Adverse events, or poor patient outcomes resulting from medical care,1 are common. Studies from several countries,2–6 including Canada,7,8 indicate that 2.7–12.5% of hospitalized patients experience an adverse event. Although these studies evaluated a broad range of hospitalized patients, relatively little is known about adverse events in obstetric patients. The paucity of data describing adverse events in obstetrics reflects exclusion of obstetric patients in some studies7 and low numbers of events in others.2–6,8 These low event rates would be reassuring were it not for the exceedingly high costs of litigation in obstetrics,9 the potential for a small but distressing subset of adverse events to produce catastrophic harm to mother or child,10 and the profound distress that the health care team often experiences when their patients experience complications.10–12 Such issues necessitate more research evaluating obstetric adverse events.
There are several challenges to conducting adverse event research in obstetrics. The first relates to the rarity of these events. Most adverse event research focuses on poor outcomes.2–8 Because there are relatively few poor obstetric outcomes, there are few opportunities to investigate them. A second challenge is the discordance between what is documented in the chart and what actually occurs. Clinicians may not describe poor outcomes in the medical record13; and even if they do, there may be insufficient information recorded to allow an investigator to determine the cause of a poor outcome. The third challenge is the anxiety created by the high risk of litigation. Obstetricians are more likely to be sued than other physician types (Canadian Medical Protective Association. Risk of litigation in Canada. Personal communication, September 29, 2005). Because fear of litigation is one barrier to reporting errors, obstetrics will be particularly affected by this issue.
We designed this study to estimate the risk of experiencing an adverse event or a potential adverse event for women admitted to the labor and delivery unit. To meet this objective, we designed and implemented a prospective method of adverse event surveillance that addresses the challenges described above.
PARTICIPANTS AND METHODS
We modified an existing method of adverse event measurement for obstetric care.13,14 A trained observer based in the labor and delivery unit monitored the unit for poor patient outcomes, procedural errors, and unsafe working conditions. When the observer identified such an occurrence, she performed a standard investigation in a timely manner. A multidisciplinary team reviewed the case to determine clinical importance.
Because we did not focus exclusively on outcomes but also processes of care and working conditions, we were able to identify so-called potential adverse events, which are defined as errors that have the potential to cause harm.14 In the context of obstetrics, potential adverse events are especially important because often the only difference between them and actual adverse events is good fortune15 and because they occur much more frequently than actual adverse events.14,15 Furthermore, by maintaining a presence in the labor and delivery unit, we hoped to remove some of the anxiety related to reporting errors and circumvent the problems related to poor documentation.
We used a prospective cohort study design. We included consecutive patients admitted to the labor and delivery unit of the Ottawa Hospital, General Campus, between July 1 and August 12, 2005. We excluded patients if they were admitted to the labor and delivery unit as a result of overflow from our triage unit or if they were admitted directly into the operating room for a dilation and curettage procedure. Patients and their infants were followed until hospital discharge. The Ottawa Hospital Research Ethics Board approved the protocol.
The study hospital is the only tertiary care center in Eastern Ontario, Canada. The labor and delivery unit at the General Campus is a level III unit that has 13 beds, three operating rooms, and an affiliated neonatal intensive care unit (NICU). In-house medical staff includes an on-call attending obstetrician, an obstetrics resident, the on-call family medicine resident, and one medical student. The nurse-to-patient ratio is 1:1. As a level III unit, it performs high-risk as well as normal deliveries. Ten percent of the 3,500 deliveries performed annually on the unit are high risk. Postpartum patients receive care on the Mother-Baby Unit, which is physically adjacent to the labor and delivery unit.
We defined an adverse outcome as the development of new symptoms, physical signs, or laboratory test abnormalities. An adverse event was defined as an adverse outcome due to health care management as opposed to progression of natural disease. A preventable adverse event represents an adverse event judged to be avoidable by means available in routine practice. We defined a defective process as a procedural error. Potential adverse events are defective processes that have a high likelihood of causing harm. We considered the composite end point of adverse events and potential adverse events as quality problems.
We generated a list of adverse outcomes and defective processes that we used to identify adverse events and potential adverse events. These trigger events prompted a thorough assessment of a patient’s care and a peer review. During peer review, a team of physicians and a risk manager rated cases as adverse events, preventable adverse events, and potential adverse events.
We performed a literature review to identify previously published risk factors for adverse obstetric outcomes or indicators for poor quality obstetric care.16–25 A multidisciplinary group consisting of three obstetricians, a clinical risk manger, a clinical epidemiologist, and two obstetrics nurse managers reviewed this list to determine each trigger’s face validity and to suggest additional triggers.
It is important to note that these triggers were meant to capture clinical events for subsequent review. As such, they did not represent outcomes on their own; rather they were indicative of possible safety concerns that required further expert assessment. Many of the triggers used cut-points with somewhat arbitrary values (such as, decision-to-delivery time of more than 30 minutes for an emergency cesarean delivery). To ensure that our study was clinically relevant yet inclusive of all important events, we used a large number of triggers, and we relied on the clinical judgment of our experts to rate how important the occurrences of these triggers were to the outcome of a particular case. We achieved consensus for 72 triggers (Appendix 1).
A trained observer performed all data collection between Monday and Friday, from 8:30 am to 4:30 pm. Standard information was collected on all patients admitted to the labor and delivery unit. These data included important baseline information such as demographics, comorbidities, and indication for admission; delivery outcome data such as duration of labor, birth weight, and Apgar scores; and hospital data such as hospital length of stay.
Clinical surveillance consisted of three activities designed to identify any patient experiencing a trigger event. We designed these activities to account for the inability of our observer to be in all places at all times and to be on the unit 24 hours a day. First, the observer reviewed medical records for all patients currently admitted to the labor and delivery unit and those admitted to the postpartum ward. Second, the observer also spent considerable portions of each day stationed at the front desks of the labor and delivery unit, the NICU, and the postpartum ward. While in these locations, the observer asked care facilitators and health care providers if any triggers had occurred. Third, the observer reviewed any spontaneously reported incidents that providers generated using the hospital incident reporting system. To help the observer and the care providers recall all of the triggers, she carried a list with her, and we posted lists of the triggers in prominent locations in the unit.
If the observer identified a trigger, she then collected standard information. This included a full description of the trigger, the response to the trigger by the health care team, the impact that the trigger event had on the patient, whether the trigger itself or its cause was documented in the medical record, and whether the trigger event resolved with health care intervention. This information was collated along with the baseline patient data to create case report forms. We collected all patient and trigger data using a database (Pendragon Forms 3.0; Pendragon Software, Libertyville, IL) stored on a handheld personal digital assistant.
On a weekly basis, a peer review team evaluated case report forms. The review team consisted of a patient safety researcher or hospitalist (A.J.F.), two obstetricians (S.C., L.O.), one obstetric clinical risk manager (C.B.), and the clinical observer (I.F.). For each case, the clinical observer first presented the data from the case report form. The group discussed the case while emphasizing particular aspects of the care that could impact subsequent ratings. The reviewers then assessed whether the trigger was an adverse outcome (eg, trauma to the newborn or postpartum hemorrhage) or defective process of care (eg, physician not following a standard protocol or staff unavailability). Next, they decided whether they had enough information to further classify the case as an adverse event, potential adverse event, or neither. When they required further information, they deferred classifying the case to a future meeting so that this information could be collected. An example of the type of information needing further collection was a patient’s fetal heart rate monitor strip.
If the group decided that there was sufficient information to classify the case, then they performed ratings to characterize adverse outcomes as adverse events and preventable adverse events and to determine whether deficient processes represented potential adverse events. Each member of the group was encouraged to provide his or her rating and the rationale for the rating, and then a final rating was achieved through consensus. The group performed ratings for these determinations using a previously published and widely used scale.2–5,7,8 Adverse outcomes were rated an adverse event if reviewers agreed that the outcome was more likely to have been caused by treatment than disease, a preventable adverse event if they agreed that the adverse event was more likely to have been caused by an error, and a potential adverse event if they agreed that the defective process was likely to cause harm. The scales used by the reviewers to make their judgments are reproduced in Appendix 1. Note that in determining whether an adverse event was preventable, we asked the reviewers to consider such factors as whether it appeared that standard protocols were followed, whether the care provider appeared to identify all risks in a timely manner and mitigate them appropriately, and whether there were any obvious errors in performance. If there was an inability to achieve consensus regarding their ratings, then the outcome was judged to be due to disease and the procedure was judged to be unexpected to cause harm.
We report our outcomes in terms of the risk per encounter and the rate per patient day. We calculated risk using counts of events as the numerator and total number of encounters as the denominator. For risk, we calculated 95% confidence intervals using the Wilson score method.26 We calculated rates using counts of events as the numerator and length of stay in days as the denominator. For rates, we calculated 95% confidence intervals assuming a Poisson distribution.
We described the population characteristics using median and interquartile range for continuous variables and frequency for categorical variables. We tested for an association among clinical factors and important quality problems. For univariable analyses, we used the Wilcoxon rank sum test for continuous variables and a χ2 test for categorical variables. For multivariable analyses, we used a multivariable Poisson regression model.27 We selected variables for this model using standard methods.27 We used SAS 9.0 (SAS Institute, Cary, NC) for all analyses.
We studied 425 patient encounters (Table 1). Patients were, on average, 31 years of age and presented at term. Twelve percent of patients had a previous cesarean delivery, and virtually all of them received prenatal care in the first trimester. Most patients had a singleton pregnancy, and the majority had been previously pregnant. Forty-seven percent of patients were admitted in labor, with the rest being admitted for an assessment of a potential adverse pregnancy outcome (such as bleeding, decreasing fetal heart rate, or hypertension), ruptured membranes, induction, or an elective cesarean delivery. A total of 372 (88%) patients were delivered during the index encounter. Most infants were delivered vaginally (73%) and had good Apgar scores. The average birth weight was 3,340 g (interquartile range 2,848–3,706 g). Deliveries occurred evenly distributed between night and day. The median hospital length of stay for all patients, whether or not they delivered, was 2 days (interquartile range 2–3 days).
Our clinical surveillance program identified 110 separate trigger events (risk 26%, 95% confidence interval [CI] 22–30%). Triggers were approximately evenly distributed between regular hours and “off hours,” with 63 (57%) occurring on weekdays between 8:30 am and 4:30 pm and the remaining 47 events occurring at other times. Only 25 of the triggers (23%) were documented in the medical record.
The types of problems identified are described in Table 2. The most common triggers were “system problems” (n=41, 37%). These included such events as unavailable staff, protocol violations, and cancellations of cesarean deliveries due to operating room unavailability. “Maternal events” were the next most common type of trigger (n=37, 33%). These included extreme hypertension, fever, and placental abruption. “Fetal events” (n=16, 15%) and problems related to “interventions” (n=16, 15%) occurred less frequently. Examples of fetal events included admissions to neonatal intensive care (either unplanned or in a newborn delivered more than 36 weeks of gestation) and birth trauma, whereas examples of intervention-related triggers included stat cesarean delivery and T-incisions during cesarean delivery.
Of the 110 trigger events, 67 (61%) were rated adverse patient outcomes. Nine of these were determined to be a result of medical care (overall adverse event risk 2%, 95% CI 1–4%, overall adverse event rate 0.8 events per 100 patient days, 95% CI 0.4–1.4 events per 100 patient days), including six that were considered due to errors (overall preventable adverse event risk 1%, 95% CI 0–3%, overall preventable adverse event rate 0.5 per 100 patient days, 95% CI 0.2–1.0 event per 100 patient days).
Forty-three trigger events were rated adverse care processes. Fourteen were determined to be errors that had the potential to cause harm (overall potential adverse event risk 3%, 95% CI 2–5%, overall potential adverse event rate 1.3 per 100 patient days, 95% CI 0.7–2.0 events per 100 patient days). Thus, in total, 23 patients (overall important quality problem risk 5%, 95% CI 4–8%, overall important quality problem rate 2.1 events per 100 patient days, 95% CI 1.3–3.0 events per 100 patient days) experienced an adverse event or a potential adverse event.
Of the patients with adverse events, most had transient symptoms or temporary disabilities that were not expected to result in permanent functional limitation. Examples of preventable adverse events included trauma to the newborn, emotional distress in parents as a result of a failure to perform autopsy, and delays in treatment resulting in worsening newborn illness. Examples of nonpreventable adverse events included a third-degree perineal tear after an instrumental delivery and a postlumbar puncture headache. Examples of potential adverse events include conflicts among staff, patient care responsibilities exceeding the capacity of staff (either physician or nurse) to provide adequate level of care, mislabeled or lost laboratory specimens, and protocol violations. Of particular note, we found two incidents that affected health providers rather than patients. These were considered potential adverse events because the reviewers felt that in these cases the problem impaired the provider’s ability to function during critical aspects of the patient’s care. A brief description of every adverse event and potential adverse event is listed in Appendix 2.
We evaluated the probability that each type of trigger was rated an adverse event or potential adverse event (Table 2). Overall, 23 of 110 triggers were rated as adverse events or potential adverse events (P=21%). Adverse fetal outcomes (n=16, P=38%) and system problems (n=41, P=34%) were most likely to be rated as a quality problem, whereas interventional (n=16, P=6%) and maternal outcomes (n=37, P=5%) were less likely to be rated as such. We also assessed the probability that specific types of trigger were classified as quality problems, although these probabilities must be interpreted with extreme caution given the small numbers of events experienced. The following triggers were most likely rated as a quality problem: birth trauma (n=2, P=100%), admission to NICU (n=4, P=50%), equipment failure (n=2, P=50%), delay after call for help (n=2, P=50%), communication error (n=2, P=50%), protocol violation (n=5, P=40%), staff unavailable (n=8, P=50%), and third- or fourth-degree perineal tear (n=2, P=50%).
By univariable analyses, we found that there appeared to be an increased likelihood of experiencing a quality problem for older patients and women who were nulliparous, although these associations were not statistically significant at a P<.05 level. While simultaneously adjusting for these factors in a Poisson regression model, we found that maternal age and nulliparity were significantly associated with risk of experiencing a quality problem (rate ratio 2.9, 95% CI 1.0–8.5 and rate ratio 2.4, 95% CI 1.0–5.4, respectively)(Table 3).
Our work confirms previous research demonstrating that obstetric care is associated with a low risk of severe adverse event.2–5,7,8 Overall, 2% of patients experienced an adverse event. All of the adverse events we identified had relatively minor impact, and no patient experienced permanent disability or death as a result of medical care, despite our evaluation of teaching hospital patients, generally considered a high-risk population.7 Based on this sample of 425 patients in which we observed no severe events, the upper end of the 95% confidence limit for the risk of severe adverse event in obstetrics patients is 0.7%.28 Despite these encouraging findings, we did observe that 5% of patients experienced an important quality problem (ie, an adverse event or potential adverse event). These occurrences concern us because many had the potential to cause severe harm to the mother or baby or both. It is even more worrying that 87% of these quality problems were due to errors (preventable adverse events and potential adverse events). System problems such as poor team work, protocol violations, and staff unavailability appear to be the most important types of problems, whereas technical proficiency and therapeutic decision making seem relatively less important.
These findings have three important implications. First, they suggest that strategies to improve patient safety on the obstetric ward will need to be different from interventions elsewhere. Given the types of quality problems we identified, interventions that focus exclusively on system issues, such as team work, communication skills during high-stress situations, and work flow (eg, laboratory specimen processing), are likely very important to improving safety in obstetrics. On the other hand, safety interventions such as physician order entry with decision support will likely not be helpful, given the lack of adverse drug events.29 In fact, this particular intervention might actually cause problems by increasing the complexity of the work flow. Similarly, although technical skills are obviously very important and should be the primary focus of undergraduate and graduate training, additional interventions to improve these clinical skills may not be particularly helpful from the patient safety perspective.
Second, traditional methods of adverse event detection, like structured chart reviews,16 will be ineffective in monitoring safety or guiding safety interventions in the obstetrics service. Three quarters of the trigger events that we investigated were not even documented. Therefore, risk management personnel will have a difficult time identifying quality problems, let alone investigating them, if the medical record is the source of information for their analyses. Our process of clinical surveillance has three advantages. Firstly, we found a large number of clinically relevant poor outcomes and defective processes of care. Secondly, we had a standard method of investigating triggers in a timely manner. This ensures that all involved parties had good recollection of the event. Lastly, we had a standard method of rating the case, which promotes objectivity in ratings. We did not experience any resistance to our presence in the labor unit. We did not identify any changes in behavior or attitudes as a result of our being on the ward although we did not specifically measure this. Future studies will evaluate this.
Third, efforts to manage risk cannot focus solely on severe events. As we have shown, these events are rare. Thus, they are unlikely to be representative of the true quality problems existing on the ward. Although some could argue that the problems we identified were of little clinical significance and therefore do not warrant urgent attention, we disagree. For example, we found three problems related to inappropriate handling of stillbirths. Although these events had no direct impact on the physical health of the patients, they did cause substantial psychological impact. In fact, similar events have occurred elsewhere and attracted widespread attention, including a case in England in which the Prime Minister publicly expressed his condolences to the parents of a deceased child whose body had been mistakenly sent to the laundry and could not be found for several hours.30
Our study’s conclusions must be interpreted cautiously because our methods had limitations. The study was conducted at a single center, and it had a relatively small sample size. Thus, the low adverse event rate we observed may be a result of chance or bias. Our center may be expected to have a higher rate of adverse events given that it is a referral center and a teaching hospital.7 On the other hand, the obstetrics service we studied has a proactive risk management program and has been very involved in the MoreOB program (Society of Obstetricians and Gynecologists of Canada), which is designed specifically to improve patient safety.31 These factors might suggest that we would have a lower adverse event rate than other institutions, assuming that risk management programs are effective. A larger, multicenter study would shed more light on this issue. Another potential limitation is that active surveillance occurred only during normal working hours, suggesting that we may have underestimated the true adverse event rate. However, 43% of the triggers occurred at times other than the periods of active surveillances. Even though we were not physically present on the ward 24 hours a day, we found that our presence on the ward prompted clinicians to provide us information on quality problems that they experienced at other times. A third limitation is the subjectivity of the review process, which some investigators have criticized.32 We tried to overcome this by having a multidisciplinary review team and by presenting standardized, anonymous case summaries. Finally, the fact that we identified nulliparity and age as risk factors for quality problems may suggest that our method is unable to separate quality issues from underlying clinical problems because these are risk factors for poor clinical outcomes. An alternative explanation for this finding is that the higher risk of quality problems exists simply because more care is delivered in high-risk patients. A similar relationship is found in studies of adverse drug events, in which patients prescribed more medications have a higher risk of adverse drug events.14 Future studies with larger samples are needed to explore this issue.
In conclusion, we found that 2% of obstetrics patients are affected by adverse events, and another 3% of patients had important errors occur that had the potential to cause harm. These data confirm that efforts to improve safety through team building and improved communication strategies are appropriate targets for quality improvement. We suggest that in-depth studies of work flow using human factors analysis may be a fruitful next step in designing such interventions.33 Finally, our methods can be reproduced by other institutions interested in managing their risk. This will likely be more informative than relying on traditional methods of collecting information on adverse events.
1. Institute of Medicine (US). To err is human: building a safer health system. Washington, DC: National Academy Press, 2000.
2. Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, et al. Incidence of adverse events and negligence in hospitalized patients: results of the Harvard Medical Practice Study I. N Engl J Med 1991;324:370–6.
3. Thomas EJ, Studdert DM, Burstin HR, Orav EJ, Zeena T, Williams EJ, et al. Incidence and types of adverse events and negligent care in Utah and Colorado. Med Care 2000;38:261–71.
4. Vincent C, Neale G, Woloshynowych M. Adverse events in British hospitals: preliminary retrospective record review [published erratum appears in BMJ 2001;322:1395]. BMJ 2001;322:517–9.
5. Wilson RM, Runciman WB, Gibberd RW, Harrison BT, Newby L, Hamilton JD. The Quality in Australian Health Care Study. Med J Aust 1995;163:458–71.
6. Davis P, Lay-Yee R, Briant R, Ali W, Scott A, Schug S. Adverse events in New Zealand public hospitals I: occurrence and impact. N Z Med J 2002;115:U271.
7. Baker GR, Norton PG, Flintoft V, Blais R, Brown A, Cox J, et al. The Canadian Adverse Events Study: the incidence of adverse events among hospital patients in Canada. CMAJ 2004;170:1678–86.
8. Forster AJ, Asmis TR, Clark HD, Al Saied G, Code CC, Caughey SC, et al. Ottawa Hospital Patient Safety Study: incidence and timing of adverse events in patients admitted to a Canadian teaching hospital. CMAJ 2004;170:1235–40.
9. Griffin LP, Heland KV, Esser L, Jones S. Overview of the 1996 Professional Liability Survey. Obstet Gynecol Surv 1999;54:77–80.
10. Sachs BP. A 38-year-old woman with fetal loss and hysterectomy. JAMA 2005;294:833–40.
11. Wu AW. Medical error: the second victim. The doctor who makes the mistake needs help too. BMJ 2000;320:726–7.
12. Aasland OG, Forde R. Impact of feeling responsible for adverse events on doctors’ personal and professional lives: the importance of being open to criticism from colleagues. Qual Saf Health Care 2005;14:13–7.
13. Andrews LB, Stocking C, Krizek T, Gottlieb L, Krizek C, Vargish T, et al. An alternative strategy for studying adverse events in medical care. Lancet 1997;349:309–13.
14. Bates DW, Cullen DJ, Laird N, Petersen LA, Small SD, Servi D, et al. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group. JAMA 1995;274:29–34.
15. Bates DW, Boyle DL, Vander Vliet MB, Schneider J, Leape L. Relationship between medication errors and adverse drug events. J Gen Intern Med 1995;10:199–205.
16. Kaushal R. Using chart review to screen for medication errors and adverse drug events. Am J Health Syst Pharm 2002;59:2323–5.
17. Drife J. Quality measures for the emergency obstetrics and gynaecology services. J R Soc Med 2001;94 suppl:16–9.
18. Luckas M, Walkinshaw S. Risk management on the labour ward. Hosp Med 2001;62:751–6.
19. Michel P, Quenon JL, de Sarasqueta AM, Scemama O. Comparison of three methods for estimating rates of adverse events and rates of preventable adverse events in acute care hospitals. BMJ 2004;328:199.
20. Robbins D. Incident report analysis: the experience of one large labor and delivery unit. J Perinat Neonatal Nurs 1987;1:9–18.
21. Holden DA, Quin M, Holden DP. Clinical risk management in obstetrics. Curr Opin Obstet Gynecol 2004;16:137–42.
22. Lakasing L, Spencer JA. Care management problems on the labour ward: 5 years’ experience of clinical risk management. J Obstet Gynaecol 2002;22:470–6.
23. Mantel GD, Buchmann E, Rees H, Pattinson RC. Severe acute maternal morbidity: a pilot study of a definition for a near-miss. Br J Obstet Gynaecol 1998;105:985–90.
24. Ashcroft B, Elstein M, Boreham N, Holm S. Prospective semistructured observational study to identify risk attributable to staff deployment, training, and updating opportunities for midwives. BMJ 2003;327:584.
25. Stanhope N, Crowley-Murphy M, Vincent C, O’Connor AM, Taylor-Adams SE. An evaluation of adverse incident reporting. J Eval Clin Pract 1999;5:5–12.
26. Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998;17:857–72.
27. Stokes ME, Davis CS, Koch GG. Poisson regression. In: Categorical data analysis using the SAS system. 2nd ed. Cary (NC): SAS Institute Inc. 2000. p. 347–62.
28. van Belle G. Sample size. In: Balding DJ, Bloomfield P, Cressie NAC, Fisher NI, Johnstone IM, Kadane JB, et al, editors. Statistical rules of thumb. New York (NY): John Wiley & Sons; 2002. p. 29–51.
29. Kaushal R, Shojania KG, Bates DW. Effects of computerized physician order entry and clinical decision support systems on medication safety: a systematic review. Arch Intern Med 2003;163:1409–16.
30. Lyall S. New British hospital scandal: dead baby put in hamper. New York Times, January 31, 2002.
32. Hayward RA, Hofer TP. Estimating hospital deaths due to medical errors: preventability is in the eye of the reviewer. JAMA 2001;286:415–20.
33. Christian CK, Gustafson ML, Roth EM, Sheridan TB, Gandhi TK, Dwyer K, et al. A prospective study of patient safety in the operating room. Surgery 2006;139:159–73.
APPENDIX 1: TRIGGERS USED IN THIS STUDY
Prolonged second-stage labor (4-h primapara, 3-h multipara with epidural, 3-h primipara with epidural, 2-h multipara without epidural)
Major placental abruption
Hypertension greater than 180/110 or requiring intravenous treatment
Hypotension with systolic pressure less than 90
Admission to intensive care unit
Postpartum length of stay more than 3 days for vaginal delivery, more than 5 days for a cesarean delivery
Fever (temperture more than 38.5°C)
Blood loss of more than 500 mL for vaginal birth, more than 1,000 mL for cesarean delivery
Disseminated intravascular coagulation
Cardiorespiratory arrest, including suspicion of amniotic fluid pulmonary embolus
Subarachnoid or intracerebral hemorrhage
Intubation and ventilation other than for general anesthesia
Pulmonary edema necessitating furosemide or intubation
Platelets less than 100×109
Death or coma
Other maternal trauma of concern
Major patient complaints
Delayed arrival of anesthesia
Use of a general anesthesia
Anesthesia problem (inadequate amount, awareness, hypotension, difficult intubation)
Hemorrhage requiring transfusion
Failed interventional vaginal delivery followed by a cesarean delivery
Complete anal sphincter or rectal mucosal tear
Other soft tissue damage (bladder, ureter)
Unplanned removal, injury or repair of organs during operative procedure
Stat cesarean delivery
Return to operating room after delivery
Use of more than one instrument in a vaginal delivery
Forceps or vacuum use protocol violations (eg, three pop-offs)
Vaginal breech delivery
Retained or lost swab or instrument
Manual removal of placenta
Decision-to-delivery time of more than 30 minutes for an emergency cesarean delivery
System or Environmental Factors
Missing or incorrect blood work on charts
Outpatient assessment unit or inpatient charts unavailable
Failure of medication delivery or order implementation
Delay after call for assistance
Interpersonal conflict over case management
Failure to use continuous electronic fetal monitoring
Failure to respond to nonreassuring fetal heart rate
Transfers from other units or hospitals
Delivery outside of birthing unit
Cancellation of booked operation procedure (cesarean delivery or dilation and curettage)
Patient injury unrelated to procedure or medications
Fetal or Neonatal
Neonatal death or stillbirth at 24 weeks or later
Cord pH less than 7.1 OR base excess less than –12 AND 5-minute Apgar less than 5
Admission to neonatal intensive care unit (NICU) after 36 weeks
Unplanned admission to NICU
Unexpected multisystem organ dysfunction
Unexpected respiratory distress syndrome
Other: Determination of Adverse Events and Potential Adverse Event Situations by Physicians
Each incident was determined to be either an adverse outcome or an adverse process. If deemed an outcome, a 6-point scale was used to determine the degree to which the outcome was influenced by medical management:
- It was definitely due to the patient’s underlying condition
- It was most likely due to the patient’s underlying condition
- It was a close call but more likely due to the underlying condition
- It was a close call but more likely due to the medical management
- It was most likely due to the medical management
- It was definitely due to the medical management
If an outcome received a rating of 4 or more, it was deemed an adverse event and further rated by a 6-point scale to determine how likely the outcome was caused by an error:
- Definitely not due to error
- More than likely not due to error
- It was a close call but more likely not due to error
- It was a close call but more likely due to error
- More than likely due to error
- Definitely due to error
If the incident was deemed an adverse process, a potential adverse event rating was given using a 6-point scale to determine its potential for causing harm:
- Definitely would not cause harm
- More than likely would not cause harm
- It was a close call but more likely would not cause harm
- It was a close call but likely could cause harm
- More than likely could cause harm
- Definitely could cause harm
APPENDIX 2: DESCRIPTION OF QUALITY PROBLEMS
Preventable Adverse Events as Rated by Review Team
Case 1. Inappropriate response to postpartum hemorrhage leading to increased morbidity.
Case 2. A neonate with an important illness missed having blood work drawn to assess his condition. This resulted in a delay in treatment.
Case 3. Full autopsy performed in setting of a stillbirth when parents had consented to a partial autopsy only. Parents experienced significant emotional distress.
Case 4. Minor trauma to infant during cesarean delivery.
Case 5. Minor trauma to infant during cesarean delivery.
Case 6. No autopsy done in setting of a stillbirth when parents had requested one. Parents experienced significant emotional distress.
Nonpreventable Adverse Events as Rated by Review Team
Case 7. Complications of epidural analgesia. Patient developed a postepidural headache.
Case 8. Delay of emergency cesarean delivery of more than 20 minutes. Indication for cesarean delivery was a nonreassuring fetal heart rate, and delay resulted in an admission to NICU.
Case 9. A patient received a third-degree perineal tear during an instrument-assisted delivery.
Potential Adverse Events as Rated by Review Team
Case 10. Infant with thick meconium was delivered and not suctioned or given to NICU staff for assessment. Stimulation nearly given by resident physician, but was stopped.
Case 11. A nurse mistakenly recorded a patient as Rh-negative in the obstetric nursing history. This error was transferred onto the delivery record and a laboratory report. The patient was given WinRho (Cangene, Winnipeg, Canada) unnecessarily.
Case 12. Newborn not assessed on day 1 by pediatrician due to clerical error.
Case 13. A resident got amniotic fluid inside his glove while performing cesarean delivery with a patient known to be infected with hepatitis C virus.*
Case 14. Resident ordered and administered a medication by intravenous route despite warnings on packaging and medication manuals that this route was associated with risk of reaction. The nurse questioned the resident, but he ignored the advice.
Case 15. Patient with oligohydramnios placed on her back for more than 1 hour during an ultrasound examination. Fetal heart monitor showed variable decelerations that resolved when the patient was placed in the appropriate position.
Case 16. A delay in physician’s response to a nurse who was concerned about a patient.
Case 17. Blood work was lost.
Case 18. A patient in premature labor was accepted from an outside facility despite the NICU being over capacity.
Case 19. Stillbirth was not handled appropriately.
Case 20. Nurse suffered a needle-stick injury.*
Case 21. Inappropriate number of nurses available for the workload in the labor and delivery unit.
Case 22. An important congenital heart defect went unnoticed until day 2 of life. Infant was hypoxic and had to be referred to a specialized pediatric hospital for management.
Case 23. Late notification of ABO blood type status delayed escalation of phototherapy.
* Event led to a temporary interruption of health provider’s capacity to perform work.