Ab s t r a c t:-Article history:Problem: Patient fall prevention begins with accurate risk assessment. However, sustained improvements in prevention and quality of care include use of validated fall risk assessment tools (FRATs). The goal of FRATs is to identify patients at highest risk. Adult FRATs are often borrowed from to create tools for pediatric patients. Thoughfactors associated with pediatric falls in the hospital setting are similar to those in adults, such asmobility, medication use, and cognitive impairment, adult FRATs and the factors associatedwith themdo not adequately assess risk in children.
Eligibility Criteria: Articles were limited to English language, ages 0–21 years, and publish date 2006–2015.
Sample: The search yielded 22 articles. Tenwere excluded as the populationwas primarily adult or lacked discussion of a FRAT. Critical appraisal and findings were synthesized using the Johns Hopkins Nursing evidence appraisal system.
Results: Twelve articles relevant to fall prevention in the pediatric hospital setting that discussed fall risk assessment and use of a FRATwere reviewed. Comparison between and accuracy of FRATs is challenged when different classifications, definitions, risk stratification, and inclusion criteria are used.
Conclusions: Though there are several pediatric FRATs published in the literature, none have been found to be reliable and valid across institutions and diverse populations.
Implications: This integrative review highlights the importance of choosing a FRAT based on an institution’s identified risk factors and validating the tool for one’s own patient population as well as using the tool in conjunction with nursing clinical judgment to guide interventions.
Falls are a nursing sensitive indicator tracked by Magnet® designated institutions, all of which are seeking to better understand and prevent falls in their specific setting. The Joint Commission’s National Patient Safety Goals deem inpatient falls as a significant patient safety risk and require organizations assess fall risk and implement interventions to reduce fall risk (The Joint Commission, 2015). A commonly used definition of a fall comes from the National Database for Nursing Quality Indicators (NDNQI) as “a sudden, unintentional descent, with or without injury to the patient that results in the patient coming to rest on the floor, on or against some other surface, on another person, or an object” (NDNQI, 2016).
Adult studies report that fall prevention programs with sustained improvement include the use of validated fall risk assessment tools (FRATs). Extensive literature exists on fall risk assessment and the impact of fall prevention programs in the adult population (Matarese, Ivziku, Bartolozzi, Piredda, & De Maranis, 2015); however, we have begun to fully explore the unique risks of the pediatric patient just over the past several years. Several pediatric FRATs exist; though, more research is needed to establish their reliability and validity (Harvey, Kramlick, Chapman, Parker, & Blades, 2010; Jamerson et al., 2014; Ryan-Wenger, Kimchi-Woods, Erbaugh, LaFolette, & Lathrop, 2012). Due to patient variability across pediatric hospitals, no single FRAT has been found to reliably and accurately assess every type of patient. The purpose of this integrative reviewis to synthesize the existing literature on pediatric FRATs and provide recommendations for choosing the best tool that meets your institution’s needs.
Falls are the leading cause of childhood injury with as many as nearly three million children experiencing fall related injuries annually (Centers for Disease Control and Prevention, 2016). Recent studies report the incidence of pediatric falls in the hospital inpatient setting ranging from an incidence rate of 0.51 to 1.0 per 1000 patient days (Cooper & Nolt, 2007; Hill-Rodriquez et al., 2009; Jamerson et al.,2014; Kingston, Bryant, & Speer, 2010; Neiman, Rannie, Thrasher, Terry, & Kahn, 2011; Schaffer et al., 2012). Though these rates are low when compared to adult rates, the incidence of injury is significant at 30–35% (Jamerson et al., 2014; Kingston et al., 2010). Comparing the incidence of pediatric falls between hospital settings is challenging due to the differences in fall definitions, patient populations, classifications, and measurement across institutions. Pediatric falls are often classified as one of four types: accidental, anticipated physiologic, unanticipatedphysiologic, and developmental. Prior to the NDNQI definition for a fall, institutions typically defined a fall simply as a descent to the floor. The Child Health Corporation of America (CHCA) Nursing Falls Study Task Force recently conducted a multi-site study of CHCAmember hospitals to address this issue by using a consistent definition and data collection tool. Although the authors found that fall rates in hospitalizedchildren is low at 0.88 per 1000 patient days compared to adult rates at 3.2–10.7 per 1000 patient days, they acknowledged the study did not control for factors that can vary across institutions such as how each institution identified the occurrence of a fall event (i.e., formal tracking system, verbal report) and determination of fall risk (Jamerson et al., 2014).
Assessing risk factors for pediatric falls differs from adults. Though factors associated with pediatric falls in the hospital setting are similar to those in adults, such as mobility, medication use, and cognitive impairment, many experts agree that adult FRATs and the factors associated with them do not adequately assess risk in children (Kingston et al., 2010; Ryan-Wenger et al., 2012; Razmus, Wilson, Smith, & Newman, 2006). Historically, falls in children have been considered a normalpart of childhood growth and development (Harvey et al., 2010). For example,falls that happen during the time a child is learning to walk are classified as developmental and do not count negatively towards a fall rate unless an injury occurs. However, we must be careful to not deem all children at risk so that we can focus prevention efforts on those at highest risk such as children with physiological, behavioral, or mobility issues. The goal of all FRATs is to identify the patient at highest risk (Razmus & Davis, 2012); however, there is limited literature available that identifies a comprehensive list of risk factors associated with falls in the pediatric patient.
Significance of the Clinical Problem:-
To implement appropriate fall prevention interventions specific to the pediatric patient population, we must begin with the identification of variables that increase a pediatric patient’s fall risk. Studies have found that medication use, an unfamiliar environment, and underlying medical conditions can hinder a child’s orientation and understanding, thus increasing their risk for falling (Cooper & Nolt, 2007; Hill-Rodriquez et al., 2009). Other studies show that common pediatric inpatient fall risk factors include: mobility and mental status impairment, increased length of stay, and history of falls (Cummings, 2006; Graf, 2011; Razmus et al., 2006), but less is known about the correlation between a child’s developmental level, illness, hospital environment, and the risk for falling. Some hospitals have developed their ownFRAT, many based on adult tools, to assess risk in their specific patient population. In 2009, a CHCA sponsored study revealed their member hospitals were using a variety of FRATs,with only six hospitals reporting use of a validated pediatric tool (Child Health Corporation of America (CHCA), Nursing Falls Study Task Force, 2009). Themajority of hospitals were using internally created tools developed from retrospective medical record reviewand analysis of fall events in their population. In a follow- up study, CHCA examined the sensitivity of the FRATs used at the 26 member hospitals and raised further concern as to whether these site specific tools accurately assessed risk of the children who fell (Jamerson et al., 2014). Though creation and use of these tools may have led to initial reduction in fall rates for these institutions, manyhave not been validated beyond the initial testing and are restricted to that single institution’s population.
An extensive literature search was conducted for pediatric FRATs. Online databases were searched including MEDLINE and CINAHL. The search was limited to English language, ages 0–21 years of age,and included evidence published between 2006 and 2015 to ensure the most up-to-date information. Relevant search terms included:fall prevention, fall tool, fall assessment tool, fall risk assessment tool, pediatric fall, pediatric fall tool, pediatric fall risk assessment tool, fall risk, pediatric fall prevention, and fall prevention interventions. Wildcards, truncations and adjacencies were also used. Due to a lack of published literature, we also conducted an internet search using the relevant search terms to locate any pertinent grey literature such as conference proceedings, organizational documents, and abstracts. We also used reference lists from each article to identify additional literature.
The search yielded 22 articles relevant to fall prevention in the hospital setting. Articles were excluded if the population was primarily adult (N21yo) (5 articles) or lacked discussion of fall risk assessment using a FRAT (5 articles). Ultimately, 12 articleswere included in this integrative review and were evaluated using the Johns Hopkins Nursing evidence appraisal system (Dearholt & Dang, 2012). Strength of evidence can range fromLevel I (highest) to Level V (lowest). Quality of evidenceratings of A = High, B = Good, and C = Low/Major flaw were also assigned. All articles reviewed were of high to good quality (Table 1). The authors critically appraised and agreed on the reviews.
When evaluating the extent towhich scores on these FRATs are predictive of children at highest risk to fall, precision, accuracy, and error are often discussed. Precision is a measure of consistency among raters (reliability). Accuracy is a measure of the degree to which observed scores are in agreement with scores from an established standard (validity) (Nunnally & Bernstein, 1994). In the case of falls, accuracy isthe degree to which the patient’s risk to fall (low/high) compares to whether or not the patient actually fell. Accuracy is further evaluated by sensitivity, specificity, and positive and negative predictive values. Sensitivity is a measure of how well a tool correctly classifies or predicts a patient as high risk to fall and the patient fell. Specificity is a measure of how well a tool correctly classifies the patient as low risk to fall and the patient did not fall (Waltz, Strickland, & Lenz, 2010). Positive predictivevalue (PPV) is the probability the patient who scores as high risk will actually fall. Negative predictive value (NPV) is the probability the patient who scores as low risk will not fall (Hennekens & Mayrent, 1987). A high PPV indicates likelihood a person with a high risk score will fall. Systematic measurement error occurs when a FRAT score under- or over-estimates the actual risk resulting in false positives (patientsscored as high risk who do not fall) and false negatives (patients scored as low risk who fall) (Ryan-Wenger et al., 2012).
Synthesis of the Literature:-Pediatric FRAT development began in the early 2000’s, pioneered by ElaineGraf. Graf reported on a retrospective case-matched control study in hospitalized children and found five significant risk factors for falls: length of stay (LOS), orthopedic diagnosis, physical or occupational therapy, seizure medication, and being IV/heparin lock free. These five factorswere found to have a PPV of 84%. Grafwent on to create the General RiskAssessment For Pediatric In-patient Falls (GRAF PIF) scale usingthese factors. A cut-off score of two yielded overall sensitivity of 0.75 and specificity of 0.76 (Graf, 2008). Graf’s studies showed anticipated physiologic falls accounted for 45–61% of all falls (Graf, 2008, Graf, 2011).
In 2006, Cummings sought to determine risk factors for inpatient pediatric falls for development of a FRAT (Cummings, 2006). A case matched study of 78 inpatients who fell and 78 controls was conducted and 28 factors were identified and incorporated into a 6-item tool called The Cummings Scale. The tool evaluates fall history, physical function, cognitive/psychological impairment, equipment need, and medication altering equilibrium using a 3-point cut-off score (no risk, low risk,high risk). Pilot testing of the tool revealed environmental and developmentalfactors were most highly associated with a fall. No data on specificity or sensitivity was provided.
Researchers often turn to adult falls work to guide development of pediatric FRATs. The Morse Fall Scale and the Hendrich II Fall Risk Model are two validated, widely-used adult FRATs that have been examined for their relevancy to the hospitalized pediatric population.Neither one has been found to be wholly applicable to children. However, the Morse Fall Scale was determined to be more accurate in predictingpediatric falls; therefore, individual components of the Morse Fall Scale were further examined to assess each component’s ability to predict pediatric falls. Analysis revealed disorientation, fall history, and impaired gaitwere statistically significant predictors (Razmus et al., 2006). From this work, these authors developed the CHAMPS pediatric FRAT which contains four risk factors (Change in mental status, History of falls, Age b 36 months,Mobility impairment) and two nursing interventions(Parental involvement and Safety). Validation of the CHAMPS was examined in a prospective cohort study matching 47 hospitalized children who fell with forty-seven who did not fall (Razmus & Davis, 2012). The authors concluded the CHAMPS scale predicted falls in hospitalized children. Researchers at a California children’s hospital set out to develop, implement, and evaluate a pediatric fall prevention program (Cooper & Nolt, 2007). Using variables identified in the adult literature, the authors developed a fall assessment form to identify additional factors. A prospective, descriptive review was used to gather relevant data on the falls occurring at their institution and validate the form with three falls that occurred over a three-month period. There was no sensitivity or specificity testing reported for the unnamed tool.The Humpty Dumpty Fall Scale (HDFS) categorizes the pediatric patient into either lowor high risk to fall based on factors that include: age, gender, diagnosis, cognitive impairment, environmental, response to surgery/sedation/anesthesia, and medication usage. The HDFS was piloted using a case-control design to assess whether a certain score on the tool correlated with an actual fall event (Hill-Rodriquez et al., 2009). Of the total 302 cases, there were 22 false negatives and 115 false positives for an overall accuracy of 59.3% of patients correctly classified as to their risk to fall. The authors acknowledged that further refinement of the tool was needed to maintain sensitivity and increase specificity.
In further testing of the HDFS, researchers at a pediatric specialty care hospital in the south central US replicated the 2009 study by Hill- Rodriguez and colleagues to determine if the HDFS accurately captured risk in their patient population (Pauley et al., 2014). This study found the HDFS lacked accuracy in identifying patients who were at high risk to fall. Their study also found a cut-off score of 15 better reflectedthe balance between sensitivity and specificity,whereas the original authorsused a cut-off score of 12. However, even with a cut off of 15, there were 54.1% false negatives and 39.2% false positives. The only commonality between the two study populations was a high percentage of patients with a neurologic diagnosis (Pauley et al., 2014).
The I’M SAFE tool, developed by Neiman et al. (2011) was modeled after the GRAF PIF. This tool assesses Impairment, Medications, Sedation/ anesthesia, Admitting diagnosis, Fall history, and Environment of care. Variables used in the tool were those found to be most highly correlated with falls, those identified by clinical experts, aswell as data that were readily available in the electronic medical record. Inclusion criteria was specific to intrinsic falls defined as those falls that occurred due tofactors associated with the patient’s mental or physical condition.
Though a decrease in the intrinsic fall rate was observed post implementation,sensitivity and specificitywas not discussed. The authors acknowledged further research was necessary, but to date no additional literature has been published to validate this tool. Although some FRATs share similar factors that assess the patient’sfall risk,manywere created based on factors specific to their population of patients who fell making comparisons challenging. Harvey et al. (2010)were the first to compare and analyze the validity of all available pediatric FRATs. Seven FRATs were identified fromthe literature: HDFS, GRAF PIF, Cummings, I’M SAFE, CHAMPS, an unpublished tool from Children’s Hospital of Central California and an unpublished tool from The Children’s NationalMedical Center (CNMC). The assessment factorswere reviewed and compared resulting in the inclusion of five tools. The authors reported all but CHAMPS had reasonably acceptable values for reliability at N60%. Of the five tools, the GRAF-PIF and Cummings were most accurate in correctly identifying patients who fell. Though errors were highest with HDFS, this was the only one of the five to correctly identify the two children who fell in the prospective sample as high risk. The CNMC tool identified the two children who fell in the prospective sample as having no risk. The authors concluded that no one tool was more effective in determining fall risk over the others, though factors fromeach tool such as LOS, bleeding disorders, and temperament/ behavior issues were identified as strong predictors of fall risk.
Similarly, Ryan-Wenger et al. (2012), in anun matched retrospective case-control study, evaluated their institution’s tool, the Pediatric Fall Risk Assessment (PFRA), compared to three published pediatric FRATs: GRAF PIF, HDFS, and CHAMPS. They discussed precision, accuracy, PPV, NPV, and error rate of the three established FRATs. Sensitivity, specificity, and error were below standard for the HDFS and GRAF PIF. Similar to Harvey et al. (2010), these authors concluded that precision and accuracy in predicting pediatric fall riskwas lacking in all four tools. One limitation of research to validate the different pediatric FRATs is that initial testing is restricted to a single institution. A CHCA Nursing Task Force conducted a multi-site study in 2014 to better understand the nature of pediatric inpatient falls. Using a unified definition across multiple pediatric inpatient settings, this convenience sample from 26 CHCA member hospitals showed that only 47% of the children who fell were identified as at risk to fall. The majority of hospitals included in the study did not use published, psychometrically sound FRATs. In the FRATS used that had published psychometrics, sensitivities were found to be less than previously reported (Jamerson et al., 2014).Their findings reinforced the need for further testing and validation of the current pediatric FRATs found in the literature.
Discussion:-We have learned from research thus far that some factors and characteristicsof patients that fall are common across pediatric institutions. However, we continue to see low predictive power, false positives, and false negatives associated with these tools that incorporate those common factors into the assessment (Harvey et al., 2010; Hill-Rodriquez et al., 2009; Pauley et al., 2014). This may be occurring because the risk factors in the FRAT do not match the clinical characteristics seen in that hospitals’ patient population that fall; therefore, it is difficult for the tool to validly measure fall risk (i.e., differentiates those children at high risk from those at low or no risk).
Comparison between and accuracy of FRATs is challenged when different classifications, definitions, risk stratification, and inclusion criteria are used. This integrative review showed some of the studies included intrinsic falls only (Cummings; Graf, 2008, Graf, 2011), some excluded accidental falls (Hill-Rodriquez et al., 2009), and some evaluated patients from birth to 21 years-old whereas others stopped at 16 years (Harvey et al., 2010). In one study, two FRATs had to be excluded because their scoring varied fromthe others,making it difficult to compareacross tools (Harvey et al., 2010).
Validation of most FRATswas not supported in later studies (Harvey et al., 2010; Pauley et al., 2014). Additionally, many lacked reports of sensitivity and specificity. Several FRATs were published on hospital websites or the study data only published in abstract or short report format making assessments and comparisons difficult (Cummings, 2006; Graf, 2008). Psychometrically sound FRATs are essential to being able to predict and prevent falls in hospitalized children. Future studies need to be multi-site and evaluate more than one pediatric population (i.e., intensive care, medical/surgical, oncology). Except for two (Jamerson et al., 2014; Pauley et al., 2014), all studies were restricted to one site and limited populations. Populations and acuity vary across sites and could play a part in the varied results we find when attemptingto validate tools.
The results of this integrative review provide evidence that fall risk assessment is a critical first step in preventing harmand improving safety for our hospitalized pediatric patients. The hospital environment is different from the home, with unfamiliar surroundings and equipment, placing the patient at risk for a fall. Nurses use FRATs in their daily clinical assessment of patient fall risk and are responsible for the implementation of evidence-based interventions to prevent the patient fromfalling. Therefore, further validation of the currently available pediatric FRATs needs to be established with input from nurses at the bedside. Though some tools have similar factors associated with risk, these tools are based on the factors significant to a particular patient population; therefore, vary in what is assessed. This highlights the importance of choosing a FRAT based on your institution’s identified risk factors andvalidating the tool for one’s own patient population and using the tool to guide interventions.We also must appreciate the limitations. FRATs aimto predict the patients thatwill fall, butwe knowfromthe literature and anecdotally that patients who score as low risk may fall and those that score as high risk may not. We should look to these tools not to be the sole source of prediction, but rather prompt action to mitigate the specific factors putting the patient at risk. Use of these tools shouldnot replace nursing clinical judgment. Although more literature is beginning to emerge on fall risk and interventions specific to the pediatric patient, research is still in the beginning stages. Development of a reliable and valid FRAT requires investigators to use study designs that can determine precision, accuracy, and error across multiple, diverse pediatric populations. Only then can we effectively compare fall rates nationally and begin to standardize our approach to prevention. Untilwe have a homogeneous tool,we suggest evaluating your chosen FRAT in your institution’s patient population to assess reliability and validity.
Acknowledgment:-Wewould like to thank Claire Toner,MSN, RN for her support in this work.
References:-Centers for Disease Control and Prevention (2016). Protect the ones you love: Child injuries are preventable. Retrieved from http://www.cdc.gov/safechild/falls (Accessed06/30/16) Child Health Corporation of America (CHCA), Nursing Falls Study Task Force (2009). Pediatric falls: State of the science. Pediatric Nursing, 35(4), 227–233.Cooper, C. L., & Nolt, J. D. (2007). Development of an evidence-based pediatric fall prevention program. Journal of Nursing Care Quality, 22(2), 107–112. http://dx.doi.org/10. 1097/01.ncq.0000263098.83439.8c.Cummings, R. (2006). An evidence-based approach to fall risk assessment. Sigma Theta Tau International Conference. Retrieved from ttp://www.nursinglibrary.org/vhl/handle/10755/155532 (Accessed 06/30/16) Dearholt, S. L., & Dang, D. (2012). Johns Hopkins nursing evidence-based practice model and guidelines (2nd ed.). Indianapolis, IN: Sigma Theta Tau International Honor Society of Nursing.Graf, E. (2011). Magnet children’s hospitals: Leading knowledge development and quality standards for inpatient pediatric fall prevention programs. Journal of Pediatric Nursing, 26, 122–127. http://dx.doi.org/10.1016/j.pedn.2010.12.007.Graf, E. (2008). Pediatric fall risk assessment & classification: Two hallmarks for a successful inpatient fall prevention program. SPN news. 17. (pp. 3–5).Harvey, K., Kramlick, D., Chapman, J., Parker, J., & Blades, E. (2010). Exploring and evaluating five paediatric falls assessment instruments and injury risk indicators: Anambispective study in a tertiary care setting. Journal of Nursing Management, 18,531–541. http://dx.doi.org/10.1111/j.1365-2834.2010.01095.x.Hennekens, C. H., & Mayrent, S. L. (1987). Epidemiology inmedicine. Boston/Toronto: Little, Brown and Company. Hill-Rodriquez, D., Messmer, P. R.,Williams, P. D., Zeller, R. A.,Williams, A. R.,Wood,M., & Henry, M. (2009). The humpty dumpty falls scale: A case–control study. Journal for Specialists in Pediatric Nursing, 14(1), 22–32. http://dx.doi.org/10.1111/j.1744-6155. 2008.00166.x. Jamerson, P. A., Graf, E., Messmer, P. A., Fields, H.W., Barton, S., Berger, A., … Lunbeck, M. (2014). Inpatient falls in freestanding children’s hospitals. Pediatric Nursing, 40(3), 127–135.Kingston, A., Bryant, T., & Speer, K. (2010). Pediatrics falls benchmarking collaborative. Journal of Nursing Administration, 40(6), 287–292. http://dx.doi.org/10.1097/nna. 0b013e3181df10d9. Matarese,M., Ivziku, D., Bartolozzi, F., Piredda, M., & De Maranis,M. G. (2015). Systematic review of fall risk screening tools for older patients in acute hospitals. Journal of Advanced Nursing, 71(6), 1198–1209. National Database of Nursing Quality Indicators (NDNQI) (2016). Guidelines for data collection and submission on patient falls indicator. Retrieved from https://members. nursingquality.org/NDNQIPortal/NDNQI/Documents.aspx (Accessed 06/30/16)
Let’s block ads! (Why?)