OUP user menu

Nurses’ and nursing assistants’ perceptions of patient safety culture in nursing homes

Carmel M. Hughes, Kate L. Lapane
DOI: http://dx.doi.org/10.1093/intqhc/mzl020 281-286 First published online: 19 July 2006


Objectives. To evaluate whether perceptions of patient safety in nursing homes vary by length of employment, type of employee, and shift worked.

Design. Cross-sectional study.

Setting. Twenty-six nursing homes in Ohio participating in a randomized trial to test the effectiveness of a clinical informatics tool to improve patient safety during the medication monitoring.

Participants. Nurses (n = 367) and nursing assistants (n = 636) employed at the time of the survey in the summer and fall of 2003.

Main outcome measurements. Resident safety questions included 34 items on different aspects of resident safety (overall safety perception, teamwork within and between departments, communication openness, feedback and communication about error, non-punitive response to error, organizational learning, management expectations, and actions promoting safety, staffing, and management support for patient safety).

Results. Overall perceptions of resident safety by employees were acceptable, with clear management communication of safety goals. Approximately 40% of nursing staff found it difficult to make changes to improve things most or all of the time; similar proportions indicated that management seriously considered staff suggestions to improve resident safety; only half reported management discussions with staff to prevent recurrence of mistakes. Regardless of staff type, one in five reported feeling punished and two in five reported that reporting of errors was seen as a ‘personal attack’.

Conclusions. Interventions to change the safety culture in nursing homes are warranted. Nursing homes need guidance on how to use information to implement safety improvement projects in the context of a strict regulatory environment which may prohibit innovative system change.

  • nursing assistants
  • nursing homes
  • nursing
  • patient safety
  • safety culture

Over 1.8 million people reside in approximately 17 000 US nursing homes [1]. Despite sweeping reform which occurred with the Nursing Home Reform Act embedded in the Omnibus Budgetary Reconciliation Act (OBRA) of 1987 [2], concerns regarding patient safety remain [3]. Untoward events occurring in this health care setting include pressure ulcers, adverse drug events, and falls.

OBRA legislation has sought to improve the quality of nursing home care through an adversarial regulatory and inspection approach. The Centers for Medicare and Medicaid Services (CMS; the main administrative agency for health in the United States) developed a set of regulations, outlining all aspects of nursing home operation [4]; this has also extended to the use of quality measures that are posted on a publicly accessible Website [5]. If a nursing facility is deemed not to be attaining the required standard, then a number of sanctions are available including termination of provider agreement, transfer of residents with closure of facility, and a directed plan of correction. Such negative incentives may prevent the reporting and disclosing of errors.

Assuring the safety of nursing home residents is compounded by restrictive reimbursement systems, increasingly frail residents, and poor staffing levels. An astounding 92% of nursing homes do not have sufficient staff to provide levels of care to meet the federal regulations and practice guidelines [5,6]. The average one-year turnover rate is high for nursing assistants and licensed practical nurses (85.8%) and registered nurses (55.4%) in US nursing homes [7]. The instability of the nursing staff has been correlated with decreases in quality of care [7,8]. We sought to document the perceptions of patient safety among nurses and nursing assistants in nursing homes and to evaluate the extent to which perceptions of patient safety in nursing homes varies by length of employment, type of employee, and shift worked.


The study protocol was approved by the Brown Medical School Institutional Review Board. Twenty-six nursing facilities included in this study were participating in a large randomized clinical trial funded by the Agency for Healthcare Research and Quality, testing the effects of a unique clinical informatics tool on patient safety outcomes in nursing homes. Nursing homes provided the research team with an enumeration of nurses (registered and licensed practical nurses, n = 721) and nursing assistants (n = 1233) as well as a preference for distribution method (either directly to staff members at their home address or mailed to the nursing facility for distribution at work). We performed four mailings spaced 2 weeks apart in the following sequence: an initial mailing of the survey packet, a reminder postcard, a re-mail of the survey packet to non-respondents to the initial survey, and a final reminder postcard. The survey packets consisted of a cover letter explaining the survey and including the elements of informed consent, as well as the procedures necessary to receive the incentive for survey completion, the survey with the unique identifier included on the survey (but not the respondent’s name), and a postage paid return envelope. Return envelopes were addressed to the research team at Brown Medical School. The respondents were asked not to complete the survey at work. Mailings began in August 2003 and continued throughout the fall of 2003, before the initiation of the randomized trial. After we mailed $15 incentive checks to respondents, data were de-identified.

The nurse questionnaire and nursing assistant questionnaire contained questions that were adapted from a safety culture assessment tool [9,10] modified by Emory and Morehouse Universities (Joseph Ouslander, personal communication). Resident safety questions included 34 items on different aspects of resident safety. Domains included overall safety perception, teamwork within and between departments, communication openness, feedback and communication about error, non-punitive response to error, organizational learning, management expectations, and actions promoting safety, staffing, and management support for patient safety.

Descriptive analyses were performed by nursing staff type (nurses and nursing assistants), length of employment (≤1 year, 1+ years), and shift worked (first, second, third, or rotating across shifts as needed). Typically, first shift is 7 a.m.–3 p.m., second is 3 p.m.–11 p.m., and third is 11 p.m.–7 a.m. We hypothesized that those in longer employment would have had a clear understanding of safety issues within the nursing home, whereas those who worked later shifts (particularly the third shift) may have been more professionally isolated and less exposed to communication and discussion on resident safety. We compared the distributions of categorical variables using chi-square tests and Fisher’s exact test when the cell sizes were less than five.


We received 367 completed nurse surveys and yielding a 56% response rate and 636 nursing assistant surveys yielding a 60% response rate. Over 90% of nurses and nursing assistants were women, with nurses reporting higher education levels than nursing assistants (Table 1). Nursing assistants were more likely to report being of a racial/ethnic minority (32%) than nurses (12%).

View this table:
Table 1

Sociodemographic characteristics of nurses and nursing assistants employed at 26 nursing homes participating in Ohio

CharacteristicsNurses (n = 367)Nursing assistants (n = 636)
Women (%)9594
Racial/ethnicity minority (%)1232
Education (%)
    Less than high school115
    High school or general equivalency diploma138
    Vocational/trade school2212
        Some college3829
        Associate degree274
        Bachelors degree102
    Postgraduate degree20.3
Shift worked (%)
Years worked for nursing home [mean (SD)]6.7 (6.6)5.4 (6.3)
Years worked in present position [mean (SD)]6.9 (7.3)6.7 (7.2)

In terms of overall resident safety ratings given by staff, excellent grades were given by 11% of nurses (n = 40) and 13% of nursing assistants (n = 83), whereas a poor/failing grade was given by approximately 5% of both nursing types. Length of employment (less or more than 1 year) had little impact on safety ratings. Third shift employees were the least likely to give their departments an overall very good or acceptable grade. Approximately 10% of nursing staff reported most/or all of the time resident safety problems occur on their unit, with ∼15% reporting that they feel most or all of the time that it is pure luck that more serious mistakes did not happen in their department. These estimates did not vary appreciably by nursing grade, length of employment, and shift worked.

Table 2 summarizes the resident safety items stratified by nurse grade (nurse or assistant) and duration of employment. In terms of nurse grade, nurses were more likely to report that when a lot of work needed to be done quickly, staff worked together as a team to get the work done most or all of the time compared with nursing assistants. However, more nursing assistants (30%) compared with nurses (23%) reported that staff worked on their own and did not want to help others most or all of the time (P = 0.02). High proportions of both staff categories had reported a safety problem in the last month to someone at least once (81% nurses and 62% assistants; P < 0.0001), with more nurses than assistants having written at least one incident report in the last month (68% versus 19%; P < 0.0001). Nursing assistants (25%) were more likely to report that when a safety incident was being reported, the person was being written up rather than the problem, but nurses rather than assistants were more likely to indicate that the reporting of errors of another staff member was seen as a personal attack against them.

View this table:
Table 2

Opinions of resident safety according to staffing category and length of employment

% who answered ‘most or all of the time’Nurses (n = 367)Nursing assistants (n = 636)P valueEmployed ≤1 year (n = 259)Employed >1 year (n = 744)P value
Team work within and between departments
    When a lot of work needs to be done quickly, how often do staff work together as a team to get the work done?7358<0.00158660.02
    When one person gets really busy, how often do others not want to pitch in?29340.0932320.98
    How often do staff work on their own and do not want to help others much?23300.0229270.47
    How often do departments work together to give the best care for residents?65550.00263570.09
    How often do departments fail to coordinate well with each other?17210.1520200.77
    How often is there good cooperation among departments that need to work together?52520.8546540.03
    How often is it unpleasant to have to work with staff from other departments?9120.2211110.84
    When staff see a coworker not following standard practices, how often do they point it out to him/her? (O)50500.9249500.93
    How often do staff check the work of others when they are supposed to? (O)56480.0152500.68
    When a mistake happens, how often do staff avoid reporting it? (O)890.611080.56
    How often do supervisors/managers and staff discuss mistakes to keep them from happening again? (F)53550.5351560.25
    How often do you talk openly about resident safety problems that exist in your department? (F)5946<0.000148520.35
    How often are staff told about what happens as a result of an incident report? (F)42430.6840420.54
    How often do staff feel like they are being punished when an incident report is written up on a mistake they have made? (N)27220.05926230.25
    When a safety incident is reported, how often does it feel like the person is being written up, not the problem? (N)18250.00921220.62
    How often is reporting the errors of another staff member seen as a personal attack against them? (N)42360.0436390.47
Management priorities and expectations and actions
    How often does the behavior of management show that resident safety is a top priority? (P)60610.8355580.49
    How often does management clearly tell staff what the resident safety goals are? (P)565857610.200.48
    How often does your supervisor/manager say a good word when he/she sees a job done by the right procedures and rules? (E)27290.5829280.81
    How often does your supervisor/manager seriously consider staff suggestions for improving resident safety? (E)42380.1939400.82
    How often does your supervisor/manager overlook resident safety problems that happen over and over? (E)7100.11990.89
    How often does your supervisor/manager neglect to pay enough attention to resident safety problems? (E)570.09860.19
    Whenever pressure builds up, how often does your supervisor/manager want you to work faster even if it means taking shortcuts? (E)18160.4315170.30
    How often does management help you feel good about giving residents safe care? (E)40340.0536360.94
Organizational learning
    How often is it difficult to make changes to improve things in your department?42300.000332350.49
    How often do staff not seem to learn from mistakes?11150.1212140.44
    How often do staff receive enough training to provide safe care to residents?69700.9561720.0006
    When the same mistake keeps happening, how often do staff look at procedures to see if they need to make changes?65590.0757620.15
  • E, expectations and actions; F, feedback about error; N, non-punitive response to error; O, openness; P, management priorities.

In relation to opinion on resident safety according to length of employment, employees working for the facility greater than 1 year were more likely to report that staff worked as a team when a lot of work needed to be done quickly and that there was a good deal of cooperation among departments relative to newly hired nursing staff. Newly employed nursing staff were less likely to report that training was provided most or all of the time relative to those employed greater than a year (P = 0.0006).

Although comparisons of resident safety items by shift of employee revealed few differences in perceptions of resident safety items by shift, third shift employees gave their departments less favorable ‘overall grades’ on resident safety. When analyses were further stratified by staff type (data not shown), second and third shift nursing assistants were much less likely to report adequate training had been provided (P = 0.0165).


To our knowledge, this is the first article that has examined safety issues in nursing homes from the perspective of the nursing staff. Although this has been done in hospitals, e.g. [11], the results are not likely generalizable to the nursing home setting for several reasons. A nursing home is supposed to be a home environment with medicalization of the environment kept to a minimum. Resident autonomy, dignity, and participation in decision-making have been reported to be the most important quality of life attributes for nursing home residents [12]. Ironically, these attributes may be difficult to implement if safety is considered the number one priority. Kapp [3] has noted that safety does not represent all the residents’ expectations and preferences concerning the quality of care and quality of life. Despite the complex milieu in which nursing home care is provided, a consideration of how safety can be maximized in this environment should not be precluded. These data provide the perspective of the health professionals on the front line in nursing homes—nursing staff.

Leadership in safety issues has been shown to be influential in creating a positive safety culture [13–15] and has also been shown to be critical in high quality care [16]. In this study, ∼60% of nursing staff reported that management showed resident safety was a top priority and safety goals were clearly articulated most or all of the time. Despite such clarity, indicators of ineffective leadership are provided. Approximately 40% of nursing staff found it difficult to make changes to improve things most or all of the time; similar proportions indicated that management seriously considered staff suggestions to improve resident safety; only half reported management discussions with staff to prevent recurrence of mistakes. Our findings are consistent with a report from the Office of the Inspector General in the United States, which found that ‘while quality assurance committees have an array of information to help them pinpoint problems in nursing homes, knowledge of how to use this information to execute projects remains a key barrier’ [17]. Typically, nursing homes use the traditional quality assurance models, which retrospectively monitor aspects of care, addressing problems on an individual basis rather than on a systems level, and minimizing staff input. Nursing homes have neither the staff nor the expertise in monitoring systems of care and work environments [18] and are challenged by strict regulatory environment that may prohibit innovative system change [19]. Individual State Quality Improvement Organizations administered by the CMS are uniquely positioned to provide direct guidance sorely needed in this setting [18].

Organizations with a positive safety culture set the tone for acknowledgement of error (e.g. communications founded on mutual trust), which in turn permits learning from experience and mitigating further errors [20]. Unfortunately, data from this study suggest that a ‘blame and shame’ culture predominates in the nursing home setting. Regardless of staff type, one in five reported feeling punished and two in five reported that reporting of errors was seen as a ‘personal attack’. The negative patient safety culture that exists in nursing homes may be a function of the adversarial and punitive nature of US nursing home regulation [3,21,22]. Nursing homes are under constant scrutiny and subject to detailed inspections (or surveys) and those which do not meet the regulations may be subject to a number of penalties, the most severe being closure of the facility. Staff may feel reluctant to report safety issues that may draw attention to individuals and to the nursing home. It would be interesting to carry out a similar study in nursing homes in a different regulatory setting outside the United States. Despite the consequences, we found that at least 60% of nursing assistants and 80% of nurses reported a safety problem at least once in the past month, with nurses being primarily involved in writing incident reports. The usefulness of the reporting systems may be limited as only two in five staffs reported that they were told what happens as a result of incident reports most or all of the time.

Training is clearly an important aspect of safety improvement [20]. The logistic challenges of provision of training equally across shifts and continuously (given the poor retention of staff) is non-trivial. Training programs have been noted to help facilities achieve the ‘three R’s of retention: relationships, respect, and recognition’ [23]. Online resources are available for ideas on how to ‘build and maintain a stable, high-quality workforce’ [24]. Innovative computer-based comprehensive curriculum of individual, self-paced education through interactive documentaries may provide solutions to the practical training challenges of the nursing home environment [25]. Widespread dissemination of effective training modalities is hampered by the lack of computer equipment.

We interpret these data with caution. Although the original tool has good psychometric properties [10], it was not designed for use in nursing homes. Despite our inclusion of a survey incentive and multiple mailings, our response rate was not optimal. Our data are cross-sectional and thus provide a snapshot of the perceptions of nursing staff. The results may also not be generalizable to all nursing homes in the United States. All participating homes were in one state (Ohio) although not owned by the same proprietor. The homes had at least 50 residents each and had stable contracts with one pharmacy services provider.

To date, the nursing home has largely been invisible in the safety debate. Yet, the older and frail residents in nursing home settings may be at greater risk for adverse safety events than individuals in other health care settings. A greater focus on these residents should be more clearly articulated by policy makers, managers, and practitioners.


This study was supported by a grant from the Agency for Health Care Research and Quality.


View Abstract