Nurses' disclosure of error scenarios in nursing homes - Nursing Outlook

Available online at www.sciencedirect.com

Nurs Outlook 61 (2013) 43e50

www.nursingoutlook.org

Nurses’ disclosure of error scenarios in nursing homes Laura M. Wagner, PhD, RN, GNP-BCa,*, Kimberley Harkness, BScN, RNb, Philip C. He´bert, MD, PhD, FCFPCc, Thomas H. Gallagher, MD, FACPd a

University of California, San Francisco, School of Nursing, Community Health Systems, San Francisco, CA b Childbirth Centre, Markham Stouffville Hospital, Markham, Ontario, Canada c Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada d University of Washington Medical Center, Seattle, WA

article info

abstract

Article history: Received 13 February 2012 Revised 6 May 2012 Accepted 29 May 2012

Background: Little work has explored the disclosure of errors in nursing homes

Keywords: Error Nursing home Disclosure Patient safety

(NHs). Purpose: This paper reports how nurses would disclose hypothetical errors that

occur in NH settings. Method: A cross-sectional survey was given to a randomly selected sample of registered nurses (RNs) and registered practical nurses (RPNs) working in Ontario, Canada NHs. Results: Of 1,180 respondents, only half might provide full details and the cause of the error and provide steps in how the error would be prevented if they were in situations described by the hypothetical scenarios. Scenarios that were less serious had an almost 3 times higher likelihood of an explicit apology (OR 2.97; 95% CI 1.36-6.51; P ¼ 0.007). Nurses who were RNs, had more education, had a prior history of disclosing a serious error, and agreed with full disclosure were more likely to respond to disclosing more information about the error. Nurses also reported numerous barriers to effective disclosure in their workplace. Conclusion: Improvements in NH safety culture are necessary to enhance the error disclosure process. Cite this article: Wagner, L. M., Harkness, K., He´bert, P. C., & Gallagher, T. H. (2013, FEBRUARY). Nurses’ disclosure of error scenarios in nursing homes. Nursing Outlook, 61(1), 43-50. http://dx.doi.org/10.1016/ j.outlook.2012.05.008.

Introduction Background A considerable amount of health care is provided to 1.5 million residents in the 16,100 U.S. nursing homes (NHs) (Jones, Dwyer, Bercovitz, & Strahan, 2009). Organizational conditions prevalent in NHs make errors more likely, such as high nursing staff turnover,

inadequate nursing supervision, poor communication and teamwork among nursing staff and managers, and frequent changes in nursing administration (Rosner, Berger, Kark, Potash, & Bennett, 2000). Low NH staff education levels and staff who are hurried and distracted are additional organizational factors that contribute to medical errors in the NH setting (Penson, Svendsen, Chabner, Lynch, & Levinson, 2001). NH residents are particularly vulnerable to errors because of their complex physical and mental health care

The Canadian Institutes for Health Research funded this study. * Corresponding author: Dr. Laura M. Wagner, Assistant Professor, University of California, San Francisco, School of Nursing, Community Health Systems, 2 Koret Way, N511R, San Francisco, CA 94143. E-mail address: [email protected] (L.M. Wagner). 0029-6554/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.outlook.2012.05.008

44


needs and multiple comorbidities. NH residents and their families may also be less likely to provide an accurate and complete health history because of sensory and cognitive impairments. As a result of these multiple conditions, NH residents use a disproportionately higher amount of health care services and medical procedures, which further increases their risk for medical error (Kapp, 2001). High-quality nursing care includes disclosure of an error, if one occurs, to the NH resident or their family member. An important part of the nursing process includes informing the residents and families about nursing errors that have occurred. The responsibility of disclosure falls upon nurses in the NH setting because of the limited day-to-day contact medical providers have with NH residents and their families. The level of error disclosure is dependent on multiple factors including the error seriousness, with providers being more likely to disclose errors that are more serious (Gallagher et al., 2006b; Garbutt et al., 2007). The few studies conducted on the role of nurses in disclosure have focused primarily on reporting adverse events to supervisors (Balas, Scott, & Rogers, 2004; Hobgood, Xie, Weiner, & Hooker, 2004; Jeffe et al., 2004; Osmon et al., 2004) but have had less emphasis on the process of disclosure to the patient and family. The research conducted thus far on error disclosure has primarily centered on the physician’s role (Gallagher et al., 2006a; Gallagher et al., 2006b). One article explored the perspectives of registered nurses (RNs) with regard to disclosing errors to patients in acute care settings (Shannon, Foglia, Hardy, & Gallagher, 2009) and found that nurses often feel excluded from team discussions regarding disclosure of adverse events to patients and favor hospital policies that promote a collaborative disclosure process.

The Study Aim The role of nurses in disclosing nursing errors has not been described in the NH setting and it is unknown how much content nurses would disclose to residents and families when an error occurs. The goal of this paper is to better understand how nurses have disclosed and would disclose an error in the NH setting.

Design A cross-sectional descriptive study was used based on a well developed, validated (Gallagher et al., 2006a; Gallagher et al., 2006b), and self-administered mailed survey among a randomly selected sample of RNs and registered practical nurses (RPNs) in Ontario.

Sample Using a random sampling method, we surveyed a subset of the approximately 14,500 full and part-time employed RNs and RPNs working in Ontario NHs. Participants (n ¼ 2,800) were obtained from a mailing list provided to us through the College of Nurses of Ontariodthe mandatory nursing regulatory body for all RNs and RPNs in Ontario, Canada. Follow-up reminders were mailed to nurse respondents two and four weeks after the survey was sent out. Nurses also had the option of completing the survey online via SurveyMonkey. Anonymity and confidentiality were maintained through the online response method as with the hard copy of the survey.

Survey The Communicating about Nursing Errors (CANE) survey was adapted from the Communicating about Medical Errors with Physicians survey created by Dr. Thomas Gallagher and colleagues (Gallagher et al., 2006a; Gallagher et al., 2006b). A six-item list of demographic and workplace information developed by the investigators included questions at the beginning of the survey regarding the respondent’s characteristics, such as position (e.g., RN, RPN, Administration); education level; hours worked per week; and total years of nursing work experience. The final version of the CANE was a 60-item survey with six defined sections (Wagner, Harkness, He´bert, & Gallagher, 2012). This paper examines nurse respondents’ past experiences with errors (i.e., disclosing of a serious error to a resident and their family) and responses to hypothetical scenarios on what the nurse might do if they were confronted with the same situation. Each nurse respondent randomly received one of the four versions of the CANE, the only variation being the clinical scenarios: (1) More serious error, cognitively impaired; (2) more serious error, cognitively intact; (3) less serious error, cognitively impaired; and (4) less serious error, cognitively intact (Table 1). Seriousness in this paper relates to the level of harm or potential harm that the error could have on the NH resident. The clinical scenarios were adapted to situations that a nurse might encounter in the NH setting to explore how differences in nursing errors and residents’ mental status might affect disclosure. Using scenarios allowed us to study the way in which nurses would disclose an error in a standardized situation and explore variations in disclosure (Peabody, Luck, Glassman, Dresselhaus, & Lee, 2000). The errors were constructed to be comparable in type (e.g., medication) and to vary along two primary axes: how serious the error would be categorized and the resident’s level of cognition. We constructed realistic, detailed, and common case scenarios. For each of the four scenarios, respondents received a comparable set of questions to answer according to what they might do if they were confronted with the


45

Table 1 e Clinical Scenarios Type of Scenario Insulin overdose (more serious error)

Insulin overdose (more serious error)

Incorrect medication administration (less serious error)

Incorrect medication administration (less serious error)

Description Mrs. Smith is a new diabetic resident admitted to your facility. She is cognitively intact and able to make her own care decisions. You handwrite a telephone order from the physician for the resident to receive “10U” of insulin. The “U” in your order looks like a zero. The following morning the resident is given 100 units of insulin, ten times the resident’s normal dose, and is later found unresponsive with a very low blood sugar level. The resident is resuscitated and transferred to the hospital. You expect the resident to make a full recovery. Mrs. Smith is a new diabetic resident admitted to your facility. She is cognitively impaired and unable to make her own care decisionsdher husband is your contact. You handwrite a telephone order from the physician for the resident to receive “10U” of insulin. The “U” in your order looks like a zero. The following morning the resident is given 100 units of insulin, ten times the resident’s normal dose, and is later found unresponsive with a very low blood sugar level. The resident is resuscitated and transferred to the hospital. You expect the resident to make a full recovery. Mr. Wilson has recently been admitted to your facility for rehabilitation after a hip fracture caused by a fall. He is cognitively intact and able to make his own care decisions. The physician ordered a “one time dose now of Vitamin K by mouth” to correct an elevated INR of 6.5. You administer the medication and realize later that you have incorrectly given the patient Vitamin C 5 mL liquid dose by mouth (dispensed as 500 mg/5mL). The resident is expected to make a full recovery after receiving the correct medication. Mr. Wilson has recently been admitted to your facility for rehabilitation following a hip fracture due to a fall. He is cognitively impaired and unable to make his own care decisionsdhis wife is your contact. The physician ordered a “one time dose now of Vitamin K by mouth” to correct an elevated INR of 6.5. You administer the medication and realize later that you have incorrectly given the patient Vitamin C 5 mL liquid dose by mouth (dispensed as 500 mg/5mL). The resident is expected to make a full recovery after receiving the correct medication.

same situation including: (1) How serious they perceived the error to be, (2) how responsible the nurse respondent was for the error, (3) how upset the respondent would feel about causing the error, (4) how concerned they would be about their reputation being damaged, (5) how likely the respondent thinks they would be reprimanded, and (6) how likely the respondent would be to disclose the error to the resident and their family. In addition, there were five disclosure content questions used to measure what level of information the nurse would disclose to the resident and/or family, including: (1) “What would you most likely say about what happened?”; (2) “How much detail would you most likely give the resident about the error?”; (3) “What most closely resembles what you would say about the cause of error?”; (4) “What would you say regarding an apology?”; and (5) “What would you most likely say about how the error would be prevented in the future?” The responses included three items representing increasing amounts of disclosed information: (1) No disclosure, (2) partial disclosure, and (3) full disclosure.

Ethical Considerations An academically affiliated research ethics board for the protection of human subjects approved this study. Informed consent was implied by return of the survey, and participants were free to withdraw from the study at any time.

Data Analysis Descriptive statistics were presented as means, standard deviations, and medians and ranges for continuous variables and as counts and percents for categorical variables. Chi-square analyses were used to determine whether the clinical scenario responses varied based on: (1) The seriousness of the scenario to the resident (more or less serious) and (2) whether the resident was cognitively intact or impaired. Next, univariate and multivariate logistic regression models were used to explore seriousness, cognitive status, and the possible interaction between the two factors in their effects on the error disclosure outcomes. Finally, a single dependent variable was created using factor analysis to represent the five disclosure content questions. Therefore, the dependent variable in the multivariable modeling was a cognition-specific scaled disclosure score. Multivariable linear regression was used on the new dependent variable (a composite score of the five factors, with higher values indicating more disclosure). Possible independent variables included type of nurse, employment position, highest level of education completed, number of years in current long-term care setting, whether the respondent has ever disclosed a serious error, perceived seriousness of the situations, how responsible the nurse is for the error, how upset the respondent would be, how concerned the respondent would be about their reputation being damaged,

46


Table 2 e Characteristics of the 1180 Nurse Respondents* Characteristic Frequency Type of nurse Employment position

Highest level of completed education

Years at current long term care settingy Workplace location

RN RPN Staff nurse Manager (role in performance evaluations) Other High school, vocational, or certificate school Diploma/Associate’s degree Bachelor’s, Master’s, or Doctorate degree 5 years or fewer 6-10 years 11 years or more Urban Suburban Rural

49.6 50.4 77.7 16.4

5.9 16.2 70.0 13.8 33.5 25.1 41.4 24.4 57.9 17.7

* Data are presented as percentage for each group unless otherwise indicated. y Data have been dichotomized.

how likely the respondent thinks they will be reprimanded, and how likely the respondent would disclose the error to the resident/family. Full models were reduced using backward deletion with a P > 0.05 criterion for deletion. Analyses were performed using SAS 9.1 (SAS Institute Inc., Cary, NC).

the NH environment (e.g., “physician” was changed to “nurse”; “hospital” was changed to “NH,” etc.). Nurses provided initial feedback on the scenarios in our pilot work to ensure key “disclosure” and “nursing error” terms were understood clearly. Two rounds of pilot work in 2007-2008 among 79 racially and ethically diverse licensed nurses from two Ontario, Canada NHs occurred to adapt this reliable and valid questionnaire to the NH setting. Content validity testing of the CANE, including the four scenarios, occurred during these phases (Content Validity Index >0.80) (Waltz, Strickland, & Lenz, 1991). After our first round of pilot testing, personal support workers (unlicensed aids) were excluded from further pilot testing based on post-debriefings findings that they did not play a significant role in error disclosure. In a third round of testing, we retested the CANE on 19 RNs and RPNs to obtain further data on the readability and validity of the questionnaire. The nurses commented on the clarity of questions, plausibility, and how realistic the scenarios and scenario responses were. In general, the questionnaire was viewed as unambiguous and the scenarios clear and credible. Overall, more than 75% of the sample reported adequate clarity (e.g., options easily understood); utility (e.g., likely to elicit candid information); face validity (e.g., questions accurately reflected factors influencing disclosure); and content validity (e.g., degree of gaps in the items in the questionnaire). Additional minor revisions (e.g., rewording items) to improve clarity of the items were made to the CANE as a result of the pretesting.

Rigor

Results

Psychometric pilot testing ensured adequate validity (e.g., content and predictive) and reliability. Initial modifications to the CANE survey were primarily made to ensure that terminology remained consistent with

A total of 2,800 surveys were mailed and 1,389 were returned, resulting in a 49.6% response rate. Of these, 209 reported they were no longer working in long-term

Table 3 e Respondents’ Attitudes Regarding Scenarios Statement

As the nurse, how responsible are you for this error?* How upset would you be about making this error?y How concerned would you be about your reputation being damaged?z How likely do you think it is that you would be reprimanded?z How likely would you be to disclose the error to the resident/family?x * y z x

More Serious Error

Less Serious Error

P value

Cognitively Intact

Cognitively Impaired

P value

87.6

94.6