Cesarean delivery is the most common operation in the United States, the rate having increased from 5.8% in 1970 to 32.3% in 2008.1 In private hospitals of Latin American countries, it has reached a median rate of 51%.2 In 1975 in hospitals in England, only 32% of women were discharged within 3 days of delivery and 42% had a hospital stay of 1 week or more. By 2011 90.9% were discharged within 3 days and only 1.9% had a hospital stay of 1 week or more; 62% were discharged within 2 days after an elective cesarean delivery under spinal anesthesia.3 A systematic literature review on early postpartum discharge states that hospital stays after birth have shortened over the past 30 years but that practice has contributed to the controversy about the safety of this policy for mother and newborns.4 Another review concludes that although evidence is sparse, early postpartum discharge appears safe for carefully selected and consenting patients.5 Early hospital discharge after delivery may affect breastfeeding.6-8
In our center, women were routinely informed to expect hospital discharge on postcesarean day 2 (2 days after). Recently, some of our health care providers were discharging well-motivated women on day 1 (next day).
We anticipate many women and their newborns will be physically fit for discharge on postcesarean day 1. We hypothesize that women will be equally satisfied with either day 1 or day 2 discharge and exclusive breastfeeding rates will also be similar. We performed a randomized trial to test our hypotheses of equivalence.
PATIENTS AND METHODS
The trial was conducted in a university hospital in Kuala Lumpur, Malaysia. Our hospital offered comprehensive healthcare to the public at subsidized rates or for free. Our center is accredited by the Malaysian Ministry of Health as a Baby Friendly Hospital on breastfeeding promotion based on the World Health Organization/UNICEF Baby Friendly Hospital Initiative program. Ethics approval was obtained from the University of Malaya Medical Centre Medical Ethics Committee (approval number 811.7 dated September 22, 2010). All participants gave their written consent. Recruitment was from November 5, 2010, to February 15, 2012.
Inclusion criteria were planned cesarean delivery, age 18 years or older, and gestation 37 weeks or greater with a singleton pregnancy. Exclusion criteria were grossly abnormal fetus, established medical problems, or operative factors likely to preclude day 1 discharge (eg, pregestational diabetes, preeclampsia, epilepsy, cardiac disease, renal disease, connective tissue disease, antiphospholipid syndrome, two or more previous cesarean deliveries, and major previa).
Health care providers reviewed the case notes of women admitted to the antenatal ward for a planned cesarean delivery. Potential recruits were approached, counseled, and given the patient information sheet. Eligible recruits who declined participation were recorded. Women were told that they would be randomized to day 1 (next day) or day 2 (2 days after) postcesarean discharge. Also, regardless of randomization, participant and newborn would only be discharged when both were assessed to be clinically fit by health care providers. Discharge per protocol would be predicated on predefined criteria of the cesarean delivery having been surgically uncomplicated, blood loss was 800 mL or less, postoperative hemoglobin was 80 g/L or greater, unassisted ambulation to the bathroom, established urination, tolerated at least one normal meal, afebrile (temperature less than 38°C in the last 24 hours), normal blood pressure and health care provider willingness to discharge. Neonatal discharge is predicated on standard pediatric practice in our center. Potential participants were also told they may withdraw from the study at any time without having to give a reason.
Participants were asked to answer a recruitment questionnaire where they were asked about their opinion on preferred length of hospital stay after cesarean delivery, breastfeeding intention, and the level of adult assistance at home in the few days after hospital discharge.
A sealed numbered opaque envelope containing the allocated discharge protocol was then attached to the chart of each participant. The envelope was opened on postcesarean day 1 morning ward round by the health care provider to reveal the allocation. If the participant was unsuitable for discharge after health care provider assessment or was allocated to day 2 discharge, she was informed of this. It was considered impossible to blind participants to the intervention.
The randomization sequence was generated using an online random number generator (www.random.org) in random blocks of four or eight on a 1:1 ratio. Numbered opaque envelopes were prepared containing the allocation to day 1 or day 2 discharge.
If per-protocol day 1 discharge was not carried out as allocated, the reasons were recorded. These patients were reviewed at least daily by health care providers with their hospital discharge based on prevailing clinical practice. Participants allocated to day 2 discharge were assessed for suitability for discharge on day 2, and, if they were not discharged as per protocol, they were reviewed at least daily subsequently until a discharge decision was reached. The reasons for protocol violation were recorded. If a participant declined to accept her discharge on day 1, she was treated respecting her wishes.
In our center, barring contraindications, planned cesarean deliveries were performed under spinal anesthesia. After completion of a planned cesarean delivery, 100–150 mg rectal diclofenac was given on the operating table followed by 500 mg oral mefenamic acid every 8 hours on return to ward with parenteral morphine or tramadol administered as required for breakthrough pain. Cefuroxime (three 8-hourly 750-mg doses) was given intravenously as antibiotic prophylaxis. After an uncomplicated cesarean delivery, women were allowed to drink and eat at their own pace on return to the ward. The indwelling bladder catheter was removed early the next morning and the women mobilized.
Before hospital discharge, women are instructed on wound care. Women were discharged with 500 mg mefenamic acid three times a day and 1 g paracetamol (acetaminophen) four times a day orally for 1 week. Home visit by health care professionals from our center were not on offer but our patients have free access to our clinics and acute care facilities, which were opened at all hours. An appointment was routinely given for maternal and newborn follow-up at 2 weeks after cesarean delivery. After this, another routine appointment would be given at 6 weeks postdelivery. Additional appointments may be scheduled as clinically required.
At 2 weeks and 6 weeks after their discharge, participants who attended their postcesarean follow-up appointments were assessed by self-administered questionnaires and for other outcomes as listed subsequently. Participants who did not attend follow-up were contacted by telephone within 1 week of their missed appointment to get their responses.
Predefined primary outcomes were satisfaction with their discharge protocol by means of a 5-point Likert scale (top two responses of strongly agree and agree taken as demonstrating satisfaction and other three responses taken as nonsatisfaction) assessed 2 weeks after discharge and the self-reported exclusive breastfeeding rate at 6 weeks. Secondary outcomes assessed at 2 weeks were satisfaction with discharge protocol based on the full 5-point Likert scale, general well-being score using a 10 point numerical rating scale, recommendation of their timing of discharge after cesarean delivery to a friend (5-point Likert scale), preferred length of hospital stay after cesarean delivery, newborn feeding status, maternal or newborn unscheduled medical consultation, maternal antibiotics and cesarean wound condition (observed or self-report of skin reddening around the wound, purulent wound discharge, or wound gaping of any severity). Secondary outcomes assessed at 6 weeks after discharge were general well-being score, maternal or newborn unscheduled medical consultation, maternal antibiotic, cesarean wound condition, and assessment of maternal anxiety and depression using the Hospital Anxiety and Depression scale. Participants were subsequently also contacted to establish whether readmissions within the first 6 weeks after discharge had occurred for mother or newborn.
The Hospital Anxiety and Depression scale comprised two subscales for anxiety and depression caseness with 14 stems and four possible answer choices to each stem. Each answer is scored 0–3. It is validated tool for the identification of anxiety and depression cases in antenatal patients and gynecology clinic patients.9 The subscale score was summated, and a threshold of 7/8 was used to differentiate between caseness and noncases, respectively.10 the Hospital Anxiety and Depression scale is particularly apt to investigate potential psychological distress after cesarean delivery because there are no questions on somatic symptoms.
A direct study comparing postcesarean discharge on day 1 compared with day 2 was not available to guide sample size calculation. We based our sample size calculation on the assumptions that 90% of patients are satisfied with the standard day 2 discharge after a planned cesarean delivery. We took a margin of 10% as acceptable for demonstrating equivalence. Applying alpha 0.05 and 80% power, 155 women are required in each arm (calculator available on http://www.sealedenvelope.com/power_binary_equivalence.php). Factoring in a 10% increase in sample size to account for postrandomization dropout resulting from nonadherence to discharge protocol because of clinical factors and a further 5% sample size increase to accommodate postdischarge dropout, 180 women are required in each arm for a suitably powered equivalence trial.
Data were entered into SPSS 17. Primary analysis was by intention to treat. As a result of expected postrandomization exclusions, dropouts, and crossover, we also planned a per-protocol analysis to demonstrate efficacy. Analysis was with Fisher’s exact test for 2×2 categorical data, χ2 for appropriate larger categorical data sets, Student’s t test for normally distributed continuous data, and Mann-Whitney U test for ordinal or nonnormally distributed data. Normality of data distribution was checked with the Kolmogorov-Smirnov test. All tests were two-sided and P<.05 was considered significant.
The recruitment flow chart of participants through the trial is shown on Figure 1. Three hundred sixty women were recruited as planned and their randomization envelopes were opened (179 and 181 women were randomized to day 1 or day 2 discharges, respectively). All eligible women who we approached agreed to participate. The total number of women admitted for planned cesarean delivery over the trial recruitment period was extracted from ward records. We did not keep records that differentiated between those women who were ineligible after scrutiny of her chart and those whose health care providers did not approach. As a result of criteria infringements or loss to follow-up postrecruitment, 170 and 172 women, respectively, were available for intention-to-treat analysis. There was further attrition to 142 and 148 women available for discharge on allocated day (per protocol) analysis as a consequence of participants’ lack of adherence to allocated discharge protocol or because either mother or newborn was determined not to be suitable for discharge on allocated protocol days.
Per-protocol discharge on day 1 was not achieved in 28 of 170 (16.5%). In 14 of 170 (8.2%), the mothers were unsuitable for discharge: eight as a result of excessive blood losses greater than 800 mL or need for blood transfusion and six as a result of operative or recovery issues involving bowel injury (one), adherent placenta (one), extensive dissection for adhesions with a surgical drain left in situ (one), and excessive abdominal distension on day 1 (three). In 5 of 170 (2.9%), the newborns were not suitable for discharge as a result of glucose-6-phosphatase deficiency (three), neonatal jaundice (one), and newborn not been observed to have passed urine (one). In another nine cases, although both mother and newborn were fit, day 1 discharge was declined, a decision we accepted without question. All nine were discharged on day 2. Postcesarean day 1 discharge seemed to be feasible and well accepted.
No woman declined day 2 discharge when offered as allocated. Per-protocol discharge on day 2 was not achieved in 24 of 172 (14.0%). Only 1 of 172 (0.6%) was the result of maternal reason (excessive blood loss), 11 of 172 (6.4%) for neonate issues (neonatal jaundice, nine; glucose-6-phosphatase deficiency, one; and special care nursery observation, one), and 12 of 172 (7.0%) requested day 1 discharge and were discharged on day 1 as requested after they were determined to be fit for discharge. Of the nine neonates kept in because of neonatal jaundice, four eventually received phototherapy and six were subsequently discharged on day 3, two on day 4, and one on day 7.
Table 1 shows the characteristics and recruitment questionnaire responses of all recruits randomized. The participants were not different in their basic demographics, indications for planned cesarean delivery, perioperative findings, and recruitment questionnaire responses. All the participants received spinal anesthesia for their cesarean delivery. Randomization envelopes were all opened on postcesarean day 1 to reveal the allocated intervention.
Table 2 shows the primary outcomes analyses according to intention-to-treat analysis or per-protocol analysis. Whether according to intention-to-treat or per-protocol analysis, satisfaction (predefined as strongly agree or agree responses of a 5-point Likert scale) with discharge protocol when assessed at 2 weeks after discharge was similar across day 1 compared with day 2 arms: 148 of 170 (87.1%) compared with 147 of 172 (85.5%) (relative risk [RR] 1.1 95% confidence interval [CI] 0.6–2.1 P=.75 [intention to treat] and 127 of 142 (89.4%) compared with 128 of 148 (86.5%) (RR 1.3 95% CI 0.6–2.7 P=.48 [per protocol]). Equivalence between the interventions on this outcome was thus demonstrated. Assessed at 2 weeks after hospital discharge, 86 of 172 (50.6%) of participants allocated to day 1 discharge expressed strong agreement with the statement of satisfaction with day 1 discharge, whereas only 57 of 172 (33.1%) of participants allocated to day 2 expressed strong agreement with day 2 discharge (χ2P=.008 [intention-to-treat analysis]) and similarly 80 of 142 (56.3%) compared with 51 of 148 (34.5%; P=.001 [per protocol]). This finding demonstrated a greater passion for day 1 discharge among those who had experienced it. Self-reported exclusive breastfeeding at 6 weeks after hospital discharge were also equivalent: 76 of 170 (44.7%) compared with 77 of 172 (44.9%) (RR 1.0, 95% CI 0.7–1.5, P=.99 [intention to treat]) and 63 of 142 (44.4%) compared with 67 of 148 (45.3%) (RR 1.0, 95% CI 0.6–1.5, P=.91 [per protocol]) for day 1 compared with day 2 discharges, respectively.
Table 3 displays the secondary outcomes of the trial analyzed according to intention to treat and Table 4 the same secondary outcomes analyzed as per-protocol discharge. At 2 weeks postdischarge, there was no difference demonstrated for maternal general well-being score, 5-point Likert score recommendation of discharge protocol to a friend, self-reported neonate feeding status, maternal or neonate unscheduled medical consultation, maternal antibiotic use, and cesarean wound complaints. The only difference at 2 weeks postdischarge was in the opinion elicited about ideal length of hospital stay after an uncomplicated cesarean delivery: 94 of 170 (55.3%) of women allocated day 1 discharge expressed preference for day 1 discharge, whereas 106 of 172 (61.6%) of women allocated day 2 discharge expressed preference for day 2 discharge (intention-to-treat basis) and on a per-protocol basis, the result was similar in 87 of 142 (61.3%) allocated day 1 expressed preference for day 1 discharge and 94 of 148 (63.5%) allocated day 2 expressed preference for day 2 discharge. In our view, this difference reflects trial participants’ satisfaction with their recent personal experiences.
At 6 weeks after discharge, there was also no difference demonstrated for general well-being score, self-reported neonate feeding status, maternal or neonate unscheduled medical consultation (and their indications for consultation), maternal and neonate hospital readmissions, maternal antibiotic use, cesarean wound complaints, Hospital Anxiety and Depression scale crude anxiety score or anxiety caseness (anxiety score 8 or greater) rate, or depression caseness (depression score 8 or greater) rate. Hospital Anxiety and Depression scale depression caseness score was significantly different across trial arms but the score value was very low (median [interquartile range, full range]: 0 [0–2, 0–8] compared with 0 [0–1, 0–21] P=.025 [intention-to-treat analysis] and 1 [0–2, 0–8] compared with 0 [0–2, 0–21] P=.014 [per-protocol analysis] for day 1 compared with day 2 discharge, respectively, favoring day 2 discharge). These median scores are well below that considered to be qualified for depression caseness10 and the difference in crude Hospital Anxiety and Depression scale depression subcomponent score, although statistically significant, is not of clinical relevance.
In our trial, satisfaction with either day 1 or day 2 postcesarean discharge when elicited soon after discharge at 2 weeks was similarly high with over 85% satisfied with their discharge timing. A previous trial has found greater satisfaction in the postnatal care in the first 10 days postpartum by the early discharge group (at 6–36 hours compared with 48–72 hours postpartum) after vaginal delivery.11 Participants’ stated preference for day 1 discharge fell from 203 of 347 (58.5%) to 154 of 347 (44.4%; P<.001) from the time of their recruitment to at 2 weeks after their discharge. These interesting observations of high stated satisfaction rates with discharge protocols and a drop in preference for day 1 discharge after the event suggests a degree of ambivalence but wide acceptance to next-day discharge in our trial population.
We did not find any effect on reported neonate breastfeeding between the trial arms. A Swedish national study also finds no correlation between length of postpartum stay and duration of breastfeeding.12 It has been demonstrated that although peripartum factors may contribute to early lactation failure, long-term success of breastfeeding was predominantly determined outside the hospital.13 Our participants were generally well supported after discharge with 85% expecting at least an adult to be available at home to assist. In our newly urbanized community, new mothers often returned to their hometowns for confinement and to get support from their own mothers.14 These factors might have played a significant role on breastfeeding outcomes.
There was no difference in unscheduled medical consultation and rehospitalization rates at 6 weeks for either mother or their neonates across the trial arms. It has been shown that one-night stays for healthy, vaginally delivered neonates do not increase the risk of readmission but that day-of-birth discharge increased the risk of neonatal readmission.15 In contrast, early (less than 30 hours after birth) discharge after vaginal birth has been shown to be associated with an increased risk of readmission for jaundice, dehydration, and sepsis.16 Early hospital discharge of women delivered by unplanned cesarean using the model of nurse specialist transitional home care has been shown to be safe, feasible, and cost-effective.17 Early (on the second day after cesarean delivery) discharge was acceptable to women, reduced hospital stay, and was not associated with increased maternal morbidity.18
At 6 weeks after discharge in our trial, the anxiety caseness rates using the Hospital Anxiety and Depression scale ranged from 20% to 24%, whereas depression caseness rates were remarkably low at 2% and similar across the trial arms. In an Australian study, length of postpartum stay was also not associated with test scores indicating probable depression.19
It has been shown for medical patients that a hospital stay carries a 5.5% risk of an adverse drug reaction, 17.6% risk of infection, and 3.1% risk of ulcer for an average episode, and each additional night in hospital increases the risk by 0.5% for adverse drug reactions, 1.6% for infections, and 0.5% for ulcers.20 In our healthy population of women after planned cesarean delivery, day 1 or day 2 discharge had no effect on subsequent maternal antibiotic use, additional medical consults, or cesarean wound issues.
Day 2 discharge protocol violation resulting from neonatal jaundice occurred in 9 of 181 (5%); four of these nine neonates had phototherapy. With day 1 discharge and without equivalent vigilance with home or outpatient neonatal monitoring, it is possible that treatment for physiologic jaundice might be delayed or even missed altogether. However, we are not aware of major issues arising from this situation; day 1 or even same-day discharge had been standard for many years after vaginal birth in our center so this is not a new concern.
There are limitations to the trial. As a result of higher-than-expected protocol violations, particularly those arising from participant initiated crossover, we did not achieve the targeted sample size of 155 women in each arm on the per-protocol discharge efficacy analysis. However, given the primary outcome satisfaction rates of 127 of 142 (89.4%) compared with 128 of 148 (86.5%) for day 1 compared with day 2 from per-protocol analysis, further recruitment to reach 155 women in each arm would be futile because the finding of equivalence will not be altered. The outcome measures in the trial were largely self-reported and were not independently corroborated. However, the outcome measures were uncontroversial and there was no reason for the participants to be disingenuous in their responses. The generalizability of our finding on postcesarean day 1 discharge is limited for outcomes like breastfeeding in which prolonged home support by grandmothers can be expected to play a significant role. However, it is noteworthy that although our women did not receive routine home support by health care professionals, they still expressed high satisfaction with very early postcesarean discharge.
There is a significant and apparently realistic patient demand for next-day hospital discharge after a planned cesarean delivery in our population. Day 1 compared with day 2 discharge is associated with equivalent high patient satisfaction, subsequent exclusive breastfeeding rates, and no increase in unscheduled medical consultation for mother or neonate. Women admitted for a planned cesarean delivery should be informed that next-day discharge is feasible for most and this discharge timing could be considered after appropriate postoperative assessment.
1. Blanchette H. The rising cesarean delivery rate in America: what are the consequences? Obstet Gynecol 2011;118:687–90.
2. Villar J, Valladares E, Wojdyla D, Zavaleta N, Carroli G, Velazco A, et al.. Caesarean delivery rates and pregnancy outcomes: the 2005 WHO global survey on maternal and perinatal health in Latin America. Lancet 2006;367:1819–29.
4. Jonguitud-Aguilar A, Tomasso G, Cafferatta ML. Early post-partum discharge: systematic review of the literature [in Spanish]. Ginecol Obstet Mex 2003;71:143–51.
5. Grullon KE, Grimes DA. The safety of early postpartum discharge: a review and critique. Obstet Gynecol 1997;90:860–5.
6. Brown S, Lumley J. Reasons to stay, reasons to go: results of an Australian population-based survey. Birth 1997;24:148–58.
7. Heck KE, Schoendorf KC, Chavez GF, Braveman P. Does postpartum length of stay affect breastfeeding duration? A population-based study. Birth 2003;30:153–9.
8. De Lathouwer S, Lionet C, Lansac J, Body G, Perrotin F. Predictive factors of early cessation of breastfeeding. A prospective study in a university hospital. Eur J Obstet Gynecol Reprod Biol 2004;117:169–73.
9. Abiodun OA. A validity study of the Hospital Anxiety and Depression Scale in general hospital units and a community sample in Nigeria. Br J Psychiatry 1994;165:669–72.
10. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70.
11. Gagnon AJ, Edgar L, Kramer MS, Papageorgiou A, Waghorn K, Klein MC. A randomized trial of a program of early postpartum discharge with nurse visitation. Am J Obstet Gynecol 1997;176:205–11.
12. Waldenstrom U, Aarts C. Duration of breastfeeding and breastfeeding problems in relation to length of postpartum stay: a longitudinal cohort study of a national Swedish sample. Acta Paediatr 2004;93:669–76.
13. Sievers E, Haase S, Oldigs HD, Schaub J. The impact of peripartum factors on the onset and duration of lactation. Biol Neonate 2003;83:246–52.
14. Tan PC, Mubarak S, Omar SZ. Absorbable versus nonabsorbable sutures for subcuticular skin closure of a transverse suprapubic incision. Int J Gynaecol Obstet 2008;103:179–81.
15. Danielsen B, Castles AG, Damberg CL, Gould JB. Newborn discharge timing and readmissions: California, 1992–1995. Pediatrics 2000;106:31–9.
16. Liu LL, Clemens CJ, Shay DK, Davis RL, Novack AH. The safety of newborn early discharge. The Washington State experience. JAMA 1997;278:293–8.
17. Brooten D, Roncoli M, Finkler S, Arnold L, Cohen A, Mennuti M. A randomized trial of early hospital discharge and home follow-up of women having cesarean birth. Obstet Gynecol 1994;84:832–8.
18. Bossert R, Rayburn WF, Stanley JR, Coleman F, Mirabile CL Jr. Early postpartum discharge at a university hospital. Outcome analysis. J Reprod Med 2001;46:39–43.
19. Brown S, Lumley J, Small R. Early obstetric discharge: does it make a difference to health outcomes? Paediatr Perinat Epidemiol 1998;12:49–71.
20. Hauck K, Zhao X. How dangerous is a day in hospital? A model of adverse events and length of stay for medical inpatients. Med Care 2011;49:1068–75.