Background: Falls are the most frequently reported adverse event among frail nursing home residents and are an important resident safety issue. Incident reporting systems have been successfully used to improve quality and safety in healthcare. The purpose of this study was to test the effect of a systematically guided menu-driven incident reporting system (MDIRS) on documentation of post-fall evaluation processes in nursing homes.
Methods: Six for-profit nursing homes in southeastern USA participated in the study. Over a 4-month period, MDIRS was used in three nursing homes matched with another three nursing homes which continued using their existing narrative incident report to document falls. Trained geriatric nurse practitioner auditors used a data collection audit tool to collect medical record documentation of the processes of care for residents who fell. Multivariate analysis of covariance was used to compare the post-fall nursing care processes documented in the medical records.
Results: 207 medical records of resident who fell were examined. Over 75% of the sample triggered at high risk for falls by the minimum data set. An adequate neurological assessment was documented for only 18.4% of residents who had experienced a fall. Although two-thirds of the sample had a diagnosis of incontinence, less than 20% of the records had incontinence-related interventions in the nursing care plan. Overall, there was more complete documentation of the post-fall evaluation process in the medical records in nursing homes using the MDIRS than in nursing homes using standard narrative incident reports (p<0.001).
Conclusion: Further improvements are necessary in reporting mechanisms to improve the post-fall assessment in nursing home residents.
Statistics from Altmetric.com
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.
Falls are a serious and complex safety issue among nursing home residents. As the most frequently reported adverse incident, approximately half of the estimated 1.5 million US nursing home residents fall annually, about twice the rate in elderly people living in the community.1–7 Nursing homes are required to conduct an investigation when a fall or other incident has occurred.8 Incident reports are the primary means to document this investigation in healthcare settings, although there are no universal or legal guidelines on what this report should include.9–11 The evaluation should provide information such as the resident’s clinical status, circumstances regarding the incident and interventions to prevent future incidents. Most nursing homes use a semi-structured narrative incident report in which nurses write an open-ended summary of the incidents. This lack of structure limits the quality of information obtained about the event. These reports, kept separate from the medical record, are used to fulfil nursing home regulatory and malpractice insurance requirements rather than for quality improvement purposes.9 Reports concerning medical errors and adverse events call for a variety of safety initiatives, including enhancements in incident reporting1213 because of their role in improving the quality of healthcare.1415
Incident reports generally are not designed to provide adequate guidance for a comprehensive post-fall assessment.16 In a previous study we found that by providing nursing staff with a structured menu-driven falls incident report, more complete fall incident information was elicited from the incident reports to guide quality improvement activities in the nursing home.17 As part of any quality improvement programme, incident report data are used in trend analyses to identify patterns of concern.18 In addition, incident report data can also guide root cause analyses, a process that also may help to reduce falls and major fall-related injuries.19
Although a form itself will not directly increase falls prevention, it should guide nurses to accurately assess and evaluate underlying causes of falls with the goal of identifying interventions to prevent future falls. Therefore, the purpose of this study was to evaluate whether using a menu-driven incident reporting system (MDIRS) would result in improved medical record documentation of post-fall evaluation processes in nursing homes.
Using a quasi-experimental study design, six southeastern US for-profit nursing homes were recruited with a non-random sampling approach. The Emory University Institutional Review Board approved the study. Eligibility criteria included having greater than 100 beds, and currently using a narrative incident report (as opposed to a computerised system) to document falls. After matching into three pairs based on the facilities’ bed size, the nursing homes were randomised into intervention and control groups. The three intervention nursing homes replaced the narrative incident report with the MDIRS, whereas the three control nursing homes continued using their existing narrative incident report. Each month during the 4-month intervention period, incident reports were reviewed to identify residents who fell. Residents who experienced a fall during the intervention period were eligible if they were aged 65 and older, non-comatose, and not receiving hospice services. A total of 259 residents in the six nursing homes fell during the 4-month study period. Of these, 207 residents met the eligibility criteria.
The MDIRS intervention
The MDIRS is a systematically guided assessment derived to improve current methods of evaluating fall incidents. It provides a series of close-ended questions with multiple-choice alternatives prompting the nurse to consider risk factors and assessment information related to the incident. Face validity of the MDIRS has been established through a review by a panel of experts.17 In the present study, a nurse filled the form by checking the appropriate boxes in several key areas including location, time, activity, possible causes, footwear, detailed physical assessment information, fall outcome and identified possible interventions added to the care plan to prevent future falls. The form was then routed to the care planning nurse to update the resident’s care plan. The data were then entered into a software program for nursing home staff to generate quality improvement reports.17
Trained research assistants reviewed each subject’s medical record for demographic and pertinent medical data from the minimum dataset. The Falls Management Audit Tool (FMAT)20 was used to quantify documentation of the post-fall evaluation processes from each subject’s medical record. The post-fall evaluation process is a series of actions conducted by nurses to evaluate and treat falls and monitor fall risk. The concepts that comprise the post-fall evaluation process are the diagnosis, management and monitoring stages of the American Medical Directors Association (AMDA) and American Health Care Association (AHCA) falls management guideline.21 The diagnosis stage occurs when nursing staff conduct an assessment at the time of fall. Management refers to the planning and implementation of individualised fall prevention interventions. Monitoring occurs when staff observe and document the resident’s response to an intervention by reassessing, evaluating or adjusting the care plan.21
Details regarding the development and testing of the FMAT are described elsewhere.20 Briefly, the AMDA and AHCA14 guideline provided the framework for the items on the FMAT. The FMAT was developed using a multistep content analysis procedure and was reviewed by national experts until content validity was deemed acceptable (content validity index >0.88).
Two masters’ prepared gerontological nurse practitioners (GNPs) were taught how to conduct the medical record audits. Chart audits (ie, extraction of data from medical records) were conducted and repeated after 14 days on the same records to assess intra-rater and inter-rater reliability. We established an acceptable level of reliability with all κ values greater than 0.70.20
The trained GNP auditors reviewed each subject’s medical record. Baseline chart audit data (pre-intervention audit) up to 6 months prior to the start of the intervention period, and data during the study period (post-intervention audit), were collected. The exception to pre-intervention chart audit data collection was if the resident was admitted to the nursing home after the start of the study period.
All items on the 57-item FMAT were scored on a dichotomous (0 = not documented and 1 = documented) scale. A composite score for each stage was computed. For example, if there were 10 items in the diagnosis stage and 8 (numerator) out of 10 (denominator) items were documented, then their score for that stage was 0.80. In subjects who fell recurrently, the total number of fall items was used as the denominator, when applicable. With this same example, in a resident who experienced three falls, the denominator was 30. Auditors used a detailed data dictionary to maintain consistency of information coding. Nursing staff were blinded to all care process items that were being collected by the research team. Figure 1 shows the MDIRS intervention and data collection methods.
All analyses were performed using SPSS version 11.5. The unit of analysis was the individual resident’s medical record. Baseline descriptive data were compared between the intervention and control groups to characterise significant differences using independent samples t tests, the Mann–Whitney U test and χ2 tests.
To test the effect of the MDIRS intervention on documentation of post-fall evaluation processes, we conducted a multistep analysis. First, covariates were chosen based on salient resident demographic and physical characteristics for which significant differences were observed between the intervention and control groups. The significant covariates were modelled in univariate step-wise regression analyses with each of the dependent variables (diagnosis, management and monitoring). Variables in the regression model that remained significant at the p<0.05 level were included in the final model. The number of falls during the study period and fall history were also included as covariates. To compare mean differences of the dependent variables between intervention and control groups, a multivariate analysis of covariance (MANCOVA) was used (Hotelling trace). Two groups were analysed: residents who fell during the intervention period (n = 104 intervention group; n = 101 control group); and a subsample of residents who fell prior to the study (n = 47 intervention group; n = 43 control group). These data were used as covariates to control for differences in fall documentation between the two groups at baseline. Of the 207 chart audits conducted, two records were not included in both MANCOVA analyses due to missing data.
A sample of at least 20 residents for each dependent variable was suggested to ensure robustness.2223 Since there were three dependent variables for the primary analysis, at least 60 residents were needed in each group (eg, intervention and control). On the basis of an estimate of fall rates and feasibility of the study, it was determined that a total of six nursing homes and study length of 4 months would be needed for sufficient power.
Setting and subjects
The number of beds in the nursing homes ranged from 120 to 186. Five of the six nursing homes were part of a corporation and one was a freestanding facility. Four nursing homes were located in an urban setting with high minority representation (60–95%). The tenure of the director of nursing at the beginning of the study ranged from 1 month to 2 years. Three of the six directors resigned during the study period; two were from the control group. The time of administrator service ranged from 3 months to over 11 years. During the 4-month study period, two of the six administrators resigned; both were from the control group.
Table 1 shows characteristics of the intervention (n = 104) and control groups (n = 103) collected from the medical records. There were significantly more older (t = 2.18; df = 205; p = 0.03), female (χ2 = 8.241; df = 1; p = 0.004) and black (χ2 = 6.107; df = 2; p = 0.047) participants in the intervention group. In addition, the intervention group experienced greater use of physical restraints (χ2 = 5.686; df = 1; p = 0.017). Over 75% of all subjects were identified as high risk for falls by the MDS Resident Assessment Plan. Overall, 57% of the sample fell in the 6 months prior to the intervention; this rate was higher in the intervention group than the control group (χ2 = 7.441; df = 1; p = 0.006).
Also included in table 1 are selected examples of falls-specific care processes that should be documented in the medical record. Overall, only 18.4% of residents who fell had an adequate neurological/mental status examination documented in the nursing progress notes following the fall. Although 60–70% of the residents had a diagnosis of bowel or bladder incontinence noted in the medical record, only 17.4% of the medical records had incontinence-related interventions (eg, toileting plan) documented in the falls care plan.
Post-fall process documentation for all residents
For the first analysis, 104 and 101 residents were included from the intervention and control groups, respectively (table 2). After controlling for covariates, the mean diagnosis, management and monitoring audit scores together were higher in the intervention group compared with the control group (Hotelling trace 0.504, F(3,196) = 32.96, p<0.001). These results suggest that the MDIRS intervention had an overall effect on improving nurses’ documentation of the post-fall evaluation process in medical records.
Post-fall process documentation for residents with a fall history prior to the study period
In this analysis, only residents falling within a 6-month period before the study period, and during the intervention period, were included to permit a true pre–post comparison of documentation processes. A total of 90 residents were included, 43 from the control group and 47 from the intervention group (table 2).
Diagnosis, management and monitoring of audit scores together were significantly higher in the intervention group compared with the control group. After controlling for covariates, the intervention group had significantly more complete documentation of the post-fall evaluation process (Hotelling trace 0.537, F(3,76) = 13.60, p<0.001). These results suggest that although the management stage was not significant in the univariate analysis, taken together, the MDIRS intervention still had an overall effect on documentation of post-fall evaluation processes in the subgroup of those with a prior history of falls who fell during the study period.
Thorough post-fall assessments often detect and reduce precipitating factors for future falls.24 Nurses have a central role in this process and work closely with other healthcare providers to ensure that care processes are carried out. This is the first study to demonstrate that structured incident reports can improve assessment and documentation of care processes in medical records following a fall in nursing homes.
The use of systematic assessment forms has the advantage of obtaining readily available information which may not be included in a narrative form. MDIRS enhances a process of care that is already mandated in nursing homes, for which procedures are already in place in all nursing homes. Improved documentation also has the potential to protect nursing homes from legal liability.25 Nursing homes could use MDIRS to measure staff compliance with policies related to fall assessment. In addition, MDIRS could be used to measure the success of particular fall interventions by identifying a reduction of falls related to a particular intervention (eg, bed alarm).
Resnick and colleagues examined care processes following introduction of the AMDA/AHCA falls management guideline. In 40 nursing homes participating in a clinical practice guideline implementation programme, only 33% (n = 13) implemented the guideline.2627 Problems with burdensome documentation were identified as a major challenge to successful implementation. These results suggest the need to implement a falls management programme that does not greatly increase staff workload. Formatted documentation tools were noted as a potential solution to improve programme success.26
Although we did not collect data on staffing ratios, staff education and their familiarity with technology, several organisational factors, including the high rate of turnover of administrators and directors of nursing may also have affected documentation. Structured computerised systems may be helpful for maintaining quality improvement even with frequent staff turnover. High turnover in nursing homes28–30 further supports the need for standardised fall assessment tools that help guide processes of care.31
A limitation of our study was that the control group did not receive any training related to falls or documentation. The research staff attempted to maintain minimal presence on the nursing units by conducting the majority (>80%) of chart audits after the intervention period was completed. We also did not collect data on the appropriateness of the post-fall nursing assessment process during the study period since we were not present when the falls occurred.
Research-based interventions that improve the processes of care are the key to preventing future falls and improving the quality of resident safety in nursing homes. Findings from this study support the argument that a systematically guided incident reporting system improves how nursing home staff assess and document post-fall evaluation processes in the medical records of nursing home residents. Additional studies are needed to identify effective and efficient methods of conducting post-fall assessments.
We are indebted to Dr M H Kutner, and G Cotsonis for statistical guidance and assistance, and C Clevenger and S Asbell for their enormous help with data collection.
Funding: Sources and related paper presentations: This research was supported by the John A Hartford Foundation “Building Academic Geriatric Nursing Capacity” Scholars Program and the Agency for Healthcare and Quality Grants for Health Services Dissertation Research (R03 HS 14663–01).
Competing interests: None.
Ethics approval: The Emory University Institutional Review Board approved the study.