Objective To assess parental preferences for medical error disclosure and evaluate associated factors.
Design Prospective survey.
Setting Hospital for Sick Children, Toronto, Canada.
Participants Parents of inpatients and outpatients.
Main exposure Anonymous questionnaire administered on May to August 2006, surveying demographic characteristics and identifying parents' thresholds for disclosure using a vignette with six levels of harm.
Main outcome measures Preferred thresholds for parent and patient disclosure and associated factors.
Results 99% of 431 parents (181 inpatients, 250 outpatients) wanted disclosure if there was potential or actual harm versus 77% if there was none (p<0.0001). Most parents (71% vs 41%) wanted their child similarly informed (p<0.0001). Parental age, education, experience with error and child's age did not affect preferences for disclosure. Parents of inpatients (p=0.03, OR 1.65, 95% CI 1.04 to 2.62) and those born in Asia (vs North America) had a lower disclosure threshold (p=0.014, OR 2.4, 95% CI 1.2 to 4.9), and administering the survey with increasing harm had a higher disclosure threshold (p<0.0001, OR 2.46; 95% CI 1.58 to 3.83).
Experience with error (p=0.05, OR 1.5, 95% CI 1 to 2.2) and child age (eg, ≤6 years vs ≥ 11 years (p<0.0001, OR 2.74, 95% CI 1.73 to 4.33)) directly affected preferences for informing the child. Asian parents had a lower threshold for informing the child than North American parents.
Conclusions Most parents want disclosure and want their child informed of errors with harm. While parental birth country, experience with error and patient age influenced parents' desire for disclosure to their child, the details of disclosure warrant study.
Statistics from Altmetric.com
In the wake of an adverse event (AE) due to medical error, there is an imperative for open and transparent communication or disclosure.1 In the USA and Canada, the duty to disclose is integral to the American and Canadian Medical Associations' code of ethics; it is often legislated and institutional requirements for disclosure and a disclosure policy are directed by the respective accrediting bodies.2,3 However, despite studies and testimonials reinforcing patients' desires to hear about errors in their own, or in their family members' care, particularly when harm has resulted,4,–,8 disclosure is not yet the norm.
The key elements of disclosure are based on studies in adults that have identified that following a medical error, patients want: (1) an explicit statement that an error occurred, (2) to be told what the error was, (3) to be told why the error occurred, (4) to know what will be done to prevent a recurrence and (5) an apology.9 Litt le work has addressed the issue of disclosure in a paediatric setting, where the frequent and integral role played by a substitute or proxy decision maker might have an impact.10 A survey of paediatricians' attitudes and experiences about disclosure of error found that while most endorsed disclosing errors with AEs to the families, fewer than half had actually disclosed serious AEs.11 Recently, Loren et al reported the results of surveying paediatricians on how and in what circumstances they would disclose errors to parents.12 Overall, just over half the respondents (109/205, 53%) said they would disclose a serious error and usually with full details about what transpired (108/109, 99%). Disclosure was more likely if the error and AE were readily apparent. In contrast, the minority of paediatricians said they would apologise (53/209, 26%).
What is already known on this topic
Parents want to be told when an adverse event happens to their child in hospital.
Parents have a low threshold for disclosure of near misses.
What this study adds
Parents often want their child told of the adverse event.
There may be cultural differences to the threshold for disclosure.
Notwithstanding these paediatricians' varied responses to hypothetical scenarios, the scant data available in paediatrics suggest that parents want to know when an error has occurred. In one report, 500 parents of children seeking care in the emergency department for illnesses which weren't life-threatening were surveyed about error disclosure; 99% of respondents wanted full and timely disclosure irrespective of the severity of the event, and the desire for disclosure was not affected by parental age, gender, race, ethnicity or insurance status.7 No further published studies have examined the needs of parents in these situations, nor are there published reports on parents' attitudes about disclosure to the affected child.
The purpose of this study was to assess parents' or substitute decision makers' attitudes toward disclosure of medical error related to the care of their child and examine factors influencing these attitudes.
Setting and population
The Hospital for Sick Children is a 300-bed quaternary care academic paediatric health sciences centre in Toronto, Canada. From May to August 2006, Monday through Friday, all parents of general paediatric patients in hospital and in the outpatient clinic were assessed for inclusion in the study. Parents of inpatients were eligible if their child had been hospitalised for at least 3 days. Parents were excluded if their English comprehension and language ability interfered with conducting the interview as determined by the research assistant; if they were foster care givers; if they were unavailable when the research assistant came by; and/or if the responsible physician felt that administration of the questionnaire would be ‘inappropriate’ for various reasons, for example, fragile mental or emotional state. The protocol was approved by the Research Ethics Board of the Hospital for Sick Children.
After obtaining informed consent, a two-part questionnaire was administered to one parent, even in cases where both were present. The tool was first piloted with six parents and modified accordingly. Part 1 of the tool explored the respondent's sociodemography as well as their previous experience with medical error (personal or close family or friend). Countries of birth were amalgamated by continent or region. Chronic disease was defined as one that was multiorgan and/or the patient was on at least three medications (excluding antibiotics) and/or at least two hospital services were involved in an ongoing fashion in their care. Part 2 was adapted from a published vignette13 and consisted of a clinical vignette with six possible outcomes, characterised by increasing severity of the resultant AEs (Appendix A). The parent was told to consider the child in the vignette to be the same age as the child that prompted their visit to the hospital. Block randomisation was used to determine whether the scenario would be administered in ascending or descending order of outcome severity. Of the severity index values, levels 2 and above were considered real or potential harm. At levels 0 and 1, the error was considered to have been intercepted before reaching the child. For each of the six possible outcomes, parents were asked: (1) whether they themselves would want full disclosure and (2) whether they would want their child informed of the error and AE. ‘Inform your child’ was defined as the medical team, along with the parent, would tell the child about the event as soon as possible after the event occurred, or when the child reached an age when he/she could understand what happened.
The data were first explored and summarised using means and SD, medians and ranges, and counts and percentages as appropriate. A weighted generalised estimating equations model (GEE, proc genmod; SAS V.9.1) was used to model the disclosure outcome and test its association with potential Table 1 Demographics of respondents explanatory variables. All variables were first explored in a univariate set-up. The ones with a level of significance below 0.2 were candidates for the multiple regression model. The final model was developed through backward elimination of nonsignificant factors. Separate models were fitted to the ‘disclose to parent’ and ‘disclose to child’ data.
A total of 431 parents were surveyed, including parents of 181 inpatients and 250 outpatients. Table 1 describes the characteristics of the parents and patients. The majority of parents interviewed were women, born in North America, Caucasian (data not shown) and under age 40 years. Sixty-two per cent of patients were under age 6 years, with a mean age of 5.9 years and a median age of 4 years. Thirty-five per cent of parents had had previous experience with error. Not all parents answered every question.
Disclosure to the parent
Ninety-nine per cent of parents wanted to be informed if there was any real or potential for harm (levels 2–5 combined) to their child related to the error, versus 77% if their was no possibility of injury (p<0.0001, see figure 1). Parents born in Asia were more likely than those from North America to want disclosure at lower levels of harm (p=0.014, OR 2.4, 95% CI 1.2 to 4.9). Parents born in Central and South America, Africa and the Middle East did not differ significantly from North Americans in their threshold for wanting disclosure. Parents of children who were inpatients at the time of the interview had a lower threshold for disclosure than those of outpatients (p=0.03, OR 1.65, 95% CI 1.04 to 2.62), as did parents who were administered the survey with increasing versus decreasing severity of harm (p<0.0001, OR 2.46; 95% CI 1.58 to 3.83). Parental age, education level, religion, age of the child, whether the child had a chronic illness and previous experience with error did not significantly affect attitudes on disclosure to themselves.
Disclosure to the child
Most parents (71%) wanted medical errors disclosed to their child if there was any potential or real harm, versus 41% if the error was intercepted with no possibility of harm (p<0.0001). This was particularly the case for parents born in Asia, compared with those born in North America (p<0.0001, OR 5.3, 95% CI 2.7 to 10.6). Previous experience with error (p=0.05, OR 1.5, 95% CI 1 to 2.2) and older age of the child (figure 2, not all data shown) (≤ 6 years vs ≥ 11 years (p<0.0001, OR 2.74, 95% CI 1.73 to 4.33); ≤6 years vs 6–10 years (p=0.001, OR 2.47, 95% CI 1.41 to 4.30)) resulted in a significantly increased likelihood of wanting the child to be informed of the event. Here, too, administering the survey with severity of harm in ascending versus descending order affected the threshold for disclosure (p=0.04, OR 1.46, 95% CI 1.01 to 2.10). However, parent age, education level, religion and whether the child had a chronic illness did not significantly affect attitudes on disclosure to their child.
The results of the current study support previous evidence that the majority of parents want to know if there has been an error in the care of their child, particularly when there is potential for harm or when there is actual harm. This study is unique, however, in its exploration of parent preferences towards disclosure to the child. Significantly, most parents wanted their child informed and this preference was directly related to the age of the child. Parents' desire for disclosure to their child is complementary to the notion that every child has a right to information about their care, as articulated in the United Nations Convention on the Rights of the Child.14 Although there are no paediatric-specific data to inform the process of disclosure in the wake of medical error, research in other contexts such as the importance of communicating with paediatric patients about a diagnosis of cancer,15,16 and the need for age-appropriate information when disclosing HIV status,17 may be informative. However, the optimal process for disclosing medical error to children, particularly the ageappropriate content and identification of the agent most suited to disclose, requires formal study.
Despite the dearth of information on the impact of race and ethnicity on patient safety in general,18 there is some evidence that perceptions of healthcare safety are influenced by patients' racial and ethnic backgrounds. In a previous report, parents' preferences for disclosure were not found to be influenced by race or ethnicity.7 In the current study, we found that parents born in Asia wanted errors disclosed to themselves and to their child at a lesser level of consequence compared with other parents. The distribution of our sample, including close to 20% Asian, may have facilitated recognition of this association. While these data are exploratory, they provide some evidence that race and ethnicity might indeed have impact and suggest that further research is required to explore the nature, if any, of the association between race and ethnicity and disclosure of AEs in paediatrics.
A lowered threshold for parent disclosure was noted when parents of inpatients (vs outpatients) were surveyed and when the vignette was administered with increasing versus decreasing levels of harm. A better understanding of the heuristics individuals bring to bare in decision-making in these circumstances may inform physicians' approaches to disclosure of AEs. The anxiety and vulnerability felt by parents of inpatients may make them more sensitive to the possibility of adverse outcomes, which may affect their threshold for disclosure; however, this hypothesis is merely speculative in the context of this study. Similarly, the anticipation of escalating harm may have prompted a more defensive, knowledge-seeking response (lower threshold for disclosure), when compared to presenting the consequences in descending order. Understanding these differences will likely prove useful in guiding the practice of disclosure.
There are a number of limitations to the current study. The questionnaire did not accommodate free text nor was it designed to explore parents' rationale for selecting a particular level for disclosure; thus, the results are limited in their ability to further inform the process of disclosure. The roles of race and ethnicity were not robustly explored; although country of birth was used as a proxy for race and ethnicity, this single characteristic does not take into account other attributes that may define ethnicity, including birth place of parents, language spoken at home, religion or nationality. This limitation was potentially further amplified by the method that excluded parents who were not sufficiently proficient in English.
The inclusion of only one vignette in our tool may restrict the generalisability of the findings, as does having carried out the study at a single centre. Additionally, any conclusion regarding disclosure of medical error to the paediatric patient is incomplete without surveying the paediatric patients themselves.
Despite these limitations, our study confirms that parents want to be informed of medical error, particularly when harm may, or has already occurred. Although more in-depth study is required to understand parents' attitudes on the timing and nature of disclosure to the child, our findings enrich the dialogue by considering the rights of the paediatric patient to disclosure. The importance of the link between culture, language and patient safety has recently been highlighted: “in order to ensure that minority as well as majority patient interests in receiving safe and quality care are properly protected, the culture–language–patient safety link needs to be formally recognised and the vulnerabilities of patients from minority cultural and language backgrounds explicitly identified and actively addressed in patient safety systems and processes”.18 Clearly, further study on the impact of culture, race and ethnicity on disclosure of medical error, in adults and in children, is required to inform these systems and processes.
Although much remains to be learnt about individuals' preferences regarding the process and act of disclosure, the need for transparency around medical error, particularly where there is real or potential harm, is fundamental. Parents want to know, and most parents want their children to know as well. The impact of race and ethnicity on disclosure is an area warranting further study. Finally, respect for the child as an individual with rights compels us to better understand the child's needs for information in the wake of medical error and the optimal mode of communication.
One year ago, your previously healthy ___________ month/year old child was treated for strep throat with antibiotic A and broke out in hives. The doctor said your child should never be given antibiotic A again. Your child was just hospitalised with pneumonia and needs intravenous (IV) antibiotics. You tell the admitting nurse and physician that your child is allergic to antibiotic A. By mistake, an antibiotic related to antibiotic A (antibiotic C) is ordered.
There are six possible outcomes from this scenario and each level is a ‘severity index value’. Now please review the events below and circle ‘yes’ or ‘no’ by each item to let us know whether you would expect the doctor to inform you and/or your child that an error was responsible for the event. What we mean when we say ‘inform your child’ is that the medical team, along with yourself, would tell the child about the event as soon as possible after the event occurred, or when he/she reached an age when they could understand what happened.
Level 0: Before dispensing antibiotic C, the pharmacist notices that your child is allergic to antibiotic A. She tells the nurse and your child receives no antibiotic C.
Inform you? Yes No
Inform your child? Yes No
Level 1: A bag of IV antibiotic C is hung and is ready to infuse. The nurse checks your child's armband and notes the band says ‘allergic to antibiotic A’. She removes the bag of antibiotic C before any goes in.
Inform you that an antibiotic A-like drug was ordered? Yes No
Inform your child? Yes No
Level 2: After IV antibiotic C has begun to infuse, the nurse checks your child's armband and sees the band says ‘allergic to antibiotic A’. She immediately stops the infusion and tells you she will check every 15 min for the next 2 h to make sure no rash develops. No rash develops.
Inform you that an antibiotic A-like drug was ordered and started? Yes No
Inform your child? Yes No
Level 3: IV Antibiotic C has begun to infuse and your child starts to feel itchy and develops a rash all over your child's body. The infusion is stopped. The doctor is called and orders an antihistamine. The rash settles down in a few hours.
Inform you that an antibiotic A-like drug was infused ? Yes No
Inform your child? Yes No
Level 4: IV Antibiotic C has begun to infuse and your child is having difficulty breathing. The infusion is stopped and the doctor gives your child epinephrine immediately. Your child settles down. The team decides they should treat your child for a full week in hospital (rather than just 3 days) to make sure their recovery is complete.
Inform you that an antibiotic A-like drug was infused? Yes No
Inform your child ? Yes No
Level 5: IV Antibiotic C is hung and your child goes into allergic shock. Your child is transferred to the ICU where he/she is resuscitated. Unfortunately your child has suffered permanent kidney damage and will require dialysis for life.
Inform you that an Antibiotic A-like drug was given ? Yes No
Inform your child? Yes No
Funding Funding for this study was provided by a grant from Paediatric Consultants, Hospital for Sick Children.
Competing interests None.
Ethics approval This study was conducted with the approval of the Hospital for Sick Children.
Provenance and peer reviewed Not commissioned; externally peer reviewed.
Patient consent Obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.