth weight infants in the Washington study was 2.1%, whereas the percentage of such infants in the California healthy infant population was 1.0%. The existence of low birth weight infants in the healthy population may be explained by the fact that some ethnic groups have smaller infants who are functionally mature.16–18 In our initial analyses, we also used prematurity as a predictor variable and found the effects of prematurity and low birth weight to be similar. Because gestational age had far more missing values than did birth weight, we decided to only include birth weight in our model. We also tested several interaction terms, including multiparity and inadequate prenatal care and multiparity and intermediate prenatal care, but none of these interactions were statistically significant at the .05 level for any of the years.
With regard to insurance status, the study found that managed care enrollees were no more likely to be rehospitalized than were patients enrolled in fee-for-service plans. Of some concern was the finding that the uninsured had a significantly lower risk of readmission than did fee-for-service patients (OR: .74). It is feasible that economic considerations may discourage parents from pursuing needed follow-up hospital care for an ill newborn that is uninsured. The potential danger to the newborn in such a scenario should motivate additional investigation of this issue.
The higher readmission rate of 1995 may reflect a change in the infants making up the study population, which is composed solely of vaginally delivered, healthy newborns. For instance, while the prevalence of most factors found to increase the likelihood of readmission stayed constant over time, the percentage of mothers experiencing a complication during pregnancy, labor, or delivery increased from 35.4% to 43.2%, and the percentage of mothers receiving adequate plus prenatal care increased from 19% to 23%. Therefore, the changing nature of the study population could be a factor in the overall readmission rate increase.
Our analysis showed a pattern in the prevalence of specific clinical conditions among rehospitalized infants by discharge-timing group. Except for moderate- and high-risk infections, for which a similar prevalence was observed among infants in all discharge-timing groups, infants discharged very early exhibited statistically significantly higher proportions of readmissions for jaundice, dehydration, congenital lower bowel obstruction, left-sided and right-sided cardiac defects, low-risk infections, and other causes, compared with infants discharged early or later. Our research clearly shows that discharge on an infant's birthday leads to an increased risk of readmission from most of the clinical conditions studied and strongly suggests that same day discharge is not optimal for an infant's health.
Although infants discharged very early were found more likely to require readmission attributable to a range of clinical conditions, our research also suggests the need for improved discharge planning and follow-up processes for all newborns. This is made evident by the findings around dehydration. Over the course of the study years, the prevalence of readmissions with dehydration as a contributing factor increased from 2.91‰ in 1992 to 4.53‰ in 1995. We hypothesize that this increase may be secondary to an increase in breastfeeding or a change in supplementation practices for breastfeeding mothers. Because breastfeeding can take up to 5 days to establish, dehydration will remain a potential risk as long as hospital stays remain at 3 nights or less. This would suggest the need to enhance efforts around inhospital parental education—to increase knowledge of the warning signs and risks of infant dehydration—and to ensure that new mothers have access to lactation specialists who can evaluate and facilitate progress with breastfeeding.
Our study had several limitations. Although we were able to study a complete cohort of infants rather than a select population, and although we had a large population to study, we lacked data on both inhospital and postdischarge practices that could influence the likelihood of readmission. For example, the extent of predischarge parental education to recognize and respond to the early signs of jaundice and dehydration, the availability of routine home health service visits by a trained nurse, and the facilitation of access to early neonatal health care are all practices that would be expected to decrease readmission. Lack of data to ascertain and control for these practices may, therefore, have the effect of underestimating the odds of readmission after very early and/or early discharge in our analysis.
As is common in retrospective medical outcome studies, selection bias might have affected our findings. For instance, under the current AAP guidelines, infants who are discharged very early undergo additional follow-up visits. Medical conditions are, therefore, more likely to be detected and an infant is more likely to be readmitted. It is also possible that attributable to additional follow-up, medical conditions were diagnosed in infants discharged very early and then treated on an outpatient basis. With respect to restricting the study population to infants assigned to DRG 391, selection bias might have led to an underestimation of the risk of readmission associated with very early discharge, because infants in other DRGs who were discharged very early are likely to be at higher risk of readmission.
Data source limitations also restricted the available outcome measures to readmission. Although some readmissions represent an appropriate and potentially life-saving course of action, other readmissions are potentially avoidable. Unfortunately, our dataset does not allow for this distinction. This study also left for future investigation a number of questions. For example, unique hospital characteristics may influence the risk of infant readmission. A next step will be to examine the impact of hospital-specific variables, including the average number of beds in the well-baby nursery, average census in the well-baby nursery, rural/urban location, teaching status, etc. Finally, we have not fully investigated the severity of the illnesses of rehospitalized infants. Case–mix adjustment might be applied to rehospitalized infants in terms of complexity of diagnoses, comorbidities, number of readmissions, reasons for several readmissions of the same infant, length of readmission stay, and charges. We are currently investigating these factors and analytic changes.
This study provides an important window on the quality of care implications of current discharge policies and practices. Population-based research on this topic is particularly important in light of the legislative actions that have been taken following widespread early discharge of newborns. Although the study suggests that same-day discharge is not a safe option for infants, mandates of hospital coverage for a full 48 hours may not be necessary in terms of the risk of readmission within 28 days of life. One-night hospital stays, when applied to eligible populations seem to be a safe and potentially lower cost option. However, it is important to note that the successful implementation of 1-night discharges likely depends on 2 factors: 1) that discharge timing is the independent result of a mutual decision of the physician and mother, and 2) that appropriate support mechanisms are likely to play a key role in ensuring the safety of early discharge. In the future, a greater emphasis should be placed on the application of scientific research methods for the development and refinement of guidelines around neonatal care.
We thank Thomas S. Nesbitt, MD, MPH, University of California, Davis, Departments of Telemedicine and Family and Community Medicine; Paula Braveman, MD, MPH, University of California, San Francisco, Departments of Family and Community Medicine, Epidemiology, and Biostatistics; and Susan Egerter, PhD, University of California, San Francisco, Department of Family and Community Medicine, for reviewing and commenting on this work.
- Received February 19, 1999.
- Accepted August 6, 1999.
Reprint requests to (B.D.) Health Information Solutions, 2826 McGarvey Ave, Redwood City, CA 94061. E-mail:
↵FNa The following complications were included: 640, hemorrhage in early pregnancy; 641, antepartum hemorrhage, abruptio placentae, and placenta previa; 642, hypertension complicating pregnancy, childbirth, and the puerperium; 644, early or threatened labor; 645, prolonged pregnancy; 646, other complications of pregnancy, not elsewhere classified; 647, infectious and parasitic conditions in the mother classifiable elsewhere, but complicating pregnancy, childbirth, or the puerperium; 648, other current conditions in the mother classifiable elsewhere, but complicating pregnancy, childbirth, or the puerperium; 652, malposition and malpresentation of fetus; 643, disproportion; 654, abnormality of organs and soft tissues of pelvis; 655, known or suspected fetal abnormality affecting management of mother; 656, other fetal and placental problems affecting management of mother; 657, polyhydramnios; 658, other problems associated with amniotic cavity and membranes; 659, other indications for care or intervention related to labor and delivery, not elsewhere classified; 666, postpartum hemorrhage; 670, major puerperal infection; 672, pyrexia of unknown origin during the puerperium; 673, obstetrical pulmonary embolism; and V252, sterilization.
FNb For the years considered in this study, marital status was inferred from the last names of the mothers and fathers rather than asked as a direct question on the birth certificate.
↵FNc Typical conditions in the group of other causes of readmissions were postbirth respiratory problems, fever, acute upper respiratory infection, esophagitis, other and unspecified noninfectious gastroenteritis and colitis, newborn temperature regulation disorder, congenital pyloric stenosis, urinary tract infections, convulsions, hyperpotassemia, neonatal conjunctivitis, dyspnea, diaper or napkin rash, omphalitis, esophageal disorder, and newborn transitory tachypnea.
- Centers for Disease Control and Prevention
- Patterson PK
- Britton JR,
- Britton HL,
- Beebe SA
- Braveman P,
- Egerter S,
- Pearl M,
- Marchi K,
- Miller C
- ↵Herrchen-Danielsen B, Gould JB. User Manual and Technical Report: Linkage of Vital Statistics Linked Birth/Infant Death, Infant, and Maternal Hospital Discharge File. Berkley, CA: California Health Information for Policy Project and University of California, Berkeley, School of Public Health; 1996
- ↵American Academy of Pediatrics and American College of Obstetrics and Gynecology. Postpartum and follow-up care. In: Guidelines for Perinatal Care. 3rd ed. Elk Grove Village, IL: American Academy of Pediatrics; Washington DC: American College of Obstetricians and Gynecologists; 1992:91–116
- ↵SAS Institute Inc. SAS/STAT Software: Changes and Enhancements Through Release 6.12. Cary, NC: SAS Institute Inc; 1997:235–236
- Copyright © 2000 American Academy of Pediatrics